Unraveling Confined Dynamics of Guests Trapped in Self-Assembled Pd₆L₄ Nanocages by Ultrafast mid-IR Polarization Dependent Spectroscopy

Abstract

Self-assembled coordination cages form host-guest complexes through weak non-covalent interactions. Knowledge of how these weak interactions affect the structure, reactivity, and dynamics of guest molecules is important to further the design principles of current systems and optimize their specific functions. In this manuscript, we apply ultrafast mid-IR polarization dependent pump-probe spectroscopy to probe the effects of two Pd₆L₄ self-assembled nanocages on the properties and dynamics of fluxional group VIII metal carbonyl guest molecules. We find that the interactions between the Pd₆L₄ nanocages and guest molecules act to alter the ultrafast dynamics of the guests; restricting rotational diffusional motion and decreasing the vibrational lifetime.

Graphical Abstract

Metal directed self-assembled coordination cages can form nanosized cavities that can act to trap and stabilize guest molecules through non-covalent interactions.^1-5 The physical and chemical properties of these host-guest complexes are dictated by the interplay of many factors including steric restrictions, electrostatics, and weak interactions such as π-stacking.^{1-3, 6} One such example of the self-assembled coordination cages is the octahedral Pd₆L₄ nanocage first synthesized by Fujita and coworkers.⁷ This nanocage has been successfully employed over the last two decades to carry out selective ground and excited reactions and trap and stabilize reactive species.^8-11 Recently, Fujita and co-workers have demonstrated that the Pd₆L₄ nanocages can trap group VIII di-ruthenium carbonyl complexes.^{9, 11} The group VIII di-metal carbonyl complexes are well known for their interesting photochemistry^12-14 and catalytic properties¹⁵ in addition to their fluxional behavior where they exist as multiple isomers in dynamic equilibrium.^15-19 When encapsulated by the Pd₆L₄ nanocage, the di-ruthenium carbonyl complexes were stabilized in their cis-bridging isomeric form⁹ and were shown to undergo carbonyl photosubstitution without cleavage of the metal-metal bond¹¹. From analysis of the X-ray crystal structures the stabilization was attributed to π-stacking interactions between the triazine walls of the cage and the cyclopentadienyl ligands of the metal carbonyl complexes⁹ and the photochemical products were suggested to arise from the restricted dynamics of the guests within the cage¹¹. Even though extensive information can be extracted from the X-ray crystal structure, the crystal structure does not resolve the ultrafast dynamics of the encapsulated guest. A better understanding of how the ultrafast dynamics of the guest molecules are altered by the Pd₆L₄ hosts could lead to further insight into the rational design of Pd₆L₄ nanocages to stabilize and optimize various properties of the guests.²⁰ Though there has been many studies focusing on Pd₆L₄ host-guest complexes there have been limited studies characterizing the ultrafast dynamics of the encapsulated guests.^{8, 21-22} Ultrafast vibrational spectroscopy has proven to be a powerful tool for elucidating structural changes and dynamics of structurally confined systems; including solvent trapped in metal organic frameworks,²¹ cyclodextrin host-guest complexes,²³ and micelles containing trapped water molecules^24-27 . In this manuscript, we apply ultrafast vibrational spectroscopy to two Pd₆L₄ nanocages that encapsulate fluxional group VIII di-metal carbonyl guest complexes 1 and 2 (Figure 1). From the polarization dependent pump-probe spectra, we decipher how the ultrafast dynamics of the trapped guests are altered upon encapsulation. We find that encapsulation results in a decrease in the vibrational lifetimes of the metal carbonyl complexes and a restriction in their diffusional rotational motion. The change in rotational diffusion confirms that the guests are tethered to the nanocage over the course of the vibrational lifetime. Applying a wobbling-in-a-cone analysis we extract the maximum possible cone angles sampled by each metal carbonyl complex, finding that the larger di-ruthenium complex samples a smaller volume than the di-iron complex. As the restricted dynamics and volume sampled by the encapsulated guests are reported to be important for their reactivity¹¹, the measurements and characterizing of the restricted dynamics presented in this manuscript could lead to further insight into how to modify the Pd₆L₄ hosts to alter the reactivity of the metal carbonyl guests.

Figure 1.

A schematic representation of the coordination cages (blue, grey) with a trapped guest (pink). The structure of the two guests 1 and 2 are shown in pink. The components of the Pd₆L₄ nanocage are shown in blue and grey. The blue triangles (L) form the tetrahedral faces of the nanocage and the grey spheres represent the Pd²⁺ ion with different chelating ligands for cages A and B.

