The core problem is one of geometry. Fourier transform infrared spectroscopy, or FTIR, is the workhorse technique for detecting structurally bound hydroxyl in minerals. An infrared beam passes through a crystal, and the hydroxyl groups absorb at characteristic wavenumbers between roughly 3000 and 3800 centimetres to the power of minus one. In an isotropic mineral such as garnet, absorption is identical in every direction, so a single unpolarised measurement, multiplied by three, yields the total absorbance needed to calculate a water content via the Beer-Lambert law. But pyroxene is anisotropic. The hydroxyl dipoles are oriented in specific crystallographic directions, and the amount of light absorbed depends dramatically on how the crystal is oriented relative to the polarisation of the incident beam. Measuring along three mutually perpendicular directions gives the total absorbance, but in low-symmetry crystals, knowing which three directions to choose is anything but straightforward.

For orthorhombic orthopyroxene, the crystallographic axes and the optical indicatrix axes coincide, so randomly oriented grains can be oriented using cleavage, morphology, or optical extinction, and the three principal polarised spectra can be acquired along the a, b, and c axes. Monoclinic clinopyroxene is a different beast entirely. Here, the angle between the crystallographic a axis and the plane containing b and c deviates from ninety degrees, and, crucially, the infrared absorption indicatrices of different hydroxyl-related defects are not always aligned with the axes of the optical indicatrix. As Strens and colleagues established as early as 1982 for polarised visual light spectroscopy, the direction of principal absorption in monoclinic and triclinic crystals does not necessarily coincide with either the crystallographic axes or the principal refractive index directions. Three measurements along the optical indicatrix are therefore insufficient to fully describe the absorption behaviour, even for a single defect type. Four independent directions are needed for monoclinic crystals, and six for triclinic ones.

The first innovation from the Bern group is an automatic polariser, built from an open-source, customisable, low-cost, 3D-printed design. Attached to a Bruker Hyperion microscope, the device uses a stepper motor controlled by an Arduino microcontroller to rotate an infrared polariser in equal-angle increments between 0 and 180 degrees. A limit switch enables automatic incrementation each time the measurement stage returns to its background position, so a full angular polarisation series, comprising 36 to 90 spectra per measurement point, can be acquired without any manual intervention. This eliminates human bias in setting the polarisation angle and allows large numbers of spots to be measured with dramatically less effort. The design, including wiring diagrams, 3D models, and source code, is freely available on GitHub, meaning any laboratory with modest resources can replicate the set-up.

Complementing the hardware is a new software solution called PolPlotModule, an add-on for the open-source spectroscopy processing package SpecXY. The module divides angle-resolved polarised measurements into individual sets of spectra for each measurement position and visualises them as polarisation plots. A MultiSpectraDeconvolution routine performs a Simplex inversion, minimising the global peak deconvolution error across the full set of angle-resolved spectra simultaneously. A two-step Spectral Endmember Estimation routine, first using the N-FINDR algorithm and then a non-negative linear least-squares fit, extracts the pure spectral endmember signatures from large data sets. Together, these tools allow researchers to identify the maximum and minimum absorption directions for each absorption band, which is the prerequisite for accurate water quantification in low-symmetry crystals.

The key insight enabling the new workflow is the use of silicon-oxygen overtone bands in the wavenumber range between 1250 and 2300 centimetres to the power of minus one as orientation fingerprints. These overtone bands, which arise from the fundamental silicon-oxygen stretching vibrations, are present in every pyroxene spectrum regardless of hydrogen content, and their polarisation behaviour is characteristic of specific crystal directions. From a large data set covering diopside, augite, and omphacite, the researchers recognised five distinct spectral endmembers aligned with the optical indicatrix axes X, Y, and Z, the crystallographic b and c axes, and the orthogonal direction a-star, defined as the normal vector to the b-c plane. Once these overtone endmembers are identified, groups of mutually perpendicular spectra can be found, reducing the uncertainty in the total integrated absorbance of the hydroxyl region and thus improving water content determination.

