Kontrolle der spektralen Phase von Femtosekunden-Laserpulsen zur qualitativen und quantitativen Unterscheidung von Strukturisomeren

Im Rahmen dieser Doktorarbeit wurde das Ionisations- und Fragmentierungsverhalten organischer Moleküle mit Hilfe von Femtosekunden-Laserionisations- Massenspektrometrie erforscht. In zwei unterschiedlichen Projekten wurde dabei der Einfluss zweier Laserpulsparameter auf die Ionisation und Fragmenti...

Full description

Saved in:
Bibliographic Details
Main Author: Schäfer, Viola Carmen Maria
Contributors: Weitzel, Karl-Michael (Prof. Dr.) (Thesis advisor)
Format: Doctoral Thesis
Language:German
Published: Philipps-Universität Marburg 2019
Subjects:
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!

Within the scope of this thesis, the ionization and fragmentation of organic molecules were investigated by means of femtosecond-laser ionization mass spectrometry. The influence of two laser pulse parameters on the ionization and fragmentation of the analytes was explored in two projects. On the one hand, the spectral phase of the fs-laser pulses was controlled by systematically changing the linear or in part also the quadratic chirp parameter. On the other hand, a variation of the pulse energy was performed. In the first project of this thesis, the three primary alcohols methanol, ethanol and 1-propanol were investigated, while the second project dealt with the structural isomers of fluorotoluene, fluorobenzyl bromide and cineole. The use of high-intensity laser radiation led to a rich fragmentation pattern in the mass spectra for all investigated species in addition to the parent ion. In the following, the key findings of the two projects will be briefly summarized. The ionization and fragmentation of the three alcohols methanol, ethanol and 1-propanol were analyzed in fs-laser fields with imprinted linear chirp for three, or in the case of methanol for four different laser pulse energies. The parent ion yield Y(M+) as a function of the linear chirp parameter showed an intensity-driven behavior for all three alcohols, provided that pulse energies of 50 µJ or higher were applied. If the highest parent ion yield arises for the transformlimited pulse the dependence of the parent ion yield on the linear chirp parameter is called intensity-driven. In addition, the parent ion yield decreases as the absolute value of linear chirp parameter increases.[13] For the lower pulse energies of 10 µJ and 25 µJ, 1-propanol again showed an intensity-driven behavior, whereas for methanol and ethanol the maximum parent ion yield was shifted towards negative linear chirp parameters (a = -300 fs2). According to Reusch et al., this behavior is referred to as a sign-dependent chirp effect.[13] In addition to the observations made, the results were compared with the literature in which the alkanes methane, ethane and propane were investigated under similar conditions.[13,14] For different pulse energies the parent ion yield of the alcohols behaves similarly as a function of the linear chirp parameter. Such a behavior could not be observed for the alkanes. From the results obtained and the comparison with the literature, it was suggested that the hydroxy group dominates the electronic structure of methanol and ethanol. In addition, a signdependent chirp effect could be found only at much lower peak intensities. This sign-dependent chirp effect was far less pronounced than in comparison to the data of methane and ethane. Whether for 1-propanol, the alkyl chain or the hydroxy group dominates the ionization is less clear. Due to the fact that for all investigated pulse energies the maximum parent ion yield was found for a transform-limited laser pulse, it can be concluded that the longer alkyl chain gains increasing influence on the ionization. However, since in the literature the highest sign-dependent chirp effect of all investigated alkanes was found for the parent ion yield of propane (-875 fs2 at 50 µJ)[14], it is suggested that there are significant differences in the electronic structures of 1-propanol and propane. In turn, these differences must originate from the additional hydroxy group of 1-propanol. In addition to discussing the parent ion yield, different fragmentation channels of the three alcohols were analyzed as a function of the linear chirp parameter. An influence of the pulse length on the competition between C-C and C-O bond breakage was found in ethanol similar to the results of Itakura et al.[109]. The pulse length increases with an increase of the absolute value of the linear chirp parameter. This is reflected in an ion yield ratio (IYR), which rises or falls symmetrically with respect to the transform-limited pulse. Longer pulses resulted in an enhancement of C-O versus C-C bond breaking in ethanol at a constant pulse energy of 50 µJ. The discussion of C-O versus C-C bond breaking was only possible for ethanol. For 1-propanol, no signal could be clearly assigned to the C-O bond breaking channel due to the small resolution of the mass spectrometer used. Two competing bond breaking channels found in all three alcohols analyzed are the C-H and C-O bond breaking. Still, the resolution of the mass spectrometer does not allow unambiguous assignment of the signals evaluated to C-O bond breaking for ethanol and 1-propanol. However, the analogies found between the ion yield ratios Y([M - H]^+)/Y([M - OH]^+) as a function of the linear chirp parameters of the individual alcohols were interpreted as an indication that corresponding fragmentation channels are present. For the highest investigated pulse energy (50 µJ) for methanol, a qualitative trend of the IYR was found as a function of the linear chirp parameter, which was already found in ethanol at 25 µJ. The trend of the IYR of ethanol as a function of the linear chirp parameter at 50 µJ was then already detected for 1-propanol at 25 µJ. The length of the alkyl chain thus seems to shift the fragmentation behavior to other pulse energies. The second project, which was carried out in this thesis, dealt with the distinction of structural isomers. Here, a further development of the method was sought and achieved by extending the measurement range. In addition to fslaser pulses with linear chirp, also those with quadratic chirp were used. The comparison of different IYR as a function of the linear or quadratic chirp parameter of 1,4-cineole with that of 1,8-cineole allows a clear distinction between these two structural isomers. So far, only isomers of benzene derivatives have been distinguished with linear and quadratic chirped laser pulses, which is why the cineoles are a new class of compounds for these experiments. Therefore, the measurement range of fs-LIMS for distinguishing structural isomers was extended. Due to the rich fragmentation pattern, a variety