Small Structural Variations Have Large Effects on the Assembly

Article pubs.acs.org/IC

Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX

Small Structural Variations Have Large Effects on the Assembly Properties and Spin State of Room Temperature High Spin Fe(II) Iminopyridine Cages Tabitha F. Miller, Lauren R. Holloway, Phoebe P. Nye, Yana Lyon, Gregory J. O. Beran, W. Hill Harman, Ryan R. Julian, and Richard J. Hooley* Department of Chemistry, University of CaliforniaRiverside, Riverside, California 92521, United States

Inorg. Chem. Downloaded from pubs.acs.org by UNIV OF SUNDERLAND on 10/05/18. For personal use only.

S Supporting Information *

ABSTRACT: Small changes in steric bulk at the terminus of bis-iminopyridine ligands can effect large changes in the spin state of self-assembled Fe(II)-iminopyridine cage complexes. If the added bulk is properly matched with ligands that are either sufficiently flexible to allow twisted octahedral geometries at the Fe centers or can assemble with unusual mer configurations at the metals, room temperature high spin Fe(II) cages can be synthesized. These complexes maintain their high spin state in solution at low temperatures and have been characterized by Xray crystallographic and computational methods. The high spin M2L3 meso-helicate and M4L6 cage complexes display longer N− Fe bond distances and larger interligand N−Fe−N bond angles than their diamagnetic counterparts, and these structural changes invert the ligand selectivity in narcissistic self-sorting and accelerate subcomponent exchange rates. The paramagnetic cages can be easily converted to diamagnetic cages by subcomponent exchange under mild conditions, and the intermediates of the exchange process can be visualized in situ by NMR analysis.

■

INTRODUCTION Self-assembled cage complexes are seeing increased application as sensors, switches, and responsive materials.1−5 A recent focus has been on the synthesis of transition-metal-based cage complexes that show controlled, variable magnetic properties.6,7 Control of magnetic properties is important in device applications,8 and their modulation upon addition of external stimuli such as different guest molecules, other ligands, or light can provide a novel method of creating responsive small molecule magnets. There are a number of paramagnetic selfassembled cages and helicates, commonly using Fe(III),9−13 Co,14−17 or lanthanide salts18,19 as structural vertices. There are far fewer examples of stimuli-responsive paramagnetic cages, especially when compared to the wealth of knowledge on the magnetic properties of mononuclear Fe complexes.20−22 Paramagnetic cages and helicates have been shown to undergo spin crossover as a function of temperature,23,24 but the use of other chemical methods to effect magnetic changes is limited to harsh conditions and redox reactions.25 The use of Fe(II)-iminopyridine-based multicomponent self-assembly provides an opportunity to control cage magnetism via methods other than temperature. These reversibly formed systems display multiple different mechanisms of ligand exchange, including transimination as well as M−L bond cleavage, which suggests they have potential as stimuliresponsive magnets. However, there are far fewer examples of room temperature, high spin Fe(II) helicates,26−34 and especially the larger cage complexes35−41 than their low spin © XXXX American Chemical Society

counterparts. Fe-iminopyridyl cages are commonly (if not always) diamagnetic,42−47 in contrast to Fe(II) complexes with other N and O bidentate ligands.9−11,34 The strong N−Fe−N coordination that allows the formation of highly complex yet stable cage assemblies also favors low spin Fe(II). High spin Fe(II) complexes display Fe−N bond lengths longer than those of their diamagnetic counterparts,48,49 and as such, these weaker bonds are less able to counteract the strong entropic penalties of assembling multiple components into large cage complexes. Tailoring the coordination environment to provide control over the spin state of cage complexes is possible, however, as the Nitschke group recently described a high spin Fe4L4 cage assembly that could be “switched off” by subcomponent exchange and allowed triggered guest release.37 By adding steric bulk to the pyridine units, the cage displayed paramagnetism, as the increased N−Fe bond lengths favored high spin Fe(II) upon assembly. Other responsive spincrossover cages use imino-imidazole ligands that vary the N− Fe−N coordination angle.40,50 There is not yet a perfect strategy to control the spin state at Fe centers in these cages, however, as small variations in ligand can disfavor the cage assembly process. As part of our studies into functionalized self-assembled systems,51−55 we were interested in the effects of functional groups on cage paramagnetism, whether a Received: July 14, 2018

A

DOI: 10.1021/acs.inorgchem.8b01973 Inorg. Chem. XXXX, XXX, XXX−XXX

Article

Inorganic Chemistry

formylpyridine when combined with dianilines and Fe(II) salts.54 We initially focused on the ligands that form simple Fe2L3 meso-helicates. Treating these ligands with Fe(II) salts and the various aldehydes in CH3CN quickly showed how intolerant the assembly process is to steric hindrance at the 6 position of the pyridyl unit. No discrete complexes were formed in any case with BrPyCHO. Varying the Fe salt used had no effect, and all 4 ligands gave brown solids with broad NMR signals, even after prolonged reaction. Despite the fact that the reactions were all performed under an N2 atmosphere to prevent competitive Fe(II) oxidation, no assembly was observed. Similarly, self-assembly with 2-formylquinoline was limited. No discrete complexes were observed upon reaction of QnCHO with S or X. Diaminosuberol SOH was the exception: when combined with QnCHO (2 mol equiv) and Fe(NTf2)2 (0.66 mol equiv) and refluxed in CH3CN for 48 h, a paramagnetic self-assembled complex 1·Qn was isolated, albeit in relatively low yield (30%). The 1H NMR spectrum (Figure 2b) was consistent with a paramagnetic complex (and some

predictable series of paramagnetic self-assembled cage complexes could be formed, and how the nature of the ligand backbone could be used to control these properties. Here, we investigate the effects of varying the ligand coordination angle, steric bulk, and donor ability on the assembly of diastereoselective room temperature high spin Fe(II) coordination cage complexes and illustrate their applications in selfsorting and stimuli-responsive paramagnetism.

■

RESULTS AND DISCUSSION To determine the scope of paramagnetic assemblies that could be accessed, we focused on iterative variations in the ligand backbone and at the groups coordinating the Fe center. The four most successful ligand cores are shown in Figure 1: 3,7-

Figure 1. Subcomponents tested for self-assembly of high spin Fe(II) cage complexes.

diamino-suberone (S), -suberol (SOH),51 and 2,7-diaminoxanthene (X)52 are all known to form stable Fe2L3 mesohelicates upon reaction with 2-formylpyridine and Fe(ClO4)2 in CH3CN, whereas 2,7-diamino-fluorenol (FOH) forms an Fe4L6 cage complex.53 These four ligands were paired with four aldehydes containing a pyridyl nitrogen and varying steric bulk at the 6 position: 2-formylpyridine (PyCHO), its 6-methyl and 6-bromo equivalents (MePyCHO, BrPyCHO), and 2formylquinoline (QnCHO). These four aldehydes vary the steric bulk around the coordinating nitrogen in small increments: the behavior of PyCHO is well-known, and MePyCHO adds bulk without significantly varying the donor ability. BrPyCHO is a slightly weaker σ-donor with steric characteristics similar to those of MePyCHO,54 and QnCHO is highly similar to MePyCHO, except it has fewer degrees of freedom in the appended bulky group. The changes in electron donating ability have little effect on the assembly by themselves: 5-bromo-2-formylpyridine and 5-methyl-2-formyl-pyridine are very similar in self-assembly properties to 2-

Figure 2. Diastereocontrolled assembly of Fe2L3 meso-helicates. 1H NMR spectra of (a) 1·H, (b) 1·Qn, (c) 1·Me, and (d) 3·Me (CD3CN, 400 MHz, 298 K).

decomposition byproducts), and ESI-MS analysis showed peaks consistent with the expected Fe2L3 meso-helicate 1·Qn, albeit with significant fragmentation. This sensitivity was reflected in the overall stability of the compound: while it could be synthesized, it was susceptible to decomposition, and no X-ray quality crystals could be obtained for further structural analysis. The lack of success with BrPyCHO and QnCHO was discouraging, especially for systems as simple as the Fe2L3 meso-helicates. Evidently, there is a small window of B


