Utilizing Raman Spectroscopy as a Tool for Solid- and Solution-Phase Analysis of Metalloorganic Cage Host–Guest Complexes

The host–guest chemistry of coordination cages continues to promote significant interest, not least because confinement effects can be exploited for a range of applications, such as drug delivery, sensing, and catalysis. Often a fundamental analysis of noncovalent encapsulation is required to provide the necessary insight into the design of better functional systems. In this paper, we demonstrate the use of various techniques to probe the host–guest chemistry of a novel Pd2L4 cage, which we show is preorganized to selectively bind dicyanoarene guests with high affinity through hydrogen-bonding and other weak interactions. In addition, we exemplify the use of Raman spectroscopy as a tool for analyzing coordination cages, exploiting alkyne and nitrile reporter functional groups that are contained within the host and guest, respectively.


S1. Materials and Methods
All reagents and solvents were purchased from Alfa Aesar, VWR, Fluorochem, or Sigma Aldrich and used without further purification unless stated otherwise. All reactions were carried out under air, unless stated otherwise. All 1 H, 13 C and 19 F NMR spectra were recorded on either a 500 MHz Bruker AV III equipped with a DCH cryo-probe (Ava500), a 500 MHz Bruker AV IIIHD equipped with a Prodigy cryo-probe (Pro500), or a 400 MHz Bruker AV III equipped with BBFO+ probe (Ava400) at a constant temperature of 300 K. All DOSY experiments were performed on the Ava500 using bipolar gradient pulses for diffusion with two spoil gradients (ledbpg2s.compensated) pulse sequence. The sequence was carried out under automated conditions where the duration of the magnetic field pulse gradient (δ) was 1.5 ms and the diffusion time (∆) was 100 ms. Typically, in each PFG NMR experiment, a series of 16 spectra on 32 K data points were collected and the eddy current delay (Te) was set to 5 ms in all experiments. The pulse gradients (g) were incremented from 2 to 95% of the maximum gradient strength in a linear ramp. The temperature was set and controlled at 300 K with an air flow of 400 L h -1 in order to avoid any temperature fluctuations due to sample heating during the magnetic field pulse gradients. Chemical shifts are reported in parts per million. For 1 H NMR spectra chemical shifts are referenced to 2.50 and 5.32 ppm for dimethyl sulfoxide-d6 [(CD3)2SO] and methylene chloride-d2 (CD2Cl2), respectively. For 13 C NMR spectra chemical shifts are referenced to 39.52 and 53.84 ppm for dimethyl sulfoxide-d6 and methylene chloride-d2, respectively. Apparent multiplicities are reported using the following standard abbreviations: m = multiplet, q = quartet, t = triplet, d = doublet, s = singlet, bs = broad singlet. All NMR spectroscopic analysis was performed with MestReNova, Version 14.
Single X-ray quality crystals were grown from a standing solution of L in (CD3)2SO. Crystallographic details are provided in section S5.2.
Single X-ray quality crystals were grown from vapor diffusion of diethyl ether into a solution of 2 in CH2Cl2 over 2 days.

S3. Mass Spectrometry
Electrospray Ionization (ESI) mass spectra of 2 was performed on a Synapt G2 (Waters, Manchester, UK) mass spectrometer, using a direct infusion electrospray ionization source (ESI), controlled using MassLynx S5 v4.1 software. Crystals of 2 were collected and dried before the sample was dissolved in acetonitrile at 50 µM prior to the measurement. Capillary voltages were adjusted between 1.5 and 2.5 kV to optimize spray quality, while the sampling cone and the extraction cone voltage were minimized to reduce breakdown of the assemblies. Source temperature was set at 80 °C. The data was analyzed using the MassLynx v4.1 software, with predicted isotopic distributions calculated using mMass opensource mass spectrometry tool. S5

Figure S6
Mass spectrum of 2 with experimental results and modelled isotopic distributions shown in black and red, respectively.

S4. Host-Guest Studies
S4.1 Spectroscopic data for host-guest complexes 1 H NMR host-guest studies were performed on a 500 MHz Bruker AV III equipped with a DCH cryoprobe (Ava500) at 300 K. Initial sample volumes were 500 µL with a 0.45-0.50 mM concentration of 2 in CD2Cl2. The guests were added in excess as solids and the tubes sonicated for 5 minutes before 1 H NMR spectra were recorded.