of ion yield ratios can be used for discrimination. In this work, Y(m/z=139)/Y(M+), Y(m/z=43)/Y(M+) and Y(m/z=15)/Y(M+) were discussed for different pulse energies, all of which allow a distinction. For the IYR Y(m/z=139)/Y(M+) of the 1,4-cineole, significantly lower values were found for all investigated pulse energies than for the IYR of 1,8-cineole. This could be attributed to the stability of the resulting fragment ion. For example, 1,4-cineole can form a secondary alkyl radical upon formation of the ion with m/z=139, which is less stable than ternary alkyl radicals, which are exclusively present in the fragment ion of 1,8-cineole. The preparation of general trends regarding the qualitative isomer distinction is not yet possible. The behavior found depends on the species studied and therefore contains molecule-specific information. As an example, in the range of pulse energies used, no trend could be identified that allows a statement as to whether the identification of the isomers is easier for low or for high pulse energies. For cineole isomers, it was possible to determine the composition of binary mixtures in fs-LIMS experiments with imprinted linear chirp for the first time. For this purpose, a system of equations was formed based on the high data set, which was available due to the rich fragmentation pattern and the large number of chirp parameters. Six data points of the IYR Y(m/z=15)/Y(M+) at different linear chirp parameters were considered sufficient to quantify the molar fractions of four binary mixtures. In addition to these six equations, two side conditions were valid: the sum of the two molar fractions should be one and the individual molar fractions should be between zero and one. For three mixtures a difference of less than 2% between actual and experimentally determined molar fratcions, for one of the mixtures a difference of less than 6% was determined. Overall, it was an overdetermined system of equations. Further potential for improving the method is given by the rich fragmentation pattern and the resulting large number of data points. Three structural isomers (ortho, meta and para) of two benzene derivatives, namely fluorotoluene (FT) and fluorobenzyl bromide (FBB), were also qualitatively differentiated. These are benzene derivatives with two different substituents, which is an additional extension of the experimental scope, as previous studies with linear or quadratic chirp have investigated benzene derivatives with two identical substituents.[11,12] The use of fs-laser pulses with imprinted linear or quadratic chirp enhanced the small differences between the individual isomers, which become evident when considering different ion yield ratios. Another focus was on the elucidation of the fragmentation pathway and the rationalization of the results obtained. For the fluorobenzyl bromides, it was assumed that the bromine abstraction is the first step in the decomposition process. The cleavage of the bromine atom leads to a fragment ion with the mass-tocharge ratio of 109, as well as a hydrogen elimination in the fluorotoluenes. The comparison of the ion yield ratios Y(m/z=109)/Y(M+) of the benzene derivatives showed an analogy in the order of IYR of the individual isomers. From this result it was assumed that the hydrogen abstraction in the case of the fluorotoluenes is a first reaction step in a sequential decomposition pathway, as suggested, for example, by Safe et al.[114]. The order IYR(p) > IYR(o) > IYR(m) for both benzene derivatives can be explained by aspects of a thermodynamic controlled reaction. The resulting fragment ion from the p-isomer is the most stable, while the fragment ion from the m-isomer is the least stable. For the mand o-FT, the IYR aimed against a limit for high negative linear chirp parameters, as the hydrogen elimination competes with the intramolecular vibrational energy redistribution (IVR) at the parent molecular ion level. This information is a new aspect of the fragmentation pathway of fluorotoluenes, which has been extended due to this result. The H abstraction is followed by HF or C2H2 loss. The corresponding ion yield ratios (relative to the fragment ion with m/z=109) also allow a distinction of the three structural isomers of fluorotoluene. Also o-, m- and p-fluorobenzyl bromide can be differentiated by these ion yield ratios. For the isomers of the fluorobenzyl bromide, longer pulses enhance the decay of the fragment ion with m/z = 109 into the smaller fragment ions by HF or C2H2 loss, as indicated by the linear chirp dependence. The quantitative determination of the composition of binary and finally ternary mixtures of the isomers of fluorotoluene was also investigated in this project. As in the case of the cineoles, several of the numerous data points were empirically selected and converted into a system of equations. The quantitative determination of the molar fractions of binary fluorotoluene mixtures from the meta and para-isomers was achieved with only slight deviations from the weighed in amounts of substance. The absolute differences between the actual and the experimentally determined molar fractions was 1 %. In contrast, the absolute differences between the actual and experimentally determined molar fractions were about 10% in the analysis of three binary o- and p-FT mixtures. The molar frations of o-FT in the mixtures were always determined to be higher than that weighed. However, the experimental values of the IYR of the mixtures could be described better with the experimentally determined molar fractions. It was concluded that the composition of the liquid phase did not correspond to that in the gas phase. The vapor pressures of the pure fluorotoluene isomers support this theory, since the pure o-FT has a higher vapor pressure than the other two isomers (po > pm = pp). This offered the possibility of creating a vapor pressure diagram based on binary mixtures of o- and p-fluorotoluene, which makes the real partial vapor pressures of both isomers available in binary mixtures. The potential to create vapor pressure diagrams from a chirped fs-LIMS study was previously unknown and offers further promising analysis. Nevertheless, the dependence of the results on the vapor pressures made quantitative analysis of ternary fluorotoluene mixtures difficult. For the two ternary fluorotoluene mixtures investigated in this work, an experimental determination of the composition was made, whereby the system of equations was set up to optimize the result. Here, the absolute differences between actual and experimentally determined molar fractions were below 6 %. To the best of my knowledge, this is the first fs-LIMS study with variation of the linear and quadratic chirp parameter, which allows a quantitative determination of the composition in ternary isomeric mixtures.