The series investigated is schematically depicted in Figure 1 and consists of two variants of the octahedral Pd₆L₄ nanocage, A and B, that trap the fluxional metal carbonyl complexes (Bis(cyclopentadienyl ruthenium dicarbonyl) dimer) 1 and (Bis(cyclopentadienyl iron dicarbonyl) dimer) 2 as guests to form 1⊂A, 1⊂B, 2⊂A, and 2⊂B. The choice of 1 and 2 was motivated by the ability of the Pd₆L₄ nanocages to non-covalently trap the hydrophobic molecules in aqueous solutions^8-10 and the potential for the guests to serve as probes of structural changes, the local environment, and ultrafast dynamics^{23, 28-32}. To investigate the effect of the host structure on the guest’s dynamics we subtly alter the volume of the nanocavity by using different chelating groups having different bite angles to form nanocages A and B.³³

The synthesis of the Pd₆L₄ nanocages A and B, and incarceration of the metal carbonyls 1 and 2 by the nanocages A and B was carried out according to previous protocols^8-9 where we have implemented slight modifications to the procedures that are detailed in sections 2 and 3 of the Supporting Information (SI). Formation of the nanocages and host-guest complexes was confirmed through ¹H-NMR (see section 2, 3, and Figures Sl-7 in SI) in addition to comparison of UV/VIS and FTIR spectra.

Previous studies have shown that encapsulation of di-ruthenium metal carbonyl complexes by a similar Pd₆L₄ nanocage results in changes to the UV/VIS spectra, with the formation of the host-guest complex leading to the appearance of a bright charge transfer band in the visible spectral region.⁹ Comparing the UV/VIS spectra of the guests (1 and 2) and host-guest complexes (1⊂A, 1⊂B, 2⊂A, and 2⊂B) we observe the appearance of a charge transfer band, further confirming the formation of the host-guest complex. The UV/VIS spectra are displayed in Figure 2. As 1 and 2 are not soluble in D₂O, their spectra are reported in 2 chloroform, consistent with previous studies⁹.

Figure 2.

UV/VIS of 1 in chloroform (cyan) ; 1⊂A. (purple) ; 1⊂B (green) ; 2 in chloroform (pink) ; 2⊂A. (red) ; 2⊂B (orange) . Inset shows UV/VIS normalized at 440 nm.

Previous studies performed on 1 and 2 have assigned the transitions beyond 400 nm to metal-to-ligand charge transfer (MLCT) bands of the different isomeric forms.^{12, 14} Upon insertion of 1 and 2 into cages A and B, a change in this spectral region is observed, with a loss of the structured MLCT features and appearance of a broad spectral feature that extends beyond the local absorption of 1 and 2. As the steady state UV/VIS spectra of cages A and B do not extend beyond 400 nm (see figure S8-9 in SI), the change in the UV/VIS spectrum is assigned to an electronic state associated with the host-guest complex and is attributed to the formation of a host-guest charge transfer state observed in previous studies.^8-9 The UV/VIS spectra confirm that the changes in the chelating ligands of the cages and the smaller size of the di-iron metal carbonyl complex still result in the formation of the host-guest charge transfer state.

Along with changes to the electronic excited state, previous studies have observed that the Pd₆L₄ nanocage alters the dynamic equilibrium of the di-ruthenium complex (1).⁹ When free in solution metal carbonyl complexes 1 and 2 exist as different isomers in dynamic equilibrium with 1 existing in 4 forms, a cis-bridging, trans-bridging, gauche-non-bridging and trans-non-bridging, and 2 existing in two forms, a cis-bridging and trans-bridging.^{16, 18-19, 32} Each isomer has a distinct vibrational spectrum in the carbonyl stretching region which can be used as a spectroscopic reporter for a given isomer.^{12, 18-19, 32} Fujita and coworkers have demonstrated that this dynamic equilibrium is altered when the diruthenium metal carbonyl complex (1) is encapsulated by a Pd₆L₄ nanocage, with the Pd₆L₄ nanocage acting to stabilize the cis-bridging form.⁹ Here we expand on these studies through comparing FTIR spectra of 1 and 2 with their encapsulated forms, 1⊂A, 1⊂B, 2⊂A, and 2⊂B, to determine if altering the size of the nanocage cavity and the metal carbonyl guest will result in the stabilization of the cis-bridging isomer. Figure 3 displays the FTIR spectra of the free and encapsulated metal-carbonyl complexes as dashed and solid lines, respectively. It is important to note that different solvents were employed with a polar and non-polar solvent chosen for the free metal carbonyl complexes, and D₂O employed for the encapsulated host-guest complexes. Ideally one would desire to employ the same solvent for the free in solution and encapsulated measurements. However, as the metal carbonyl complexes are not soluble in D₂O we are unable to perform the free in solution measurements in D₂O. In fact, the insolubility ensures that the host-guest complexes are not contaminated by free metal carbonyl signals. Given the issues with solubility we compare the FTIR spectra of the encapsulated complexes to that of the chloroform (a polar solvent), hexane (a nonpolar solvent), DFT calculations, and previous results⁹.

