Selective separation of planar and non-planar hydrocarbons using an aqueous Pd6 interlocked cage

Polycyclic aromatic hydrocarbons (PAHs) find multiple applications ranging from fabric dyes to optoelectronic materials. Hydrogenation of PAHs is often employed for their purification or derivatization. However, separation of PAHs from their hydrogenated analogues is challenging because of their similar physical properties. An example of such is the separation of 9,10-dihydroanthracene from phenanthrene/anthracene which requires fractional distillation at high temperature (∼340 °C) to obtain pure anthracene/phenanthrene in coal industry. Herein we demonstrate a new approach for this separation at room temperature using a water-soluble interlocked cage (1) as extracting agent by host–guest chemistry. The cage was obtained by self-assembly of a triimidazole donor L·HNO3 with cis-[(tmeda)Pd(NO3)2] (M) [tmeda = N,N,N′,N′-tetramethylethane-1,2-diamine]. 1 has a triply interlocked structure with an inner cavity capable of selectively binding planar aromatic guests.


Materials and methods:
General chemicals and the solvents were purchased from commercially available suppliers and were used without further purification. All the reactions were carried out under ambient conditions in normal atmosphere. The NMR spectra of the newly prepared materials were recorded on BRUKER 400 MHz and 500 MHz spectrometers. The chemical shifts () in the 1 H NMR spectra were reported in ppm relative to the tetramethylsilane, which was used as an internal standard ( = 0.00 ppm) or the resonance of the proton resulting from partial deuteration of the NMR solvents: D2O ( = 4.79 ppm), CDCl3 ( = 7.26 ppm), CD3CN ( = 1.94 ppm) and DMSO-d6 ( = 2.50 ppm). 13 C NMR spectra were recorded using the same instruments at 100 MHz, 125 MHz and all the chemical shifts () were reported in ppm relative to external CDCl3 at 77.8-77.2 ppm, CD3CN at 1.32, 118.26 ppm and DMSO-d6 at 39.52 ppm. Electrospray ionization mass spectra were recorded using Agilent 6538 Ultra-High Definition (UHD) Accurate Mass Q-TOF spectrometer along with the use of standard spectroscopic grade solvents. Electronic absorption spectra were recorded on a LAMBDA 750 UV/Vis spectrophotometer.

Synthesis of interlocked cage 1
In a 4 mL glass vial, L·HNO3 (24.2 mg, 38.44 µmol) was added to a 2 mL Millipore water solution of cis-[(tmeda)Pd(NO3)2] M (20.0 mg, 57.64 µmol). The mixture was heated at 55 °C for 24 hours to result a turbid orange solution. The resulting mixture was centrifuged, and the clear supernatant was taken. It was further purified via crystallization by slow vapor diffusion of acetone into the aqueous solution. Yield: 7.5 µmol (78%

X-Ray crystallographic study of 1
Single crystal X-ray data were collected using Silicon Double Crystal monochromated synchrotron radiation at 100(2) K at the MX2 beamline of the Australian synchrotron. 2 Data integration and reduction was performed using XDS. 3 The structure was solved by intrinsic phasing using ShelXT 4 and refined by the full-matrix least-squares method using SHELXL through the Olex2 GUI. 5 In general, non-hydrogen atoms with occupancies of greater than 0.5 were refined anisotropically, carbon-bound hydrogen atoms were included in idealized positions and refined using a riding model. 6 The structure has a large volume of smeared electron density in the lattice corresponding to highly disordered solvents and anions which could not be successfully modelled. This region of electron density was therefore treated with the solvent masking 6b algorithm of Olex2. Crystallographic data and refinement parameter are given in Table S4. The CIF has been deposited with the CCDC number 2174542.                         The host-guest complexes of 1 were obtained in two different ways: (a) In-situ complexation: To a clean 4 mL glass vial a mixture of M (5.0 mg, 14.41 µmol), L·HNO3 (6.0 mg, 9.61 µmol) and anthracene(A) or phenanthrene(P) (or any other guest) (5.0 mg, excess) was taken followed by the addition of 0.5 mL D2O. The mixture was then stirred for 12 hours at 55 o C to get an orange turbid mixture. The excess guest was removed via centrifugation. The clear orange supernatant was then isolated and characterized by 1 H NMR spectroscopy.
(b) Complexation after the formation of the cage: To a clean 4 mL glass vial a mixture of the acceptor M (5.0 mg, 14.41 µmol) and the ligand L·HNO3 (6.0 mg, 9.61 µmol) was taken followed by the addition of 0.5 mL D2O. The mixture was then stirred for 12 hours at 55 o C to get a turbid orange solution. The solution was centrifuged to get the clear solution of the interlocked cage 1. The solution was then mixed with the guest (5 mg, excess) and stirred for overnight. The orange solution was then isolated by centrifugation and characterized by 1 H NMR spectroscopy.
The 1 H NMR spectra obtained from both the methods were identical and hence any one of the methods could be used to obtain host-guest complexes.

