Mass Spectrometric Characterization of Supramolecules
Often, the mass spectrometric characterization of supramolecules causes problems with respect to an intact ionization. Supramolecular complexes are held together by weak intermolecular forces which may change significantly upon transferring the complexes into the gas phase. Also, many common ionization techniques are not soft enough to prevent dissociation of the complexes under study. Consequently, electrospray ionization is one of the most useful and one of the softest methods available today.
Another difficulty is the often limited mass range of most mass spectrometric analyzers. Unlike many biomolecules, it is usually difficult to generate multiply charged complexes of artificial supramolecular species in the ion source. Nevertheless, the complexes of interest easily reach quite high masses so that a large enough mass range is required.
For the work done in the Schalley group, electrospray ionization (ESI) coupled to a Fourier-transform ion-cyclotron-resonance (FTICR) analyzer is the best way to analyze supramolecules in the gas phase. Besides high mass accuracy and excellent resolution, the FTICR technique enables us to perform gas-phase chemistry experiments by utilizing the tandem MS and MSn capabilities, double resonance or exchange experiments. In the following, a short overview of our research interests in the field of mass spectrometry of supramolecules is given.
1. Analytical Characterization: Organometallic Gelators
Mass spectrometry first offers analytical data about exact masses, isotope patterns, elemental composition, stoichiometries or impurities in the sample. Just as an example let's take a look at Fischer carbene complexes which have been investigated in collaboration with the group of Karl Heinz Dötz at the University of Bonn. They are useful compounds for organic synthesis. Besides this, some of them are known as gelators for organic solvents like benzene or chloroform. Chromium and tungsten carbene complexes 1 and 2 can be easily ionized in the ESI ion source by deprotonation. Figure 1 shows the ESI-MS spectrum of 1 in the negative ion mode. Large clusters up to a triply charged icosamer can be detected.
Figure 1: Negative ESI mass spectrum of 1 showing the formation of large clusters up to the triply charged 20-mer. For clarity, the labels in the inset provide the number of monomers incorporated in the ions together with their charge state as a superscript.
2. Structure Elucidation: Hexameric Pyrogallarene Capsules
Pyrogallerenes and resorcinarenes like 1 - 3 (Figure 2) form hexameric capsules in solution and in the solid state. Their existence in the gas phase was not shown so far. A particular difficulty is that polar molecules tend to unspecifically dimerize under the conditions of the electrospray ionization process. Consequently, it is necessary to provide some structural evidence for the hexamer to have a capsular structure. In our experiments, larger cations such as 7 were used to template capsule formation (Figure 3).
Figure 2: Hosts 1 – 4 and cationic guests 5+ – 72+ (left).
Figure 3: Model of the pyrogallarene hexamer (ball-and-stick representation, each monomer colored differently) and 72+ as the guest (spacefilling representation). Alkyl chains at the bottom rim have been simplified as methyl groups.
Figure 4a shows the ESI mass spectrum of 1. In this spectrum only peaks for the monomer and for smaller aggregates in low intensities can be detected. After addition of 5BF4 to the sample solution the peak for dimeric 1 including one cation 5 becomes the major peak, but larger aggregates could not be found (Figure 4b). By using 63[Fe(CN)6] as a template the hexameric signal is observable as one of the major signals (Figure 4c). But still, additional peaks for smaller and larger clusters are visible. This leads to the conclusion, that 6+ effects a to light template effect. Consequently, 7(PF6)2 should be the appropriate template to form the hexameric capsule due to his pseudo octahedral shape. Indeed, the mass spectra of either pyrogallarene 1 or resorcinarene 2 (Figures 4d/e) show the hexameric clusters as the by-far dominating signals after addition of this cation to the sample solution. Additionally, some smaller signals resulting from unspecific ESI aggregation or fragmentation are visible.
Figure 4: ESI-FTICR mass spectra of (from top to bottom): a) a 200 µM solution of 1 in CHCl3/acetone (2:1), b) after addition of 5 BF4 , or c) 63 [Fe(CN)6]. d, e) ESI-FTICR mass spectra of the same solution of 1 and 2 , respectively, with 7 (PF6)2, each optimized for hexamer intensity. f) Control experiment with tetramethylresorcinarene 4. Inset: Experimental and calculated isotope patterns of the pyrogallarene hexamer ion.
In order to further confirm the structure of the hexamers, tandem-MS experiments with mass selected hexamer/guest ions were performed. These ions were irradiated in an IRMPD (infrared multiphoton dissociation) experiment with an IR laser for different irradiation times to induce fragmentations. The spectra series presented in Figure 5 shows us that the signal for the free guest cation appears not before at least three pyrogallarene monomers have dissociated from the hexamer. This behaviour is not a definite proof for the capsule structure, but it strongly suggests the proposed structure.
Figure 5: IRMPD experiment with mass-selected [7 @16]2+: Longer irradiation times lead to consecutive monomer losses. The formation of free 72+ starts to compete with the loss of additional monomers from the trimer.
3. Reactivity in Solution: Dimeric Helicates
Dimeric helicates shown in Figure 6 are formed through a hierarchical self assembly process in solution. In the first step three catecholate ligands coordinate to one Ti(IV) cation to the monomeric helicate. In the second step two of such monomers aggregate to the dimeric complex Li3[L6Ti2] bridged by three lithiumions coordinating to the carbonyl groups and the phenolate oxygen atoms of the catechol ligands.
Figure 6: Hierarchical assembly of a triply lithium-bridged helicate-type complex with R = n-alkyl.
The negative ion ESI mass spectra of these complexes show only one signal for the monoanion Li2[L6Ti2]- generated by loosing one lithiumion from the dimer and Li[L3Ti]- formed by dissociation of the dimer. Besides analytical characterization, mass spectrometry gives us the possibility to analyze solution-phase processes kinetically. After dissolving the crystalline Li2[L6Ti2]- dimer, its slow dissociation can be easily monitored by ESI-MS. While directly after dissolving only a signal for the dimer appears, the signal for the dimer becomes smaller during a time period of two days, and after two days only a signal for the monomer is visible. To study the formation of heterodinuclear complexes, a mixture of two homodimers was sprayed in the ESI ion source. The spectra in Figure 7 show that due to the signal intensities a ligand exchange is only possible in monomeric complexes and not in the dimer. This exchange is much faster in the methanol/THF mixture then in pure THF effected by higher polarity of methanol. The exchange profiles at intermediate times thus provide insight into the exchange mechanisms operative in solution.
Figure 7: Exchange of ligands in solution as monitored by mass spectrometry. Direct exchange of individual ligands between dimers would cause an initial U-profile, which is not observed. Consequently, the dissociation of the dimer into two monomers competes with ligand exchange within these monomers. In the presence of a protic solvent, ligand exchange in monomers is faster than the monomer-dimer equilibrium (left). In the absence of a protic solvent, monomer exchange is faster than the exchange of ligands within these monomers (right).
4. Reactivity in the Gas Phase: Studying Mechanistic Details not Accessible in Solution
Examples for reactivity studies in the gas phase are given elsewhere: The analysis of a dendritic effect significantly altering the fragmentation mechanism of a tweezer-dendrimer host-guest complex and a mechanism involving the supramolecular equivalent of neighboring group assistance in the fragmentation of metallosupramolecular squares are only two of them. In both cases, the gas-phase experiment provide a completely new view on reactivity. There are no dynamic processes involved; isolated ions are studied instead. Consequently, no interfering environment is present.