Article

Inorganic Chemistry functionalization that is tolerated at the 6-position of 2formylpyridines in these assembly processes. The steric bulk that confers paramagnetism on the Fe centers also limits the assembly stability: too much bulk prevents strong coordination. Fortunately, reacting either SOH or X with 2 mol equiv MePyCHO and 0.66 mol equiv Fe(NTf2)2 in MeCN gave stable, isolable self-assembled cage complexes, each of which displayed paramagnetism in the 1H NMR spectrum at room temperature and a soft red color, rather than the deep purple exhibited by diamagnetic Fe(II)-iminopyridine assemblies. The reactions required longer time to reach completion than with PyCHO, and only after 48−72 h reflux was maximal yield achieved. The only exception was diaminosuberone (S), which did not form a stable complex with MePyCHO. The 1H NMR spectra of the suberol-based complexes 1·H, 1·Qn, and 1·Me are shown in Figure 2, and the effect of increased bulk at the aldehyde 6-position is immediately apparent. The NMR spectra for 1·Qn and 1·Me show chemical shift ranges from −100 to +190 ppm, whereas 1·H is diamagnetic and displays a normal 1H NMR spectrum with aromatic H peaks present between 6.0 and 8.2 ppm. ESI-MS analysis corroborates the Fe2L3 stoichiometry in each case (see Supporting Information). Reaction of xanthene ligand X, Fe(NTf2)2, and QnCHO gave minimal yield of assembled product; however, paramagnetic assembly 3·Me could be more readily formed when X was reacted with MePyCHO and Fe(NTf2)2 or Fe(ClO4)2. The 1H NMR spectrum was broadly similar to that of 1·Me, ranging from −60 to +180 ppm. X-ray quality crystals could be grown of the triflimide salt upon slow diffusion of ether into a CH3CN solution of 3·Me·(NTf2)4. The perchlorate salt is far less soluble, and crystal growth was unsuccessful. The solid state structure of 3·Me·(NTf2)4 was determined by X-ray diffraction analysis, and Figure 3 shows the comparison between 3·Me and a DFT-minimized structure of 3·H, as well as between X-ray structures of 1·H, 2·H, and 3· Me. The solid-state structures allow analysis of the coordination at the Fe centers, illustrating why the complexes formed with QnCHO and MePyCHO display paramagnetism. The two most immediately obvious differences in structure are in the N−Fe−N bond angles and the N−Fe bond lengths. Some of these lengths and angles are shown in Figure 3c (see Supporting Information and ref 51 for full details). High spin Fe(II) centers display longer Fe−N bond distances and a distorted ligand field compared to their low spin counterparts.56 The bond angles for the diamagnetic 1·H and 2·H are relatively consistent and are close to octahedral. The internal N−Fe−N angles (in a single iminopyridine 5-membered ring) are 81.03° and 82.07° for 2·H and 1·H, respectively, and the interligand N−Fe−N angles (between two pyridyl N) are 94.95° and 94.30°. Only two angles are discussed here, but they are representative of their other equivalents. In contrast, the coordination around the Fe centers in 3·Me·(NTf2)4 is far more distorted (see Figures 3a and c). The internal N−Fe−N angle is 75.44°, and the interligand N−Fe−N angle (between two pyridyl N) is 107.90°. This distortion in the Fe ligand field is complemented by an increase in Fe−N bond length of almost 0.3 Å. The pyridyl N−Fe bond lengths for 2·H and 1·H are 1.969 and 1.947 Å, respectively, and the imine N−Fe lengths are 1.990 and 1.974 Å. In contrast, those bond lengths are 2.234 and 2.244 Å in the more strained 3·Me. The longer bonds around high spin Fe(II) centers allow the more bulky

Figure 3. Structural variations in Fe2L3 meso-helicates. (a) Solid state structure of 3·Me·(NTf2)4 obtained via X-ray diffraction analysis (counterions and disordered solvent omitted for clarity). (b) Minimized structure of [3·H]4+ (B3LYP-D3(BJ)/6-31G(d), restricted spin). (c) N−Fe−N angle and length comparisons between the solid state structures of 2·H·(ClO4)4,51 1·H·(ClO4)4,51 and 3·Me·(NTf2)4 obtained via X-ray diffraction analysis (truncated structures shown for clarity, containing only one Fe center).

MePyCHO to be incorporated in the assemblies, groups that would be too hindered to assemble with low spin Fe(II). The solid state structures for 1·H, 2·H, and 3·Me illustrate the causes of paramagnetism in the methylpyridine assemblies but do not explain why three relatively similar diamine cores vary so much in their ability to form high spin cage complexes. The diamine that was most amenable to self-assembly was suberol SOH, which formed complexes with PyCHO, MePyCHO, and QnCHO. Unfortunately, we were unable to grow X-ray quality crystals of 1·Qn or 1·Me, so we analyzed their DFT-minimized structures and compared them to the solid state structure of 1·H obtained by X-ray diffraction.51 The structural variations between 1·Qn, 1·Me, and 1·H nicely illustrate why SOH is an effective internal core for assembly with both QnCHO and MePyCHO and why the 1·Qn and 1· Me helicates are paramagnetic. The N−Fe−N angles and N− Fe bond lengths in 1·Qn and 1·Me are consistent with those observed in the solid-state structure of 3·Me, with large (>105°) interligand N−Fe−N angles and small (2.25 Å) N−Fe bond lengths. More pertinent to this discussion, however, are the variations in structure of the suberol ligand. To allow the distortion around the Fe centers, the SOH cores are less puckered in 1·Qn and 1·Me than in 1·H: the Fe−Fe distances are measured to be 9.85 and 9.99 Å in 1·Qn and 1·Me, whereas the Fe−Fe distance in 1·H is only 9.70 Å. This increased Fe− Fe distance is explained by a lower bend angle at the CHOH center of the ligand (defined as 2θb, Figure 4d) in the distorted complexes. This ligand bending angle 2θb = 107.0° and 107.5° in 1·Qn and 1·Me, but is far wider in 1·H with 2θb = 114.4°. C


Article

Inorganic Chemistry

proper coordination to twisted octahedral high spin Fe(II) centers, and limited ligand bending leads to unsuccessful assembly. It was unclear whether this strategy would be effective in forming paramagnetic polyhedra. We attempted the assembly of rigid fluorene-based ligands such as FOH (Figure 5a), 2,7-diaminofluorene, and 2,7-

Figure 4. Structures of (a) solid state structure of 1·H·(ClO4)4 obtained via X-ray diffraction analysis (counterions and disordered solvent omitted for clarity) and the DFT optimized structures of (b) (1·Me)4+ and (c) (1·Qn)4+ (B3LYP-D3(BJ)/6-31G*, unrestricted spin (i.e., four unpaired spins per Fe)). (d) Illustration of the coordination angles used to describe the assembly structures (using ligand Py·SOH as example).

The ligand coordination angle (2θc) changes also, with 2θc = 79.9° and 81.6° in 1·Qn and 1·Me and 2θc = 79.0° in 1·H. The higher flexibility afforded by the sp3 centers and 7-membered central ring in SOH than the other more rigid diamines allows the ligand greater structural freedom. Presumably, this freedom allows SOH complexes to bend so that greater bulk can be accommodated at the metal center more easily in 1·Me and 1· Qn, something less accessible for S or X. This variable bend in the ligand has a secondary effect: the O−O distances between the alcohol groups on adjacent ligands are measured to be 4.0 and 4.1 Å in 1·Qn and 1·Me, but a far greater 4.4 Å in 1·H. This may contribute to the higher diastereoselectivity in assembly shown by 1·Qn and 1· Me when compared to 1·H. The internal CHOH group is prochiral, and four possible isomers can form upon assembly (for the mesocate structure with fac-ΛΔ stereochemistry at the metal centers57,58), with the OH groups either all-in, all-out, out2in, or in2out. Self-complementary H-bonding between the OH groups favors the all-in isomer of 1·H,51 but in solution 10% of a dissymmetric isomer is also present (Figure 2a). These minor peaks are not present in the spectra for 1·Qn and 1·Me, indicating that only one isomer is formed. The modeling suggests that the OH groups are closer and more able to effectively H-bond in 1·Qn and 1·Me, providing even greater favorability for the all-in isomer. The formation of M2L3 meso-helicates is simple and tolerant of extra bulk around the metal centers. The flexibility of the ligand backbone allows significant bending upon assembly, allowing the distorted octahedral coordination environment seen in Figures 3 and 4. However, to form more complex (and interesting) M4L6 tetrahedral cage complexes, more linear, rigid ligands are required. Our experience with (1−3)·Me suggests that flexibility in the ligand backbone is essential for