Figure 3.

FTIR spectra of 1 in chloroform (cyan); 1⊂A. (purple); 1⊂B (green) ; 2 in chloroform (pink) ; 2⊂A. (red) ; 2⊂B (orange) ; 1 and 2 in hexane are shown as black dashed lines.

For 1 and 2, free in solution, it is known that their different isomeric forms are stabilized by different solvents and changing the solvent leads to a shift in the relative population of the different isomers, resulting in a change in the amplitudes of the peaks in the FTIR spectra associated with the different isomeric forms .^18-19 This can be seen when comparing the FTIR spectra of 1 and 2 in chloroform (dashed line in cyan and pink respectively) and hexane (dashed line in black) displayed in Figure 3.

The FTIR spectrum of the encapsulated metal carbonyls (1⊂A, 1⊂B, 2⊂A, and 2⊂B) in D₂O are displayed in Figure 3 as solid lines. The spectral changes along with comparison to previous results⁹ indicate that 1 and 2 exist in their cis-bridging isomeric forms when encapsulated in A and B. The cis-bridging isomer exhibits three peaks in the carbonyl stretching region: a strong transition in the bridging carbonyl region ~1700 cm⁻¹, and in the terminal carbonyl region, a weak lower-frequency transition at ~1960 cm⁻¹ and a strong higher-frequency transition at ~2000 cm⁻¹ (see section 5, Figure S10, 11 and Table SI in the SI for DFT calculated spectra) . The FTIR spectra of 1⊂A, 1⊂B, 2⊂A, and 2⊂B indicate the cis-bridging isomer is stabilized when either 1 or 2 is encapsulated by either of the nanocages A or B. This observation is consistent with the previous results of 1 in a similar Pd₆L₄ cage⁹ and demonstrates that the smaller di-iron complex is also stabilized in its cis-bridging form when encapsulated. Our findings exemplify that changes in the size of nanocage cavity and the metal carbonyl complexes result in a similar stabilization of the cis-bridging isomer for 1 and 2.

The FTIR spectra confirm that the cages A and B stabilize the cis-bridging isomeric form of both 1 and 2, but steady state measurements lack the dynamic information of the guests encapsulated by the nanocages. To investigate the impact of the Pd₆L₄ nanocages on the dynamics of the cis-bridging isomer we applied ultrafast mid-IR pump-probe spectroscopy to 1 and 2 free in solution and encapsulated by the cage. Ultrafast mid-IR pump-probe measurements were obtained under two different polarizations schemes, parallel (I∥) and perpendicular I_┴, to extract the population relaxation, P(t) (Eq. 1), and anisotropy decay, r(t) (Eq. 2) , (see section 1 and 6, Figure S12-23 of SI) .^34-36

$$P(t)=\frac{I_{\Vert }(t)+2\ast I_{\perp }(t)}{3}$$ Eq. 1

$$r(t)=\frac{I_{\Vert }(\mathrm{t})-I_{\perp }(t)}{I_{\Vert }(\mathrm{t})+2\ast I_{\perp }(\mathrm{t})}$$ Eq. 2

Note that the population and anisotropy decays reported in the SI and Figure 4 are extracted from the highest frequency transition to ensure minimal contributions from additional excited state absorptions and ground state bleaches (see SI for additional information)

Figure 4 .

Population and anisotropy decays are plotted for the highest frequency transition of 1 in chloroform (cyan); 1⊂A (purple); 1⊂B (green); 2 in chloroform (pink); 2⊂A (red); 2⊂B (orange).