The practical consequences are substantial. For diopside and augite, crystal cuts along the axes of the optical indicatrix are best suited to determining the total integrated absorbance, whereas for omphacite, cuts along the crystallographic b and c axes and the a-star direction give the most reproducible results. The average deviation between the total integrated absorbance calculated from different orthogonal endmember combinations is smaller than 7 percent across all investigated samples, with a maximum observed deviation below 16 percent. The researchers also demonstrated that silicon-oxygen overtones can evaluate sample homogeneity: spectra with identical overtone signals are measured in crystal sections with the same orientation and polarisation direction, so hydroxyl absorption becomes directly comparable between grains. In an inhomogeneous orogenic eclogite from the western Gneiss region of Norway, this approach revealed variations in water content across different grains that would be invisible using unpolarised spectra, where it remains unclear whether absorbance variation reflects orientation or genuine differences in hydrogen content.

Beyond quantification, the angular polarisation series provides a powerful tool for understanding how hydrogen is actually incorporated into the crystal structure. In the gem-quality diopside from Binntal, Switzerland, at least seven overlapping absorption bands were identifiable, and the polarisation behaviour allowed the researchers to distinguish two families of defects corresponding to different incorporation mechanisms: hydrogen on the tetrahedral site, absorbing between 3660 and 3600 centimetres to the power of minus one, and hydrogen on the octahedral M site, absorbing between 3590 and 3300 centimetres to the power of minus one. The observed polarisation behaviour agrees remarkably well with Density Functional Theory modelling by Balan and colleagues, which predicts different pleochroism for defects involving different trivalent cations such as aluminium, iron, and chromium. In the magmatic augite from Mount Guntur in western Java, the FTIR mappings revealed that water zoning is preserved at magmatic temperatures and is strongly coupled to chemical zonation, with higher-wavenumber bands anticorrelating with iron and correlating with chromium and calcium, while M-site-related defects show similarities with aluminium, titanium, silicon, and magnesium.

The team also proposed a new nomenclature for polarised FTIR measurements in low-symmetry crystals, stated as a vector subscripted by a plane. The vector describes the polarisation direction of the incident infrared beam, set by rotating the polariser, while the plane is the crystal section orthogonal to that vector. This flexible scheme, which can combine Miller indices, optical indicatrix axes, crystallographic axes, or any other defined vectors such as a-star, makes different samples directly comparable in terms of absorption behaviour and anisotropy, something previous nomenclature constrained to the optical indicatrix could not achieve, particularly since the angle between the crystallographic c axis and the optical axes X and Z depends on major element composition and can vary from 4 to 94 degrees across clinopyroxene compositions.

The implications reach far beyond the laboratory. Water in nominally anhydrous minerals governs the viscosity and melting behaviour of the mantle, influences electrical conductivity, and controls the transport of hydrogen into the deep Earth through subduction. Accurate quantification of this hidden hydrogen has been hampered for decades by the anisotropy problem, particularly in monoclinic minerals like clinopyroxene, which is one of the most abundant phases in eclogites and mantle peridotites. By automating the acquisition of angle-resolved polarised spectra, providing open-source tools for their processing and visualisation, and establishing robust protocols for identifying the spectral endmembers needed for quantification, the Bern group has given the community a framework to measure mantle water with unprecedented precision. The work provides a foundation for investigating hydrogen incorporation mechanisms in clinopyroxene and other monoclinic or triclinic crystals, and it demonstrates that acquiring more than three independent polarised measurements is essential to fully describe absorption anisotropy in low-symmetry crystals. As the hidden water inventory of the deep Earth comes into sharper focus, these new methods promise to make the invisible measurable.

Subject of Research: New methods for Fourier transform infrared spectroscopy investigation and quantification of OH absorption in anisotropic nominally anhydrous minerals: the example of pyroxene

Article Title: New methods for Fourier transform infrared spectroscopy investigation and quantification of OH absorption in anisotropic nominally anhydrous minerals: the example of pyroxene

Article References: New methods for Fourier transform infrared spectroscopy investigation and quantification of OH absorption in anisotropic nominally anhydrous minerals: the example of pyroxene. (n.d.). https://doi.org/10.5194/ejm-38-583-2026

Image Credits: AI Generated

DOI: 10.5194/ejm-38-583-2026

Keywords: methods, Fourier, transform, infrared, spectroscopy, investigation, quantification, absorption, anisotropic, nominally, anhydrous, minerals