Ion Mobility Mass Spectrometry Uncovers Guest‐Induced Distortions in a Supramolecular Organometallic Metallosquare

Abstract The encapsulation of the tetracationic palladium metallosquare with four pyrene‐bis‐imidazolylidene ligands [1]4+ with a series of organic molecules was studied by Electrospray ionization Travelling Wave Ion‐Mobility Mass Spectrometry (ESI TWIM‐MS). The method allowed to determine the Collision Cross Sections (CCSs), which were used to assess the size changes experienced by the host upon encapsulation of the guest molecules. When fullerenes were used as guests, the host is expanded ΔCCS 13 Å2 and 23 Å2, for C60 or C70, respectively. The metallorectangle [1]4+ was also used for the encapsulation of a series of polycyclic aromatic hydrocarbons (PAHs) and naphthalenetetracarboxylic diimide (NTCDI), to form complexes of formula [(NTCDI)2(PAH)@1]4+. For these host:guest adducts, the ESI IM‐MS studies revealed that [1]4+ is expanded by 47–49 Å2.. The energy‐minimized structures of [1]4+, [C60@1]4+, [C70@1]4+, [(NTCDI)2(corannulene)@1]4+ in the gas phase were obtained by DFT calculations.Introduction


S4
Scheme S1. Schematic view of the Synapt XS High Definition Mass Spectrometer. Reprinted from the waters.com webpage with permission from the Waters Corporation.

Single-stage ESI-MS and CID experiments.
A capillary voltage was set to 1.5 kV operated in the positive ionization mode and in the resolution mode. Source settings were adjusted to keep intact the supramolecular adducts of interest. Typical values were cone voltage 20 to 40 V and source offset 4 V; source and desolvation temperatures were set to 110 and 350 ºC, respectively. Cone and desolvation gas flows were 150 and 500 (L/h), respectively. Sample solutions were prepared from stock acetonitrile 1mM solutions by 1000-fold dilution with acetonitrile to reach the 1M concentration and introduced directly to the ESI chamber through an external syringe pump at a flow rate of 5 μL·min -1 . Samples were investigated over several days in the 50 to 2500 m/z range and fragmentation of the roboust [1] 4+ skeleton was not observed at all. Calibration of the m/z axis up to m/z 3000 was performed using the routine implemented in intellistart from a mixture of sodium and cesium iodide (2 mgmL -1 in 1:1 v/v H2O:isopropanol). Comparison of experimental vs theoretical isotopic pattern was carried out S5 using Masslynx 4.2 (SCN 982). CID experiments were performed by mass selecting the supramolecular [NTCDI)2(PAH)@1] 4+ and the fullerene [fullerene)@1] 4+ ions of interest in the first quadrupole and increasing the collision voltage (V) in the trap region starting from 2 V and stepped by 3 V up to a maximum of 20 V. An isolation width of approximately 2 Da was selected (LM resolution set to 8). Collision energies were converted to the center-of-mass frame, ECM =q*m/(m+M)Elab, where q stands for the charge state and m and M stand for the masses of the collision gas and the ionic species, respectively. For the breakdown profiles representations, signal intensities were obtained from the average of 20 scans and measuring the area of the fragmentation peaks. These graphs were represented taking into account the relative abundance of the precursor and product peaks of each compound (Iprecursor ion or Iproduct ion / Iprecursor ion + I product ion) against ECM.

ESI TWIM-MS.
The same sample solutions and source settings to that described above for single-stage ESI-MS were used. The instrument was switched from TOF acquisition to mobility TOF acquisition mode and left for 30 minutes before recording IM mass spectra. The m/z 50-1500 range was investigated and ion mobility separation settings were used as follows: the traveling wave height was set to 40 V and wave velocity was set to 650 m/s. The drift gas was nitrogen (N2) at a flow rate set to 90 mL/min. The helium cell gas flow was 180.00 mL/min. IMS DC values were as follow: Entrance 20; Helium cell DC 50; Helium exit -20; Bias 3; Exit 0. Trap DC values were controlled manually; entrance, 3; trap DC bias was varied from 25 to 45 V, the 35 V value being the compromise value to visualize the intact supramolecular adducts and maximize ion transmission; Exit 0. The adjustment of the trap bias potentials (the accelerating voltage between the trap and the He cell that precedes the IMS chamber), proved to be crucial to visualize the [(NTCDI)2(PAH)@1] 4+ adducts. For example, progressive reduction of the trap bias potentials from initial 45 V to 25 V (that is, softening the ion injection conditions) allowed the [(NTCDI)2(donor)@1] 4+ species of interest to be S6 visualized and characterized at expenses of a significant reduction of ion abundances. All the investigated compounds and calibrants in the present work were recorded under these conditions.
The IM-MS data were processed using Masslynx 4.2 (SCN 982). All ions of interest displayed a gaussian-shaped arrival time distribution profile. Ion mobility spectra of the species of interest were extracted using a 0.15 Da mass window and were converted from waters.raw to .txt files. Gaussian fitting of the IM data was applied to improve the precision of the drift time measurements. The reported drift times values were obtained by Gaussian peak fitting using origin 6.0 (Microcal) rendering good correlation in all cases. Each sample was recorded by triplicate on the same day and the deviation in the drift time values was less than 0.5 %.

CCS Calibration:
The CCS calibration protocol reported by Ruotolo was followed to convert drift times into CCS, [3] using a series of similarly charged peptide ions (Melittin, Angiotensin I and a tryptic digestion of bovine hemoglobin), which cover the transit time range of the ions of interest. Calibration of the IM-MS device for determining collision crosssectional areas from drift time measurements was performed considering a series of multiply charged (M + nH) n+ (n = 3 and 4) species and their DT CCSN2 values were taken from the literature (Angiotensin I, melittin [4] and tryptic digest of bovine Hemoglobine [5] ). As the TWIMS device is operated with N2 buffer gas, the obtained TW CCS values will be noted TW CCSN2. Drift times (tD) were subjected to correction for mass-dependent and massindependent flight times according to ´= − * √ / 1000 − 0,9 (C = 1.5 and the term 0.9 ms is the mass-independent time to account for the time of transit of one wave in the IMS and the transfer region). The literature CCS values were converted to CCS´ according to ´= √ z where  and z stands for the reduced mass of the collision partners and the charge state, respectively. The calibration curve is represented as CCS´ as a function of tD´ using a power law, [10]

NMR studies on host-guest adduct formation with PAHs.
The formation of the quaternary supramolecular [(NTCI)2(PAH)@1](BF4)4 (PAH = anthracene, phenanthrene, perylene and corannulene) was performed by adapting the previous synthetic protocol described for other PAHs. [2] Typically, to an NMR tube containing a 1mM CD3CN solution of host [1](BF4)4, two equivalents of the Polycyclic Aromatic Hydrocarbon and then two equivalents of NTCDI were added. The resulting suspensions were placed for 30 minutes S10 in the ultrasonic bath before recording the spectra. Representative NMR spectra are given in the Supplementary Information file.

Trajectory method (TM) CCS predictions.
To compare empirical CCS results to DFT-derived structures, IMoS was used to calculate the average drag caused by the impinging gas molecules over the flight path. The potentials employed in this case correspond to those of the standard TM methods using a 4-6-12 potential.
How these calculations are performed is available elsewhere. [6] 1.8 Computational Details.
All calculated complexes of this paper are diamagnetic and possess a +4 net charge. The gasphase geometry optimizations were carried out with Gaussian 16, rev. c.01, [7] using the MN15-L functional [8]