CONICET | Buscador de Institutos y Recursos Humanos

Mössbauer Line Tracking (MLT) is a recently developed technique [1][2] designed to continuously locate a resonant absorption line of the spectrum while it shifts in energy as a consequence of some external parameter manipulation (e.g. temperature). In this kind of experiment a programmable-velocity scaler [3] is used to record a few spectral channels in the surroundings of the absorption line for a predetermined time interval. Based on data from this region of interest (ROI), an algorithm relocates the ROI and restarts the measure in a closed loop manner. The nature of the tracking algorithm, the number of channels in the ROI, its energy distribution and the recording time for each channel are all parameters related to the experimental design. The best choice for such parameters depends on the characteristics of the experiment. They can even be programmed to change in time or to be tied to the evolution of the externally manipulated parameter. In this work we find the theoretical distribution of points that best keeps track of the line center in a simplified case. This result can be considered as a first approximation to the optimal design. Experimental verification of the results is also presented. In order to find the optimal ROI constitution, several computational simulations were conducted for different cases. The first case to be considered was the problem to find the optimal number of channels and its spacing, taking a fixed time for each ROI recording (i.e. increasing the number of points decreases the time spent in each point collecting data). The optimal tracking implies the smallest statistical reconstruction error for a given measure time. Equally-spaced channels were considered, as well as Poisson statistical distribution and Lorentzian shaped absorption lines with known full idth at half maximum (FWHM). We found that the best tracking option consists of only two channels, separated in energy a value that ranges between FWHM and the maximum-slope points of the Lorentzian curve. As an example, Fig.1 presents relative standard deviation of the reconstruction error for 2x104 events/point as a funtion of spacing for the two-channel ROI. The study was extended for the cases where the line width is unknown and where the width changes with the external parameter. The effect of using not equally-spaced points is also considered. Finally, some experiments were conducted on FeSn2 in order to verify the results. FeSn2 is an antiferromagnet with a unique structural site for the iron probe, with no quadrupolar interaction [4]. Its Mössbauer spectrum consists of a unique sextet that collapses into a singlet at about 340 K, what makes it suitable for fast scan testing from RT. Fig. 2 presents the results for one of the experiments, where the sixth absorption line (2.5% effect) was successfully tracked by a two-channel ROI at 2 s/channel, while the sample temperature was increased from RT to 120 C at a rate of 100 C/hour. The tracking succeeds up to the magnetic collapse, where the algorithm catches the more intense paramagnetic line. The whole experiment required less than an hour. Off-line reconstruction was conducted later in order to find the line position dependance with temperature. [1] A. Veiga et al., Hyperfine Interact. 118 (2009) 137. [2] P. Mendoza Zélis et al., Hyperfine Interact. (in press). [3] A. Veiga et al., Hyperfine Interact. 167 (2006) 905. [4] P. Mendoza Zélis et al., Phys. Lett. A 298 (2002) 55

enviar mensaje