Caracterização metrológica do sistema de medição de deslocamento por interferometria de um espectrômetro FTIR

Abstract

Displacement measuring interferometry systems are widely used in many fields of science and can accurately provide information about several different quantities of interest. Because they are composed by optical elements and mechanical devices, the interferometers are affected by several types of errors. In analytical spectroscopy, the use of infrared spectrometers aided by interferometers brings considerable improvements and offer many advantages in comparison to the dispersive spectrometers, for qualitative and quantitative characterization of chemical susbstances. However, this better quality instrumentation requires mathematical and statistical models that represent not only the physical laws, but also the measurement process, including all those quantities relevant to the determination of the uncertainty of the measurement result. This work presents a methodology employed to develop the measurement model and to evaluate the measurement result for the Michelson interferometer. The device used as the reference system of the Fourier transform infrared spectrometer taking into account the variations of the laser wavelength, refractive index of the air, optical path errors, misalignment of mirrors and electronic systems. The displacement measuring interferometry system or DMI is responsible for the correct relationship between the infrared radiation intensity and the optical displacement or retardation. The value of the retardation is measured as a function of the phase-shiffting of the signals generated by two optical sensors positioned in quadrature and aligned towards the reference laser beam in an arrangement called QPSI (Quadrature Phase Shift Interferometer) The measurement result are figured-out with a computer program that implements the Monte Carlo method. The MCM is based on computer simulations, iterative algorithms and the model evaluation by drawing the most likely values that can affect a quantity of interest and evaluated by the measurement model proposed. The MCM also allow to find-out directly the probability distribution of the measurand when a large number of iterations is performed. This method can be applied to linear models as well as to nonlinear models, such as the one proposed in this dissertation.