Figure S7
1 H NMR host-guest spectra (500 MHz, CD2Cl2, 300 K) showing the shifting of Ha (labelled) of (a) 2 on the addition of (b) 1,4-dicyanobenzene (DCB), (c) 1,4-dicyanonaphthalene (DCN), (d) 9,10dicyanoanthracene (DCA), (e) 2,3,5,6-tetrachlorodicyanobenzene (TCDCB). The host and guest signals are represented by the green and orange colors, respectively. S4.2 Experimental details for association constant determination 1 H NMR titration experiments were carried out on a 400 MHz Bruker AV III spectrometer equipped with BBFO+ probe (Ava400) at 300 K. Initial sample volumes were 500 μL with 0.45-0.50 mM concentration of 2. Solutions of the guest quinones were 15-30 mM in the same stock solution of the cage. 1 H NMR spectra were recorded at 0-30 equivalents of the dicyanoarene. Association constants were obtained by analysis of the resulting titration data using the 1:1 host-guest stoichiometry equation 1 for fast exchange using the Levenberg-Marquardt Nonlinear Least-Squares Algorithm implemented in the R software and the RStudio software interface. S6 The error of the determined association constants are estimated to be less than 10%. UV-Vis spectroscopy was carried out on a JASCO V-670 Spectrophotometer running Spectra Manager II (Jasco). The data was analyzed and plotted using Origin 2018 software using equation 1. All measurements were made at room temperature (16-21 °C) at 100 μM in CH2Cl2 using a fused silica cuvette with a 10 mm path length.    Figure S15 1 H NMR (400 MHz, CD2Cl2, 300 K) titration curve of 2 (0.5 mM) with benzoquinone (25 mM). The curve was obtained by monitoring the internal cage cavity proton Ha. The solid points represent the experimental data with the continuous dashed line represents the best-fit binding isotherm.

Table S1
Comparison of the association constants (KAss) of quinone and dicyanoarene guests in 1 and 2, respectively. All KAss were obtained in CD2Cl2 with BArF as the counterion.

