Performing polarized reflectivity and absorption measurements in the experimental range 15--1500 ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}1}$, we have investigated the infrared activity of phonons in aluminum phosphate. We report both ${\ensuremath{\Gamma}}_{2}$ (E\ensuremath{\parallel}c) and ${\ensuremath{\Gamma}}_{3}$ (E\ensuremath{\perp}c) polarized components and, in order to identify all modes predicted by group theory arguments, we compare with similar measurements performed on \ensuremath{\alpha}-quartz. We have found the following. In polarization E\ensuremath{\parallel}c, we resolve from reflectivity spectra all but one infrared-active ${\ensuremath{\Gamma}}_{2}$ components. Since they are exclusively infrared active, they could not be found from any other technique. To investigate the missing mode, which corresponds to the folded acoustic branch, we performed a series of transmission measurements. In this way we could find a weak absorption structure, which appears at room temperature at 48 ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}1}$. In polarization E\ensuremath{\perp}c, we find 15 infrared-active ${\ensuremath{\Gamma}}_{3}$ components which compare satisfactorily with previously published Raman data. Two modes, at 112 and 126 ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}1}$, form a close doublet which resolves only when running absorption measurements at liquid-helium temperature. The high-energy component (126 ${\mathrm{cm}}^{\mathrm{\ensuremath{-}}1}$) comes from the folded acoustic branch and the second is the infrared counterpart in aluminum phosphate of the 128-${\mathrm{cm}}^{\mathrm{\ensuremath{-}}1}$ phonon of \ensuremath{\alpha}-quartz. This is clearly established from a comparison of both reflection and transmission experiments performed on the same sample. Finally, performing a series of oscillator fits, we get quantitative values for (i) the LO-TO splitting of the phonon modes under consideration and (ii) the corresponding oscillator strengths. The series of parameters obtained in this way compare satisfactorily with a simple model of infrared exclusive modes in \ensuremath{\beta}-quartz admixed twice: first, with silent or Raman exclusive components (this is the case in the \ensuremath{\alpha} variety) and, second, with zone-boundary modes (this appears in berlinite when doubling the length of the unit cell).
Read full abstract