Homochiral Self‐Sorted and Emissive IrIII Metallo‐Cryptophanes

Abstract The racemic ligands (±)‐tris(isonicotinoyl)‐cyclotriguaiacylene (L1), or (±)‐tris(4‐pyridyl‐methyl)‐cyclotriguaiacylene (L2) assemble with racemic (Λ,Δ)‐[Ir(ppy)2(MeCN)2]+, in which ppy=2‐phenylpyridinato, to form [{Ir(ppy)2}3(L)2]3+ metallo‐cryptophane cages. The crystal structure of [{Ir(ppy)2}3(L1)2]⋅3BF4 has MM‐ΛΛΛ and PP‐ΔΔΔ isomers, and homochiral self‐sorting occurs in solution, a process accelerated by a chiral guest. Self‐recognition between L1 and L2 within cages does not occur, and cages show very slow ligand exchange. Both cages are phosphorescent, with [{Ir(ppy)2}3(L2)2]3+ having enhanced and blue‐shifted emission when compared with [{Ir(ppy)2}3(L1)2]3+.

The structure of 1·3BF 4 ·n(MeNO 2 )w as confirmed by crystallography(Figure1). [9] There are two independentcage 1 cations that show minor structurald ifferences. Anionsa nd additional solvent were not located due to significant disorder.E ach cage has three pseudo-octahedrally coordinatedI r III centres,e ach with two ppy ligandsa nd the pyridyl groups from two L1 ligandsa re in a cis arrangement.T he two L1 ligands bridge betweent hree Ir III centres. The averaget orsion angleb etween cis-pyridyl groups is 38.048,t ypical for [Ir(ppy) 2 (pyridyl) 2 ]-type complexes [10] with the bowl shape of CTV-type ligands being able to accommodate these torsion angles within the cage structure.
Both L1 ligands within each cage 1 are the same enantiomer,g iving the chiral anti-cryptophane isomer.E ach [Ir(ppy) 2 ] unit within ac age has the same chirality,s uch that only the enantiomeric MM-LLL and PP-DDD cage isomersa re observed in the structure.G iven that the L and D enantiomers of the [Ir(ppy) 2 ] + moietiesa nd the M and P enantiomers of the L-types ligands are presenti nt he reaction mixture, there are twelve possible stereoisomers of the cage. The 1 HNMR spectra of both cages 1 and 2 undergo significant sharpening upon standing (Figures S7 and S15 in the Supporting Information), and fully equilibrate after severalm onths. The 1 HNMR spectrum of cage 1·3PF 6 ,c ollected after 3m onths of standing, is virtually identical to that of the single crystalso f 1·3BF 4 ·n(CH 3 NO 2 )r e-dissolved in [D 3 ]-MeNO 2 (Figure 2a,b). (AE)-L1 was resolved into itsc onstituent enantiomers by chiral HPLC, [11] and each L1 enantiomer reactedw ith each of L-[Ir(ppy) 2 (MeCN) 2 ]·BF 4 and D-[Ir(ppy) 2 (MeCN) 2 ]·BF 4 .A se xpected, the two combinationst hat were mis-matched pairs of enantiomers gave poorly resolved 1 HNMR spectra (Figures S10 and S11), whereas the two combinations that were matched pairs (presumably M-D and P-L)g ave sharp spectra in short timeframes that weres imilart ot he fully sorted cage mixture (Figures 2d, S12, S13). ESI-MS of matched and mis-matched pairs are similar with all combinations showing cage formation ( Figure S14). The observed 1 HNMR spectral sharpeningi st herefore indicative of equilibration involving chiral self-sorting of an initial mixture of cage stereoisomers;t his was also seen in our previous studies of a[ Pd 6 (L1) 8 ] 12 + cage but only the ligand was ac hiral component. [12] We could not resolve the sorted cages by analytical chiral HPLC.
Homochiral metallo-cages with tris-chelate metal coordination are known both from achiral [13a,b] and resolved chiral ligands. [13c-e] Metallo-cagest hat show homochiral self-sorting from ar acemic mixture of ligand enantiomers observed in so-Scheme1.Synthesis of metallo-cryptophane cage species.  lution are rare, [14] although these include Pd II metallo-cryptophanes. [8a] The simultaneous chiral self-sorting of both ligand and pre-formed inert metallo-tecton as reportedh ere have not been previously reported. In ap reliminaryi nvestigation of the influenceo fc hiral guests on the self-assembly of cage 1,g lobular additives were included in 3:2m ixtures of (L,D)-[Ir-(ppy) 2 (MeCN) 2 ]·PF 6 and (AE)-L1.A ddition of chiral R-camphoro r S-camphor led to noticeably faster sharpeningo ft he 1 HNMR spectra than in their absence, but this was not observedf or the addition of achiral adamantane (Figures S15-S20 in the Supporting Information). Interestingly,a ddition of the related anionic species R-(or S-)-10-camphorsulfonic acid to the reaction mixture prevents cage formation presumably as carboxylate is ac ompeting ligand for the iridium (Figures S21 and S22). The cages do not show self-recognition of L-ligand species. ESI-MS of aM eNO 2 solution of L1, L2 and [Ir(ppy) 2 (MeCN) 2 ]·BF 4 shows as tatistical mixtureo f1:[{Ir(ppy) 2 } 3 (L1)(L2)] 3 + :2 cage species ( Figure 3). Mixing 1·3BF 4 and 2·3BF 4 in MeNO 2 resultsi n very slow exchange between L1 and L2 with appreciable ligand exchange only observed after four weeks, and near-statistical mixingr eached after ten weeks ( Figure S6 in the Supporting Information). Thus, these cages have ah igh degree of kinetic stability but are not completely inert. It is interesting to note that this speciation behaviour is in contrastw ith recently reported[ Pd 3 L 2 ] 6 + metallo-cryptophanes, which exclusively formed homocages from two different L-type ligands, with no ligand exchange. [8a] The absorption spectra of 1 and 2 in dichloromethane (DCM) are similar to other [Ir(ppy) 2 (N _ N)] + systems, [7] and characterised by two intense ligand centred( 1 LC) transitions between 260 and 320 nm localised on the ppy and three lower intensity broad bands below 380 nm that consist of spin-allowed and spin-forbidden mixed metal-to-ligand and ligandto-ligandc harge transfer ( 1 MLCT/ 1 LLCT and 3 MLCT/ 3 LLCT,r espectively) transitions ( Figure S26 in the Supporting Information). The weak CT transition observed for 1 at 470 nm was not reported for the monomeric [Ir(ppy) 2 (4-pyCO 2 Et) 2 ] + (4-pyCO 2 Et = 4-ethyl isonicotinate), [10c] suggesting increased conjugation in 1 due to the CTV scaffold.For both 1 and 2,the excitation spectra in DCM match the absorption spectra and indicate asinglephotophysically active species.