Determination of apparent association constant of 1 with different guests:
The binding behavior of capsule 1 was studied by competitive 1 H NMR titration of 1 with anthracene (A) and phenanthrene (P). Titration experiments were carried out in D2O. The D2O stock solution of 1 with an internal standard of tetrabutylammonium nitrate [(N(Bu)4)NO3] was prepared and stock solutions of A in DMSO-d6 (0.017 M) and P in CD3OD (0.017 M) were also prepared. The 1 H NMR titration was performed by adding aliquots of guest stock solution to 0.5 mL of 1 stock solution. It was found that the guests (A and P) were bound to 1 by slow exchange on the NMR time scale at 25 °C. Thus, the host-guest concentration [HG] could be calculated by integration of Ha' proton with respect to the internal standard.
The binding constant (Ka) and Hill coefficient value (n) were calculated using the Hill equation. 7 Where θ is the fraction of binding sites (host molecule) occupied by the guest, [G] is the guest concentration, n is the Hill coefficient describing cooperativity, and Ka is the apparent association constant.
In the Hill equation, the value of θ was obtained using equation S1. The integration of the NMR peak of Ha', gave the concentrations of the host-guest complex [HG]. Knowing the concentration of [HG], the concentration of guest [G] and that of host [H] can be calculated from the mass balance equation S3 and S4. With the value of θ and [G] calculated, a linear plot can be obtained using the equation S2. From the slope and intercept of such plot n and Ka can be calculated.

Selective host-guest chemistry:
To check selective host-guest chemistry of 1, the complex was prepared as follows: To a clean 4 mL glass vial a mixture of the acceptor M (5 mg, 14.41 µmol), and the ligand L·HNO3 (6.05 mg, 9.61 µmol) was taken followed by the addition of 0.5 mL D2O. The mixture was then stirred for 12 hours at 55 o C to get a turbid orange solution. The solution was centrifuged to get the clear solution of the interlocked cage 1. To this solution was added a solid mixture of equivalent amount of planar guest (5 eqv. of L·HNO3) (A or P) and non-planar guest (5 eqv. of L·HNO3) (H2A: 9,10-dihydroanthracene or S2A: thianthrene). This mixture was then stirred for 12 hours at room temperature. This solution was then centrifuged to obtain the host-guest complex.
To extract the guest from the host-guest complex, the D2O solution of the host-guest complex was stirred with 0.5 mL of CDCl3 overnight. This solution was then centrifuged and the CDCl3 part was taken for recording 1 H NMR.       The peaks of the internal standard tetrabutylammonium nitrate are indicated by . This titration shows no change in 1 H NMR as 1 did not encapsulate MP.

Optimized structures
All the computational studies were carried out using Gaussian 09 package. 8 Full geometry of the inclusion complex A⊂1 was optimized using semiempirical method with PM6 basis set. The structure of 1 as obtained from the SCXRD was used as a starting point. As the number of atoms were high, initial optimization was done by fixing the coordinated of the host and only optimizing the guest (A) inside the host (1). Then, the whole structure was optimized by unfreezing the coordinates of the host. In all calculations, solvation was considered introducing Polarization Continuum model (PCM) choosing water as a solvent for better comparison with experimental results.