Figure 5. (a) Diastereocontrolled assembly of Fe4L6 cages 4·H and 4· Me. (b) Solid state structures of 4·H·(ClO4)853 and 4·Me·ClO4· (NTf2)7, obtained via X-ray diffraction analysis (external counterions and disordered solvent omitted for clarity). 1H NMR spectra of (c) 4· Me·ClO4·(NTf2)7 and (d) 4·H·(ClO4)8 (CD3CN, 600 MHz, 298 K).

diaminofluorenone with bulky aldehydes Qn/MeCHO as before. As expected, the assembly process was far less favorable than for the M2L3 assemblies, and only FOH yielded a discrete product. Reaction with QnCHO formed a transiently stable complex: FOH·Qn could be isolated in low yield, and a 1H NMR spectrum was obtained that showed sharp peaks and paramagnetism, but the assembly was not sufficiently stable for ESI-MS analysis, let alone X-ray diffraction. On the other hand, heating FOH with MePyCHO, Fe(NTf2)4, and NaClO4 in CH3CN for 72 h gave cage 4·Me·ClO4·(NTf2)7 in good yield, and the 1H NMR spectrum (Figure 5c) showed peaks from −100 to +200 ppm, indicative of high spin Fe(II) coordination. ESI-MS analysis corroborated the M 4 L 6 stoichiometry (Supporting Information). The presence of ClO4− anions is important for the assembly. When reacted with PyCHO and Fe(II) salts, FOH forms an unusual Fe4L6 cage complex that uses ClO4− as a template. Cage 4·H·(ClO4)8 displays the structure shown in Figure 5b,53 existing as one diastereoisomer with a mer3fac C3-symmetric structure. Hbonding between the OH groups and the templating anion D


Article

Inorganic Chemistry

The presence of mer centers, which are already more distorted than the fac centers in the pyridyl equivalent 4·H·ClO4, allows additional distortion in the paramagnetic 4·Me·ClO4 structure. The N−Fe bond distances are longer in 4·Me·ClO4, ∼2.25 Å in both the fac and mer centers, as opposed to N−Fe bond distances of ∼2.00 Å in 4·H·ClO4. The greatest distortion occurs at the mer centers in 4·Me·ClO4, with very large (116°) interligand N−Fe−N angles and very small (75°) iminopyridine angles. The fac center is still distorted (N−Fe−N = 103°, 75°), but less so. The favorable presence of mer Fe centers in the M4L6 assembly aids the packing of the bulky methylpyridine groups in the self-assembled high spin Fe(II) cage. While 1/3/4·Me are all obviously paramagnetic, determining the spin state of the Fe(II) centers (high or intermediate) requires determination of their magnetic susceptibility (see Experimental Section for definitions). Table 1 and Figures 7

favors this isomer. Synthesis of the MePyCHO equivalent was possible with Fe(ClO4)2, but the resulting 4·Me·(ClO4)8 was quite insoluble in CD3CN, and so we used Fe(NTf2)2 as metal source. In the absence of added NaClO4, no discrete product was formed: one equivalent of ClO4− is required for assembly. This also explains the lack of success with the fluorene/ fluorenone ligands, as they are incapable of H-bonding and favorable templation. The 1H NMR spectrum of 4·Me·ClO4· (NTf2)7 is complex, and a number of peaks are observed. This spectrum is consistent with that of 4·H·(ClO4)8: the mer3fac structure is quite dissymmetric, and (for example) four discrete peaks for imine CH (one fac, three mer) protons can be observed. The peak separation in the paramagnetic spectrum of 4·Me·ClO4·(NTf2)7 further supports this assignment: the peak clusters are split into sets of four, most notably at 180 to 200 ppm and −85 to −100 ppm. Fortunately, X-ray quality crystals of 4·Me·ClO4·(NTf2)7 could be grown which confirm this Fe4L6 mer3fac structure (Figure 5b, Supporting Information). Unfortunately, the crystals were extremely complex, and polymorphic. There were four 4·Me cations, 16 NTf2− and 5 ClO4− anions, plus 15 solvent molecules present in the asymmetric unit of the unit cell. The crystals display weak scattering and solvent/ counterion disorder, the crystals cracked at low temperatures (presumably due to a spin transition at low temperature), and so the acquisition was performed at 150 K. However, the basic structure and connectivity of the complex cation could be unambiguously determined: all four cations in the unit cell showed the Fe4L6 mer3fac 4·Me structure. The solid-state structure also allowed analysis of the coordination environment at the Fe(II) centers. Due to the lower resolution of 4·Me·ClO4·(NTf2)7 (as compared to 3·Me, for example), caution should be taken when comparing the bond angles and lengths. However, the measured internal/ interligand N−Fe−N bond angles and Fe−N bond lengths (Figure 6) are quite consistent with those seen in 3·Me and nicely illustrate why 4·Me·ClO4·(NTf2)7 forms a stable high spin Fe(II) assembly, despite having no flexibility in the ligand backbone, with no ligand bending observed upon assembly.

Table 1. Molar Susceptibility (χM), Corrected Molar Susceptibility (χM′), and Effective Magnetic Moment (μeff) for Cages 1·Me, 1·Qn, 3·Me, and 4·Me at 293 K cage

Δf (Hz)

χM (cm3 mol−1)

χM′ (cm3 mol−1)

μeff

1·Me 1·Qn 3·Me 4·Me

201.6 154.2 184.8 180.0

0.051 0.021 0.023 0.042

0.026 0.023 0.024 0.044

7.7 7.1 7.4 9.9

Figure 7. Downfield regions of the 1H NMR spectra of (a) 3·Me· (NTf2)4 and (b) 4·Me·ClO4·(NTf2)7 at various temperatures, illustrating the high spin nature of both complexes in solution from 313−233 K (CD3CN, 500 MHz).

and S-17−S-20 show the results of Evans method59−61 magnetic susceptibility measurements on the four paramagnetic complexes. Variable temperature NMR studies showed that the paramagnetism persisted throughout the 313−233 K range for all four complexes 1·Me, 1·Qn, 3·Me, and 4·Me in CD3CN solution. Interestingly, χMT rises slightly with decreasing temperature in this range, a relatively unusual observation, but one that is consistent with Nitschke’s related M4L4 complex.37 The observed and corrected χM values are shown in Table 1: as the Mw of the complexes is large, the diamagnetic contribution was calculated and applied in the correction, not ignored.61 In each case, these data are consistent with high spin Fe(II) centers that do not exhibit significant magnetic coupling. The corrected magnetic susceptibilities χM′ of 1·Me, 1·Qn, and 3·Me are all very similar, as would be expected for their similar coordination environment. If the assembly is successful, changing between QnCHO and MePyCHO has little effect on the spin state of

Figure 6. Structural variations in Fe4L6 cages. N−Fe−N angle and length comparisons between the solid state structures of fac and mer Fe centers in (a) 4·Me·ClO4·(NTf2)7 and (b) 4·H·(ClO4)848 obtained via X-ray diffraction analysis (truncated structures shown for clarity, containing only one Fe center). E