The population relaxation reports on the decay of the population in the excited vibrational state, serving as a measure of the vibrational lifetime. Figure 4(a) and 4(c) display the population relaxation for the metal carbonyl complexes 1 and 2 in chloroform and cages A and B. As noted above, due to issues with solubility the same solvents cannot be used for the free in solution and encapsulated measurements. As the spectral broadening and spectral shift of the metal carbonyl in the positively charged cage is more akin to the chloroform solvent compared to hexane (see Figure 3), we compare the encapsulated forms to the free forms of 1 and 2 in chloroform. This choice is also motivated by previous studies.⁹ We note that for both 1 and 2 we observe that the lifetime in chloroform solution is longer than their cage encapsulated forms. As the cages can be thought of as isolating the metal carbonyls from the bulk D₂O solvent, one may expect the lifetime of the metal carbonyls to be longer when encapsulated. However, the cage also introduces additional pathways by which vibrational energy can be dissipated due to the increase in vibrational degrees of freedom. We attribute the faster population relaxation of the encapsulated metal carbonyls to additional vibrational degrees of freedom introduced by the cage. We note that the cage may also alter the intramolecular vibrational energy redistribution among the carbonyl modes of the metal carbonyl complexes leading to changes in the population relaxation. Future work employing 2DIR spectroscopy will focus on investigating the influence of cage on intramolecular vibrational energy redistribution of the metal carbonyl complexes.

To further exemplify changes in the structural and dynamical nature of the metal carbonyls resulting from encapsulation, we compared their anisotropy decays. The anisotropy is a measure of the diffusional rotational motion of the molecule. The anisotropy decays for 1 and 2 in chloroform and cages A and B are plotted in Figure 4 (b) and 4 (d) . For all the systems, the anisotropy has initial drop from the theoretical value of 0.4 at zero time for an isotropic system, ^37-40 indicating an unresolved ultrafast component. After the initial drop in anisotropy, the metal carbonyls in solution decay to zero with a mono-exponential timescale of ~20 ps for 1, the larger metal carbonyl complex, and ~12 ps for 2, the smaller metal carbonyl complex. We note that the anisotropy decay will vary in different solvents, depending on viscosity, temperature, and solute-solvent interactions. ^{34-35, 41-43} Nonetheless, the measurements in chloroform demonstrate that the metal carbonyls decay as a mono-exponential to zero indicating that the metal carbonyls can freely rotate in solution on the picosecond timescale. When the metal carbonyls are encapsulated in cages A and B the anisotropies do not decay to zero over the course of the vibrational lifetime. This indicates a restriction of the metal carbonyl’s angular motion. We attribute this to the previously observed π-stacking interactions⁹ between the host and guest, in addition to favorable electrostatic interactions between the cyclopentadienyl groups of the metal carbonyl and the triazine walls of the cage. These interactions effectively tether the metal carbonyl complex to the triazine wall of the host over the course of the vibrational lifetime, indicating that the interactions are not breaking/reforming on the timescale of the vibrational lifetime.

To extract additional information we apply a wobbling-in-a-cone model to the anisotropy data.^38-39 Previous mid-IR pump-probe experiments performed on various probe molecules in confined environments such as membranes, polymers, or MOFs have employed a wobbling-in-a-cone model to interpret the anisotropy decay where angular motion of the probe molecule is influenced by the local environment.^{34-35, 38-39} According to the wobbling-in-a-cone model the anisotropy is given by the following expression (Eq. 3) where r₀ is the anisotropy at zero time, r∞ accounts for the non-decaying anisotropy at long times, and τ_θ is the timescale associated with the reorientational motion as the probe decays from r₀ to r∞.^38-39

$$r(t)=(r_o-r_{\infty })\ast \text{exp}\left(-\frac{t}{{\tau }_{\theta }}\right)+r_{\infty }$$ Eq. 3

With this model it is possible to calculate the orientational diffusion of the probe molecule within the boundary of a cone with a half angle given by θ_C. The initial and final anisotropy values are related to the cone semi-angle (θ_C) according to the following expression.^38-39

$$\frac{r_{\infty }}{r_0}=[\frac{1}{2}cos{\theta }_c\left(1+cos{\theta }_c\right){]}^{{}^2}$$ Eq. 4

The cone semi-angle (θ_C) reports on the extent by which the nanocages restrict the motion of the metal carbonyls. To determine (θ_C) we fit the anisotropy data using Eq. 3 to obtain r∞. The fits are plotted as black lines in Fig. S24 in the supporting information and the optimized parameters are given in Table S3 in the SI. From the optimized parameters, we extract a final anisotropy of r∞ ~0.31 for 1⊂ (A, B) and r∞~0.25 for 2⊂(A, B) . We find that 1 has a larger final anisotropy, which corresponds to a smaller region of space sampled during restricted rotational motion. To quantify the differences in the samples, we extract an upper limit for the cone semi-angle sampled as the metal carbonyls undergo restricted rotational diffusion by employing Eq. 4. As we do not completely capture the initial ultrafast decay, due to the 120 fs temporal width of the incoming mid-IR pulses, we employ Eq. 4 to extract an upper limit for the semi-cone angle. The largest possible semi-cone angle provides insight into the restricted motion of the encapsulated metal carbonyl complexes and enables a comparison of guests 1 and 2. To extract the upper limit, we assume a single exponential decaying component with an initial anisotropy of 0.4. Within this limit we extract an angle of θc = 23.2°± 0.1° for 1⊂A., θc = 22.8°± 0.3° for 1⊂B, θc = 30.0°± 0.7°for 2⊂A, and θc = 32.3 °± 0.5° for 2⊂B.

