The study of short lived species by electron spin resonance

• Main Author: Simpson, A. F.
• Published: University of Oxford 1970
• Subjects: 538
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.644645

The work described in this thesis consists primarily of the development of an Electron Spin Resonance (E.S.R.) spectrometer, having a very rapid response time and therefore capable of observing E.S.R. signals from free radicals within microseconds of their formation. The development has proceeded in three phases corresponding to response times of 1 millisecond, 50 microseconds and 1 microsecond. The special apparatus required for each phase is described and the results obtained with each on a trial system, benzophenone, are discussed. The most important requirement of a rapid response spectrometer is a wide bandwidth in all the signal handling circuitry. It is shown how this bandwidth has been achieved in practice and how such an extension of bandwidth leads to a loss in spectrometer sensitivity. The use of signal averaging to recover the lost sensitivity is described and it is shown how by this means it has proved possible to obtain usable E.S.R. signals within 1 microsecond of the radicals formation. It is seen that broadening of the magnetic energy levels in accordance with the uncertainty principle makes any attempt to pursue E.S.R. detection to times appreciably shorter than 1 microsecond pointless. The final spectrometer is described in detail and. consists of a conventional spectrometer with the normal 100 kHz magnetic field modulation replaced by a 2 MHz modulation and detection system specially designed to achieve a wide bandwidth. A highly stable magnetic field is required if efficient signal averaging is to be achieved over long periods. This is ensured by the use of a Hall-effect probe to monitor and stabilise the field. Radicals are produced by a pulsed ultra violet laser, and the information contained in the resulting E.S.R. signal is stored and averaged digitally. Three ways of operating the apparatus are described, the first providing the maximum possible information from the system but involving a considerable amount of manual data processing if a spectrum is required. The second and third methods both provide a spectrum directly and more rapidly but at the expense of all the kinetic information and a possible loss in sensitivity. The three methods largely complement each other and are all in use. For all the development work benzophenone has been chosen as the molecule for investigation. Spectra are obtained for the ketyl radical anion produced from the photolysis of this molecule in alkaline aqueous isopropanol and the ketyl radical produced from its photolysis in liquid paraffin. The latter is seen to display a previously unsuspected population inversion in the magnetic spin states when first formed, the result of which is that during the first few tens of microseconds of the radical's life it displays a totally emitting E.S.R. spectrum. Previously the only emitting spectra to be seen in the liquid phase displayed emission in only some of their hyperfine lines, an effect explicable in terms of a polarisation of the nuclear spin energy levels. Totally emitting spectra require a polarisation of the electron spin energy levels when the radical is first observed. The initially emitting signal decays to a conventional absorbing signal with. a characteristic relaxation time. This in turn decays to zero in a normally much longer tine as the radicals themselves react and are lost. Using the apparatus in the first mode of operation provides a complex decay curve which yields information on the initial spin polarisation, the relaxation time and the kinetics of the radical loss. Such curves are illustrated and a computer program is described which simulates the complex decay curves and thereby aids their analysis. Spectra are obtained for both emitting and absorbing radicals and a second program is described for simulating such spectra and thereby confirming their analysis and hence yielding hyperfine splitting constants for the radical. A possible qualitative explanation to explain the emission phenomenon is proposed and it is shown how this corresponds to the recently observed phenomenon Chemically Induced Dynamic Nuclear Polarisation (CIDNP), in which partially emitting N.M.R. spectra are observed from the products of free radical reactions. By analogy the name CIDEP, Chemically Induced Dynamic Electron Polarisation, is given to the emission effect observed here. The mechanism supposes that the radicals are initially formed in pairs of which some recombine and others escape. The relative ease of escape and recombination is dependent upon the spin state of the radical pair and thus both the recombination products and the escaped radical exhibit a polarisation of the energy level populations. In the latter this drives rise directly to the phenomenon of CIDEP whilst in the former the non-equilibrium electron population 'pumps' the nuclear spin states via an Overhauser effect thereby producing the relatively long-lived nuclear polarisation observed as CIDNP. Finally the future development and use of the apparatus is considered and various possible ways of overcoming the data handling problems of the first mode of operation are discussed.