S5.1 General experimental details
Crystals were mounted on a MITIGEN holder in Paratone or perfluoroether oil on a Rigaku Oxford Diffraction SuperNova diffractometer and were kept at a steady at T = 120 K during data collection. The structures were solved with the ShelXT 2018/2 solution program using the Intrinsic Phasing solution method and by using Olex2 as the graphical interface. S8,S9 The models were refined with ShelXL 2018/3 using full matrix least squares minimization on F 2 . S10 S5.2 Crystallographic data and special refine details The value of Z' is 0.5. This means that only half of the formula unit is present in the asymmetric unit, with the other half consisting of symmetry equivalent atoms. Benzoquinone 7.24 ± 0.10 x 10 3 Dicyanobenzene 1.10 ± 0.04 x 10 3 Naphthoquinone 3.49 ± 0.47 x 10 5 Dicyanonaphthalene 2.14 ± 0.10 x 10 4 Anthraquinone 4.89 ± 0.38 x 10 7 Dicyanoanthracene 3.66 ± 0.74 x 10 4 Benzoquinone 1.8 ± 0.03 x 10 1 S19 Some of the -CF3 groups were modelled as disordered, with appropriate similarity restraints. The resolution of the data set was cut at 0.96 Å, consistent with rapidly rising values of Rint at higher resolution. The value of Z' is 0.5. This means that only half of the formula unit is present in the asymmetric unit, with the other half consisting of symmetry equivalent atoms. Crystals of these cage-type compounds are very susceptible to rapid desolvation. In an attempt to prevent crystal decomposition, a small amount of solution was poured onto a shallow well microscope slide onto which a drop of Fomblin oil had been placed. Crystals were removed from the NMR tube, from which they had grown, and transferred into the microscope well solution before being pushed into the Fomblin oil. Crystal decomposition occurred as was evident in the diffraction pattern. The diffraction pattern was handled as a non-merohedral twin consistent with two clearly different domains being observed in a reciprocal lattice viewer. These actually correspond to a split, rather than twinned diffraction pattern with component 2 rotated by 3.1204° around [-0.21 0.93 -0.31] (reciprocal) or [-0.05 1.00 0.03] (direct). Different integration options were explored and the twinned integration, with production of an hklf5 format reflection file, gives the most acceptable refinement. The model has its limitations; it was not possible to optimize the ShelXL weighting scheme. A check with PLATON SQUEEZE (using a LIST 8-style FCF file) shows no overlooked additional solvent. The cage structures (there are two half-cages per asymmetric unit, plus two half-guests) were easily identified from solution and refine well, with no issues. Refinement of the BArFions was more problematic. Two can be modelled with essentially no restraints. The BArFions containing B401 and B501 were modelled with geometric similarity restraints relating to the BArFions containing B201. Additionally, these two anions plus the B301-containing BArFion were modelled with the RIGU restraint applied. Despite this, some -CF3 groups exhibit large displacement ellipsoids, which cannot be controlled by classical disorder modelling. They are thus left as they are. They do appear to correlate with the disordered solvent, methylene chloride-d2. There are 13 molecules of solvent in the asymmetric unit. Most are disordered to a certain degree, consistent with the rapid desolvation observed during crystal mounting. Geometric and RIGU restraints were used where applicable. One of the methylene carbon atoms was refined using an isotropic model. In some instances, the methylene chloride-d2 molecules have a close approach to -CF3 groups, which also exhibit large ellipsoids. The apparent disorder is obviously correlated but could not be adequately modelled. There is a single molecule in the asymmetric unit, which is represented by the reported sum formula. In other words: Z is 2 and Z' is 1.
The solvent masking routine of Olex2 was used to account for electron density relating to dichloromethane molecules which could not be modelled using discrete atoms. The value of Z' is 0.5. This means that only half of the formula unit is present in the asymmetric unit, with the other half consisting of symmetry equivalent atoms. The structure has been refined as far as is practical given the quality of the experimental data. ShelXspecific constraints and restraints are elaborated upon in the embedded .res file. Some particular details are: (1) The data resolution was cut at 1.4 Å. Rmerge and I/σ(I) become too high and too low respectively at higher resolutions. This is still enough to identify all non-H atoms in the cage, guest and the counterions, plus some solvent molecules. (2) The solvent masking routine of Olex2 was used to handle four molecules of dichloromethane which can be identified from a difference map, but which do not refine well.
(3) All B, C and N atoms, plus some N atoms, were refined using an isotropic model as a significant proportion became non-positive definite when modelled anisotropically. (4) Distance similarity restraints were used on disordered -CF3 groups and on the Pd-N distances. The value of Z' is 0.5. This means that only half of the formula unit is present in the asymmetric unit, with the other half consisting of symmetry equivalent atoms. The crystal was observed to desolvate on removal from the mother liquor. While the cage, guest and counterions all are reasonably well defined, the unit cell also contains a lot of dichloromethane solvent molecules. Much of this could be identified from successive difference Fourier maps and included in the model but electron density pertaining to 2.5 molecules per asymmetric unit -10 per unit cell -could not, and so were removed with the SQUEEZE routine of PLATON. This triggers checkCIF alerts, which should be ignored. Some of the dichloromethane molecules were refined using an isotropic model as an anisotropic model proved to be unstable. Similarly, some of the -CF3 groups which display large displacement ellipsoids have been modelled using restraints -RIGU in ShelXL -as disorder models were not stable. It was not possible to optimize the ShelXL weighting scheme, thus parameters which depend on this (wR2, GooF) could not be optimized. There is a single molecule in the asymmetric unit, which is represented by the reported sum formula. In other words: Z is 2 and Z' is 0.5.

Figure S17
The crystal packing of DCB⊂2 as viewed along the c-axis. The BArFions are colored grey and the protons have been removed for clarity. Color code: C: green (host) or orange (guest), N: light blue, Pd: blue, Cl: light green.

Figure S18
The crystal packing of DCN⊂2 as viewed along the a-axis. The BArFions are colored grey and the protons have been removed for clarity. Color code: C: green (host) or orange (guest), N: light blue, Pd: blue, Cl: light green.

Figure S19
The crystal packing of DCA⊂2 as viewed along the a-axis. The BArFions are colored grey and the protons have been removed for clarity. Color code: C: green (host) or orange (guest), N: light blue, Pd: blue, Cl: light green.

Figure S20
The crystal packing of TCDCB⊂2 as viewed along the a-axis. The BArFions are colored grey and the protons have been removed for clarity. Color code: C: green (host) or orange (guest), N: light blue, Pd: blue, Cl: light green.