Article

Inorganic Chemistry the Fe center. The larger 4·Me·ClO4·(NTf2)7 displays a larger susceptibility due to the greater number of Fe centers. The solid state and computational data support the argument that formation of paramagnetic assemblies is favored with ligands that can form distorted octahedral environments around the Fe centers. If the ligand is flexible enough to allow twisting around the Fe centers, successful assembly of paramagnetic helicates and cages is possible. If the ligand is too rigid (diamino-xanthene, -suberone, -suberenone) or the PyCHO is too bulky (BrPyCHO, QnCHO in certain cases), assembly fails. The large ligand bend in SOH and the fac/mer mix formed with FOH allows the bulky aldehydes to pack around the Fe center, even for the rigid QnCHO. The long Fe−N bonds and distorted coordination angles have a variety of other effects on the properties of the high spin complexes. The pyridyl meso-helicates 1−2·H are examples of self-assembled systems that show high fidelity narcissistic selfsorting when the two ligands are combined with PyCHO and Fe(II) in a one-pot reaction.55 This selectivity is mainly due to the relative energies required to distort the ligand backbone upon self-assembly, as the coordination angles around the Fe centers are quite similar in each case. Suberone meso-helicate 2· H is more stable than the suberol equivalent 1·H and so is formed selectively with no heterocomplexes observed.55 Incorporation of bulky groups at the pyridyl centers changes the nature of the assembly: the Fe−N bond lengths change, as do the N−Fe−N bond angles. It was unclear whether selective sorting would be possible between the high spin complexes 1· Me and 3·Me, but considering 2·Me is unfavorable and cannot be formed via self-assembly, it is evident that the relative stabilities of the complexes change upon modification of the pyridyl group. Figure 8 (and Supporting Information) shows the results of the self-sorting tests. As the relative favorability of xanthene 3·H with respect to 1·H and 2·H had not been previously determined, diamine X was added to a CD3CN solution of 1·H. After heating at 77 °C for 2 h, complete displacement of SOH from 1·H was observed, with only 3·H remaining (Figure 8c). No evidence of any heterocomplexes was seen in the 1H NMR after equilibration. The reverse process, addition of SOH to a solution of 3·H, showed no reaction (Figure 8e), even after 24 h reflux in CD3CN. A ligand competition experiment showed that only 3·H is formed when X and SOH are combined with Fe(NTf2)2 and PyCHO, clearly indicating that 3·H is more favorable than 1·H. Interestingly, competition experiments between 2·H and 3·H did not show self-sorting: mixtures of 2·H, 3·H, and heterocomplexes were observed (see Supporting Information), indicating that there is little difference in favorability between 2·H and 3·H. In contrast, the order of stability of the high spin complexes 1·Me, 2·Me, and 3·Me is reversed. When SOH was added to a solution of 3·Me, displacement occurred rapidly (Figure 8d) and was complete within 10 min at room temperature, yielding only peaks for 1·Me and displaced ligand X. The reverse process, addition of X to a solution of 1·Me, gave no reaction, even after heating to 77 °C for 4 h. As formation of 2·Me was unsuccessful (and addition of diaminosuberone to either 1·Me or 3·Me gave no reaction), it can be concluded that the suberol-based helicate 1·Me is most favorable, followed by xanthene 3·Me and suberone 2·Me. Changing the pyridyl coordinator still allows for high fidelity in self-sorting, but the contributing factors to this selectivity are changed. The dominant contributor to the selectivity is the ability of the ligand core to deform such that the bulky MePyCHO groups

Figure 8. Variable self-sorting behavior in paramagnetic helicates. (a) Order of favorability of 1−3·H and 1−3·Me. Selected regions of the 1 H NMR spectra of the reaction in CD3CN at the noted temperature of (b) X with 1·Me, (c) X with 1·H, (d) SOH with 3·Me, and (e) SOH with 3·H (CD3CN, 400 MHz, 298 K). See Supporting Information and refs 51 and 52 for NMR peak assignment of cages 1· H and 3·H.

can be accommodated around the Fe centers. The other notable effect of varying the aldehyde terminus is the rate of the transimination reaction: full equilibration of pyridylterminated helicates requires heating for multiple hours to effect full transimination, whereas the less strongly coordinating methylpyridyl variants react in minutes at room temperature. The longer Fe−N bond lengths and more distorted coordination environment causes the 1−3·Me complexes to be less favorable than their pyridyl counterparts. The large differences in stability between the high spin 1−4· Me complexes and the low spin 1−4·H complexes suggested the possibility of interconverting between the two complexes, effectively “switching off” the paramagnetism.37 Subcomponent aldehyde exchange is possible in iminopyridine complexes by heating the complex with an excess of aldehyde and a small amount of water to aid the imine solvolysis.37,54 Reaction of paramagnetic 1−4·Me with PyCHO would form the diamagnetic counterparts 1−4·H. This reaction also allows a view into the mechanism of the subcomponent exchange process: as the shifts of the protons in the high spin complex are highly separated in the NMR spectrum, observation of intermediates is simpler than in reactions of diamagnetic complexes. The process for 1·Me is shown in Figure 9. When PyCHO and an excess of water is added to a CD3CN solution of 1·Me and heated at reflux, aldehyde exchange occurs rapidly. After only 15 min, the symmetrical NMR spectrum of 1·Me is converted to a highly unsymmetrical spectrum, with each peak in the paramagnetic region converted to multiple equivalents. When 6 PyCHO molecules are added to 1·Me, the resulting complexes retain paramagnetism, and the dissymmetric 1· F


Article

Inorganic Chemistry

these structural changes have large effects on the stability, selfsorting, and subcomponent exchange rates. While the selective self-sorting of diamagnetic iminopyridine helicates is controlled by ligand deformation energy, selective narcissistic selfsorting of the paramagnetic equivalents is controlled by the coordination environment around the metal center. More flexible ligands are better suited for assembly into paramagnetic meso-helicates, and the observed ligand selectivity is inverted with respect to the diamagnetic counterparts. The transimination rates are highly accelerated, and the paramagnetic cages can be smoothly converted to their diamagnetic counterparts by subcomponent exchange. The intermediates of the exchange process can be visualized in situ by NMR analysis and illustrate that paramagnetism can be conferred on the assembly by the presence of a small number of bulky ligand components.

■

Figure 9. Spin state changes upon aldehyde interconversion. Downfield (paramagnetic) region of the 1H NMR spectra of (a) 1· Me; + PyCHO, H2O (6 equiv), heat at 77 °C for (b) 5 min, (c) 75 min, and (d) 120 min; (e) independently synthesized 1·H (CD3CN, 400 MHz, 298 K).

EXPERIMENTAL SECTION

General Information. 1H spectra were recorded on a Bruker Avance NEO 400 or Bruker Avance 600 MHz NMR spectrometers. The paramagnetic nature of the cages and their low solubility prevented acquisition of adequate 13C spectra.37 Magnetic susceptibility experiments were recorded on a Bruker Avance 600 MHz NMR. Proton (1H) chemical shifts are reported in parts per million (δ) with respect to tetramethylsilane (TMS, δ = 0), and referenced internally with respect to the protio solvent impurity. Deuterated NMR solvents were obtained from Cambridge Isotope Laboratories, Inc., Andover, MA, and used without further purification. Mass spectrometric samples were infused into an Orbitrap Velos Pro mass spectrometer with the standard HESI source at a flow rate of 3 μL/ min. The spray voltage was 3 kV, capillary temperature was set to 170 °C and an S-lens RF level of 45% was applied. Full FTMS were acquired with a resolution of r = 30 000, and ambient ions were used as internal lock mass calibrants. CID spectra were collected in ZoomScan mode where the isolation window = 5 m/z, normalized collision energy (nCE) = 30, and activation time = 30 ms. MS data were analyzed using Thermo XCalibur. Predicted isotope patterns were prepared using ChemCalcd All other materials were obtained from Aldrich Chemical Co., St. Louis, MO, or TCI, Tokyo, Japan and were used as received. Solvents were dried through a commercial solvent purification system (Pure Process Technologies, Inc.). Ligands S,51 SOH,51 X,52 and FOH53 were synthesized according to literature procedures. The synthesis, characterization, and NMR assignment of diamagnetic cages 1·H·(ClO4)4,51 2·H·(ClO4)4,51 3·H· (ClO4)4,52 and 4·H·(ClO4)853 can be found in the representative publications. Synthesis of New Cage Compounds. meso-Helicate 1·Me. 3,7Diaminosuberol (SOH) (170 mg, 0.71 mmol), 6-methylpyridine-2carboxaldehyde (171 mg, 1.41 mmol), and Fe(NTf2)4 (310 mg, 0.47 mmol) were combined in MeCN (10 mL) in a 25 mL round-bottom flask. The solution was then heated at 77 °C for 48 h with stirring. The reaction mixture was cooled, and the acetonitrile removed in vacuo. The red solid was sonicated with 20 mL of 3:1 Et2O:MeOH solution and filtered. After drying, the product was isolated as a red powder (572 mg, 91% yield). 1H NMR (400 MHz; CD3CN) δ 196.64 (s), 60.84 (s), 57.5 (s), 24.34 (s), 16.24 (d, J = 9.6 Hz), 14.62 (s), 6.76 (d, 9.2 Hz), 6.50 (s), 2.62 (s), 1.96 (s), −15.27 (s), −28.72 (s), −36.83 (s), −103.82 (s). HRMS (ESI) m/z calcd for C87H78Cl2Fe2N12O11 ([M2L3·(ClO4)2]2+) 824.1989, found 824.9617. Elemental Analysis: Calcd for C95H78F24Fe2N16O19S8 C: 44.36; H: 3.06; N: 8.71; Found: C: 44.68; H: 2.96; N: 8.53. meso-Helicate 1·Qn. 3,7-Diaminosuberol (SOH) (30 mg, 0.13 mmol), 6-methylpyridine-2-carboxaldehyde (58.8 mg, 0.38 mmol), and Fe(NTf2)2 (50.8 mg, 0.082 mmol) were combined in MeCN (3 mL) in a 25 mL round-bottom flask. The solution was then heated at 77 °C for 48 h with stirring. The reaction mixture was cooled, and the acetonitrile removed in vacuo. The brown solid was sonicated with 20 mL of 3:1 Et2O:MeOH solution and filtered. After drying the product