The difference in the observed cone angles is attributed to the structural differences in the metal carbonyl complexes 1 and 2. This is schematically depicted in Figure 5. Fig. 5 displays the DFT optimized structures of 1 and 2 with the nuclear displacements of the ~2000 cm⁻¹ terminal carbonyl symmetric stretching vibrational mode (blue arrows) and the corresponding dipole derivate unit vector (yellow arrow). Within the wobbling-in-a-cone model we can think of the molecular reorientation probed by monitoring the symmetric carbonyl stretching mode as rotation about the derivative dipole vector of the symmetric mode, with the center of the cone lying along the derivative dipole vector. The cone angles are schematically depicted in Fig. 5. The larger cone angle associated with 2⊂A or B is attributed to the smaller size of 2 with a Fe-Fe bond length of 2.566 Å and molecular volume of 371.45 ų/molecule for the cis-bridging isomer. The smaller size of the guest leads to less spatial constraints and more accessible volume for the metal carbonyl complex 2 to sample, hence the larger cone angle. The larger size of 1, with a Ru-Ru bond length of 2.785 Å and molecular volume of 411.35 ų/molecule for the cis-bridging isomer, renders it more sterically hindered when encapsulated by the nanocage with less accessible free volume for rotation. For a given metal carbonyl complex, the anisotropies for cage A and B are very similar, indicating that structural differences in the cage that lead to changes in the electronic spectrum (see Figure 2), do not have a large influence on the restricted rotational diffusion of the encapsulated metal carbonyl complexes.

Figure 5.

The DFT optimized structure for the cis-bridging isomers of 1 and 2 are shown. Nuclear displacements associated with the symmetric stretching mode of the terminal carbonyl groups are indicated with blue arrows and the corresponding dipole derivative unit vector is shown in yellow. Semi-cone angles (θ_C) extracted from the anisotropy decays of 1 and 2 in nanocages A and B are indicated schematically.

To summarize, we demonstrated the influence of π-stacking, electrostatics, and steric constraints imposed by Pd₆L₄ self-assembled nanocages (A and B) on the structure and dynamics of the fluxional metal carbonyl guests (1 and 2) . Previous studies have demonstrated that Pd₆L₄ nanocages can trap and stabilize the cis-bridging isomer of 1.⁹ We have expanded on these studies by performing a systematic study, varying the cavity size of the nanocage and the metal carbonyl guest, to investigate the ultrafast dynamics of the trapped metal carbonyl complexes. Comparing UV/VIS spectra we find that the presence of different coordination ligands for cages A and B results in the formation of charge transfer states with different spectral properties for the four different host-guest complexes (1⊂A, 1⊂B, 2⊂A, and 2⊂B) . From FTIR spectra, we find that both cages act to primarily stabilize the cis-bridging isomer of both guest molecules. Probing the cis-bridging isomer with ultrafast mid-IR polarization dependent pump-probe spectroscopy, we find that encapsulation leads to a faster population relaxation. We attribute this to the accessibility of additional bath modes resulting from the presence of the cage. The anisotropy decays show that the metal carbonyl complexes are free to rotate in solution; however, when encapsulated by the cage their angular motion is restricted, with the larger metal carbonyl complex sampling less volume when encapsulated. The non-decay anisotropy indicates that the π-stacking and electrostatic interactions act to effectively tether the metal carbonyl complexes to the cage over the course of the vibrational lifetime. The group VIII metal carbonyl complexes are well known for their fluxional behavior,^16-19 photophysics, and photochemistry^16,19. Recent studies have shown that the photochemistry of the group VIII di-metal carbonyl complexes can be altered through encapsulation by Pd₆L₄ based nanocages.¹¹ This was attributed to a restriction in the dynamics of the metal carbonyl complexes.¹¹ Here we have characterized the ultrafast dynamics of the metal carbonyls when encapsulated by Pd₆L₄ nanocages, which could lead to further insight into how to modify Pd₆L₄ nanocages to alter the structural properties and photochemistry of the metal carbonyl complexes, and guests in general⁴⁴.