S6. Raman Spectroscopy
Raman spectra were acquired on a Renishaw InVia Raman microscope equipped a 785 nm diode laser providing a maximum power of 300 mW using a 1200 l/mm grating. According to the Manufacturer's specification, the InVia Raman microscope has a spectral resolution of 0.5 cm -1 and high spectral stability to monitor minute shifts in Raman band position (as low as 0.02 cm -1 ). For the solid-state studies, crystals of 2 and all host-guest complexes were collected and dried, before a small amount of solid was transferred onto a CaF2 window and Raman spectra were acquired using λex = 785 nm and a 5× NA 0.12 NPlanEPI objective (Leica), a 20× NA 0.4 NPlanEPI objective or a 50× NA 0.75 NPlanEPI objective (Leica). For the solution-state studies, solutions were prepared in dichloromethane (DCM) in a quartz cuvette (up to 500 μL) in a similar manner to that as described in Section S4.2, and Raman spectra were acquired using a 20× NA 0.4 NPlanEPI objective (Leica) using a 10 s acquisition time.

Figure S21
Raman spectrum of L in solid-state. Raman spectra were acquired using 785 nm excitation for 10 s using a 50× objective lens (0.18 mW). All assignments are in cm -1 .

Figure S22
Solid-state analysis of dicyanonaphthalene (DCN) encapsulation using Raman spectroscopy. Raman spectra were acquired from the unbound guest (DCN; black) and cage (2; red) and the host-guest complex (DCN⊂2; blue). Raman spectra were acquired using 785 nm excitation for 10 s using a 50× objective lens (0.18 mW). Peak annotations are in cm -1 .

Figure S23
Solution-state analysis of dicyanonaphthalene (DCN) encapsulation using Raman spectroscopy. Raman spectra were acquired from the titration of 2 (0.50 mM) with DCN (25 mM) in dichloromethane (DCM). Raman spectra were acquired using 785 nm excitation for 20 s with a 5× objective lens (~180 mW). Peak annotations are in cm -1 . ; 2213.2 a The peaks reported here are the most intense at that region. In some cases, the ν(C≡N) may contribute to the peak shape and spectral position.

Figure S24
Raman spectral analysis of the free guests. Raman spectra were acquired in solid-state using 785 nm excitation for 10 s with a 50× objective lens (~0.18 mW). DCB (dicyanobenzene, bottom, black line), DCN (dicyanonaphthalene, middle, red line), TCDCB (tetrachlorodicyanobenzene, top, blue line). A complete Raman spectrum of DCA (dicyanoanthracene) could not be acquired due to the fluorescent nature of this guest at 532, 633 and 785 nm. A partial Raman spectrum indicated detection of the nitrile band at 2220.8 cm -1 (see Figure S25).

Figure S25
Analysis of guest encapsulation in the solid-state using Raman spectroscopy. Raman spectra were acquired from the free guest (black), the free lantern, 2 (red), and the host-guest complex (blue). Raman spectra were acquired using 785 nm excitation for 10 s using a 50× objective lens (0.18 mW). Raman spectra are presented in the range 100-2500 cm -1 (left), expanded view from 1250-1750cm -1 (center) and expanded view 2180-2280 cm -1 (right). Analysis of (a) dicyanobenzene (DCB), (b) tetrachlorodicyanobenzene (TCDCB), and (c) dicyanoanthracene (DCA). The Raman spectrum for DCA (free) is presented between 2100-2500 cm -1 only due to the broad fluorescent background associated with this compound. In addition, arrowheads are added to indicate Raman peaks likely to arise from the guest molecule. All assignments are color coded to the relevant spectrum and are presented in cm -1 .

Figure S26
Repeat analysis of guest encapsulation in the solid-state using Raman spectroscopy. Raman spectra were acquired the free lantern, 2 (red), and the host-guest complex (blue). Raman spectra were acquired using 785 nm excitation for 10 s using a 50× objective lens (0.18 mW). Raman spectra are presented in the range 100-2500 cm -1 (left), expanded view from 1250-1750cm -1 (center) and expanded view 2180-2280 cm -1 (right). Analysis of (a) dicyanonaphthalene (DCN), (b) tetrachlorodicyanobenzene (TCDCB), and (c) dicyanoanthracene (DCA). All assignments are color coded to the relevant spectrum and are presented in cm -1 .