Me6−x·Hx intermediates can be clearly observed. After 2 h, the subcomponent exchange is complete, and diamagnetic 1·H can be isolated via precipitation and washing. Monitoring the reaction at 343 K in the NMR tube (see Supporting Information) allows observation of the formation and loss of the intermediates. The number of peaks in the spectrum suggests that the addition of 3 or more PyCHO groups is required for spin-state interconversion, and that 1·Me4·H2 is still paramagnetic. This spin-state interconversion is effective for both 1·Me and 3·Me. The complex 1H NMR spectra of 4· Me and 4·H made identifying the exchange components in that case quite challenging, and so reaction monitoring was not practical, though loss of paramagnetism does occur under the same conditions.

■

CONCLUSIONS In this work, we investigated the effect of varying component structure on the assembly properties of room-temperature, high spin Fe(II)-iminopyridine self-assembled cages. Varying the steric bulk of the aldehyde component favors the formation of high spin cages. The self-assembly process with bulky aldehydes is highly sensitive to ligand coordination angle and rigidity: the coordination environment around the high spin Fe(II) centers is distorted, and only ligands that can accommodate this twisted octahedral geometry successfully form self-assembled systems. If the added bulk is properly matched with ligands that are either sufficiently flexible to allow twisted octahedral geometries at the Fe centers or can assemble with unusual mer configurations at the metals, room temperature high spin Fe(II) complexes can be successfully synthesized. Use of aldehydes that are too bulky or diamine ligands that are too rigid fails to yield discrete self-assembled complexes. Both Fe2L3 meso-helicates and Fe4L6 cages can be synthesized in good yield, and these complexes remain paramagnetic in the temperature window observable by NMR. The assemblies were characterized by X-ray crystallographic and computational methods. The high spin complexes display longer N−Fe bond distances and larger interligand N− Fe−N bond angles than their diamagnetic counterparts, and G


Article

Inorganic Chemistry

removal from the mother liquor, and rapid handling prior to flash cooling in the cryostream was required to collect data. In addition, the crystals cracked at low temperature (presumably due to a spin transition), and so the acquisition was performed at 150 K in this case. There were four cations of [C162H132Fe4N24O6]8+, 16 anions of [C2F6NO4S2]−, 5 anions of [ClO4]−, and 15 solvent molecules of C4H10O present in the asymmetric unit of the unit cell. There are a total of 11 anions and possible solvent molecules of ether/acetonitrile that cannot be identified. The basic structure and connectivity of the complex cation could be unambiguously determined, which is all that is required for the purposes of this work. The structure was deposited at the CCDC (no. 1848254), and the structural details can be found in the Supporting Information. Ligand Selectivity Experiments. Amine Displacement Procedure. All displacement experiments were performed in an NMR tube. One equivalent of preformed cage 1·Me (10 mg, 0.0039 mmol), and three equivalents of dianiline X (4.3 mg, 0.012 mmol) were placed in an NMR tube. Dry deuterated acetonitrile (400 μL) was added to the tube, and a proton spectrum of the starting mixture was obtained. The tube was heated at 25 °C for paramagnetic complexes or 77 °C for diamagnetic complexes for at least 2 h to determine whether the preformed cage was displaced by the free dianiline ligand. Competitive Assembly Procedure. All mixing experiments were performed in an NMR tube. One equivalent of dianiline SOH (0.02 mmol) and one equivalent of dianiline B (0.02 mmol) were placed in an NMR tube. Deuterated acetonitrile (400 μL) was added to the tube, and a proton spectrum of the dianiline mixture obtained. Two equivalents of 2-formylpyridine were added (0.04 mmol) followed by 0.66 equiv of iron perchlorate (0.013 mmol) Fe(NTf2)2 in CD3CN). A spectrum of the mixture was obtained. The tube was heated at 77 °C for 1 h. Another spectrum was taken after heating to show the favored cage and the unfavored dianiline ligand. Aldehyde Displacement Procedure. All displacement experiments were performed in an NMR tube. One equivalent of preformed 1·Me (10 mg, 0.0039 mmol) and 6 equiv of 2-formylpyridine PyCHO (2.22 μL, 0.0233 mmol) were placed in an NMR tube. Dry deuterated acetonitrile (400 μL) was added, and the sample heated at 70 °C for a given period of time, and subjected to 1H NMR analysis throughout the course of the experiment. Magnetic Susceptibility Measurements. Experiments were performed using a previously established method.47 A solution was made using 7.5 mg of cage in a 99:1 CD3CN:DCE solvent system. An initial 1H NMR was taken at room temperature on a Bruker Avance 600 MHz NMR. A flame-sealed melting point tube containing a solution of 99:1 CD3CN:DCE was inserted into the NMR tube and used as reference. The NMR was then cooled to −40 °C, and the temperature was raised at 10 °C increments up to 50 °C. Mass susceptibility was determined by eq 1, where Δf is peak separation (Hz), f is the NMR frequency (Hz), m is mass per cm3, and χ0 is −0.534 × 10−6cm3g−1, the mass susceptibility of CD3CN.

