Engineering of bulk properties of a material and/or study of its biological function require characterisation techniques that are capable of supplying local atomic-level structural and dynamic information. For this, the power of NMR derives from the resolution of individual resonance lines by which nuclei in different electron environments are distinguished, and modulation of signal by a distance between the nuclei. For a better resolution, anisotropy of nuclear interactions has to be suppressed. Various spin space and sample reorientation methods are used. Of latter, fast single axis and double axes mechanical rotations are chosen, depending on complicacy of the problem. Both are required to comply with coupling of rf magnetic fields to the sample. Migration of NMR to higher frequencies, combining not only various nuclear spin irradiation channels, but also electron spins, presents a serious challenge to hardware design. Accuracy of lumped element approach falls below threshold of usefulness and new, first principles design methods are required. With advances in computing hardware, in particular massive parallelization, induced by gaming industry, has created whole new potential. A full 3D electromagnetic wave propagation modeling is becoming feasible for practical structures. Despite very complicated code writing and limitations of graphics oriented hardware, numerous pilot programs have been demonstrated to bring orders of magnitude acceleration. The present project is set up to implement a suitable hardware-software platform and apply it on selected development projects of high relevancy in current analytical spectroscopy. A fast stepper-motor controlled angle modulation probe has been already tested. In response to critical points by reviewers of a basically similar application in 2009 and 2010, we have stressed impact of the research, extended duration to four years and increased staff and support by other projects, including FP7 and institutional base financing. In the technical description, current work, relevant to the application, has also been dicsussed.
Present project aims to establish foundation for computer modelling of seemingly different NMR problems, governed principally by the same physics. We expect to extend 3D EM modeling from very simple single element structures (coils and resonators) to more integrated multi-frequency constructions by massively parallel Finite-Difference Time-Domain (FDTD) or Finite Integration Technique (FIT) numerical simulations. Simulation is governed by Maxwell equations, computed on the alternating Yee cells for E and H components respectively. Project starts with the evaluation of 3D and hybrid modeling packages, offered by Remcom, CST(available), Quickwave, SEMCAD (available), HFSS and other companies. Recent development by Acceleware has brought a spectacular and inexpensive hardware acceleration, capitalizing on general purpose graphics processing units. NVIDIA has also released new GPGPU generation (“Fermi-Tesla”) with full double precision arithmetics and error correction support. This enables reliable time domain broadband description of the EM field propagation in non-trivial structures. Software assessment is followed by the modelling and experimental comparison with prototype circuits, and construction of prototype NMR probes. Analyses of the NMR signal precession, combined with the detection angle modulation is expected to provide amendments to protocols for NMR measurement. The research is divided in several comprehensive activities: dynamic coil for DOR NMR, combination and optimization of multiple RF resonances and novel composite EPR-NMR resonator. As a result, up to one order of magnitude spectrometer time saving is expected for multichannel esperiments on spin 1/2 nuclei, significantly more for quadrupole nuclei and DNP.
Projekti käigus uurisime põhjalikumalt TMR jaoks sobivaid multiresonantseid võnkeringe ja nende efeektiivsuse mõõtmist ilma magnetväljata. Täpsem kirjeldus on inglise keelses osas ja lisafailis.One method to assess the B1 inside the NMR coil is based on cavity perturbation theory, which states that inserting a conducting or dielectric object into electromagnetic field of a resonant cavity, the electromagnetic field will be perturbed. This causes the resonant frequency to shift accordingly. Same principle can be applied to the NMR coil. For perfectly conducting sphere the connection between change in resonant frequency and electromagnetic field can be approximated analytically as a function of the ball radius. It is not possible to measure measure electric or magnetic field separately using the ball shift method, but only the combination of the two. When using a balanced probe design, the B1 is at it’s highest in the centre of the coil, and E1 is at the lowest. In a case where E1 << B1, ball shift method can be applied to estimate the B1 value inside the NMR coil. If that is not the case, the frequency shift of a probe circuit can be compared to frequency shift of a minimal resonant circuit. The minimal circuit has to have similar Q-factor, exact replica of the NMR coil and it has to be tuned at the same frequency. Electromagnetic fields of a minimal circuit can be easily simulated using software such as CST. From results of the ball shift test of actual minimal circuit and simulated minimal circuit, the sensitivity (B1 value vs input current) of multiresonant circuit can be approximated without magnetic field tests. This result allows a preparation of specific probes for measurements at remote locations and optimal magnetic fields, as evidenced by publications. Better probes help to improve quality of analytical measurements at institutions where NMR is available, in Estonia NICPB, TUT, UT. Future extension of the work would be 3D EM simulation of the whole probehead.