Cages 1 and 2 are emissive in DCMs olutiona nd in the solid state. Upon photoexcitation of 1,abroad and unstructured emissioni so bserved both in DCM and in the powder (Figure 4a)d ue to emission from am ixed 3 MLCT/ 3 LLCT state. [7] The photoluminescence spectrum in the powder is red-shifted (l max = 648 nm) compared to that in DCM (l max = 604 nm);however, 1 possesses similarly low F PL of around1%a nd bi-exponentiald ecay kinetics in both media (Table 1). Due to the increasedc onjugation into the CTV scaffold, cage 1 showsr edshifted emission and similar F PL compared to [Ir(ppy) 2 (4-  (ppy) 2 } 6 (tcb) 4 ] 6 + cage also showedared-shifted emission (l max = 575 nm) when compared with the corresponding [Ir-(ppy) 2 (NCPh) 2 ]OTf complex (l max = 525 nm);h owever,u nlike for cage 1 and other Ir(ppy) 2 discrete supramoleculars ystems, [15] the F PL for the Lusby cage was enhanced comparedw ith that of the mononuclear complex (F PL = 4% vs. F PL = < 1%). [4] To mitigate non-radiative vibrational motion in the cage, we spin-coated 5wt% of 1 in polymethyl methacrylate (PMMA), which serves as an inert matrix. The emission in the thin film was blue-shifted and more structured( l max = 514 nm) compared to both the powder and solutions pectra. The F PL of 5.5 %w as enhanced as ar esult of the rigidity conferred by the PMMA host and the emission lifetimes were significantly longer (t e = 634 and2 319 ns).
The photoluminescence spectrum of cage 2 in DCM is more structured and blue-shifted (l max = 516 nm) compared to 1,i ndicatinga ne mission that is more predominantly ligand-centred ( 3 LC;F igure 4b). Theb lue-shifted emission of 2 compared to 1 was expected considering the presence of the electronwithdrawing ester moieties located on L1 in 1,w hichs tabilise the LUMO. [10c] Cage 2 showsasignificantly enhanced F PL and longer t e compared to 1 in DCM (F PL = 15 %, t e = 523, 887 ns).
Unlike 1,t he emissiono f2 as ap owder is not significantly red-shifted (l max = 519 nm), thought he emissionp rofile is less structured, showingl ess well-resolved resolved vibrational bands as shoulders of the main emission peak. The emission profile for 2 in the PMMA-doped thin film is likewise very similar to that in DCM. Although F PL values are low in the powder (F PL = 1.6 %), in the doped film they are higher (F PL = 10 %). Emission lifetimes are expectedly longeri nd opedfilms than in powder( Ta ble 1). Attempts to synthesize an analogousm ononuclearc omplex of 4-phenoxymethylpyridine for comparison were not successful due to ligand oligomerization.
In summary,p hosphorescent [{Ir(ppy) 2 } 3 (L) 2 ] 3 + metallo-cryptophanes can be synthesized in high yields, with the CTV-type ligands being able to accommodate torsion anglest ypical of [Ir(ppy) 2 (L) 2 ]c omplexes to form rare examples of 3D Ir III cyclometallated coordination cages. These cages undergo ligand exchange processes over months and show ar emarkably high degree of homochiral self-sorting of both ligand and metallotecton, but not self-recognition between similar L-type ligands. Chiral sorting is enhanced by the presence of neutralc hiral additives. For cage 1,c hiral self-sorting occurs relativelyr apidly upon crystallisation through an induced seeding effect, but on at imescale of months in solution. Luminescencep roperties of the two cages are quite distinct, pointing to an ability to tune the photophysical properties of these systems. Cage 2 showeda ne nhanced and blue-shifted emissionc ompared to 1,r eaching a F PL of 15 %i nD CM solution and 10 %i nd oped film. These are promising systems for av ariety of applicationsi ncluding semiochemical hosts, photoredox catalysts andi ne nergy conversion materials.