was isolated as a brown powder (100.7 mg, 28.9% yield). 1H NMR (400 MHz; CD3CN) δ 181.41 (s), 55.48 (s), 26.17 (s), 24.15 (s), 13.68 (d, J = 9.4 Hz), 6.58 (s), 5.77 (s), 5.64 (d, J = 9.6 Hz), 3.44 (s), 3.31 (s), 1.34 (s), 1.15 (s), −10.30 (s), −19.93 (s), −48.08 (s), −76.25 (s). HRMS (ESI) m/z calcd for C109H78F12Fe2N14O11S4 ([M2L3·(NTf2)22+) 1113.6694, found 1113.6691. Elemental Analysis: Calcd for C113H78F24Fe2N16O19S8 C: 48.68; H: 2.82; N: 8.04; Found: C: 48.32; H: 2.70; N: 7.84. meso-Helicate 3·Me. 2,7-Diamino-9H-xanthene (X) (150 mg, 0.71 mmol), 6-methylpyridine-2-carboxaldehyde (171 mg, 1.41 mmol), and Fe(NTf2)4 (310 mg, 0.47 mmol) were combined in MeCN (10 mL) in a 25 mL round-bottom flask. The solution was then heated at 77 °C for 48 h with stirring. The reaction mixture was cooled, and the acetonitrile removed in vacuo. The red solid was sonicated with 20 mL of 3:1 Et2O:MeOH solution, and the red solid was removed via vacuum filtration. After drying, the product was isolated as a red powder (592 mg, 94% yield). 1H NMR (400 MHz; CD3CN) δ 184.25, 54.01 (s), 52.89 (s), 13.56 (s), 8.41 (s), 6.87 (s), −25.49 (s), −30.94 (s), −33.24 (s), −48.88 (s). HRMS (ESI) m/z calcd for C85H66F12Fe2N14O11S4 (M2L3·(NTf2)22+) 963.1208, found 963.6368. Elemental Analysis: Calcd for C89H66F24Fe2N16O19S8 C: 42.97; H: 2.67; N: 9.01; Found: C: 42.78; H: 2.67; N: 9.11. Cage 4·Me. 2,7-Diaminofluorenol (FOH) (200 mg, 0.94 mmol), 6methylpyridine-2-carboxaldehyde (228.29 mg, 1.88 mmol), Fe(NTf2)2 (412.82 mg, 0.63 mmol), and NaClO4 (24.99 mg, 0.20 mmol) were combined in MeCN (15 mL) in a 25 mL round-bottom flask. The solution was then heated at 77 °C for 72 h with stirring. The red solid was sonicated with 20 mL of 3:1 Et2O:MeOH solution and the red solid was removed via vacuum filtration. After drying, the product was isolated as a red solid (570 mg, 76% yield). 1H NMR (400 MHz; CD3CN): δ 198.17 (s), 194.84 (s), 186.44 (s), 185.28 (s), 72.77 (s), 62.18 (s), 60.26 (s), 57.38 (s), 55.50 (s), 55.41 (s), 54.81 (s), 53.05 (s), 36.65 (s), 27.89 (s), 23.98 (s), 19.40 (s), 17.15 (s), 11.32 (s), 9.17 (s) (s), 8.99 (s), 8.55 (s), 7.96 (s), 3.47 (s), 3.02 (s), 2.60 (s), 1.14 (s), −0.18 (s), −1.54 (s), −3.55 (s), −7.10 (s), −7.57 (s), −7.72 (s), −21.24 (s), −32.07 (s), −33.02 (s), −40.82 (s), −47.73 (s), −88.59 (s), −106.85 (s), −112.94 (s), −324.92 (s). HRMS (ESI) m/z calcd for C162H132ClFe4N24O10 ([M4L6·ClO4]7+) 404.6806, found 404.8304. Density Functional Theory Calculations. Structures were optimized using the B3LYP density functional with Grimme D3(BJ) dispersion correction and the 6-31G* basis set.62−65 Acetonitrile solvent effects were mimicked using an implicit polarizable continuum model. Spin unrestricted calculations were performed for the high spin complexes (four unpaired spins per iron atom), while spin restricted calculations were performed for the diamagnetic complexes. All calculations were performed using Gaussian 09. Crystallography. Xanthene meso-Helicate 3·Me·(NTf2)4. X-ray quality crystals were grown via slow diffusion of diethyl ether into a solution of cage in CH3CN. The diffraction data were obtained, and the crystal structure of 3·Me·(NTf2)4 was solved at the UC Riverside ACIF facility. Data collection, solution, and refinement were routine. X-ray intensity data were collected at 100(2) K on a Bruker APEX2 platform-CCD X-ray diffractometer system (fine focus Mo-radiation, λ = 0.71073 Å, 50 kV/30 mA power). The structure was deposited at the CCDC (no. 1848253), and the structural details can be found in the Supporting Information. Fluorenol Cage 4·Me·ClO4·(NTf2)7. X-ray quality crystals were grown via slow diffusion of diethyl ether into a solution of cage in CH3CN. The diffraction data for 4·Me·ClO4·(NTf2)7 were obtained at the UC San Diego X-ray Crystallography Facility, and the structure was solved at the UC Riverside ACIF facility. X-ray intensity data were collected at 150(2) K on a Bruker Rotating Anode generator with APEX2 platform-CCD X-ray diffractometer system (Mo radiation, λ = 0.71073 Å, 50 kV/24 mA rotating anode power). The crystal of 4·Me·ClO4·(NTf2)7 exhibited the usual problems of this type of structure, namely, weak scattering due to a combination of poor crystallinity, extensive solvation, and disorder of anions/solvent molecules. The crystals employed immediately lost solvent after

χg =

3Δf + χ0 4πfm

(1)

Molar susceptibility, χM, is determined by eq 2, where M is the molar mass of the complex. χM = χg M

(2)

Molar susceptibility χM contains the diamagnetic correction (χMdia) which, due to the large MW of these species, cannot be ignored.17 The corrected molar susceptibility (χM′) was calculated using values determined from Pascal’s constants to correct for the diamagnetic contributions from the ligands and Fe(II) core electrons and counterions. χMdia for cages are 1·Me, 1·Qn, 3·Me, and 4·Me are −0.001210, −0.001369, −0.001141, and −0.002210 cm3 mol−1, respectively. χM′ = χM + χM dia H


Article

Inorganic Chemistry The magnetic moment was calculated from the equation μeff = 8χp T ≈ n(n + 2) , where χp is the paramagnetic suscept-

(8) Estrader, M.; Uber, J. S.; Barrios, L. A.; Garcia, J.; LloydWilliams, P.; Roubeau, O.; Teat, S. J.; Aromi, G. A Magneto-optical Molecular Device: Interplay of Spin Crossover, Luminescence, Photomagnetism, and Photochromism. Angew. Chem., Int. Ed. 2017, 56, 15622−15627. (9) Saalfrank, R. W.; Hörner, B.; Stalke, D.; Salbeck, J. The 1st neutral adamantanoid iron(III)-chelate complex - spontaneous formation, structure, and electrochemistry. Angew. Chem., Int. Ed. Engl. 1993, 32, 1179−1182. (10) Saalfrank, R. W.; Burak, R.; Breit, A.; Stalke, D.; Herbst-Irmer, R.; Daub, J.; Porsch, M.; Bill, E.; Müther, M.; Trautwein, A. X. MixedValence, Tetranuclear Iron Chelate Complexes as Endoreceptors: Charge Compensation Through Inclusion of Cations. Angew. Chem., Int. Ed. Engl. 1994, 33, 1621−1623. (11) Saalfrank, R. W.; Trummer, S.; Krautscheid, H.; Schünemann, V.; Trautwein, A. X.; Hien, S.; Stadler, C.; Daub, J. A Neutral, TripleHelical, Trinuclear, Oxo-Centered Mixed-Valence Iron Complex. Angew. Chem., Int. Ed. Engl. 1996, 35, 2206−2208. (12) Clegg, J. K.; Lindoy, L. F.; Moubaraki, B.; Murray, K. S.; McMurtrie, J. C. Triangles and tetrahedra: metal directed selfassembly of metallosupramolecular structures incorporating bis-βdiketonato ligands. Dalton Trans. 2004, 0, 2417−2423. (13) Saalfrank, R. W.; Bernt, I.; Chowdhry, M. M.; Hampel, F.; Vaughan, G. B. M. Ligand-to-Metal Ratio Controlled Assembly of Tetra- and Hexanuclear Clusters Towards Single-Molecule Magnets. Chem. - Eur. J. 2001, 7, 2765−2769. (14) Paul, R. P.; Bell, Z. R.; Jeffery, J. C.; McCleverty, J. A.; Ward, M. D. Anion-templated self-assembly of tetrahedral cage complexes of cobalt(II) with bridging ligands containing two bidentate pyrazolylpyridine binding sites. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 4883−4888. (15) Tidmarsh, I. S.; Faust, T. B.; Adams, H.; Harding, L. P.; Russo, L.; Clegg, W.; Ward, M. D. Octanuclear Cubic Coordination Cages. J. Am. Chem. Soc. 2008, 130, 15167−15175. (16) Turega, S.; Cullen, W.; Whitehead, M.; Hunter, C. A.; Ward, M. D. Mapping the Internal Recognition Surface of an Octanuclear Coordination Cage Using Guest Libraries. J. Am. Chem. Soc. 2014, 136, 8475−8483. (17) Cullen, W.; Hunter, C. A.; Ward, M. D. An Interconverting Family of Coordination Cages and a meso-Helicate; Effects of Temperature, Concentration, and Solvent on the Product Distribution of a Self-Assembly Process. Inorg. Chem. 2015, 54, 2626−2637. (18) Elhabiri, M.; Scopelliti, R.; Bünzli, J.-C. G.; Piguet, C. Lanthanide Helicates Self-Assembled in Water: A New Class of Highly Stable and Luminescent Dimetallic Carboxylates. J. Am. Chem. Soc. 1999, 121, 10747−10762. (19) Terazzi, E.; Guénée, L.; Bocquet, B.; Lemonnier, J.-F.; Dalla Favera, N.; Piguet, C. A Simple Chemical Tuning of the Effective Concentration: Selection of Single-, Double-, and Triple-Stranded Binuclear Lanthanide Helicates. Chem. - Eur. J. 2009, 15, 12719− 12732. (20) Gutlich, P.; Hauser, A.; Spiering, H. Thermal and Optical Switching of Iron(II) Complexes. Angew. Chem., Int. Ed. Engl. 1994, 33, 2024−2054. (21) Halcrow, M. A. Structure: function relationship in molecular spin-crossover complexes. Chem. Soc. Rev. 2011, 40, 4119−4142. (22) Gutlich, P.; Garcia, Y.; Goodwin, H. Spin Crossover phenomena in Fe (II) complexes. Chem. Soc. Rev. 2000, 29, 419−427. (23) Fujita, K.; Kawamoto, R.; Tsubouchi, R.; Sunatsuki, Y.; Kojima, M.; Iijima, S.; Matsumoto, N. Spin States of Mono- and Dinuclear Iron (II) complexes with Bis(imidazolylimine) Ligands. Chem. Lett. 2007, 36, 1284−1285. (24) Hagiwara, H.; Tanaka, T.; Hora, S. Synthesis, structure, and spin crossover above room temperature of a mononuclear and related dinuclear double helicate iron(II) complexes. Dalton Trans. 2016, 45, 17132−17140. (25) Burke, M. J.; Nichol, G. S.; Lusby, P. J. Orthogonal Selection and Fixing of Coordination Self-Assembly Pathways for Robust

ibility. χp was calculated from χp = χM − χsol M , where Δf is peak separation (Hz), f is the NMR frequency (Hz), and m is mass per cm3.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.8b01973. Spectral and crystallographic data (PDF) Accession Codes

CCDC 1848253−1848254 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, by emailing [email protected], or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Gregory J. O. Beran: 0000-0002-2229-2580 W. Hill Harman: 0000-0003-0400-2890 Ryan R. Julian: 0000-0003-1580-8355 Richard J. Hooley: 0000-0003-0033-8653 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank the National Science Foundation (CHE-1708019 to R.J.H., CHE-1665212 to G.J.O.B., CHE-1752876 to W.H.H., and CHE-1626673 for the purchase of Bruker NEO 600 and NEO 400 spectrometers) and NIH (NIGMS grant R01GM107099 to R.R.J.) for funding, Dr. Dan Borchardt for NMR assistance, and Dr. Milan Gembicky for acquisition of X-ray diffraction data. Supercomputer time from XSEDE (TG-CHE110064 to G.J.O.B.) is also gratefully acknowledged.

■

REFERENCES

(1) Ward, M. D.; Raithby, P. R. Functional behaviour from controlled self-assembly: challenges and prospects. Chem. Soc. Rev. 2013, 42, 1619−1636. (2) Fujita, M.; Tominaga, M.; Hori, A.; Therrien, B. Coordination Assemblies from a Pd(II)-Cornered Square Complex. Acc. Chem. Res. 2005, 38, 369−378. (3) Cook, T. R.; Stang, P. J. Recent Developments in the Preparation and Chemistry of Metallacycles and Metallacages via Coordination. Chem. Rev. 2015, 115, 7001−7045. (4) Smulders, M. M. J.; Riddell, I. A.; Browne, C.; Nitschke, J. R. Building on architectural principles for three-dimensional metallosupramolecular construction. Chem. Soc. Rev. 2013, 42, 1728−1754. (5) Saalfrank, R. W.; Maid, H.; Scheurer, A. Supramolecular Coordination Chemistry: The Synergistic Effect of Serendipity and Rational Design. Angew. Chem., Int. Ed. 2008, 47, 8794−8824. (6) McConnell, A. J. Spin-state switching in Fe (II) helicates and cages. Supramol. Chem. 2018, 30, 858−868. (7) Hogue, R. W.; Singh, S.; Brooker, S. Spin crossover in discrete polynuclear iron(II) complexes. Chem. Soc. Rev. 2018, 47, 7303. I


Article

Inorganic Chemistry Metallo-organic Ensemble Construction. J. Am. Chem. Soc. 2016, 138, 9308−9315. (26) Sunatsuki, Y.; Kawamoto, R.; Fujita, K.; Maruyama, H.; Suzuki, T.; Ishida, H.; Kojima, M.; Iijima, S.; Matsumoto, N. Structures and Spin States of Bis(tridentate)-Type Mononuclear and Triple Helicate Dinuclear Iron(II) Complexes of Imidazole-4-carbaldehyde azine. Inorg. Chem. 2009, 48, 8784−8795. (27) Sunatsuki, Y.; Kawamoto, R.; Fujita, K.; Maruyama, H.; Suzuki, T.; Ishida, H.; Kojima, M.; Iijima, S.; Matsumoto, N. Structures and spin states of mono- and dinuclear iron(II) complexes of imidazole-2carbaldehyde azine and its derivatives. Coord. Chem. Rev. 2010, 254, 1871−1881. (28) Tuna, F.; Lees, M. R.; Clarkson, G. J.; Hannon, M. J. Readily prepared metallo-supramolecular triple helicates designed to exhibit spin-crossover behaviour. Chem. - Eur. J. 2004, 10, 5737−5750. (29) Pelleteret, D.; Clerac, R.; Mathoniere, C.; Harte, E.; Schmitt, W.; Kruger, P. E. Asymmetric spin crossover behaviour and evidence of light-induced excited spin state trapping in a dinuclear iron(II) helicate. Chem. Commun. 2009, 221−223. (30) Darawsheh, M.; Barrios, L. A.; Roubeau, O.; Teat, S. J.; Aromi, G. Guest-, Light- and Thermally-Modulated Spin Crossover in [Fe(II)2] Supramolecular Helicates. Chem. - Eur. J. 2016, 22, 8635− 8645. (31) Hagiwara, H.; Tanaka, T.; Hora, S. Synthesis, structure, and spin crossover above room temperature of a mononuclear and related dinuclear double helicate iron(II) complexes. Dalton Trans. 2016, 45, 17132−17140. (32) Struch, N.; Brandenburg, J. G.; Schnakenburg, G.; Wagner, N.; Beck, J.; Grimme, S.; Lutzen. A Case Study of Mechanical Strain in Supramolecular Complexes to Manipulate the Spin State of Iron(II) Centres. Eur. J. Inorg. Chem. 2015, 2015, 5503−5510. (33) Li, L.; Craze, A. R.; Akiyoshi, R.; Tsukiashi, A.; Hayami, S.; Mustonen, O.; Bhadbhade, M. M.; Bhattacharyya, S.; Marjo, C. E.; Wang, Y.; Lindoy, L. F.; Aldrich-Wright, J. R.; Li, F. Direct monitoring of spin transitions in a dinuclear triple-stranded helicate iron(II) complex through X-ray photoelectron spectroscopy. Dalton Trans. 2018, 47, 2543−2548. (34) Stadler, C.; Daub, J.; Kohler, J.; Saalfrank, R. W.; Coropceneanu, V.; Schunemann, V.; Ober, C.; Trautwein, A. X.; Parker, S. F.; Poyraz, M.; Inomata, T.; Cannon, R. D. Electron transfer in a trinuclear oxo-centred mixed-valence iron complex, in solid and solution states. J. Chem. Soc., Dalton Trans. 2001, 3373−3383. (35) Ferguson, A.; Squire, M. A.; Siretanu, D.; Mitcov, D.; Mathoniere, C.; Clerac, R.; Kruger, P. E. A face-capped [Fe4L4]8+ spin crossover tetrahedral cage. Chem. Commun. 2013, 49, 1597− 1599. (36) Bilbeisi, R. A.; Zarra, S.; Feltham, H. L. C.; Jameson, G. N. L.; Clegg, J. K.; Brooker, S.; Nitschke, J. R. Guest Binding Subtly Influences Spin Crossover in an FeII4L4Capsule. Chem. - Eur. J. 2013, 19, 8058−8062. (37) McConnell, A. J.; Aitchison, C. M.; Grommet, A. B.; Nitschke, J. R. Subcomponent Exchange Transforms an FeII4L4 Cage from High- to Low-Spin, Switching Guest Release in a Two-Cage System. J. Am. Chem. Soc. 2017, 139, 6294−6297. (38) Ren, D.-H.; Qiu, D.; Pang, C.-Y.; Li, Z.; Gu, Z.-G. Chiral tetrahedral iron(II) cages: diastereoselective subcomponent selfassembly, structure interconversion and spin-crossover properties. Chem. Commun. 2015, 51, 788−791. (39) Zhang, F.-L.; Chen, J.-Q.; Qin, L.-F.; Tian, L.; Li, Z.; Ren, X.; Gu, Z.-G. Metal-center exchange of tetrahedral cages: single crystal to single crystal and spin-crossover properties. Chem. Commun. 2016, 52, 4796−4799. (40) Li, L.; Saigo, N.; Zhang, Y.; Fanna, D. J.; Shepherd, N. D.; Clegg, J. K.; Zheng, R.; Hayami, S.; Lindoy, L. F.; Aldrich-Wright, J. R.; Li, C.-G.; Reynolds, J. K.; Harman, D. G.; Li, F. A large spincrossover [Fe4L4]8+ tetrahedral cage. J. Mater. Chem. C 2015, 3, 7878−7882. (41) Li, L.; Neville, S. M.; Craze, A. R.; Clegg, J. K.; Sciortino, N. F.; Arachchige, K. S. A.; Mustonen, O.; Marjo, C. E.; McRae, C. R.;

Kepert, C. J.; Lindoy, L. F.; Aldrich-Wright, J. R.; Li, F. Spin-State Patterning in an Iron(II) Tripodal Spin-Crossover Complex. ACS Omega 2017, 2 (7), 3349−3353. (42) Hristova, Y. R.; Smulders, M. M. J.; Clegg, J. K.; Breiner, B.; Nitschke, J. R. Selective anion binding by a chameleon capsule with a dynamically reconfigurable exterior. Chem. Sci. 2011, 2, 638−641. (43) Clegg, J. K.; Cremers, J.; Hogben, A. J.; Breiner, B.; Smulders, M. M. J.; Thoburn, J. D.; Nitschke, J. R. A Stimuli Responsive System of Self-assembled Anion-binding Fe4L68+ Cages. Chem. Sci. 2013, 4, 68−76. (44) Bolliger, J. L.; Belenguer, A. M.; Nitschke, J. R. Enantiopure Water-Soluble [Fe4L6] Cages: Host-Guest Chemistry and Catalytic Activity. Angew. Chem., Int. Ed. 2013, 52, 7958−7962. (45) Bilbeisi, R. A.; Ronson, T. K.; Nitschke, J. N. A Self-Assembled [FeII12L12] Capsule with an Icosahedral Framework. Angew. Chem., Int. Ed. 2013, 52, 9027−9030. (46) Roberts, D. A.; Castilla, A. M.; Ronson, T. K.; Nitschke, J. R. Post-assembly Modification of Kinetically Metastable FeII2L3 Triple Helicates. J. Am. Chem. Soc. 2014, 136, 8201−8204. (47) Castilla, A. M.; Ronson, T. K.; Nitschke, J. R. SequenceDependent Guest Release Triggered by Orthogonal Chemical Signals. J. Am. Chem. Soc. 2016, 138, 2342−2351. (48) Brooker, S. Spin crossover with thermal hysteresis: practicalities and lessons learnt. Chem. Soc. Rev. 2015, 44, 2880−1892. (49) Rodriguez-Jimenez, S.; Yang, M.; Stewart, I.; Garden, A.; Brooker, S. A Simple Method of Predicting Spin State in Solution. J. Am. Chem. Soc. 2017, 139, 18392−18396. (50) Struch, N.; Bannwarth, C.; Ronson, T. K.; Lorenz, Y.; Mienert, B.; Wagner, N.; Engeser, M.; Bill, E.; Puttreddy, R.; Rissanen, K.; Beck, J.; Grimme, S.; Nitschke, J. R.; Lutzen, A. An Octanuclear Metallosupramolecular Cage Designed To Exhibit Spin-Crossover Behavior. Angew. Chem., Int. Ed. 2017, 56, 4930−4935. (51) Young, M. C.; Johnson, A. M.; Hooley, R. J. Self-Promoted Post-Synthetic Modification of Metal-Ligand M2L3 Mesocates. Chem. Commun. 2014, 50, 1378−1380. (52) Holloway, L. R.; Bogie, P. M.; Lyon, Y.; Julian, R. R.; Hooley, R. J. Stereoselective Post-Assembly CH Oxidation of Self-Assembled Metal-Ligand Cage Complexes. Inorg. Chem. 2017, 56, 11435−11442. (53) Young, M. C.; Holloway, L. R.; Johnson, A. M.; Hooley, R. J. A Supramolecular Sorting Hat: Stereocontrol in Metal-Ligand SelfAssembly by Complementary Hydrogen Bonding. Angew. Chem., Int. Ed. 2014, 53, 9832−9836. (54) Wiley, C. A.; Holloway, L. R.; Miller, T. F.; Lyon, Y.; Julian, R. R.; Hooley, R. J. Electronic Effects on Narcissistic Self-Sorting in Multicomponent Self-Assembly of Fe-Iminopyridine meso-Helicates. Inorg. Chem. 2016, 55, 9805−9815. (55) Holloway, L. R.; Young, M. C.; Beran, G. J. O.; Hooley, R. J. High Fidelity Sorting of Remarkably Similar Components via MetalMediated Assembly. Chem. Sci. 2015, 6, 4801−4806. (56) Phan, H.; Hrudka, J. J.; Igimbayeva, D.; Lawson Daku, L. M.; Shatruk, M. A Simple Approach for Predicting the Spin State of Homoleptic Fe(II) Tris-diimine Complexes. J. Am. Chem. Soc. 2017, 139, 6437−6447. (57) Meng, W.; Clegg, J. K.; Thoburn, J. D.; Nitschke, J. R. Controlling the Transmission of Stereochemical Information through Space in Terphenyl-Edged Fe4L6 Cages. J. Am. Chem. Soc. 2011, 133, 13652−13660. (58) Meng, W.; Ronson, T. K.; Nitschke, J. R. Symmetry breaking in self-assembled M4L6 cage complexes. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 10531−10535. (59) Evans, D. F. The determination of the paramagnetic susceptibility of substances in solution by nuclear magnetic resonance. J. Chem. Soc. 1959, 0, 2003−2005. (60) Löliger, J.; Scheffold, R. Paramagnetic Moment Measurements by NMR. A micro technique. J. Chem. Educ. 1972, 49, 646−647. (61) Piguet, C. Paramagnetic Susceptibility by NMR: The “Solvent Correction” Removed for Large Paramagnetic Molecules. J. Chem. Educ. 1997, 74, 815−816. J


Article

Inorganic Chemistry (62) Becke, A. D. Density-functional thermochemistry. III. The role of exact exchange. J. Chem. Phys. 1993, 98, 5648−5652. (63) Rassolov, V. A.; Ratner, M. A.; Pople, J. A.; Redfern, P. C.; Curtiss, L. A. 6-31G* basis set for third-row atoms. J. Comput. Chem. 2001, 22, 976−984. (64) Weigend, F.; Ahlrichs, R. Balanced basis sets of split valence, triple zeta valence and quadruple zeta valence quality for H to Rn: Design and assessment of accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297−3305. (65) Goerigk, L.; Kruse, H.; Grimme, S. Benchmarking Density Functional Methods against the S66 and S66 × 8 Datasets for NonCovalent Interactions. ChemPhysChem 2011, 12, 3421−3433.

K


Small Structural Variations Have Large Effects on the Assembly

Recommend Documents