Using Electron Energy Loss Spectroscopy to investigate the structure of the mirror coatings of gravitational wave detectors

Cummings, Rebecca Beatrice (2023) Using Electron Energy Loss Spectroscopy to investigate the structure of the mirror coatings of gravitational wave detectors. PhD thesis, University of Glasgow.

Full text available as:
[thumbnail of 2022CummingsPhD.pdf] PDF
Download (138MB)


Reducing the noise in gravitational wave (GW) detectors is of upmost importance for improving their sensitivity, which enables them to detect smaller GWs. There are many sources of noise in GW detectors, but the dominating source at their most sensitive frequency is the thermal noise of the coatings of their mirrors. While the relationship between the optical properties of the coatings and their manufacturing process is reasonably well understood, the link between these two things and their structure is not. Fully understanding their structure will improve the search for new coatings with lower noise contributions.

In order to study the coatings, scanning transmission electron microscopy was used, specifically dark-field imaging and electron energy-loss spectroscopy (EELS). Two different techniques were used, one on each of the two coating types (tantala and silica, which have a low and high refractive index, respectively).

The first technique was performing spatially resolved quantification using high angle annular dark field imaging and DualEELS (a form of electron energy loss spectroscopy (EELS) that takes two spectra in quick succession from the high- and low-loss region). It has been established for some time that there are small spherical voids, or bubbles, in the coatings and it has been hypothesised that they are contributing to laser damage in the coatings. It was also known that there was argon in the coatings, which is left over from the ion-beam assisted deposition process used in preparing the coatings. Semi-empirical standards were created for the quantification using argon data from the EELS Atlas and experimental data scaled using a Hartree Slater cross-section. The densities and pressures of the bubbles were then calculated. It was found that the argon coalesced into bubbles post-annealing in all samples, and that greater annealing temperatures caused the bubbles to coarsen. The bubbles’ interior pressures decreased with diameter. The technique was also applied to xenon and krypton bubbles in Zircaloy-4. While these are not GW mirror coatings, they corroborate the findings in the GW samples and show the wide-ranging applicability of the technique.

The second technique was investigating the extended electron energy-loss fine structure (EXELFS) of silica, to see if there are visible differences in the EXELFS of samples made using different techniques. While EXELFS has been around for many years, it is used much less commonly than its Xray absorption spectroscopy counterpart, EXAFS. This is because, depsite EXELFS’ many advantages over EXAFS, the signal to noise ratio in EXELFS has traditionally been much lower than in EXAFS. This chapter investigated whether recent advances in EELS data acquisition meant it would a viable analysis technique. It was determined that EXELFS could be used, and that different silica samples did show subtle differences in their EXELFS.

The two techniques used and developed in this thesis have applications far beyond GW mirror coatings and will be of use to anyone wishing to perform characterisation and quantification of thin films using EELS.

Item Type: Thesis (PhD)
Qualification Level: Doctoral
Subjects: Q Science > QC Physics
Colleges/Schools: College of Science and Engineering > School of Physics and Astronomy
Supervisor's Name: MacLaren, Dr. Ian and Martin, Dr. Iain
Date of Award: 2023
Depositing User: Theses Team
Unique ID: glathesis:2023-83489
Copyright: Copyright of this thesis is held by the author.
Date Deposited: 21 Mar 2023 10:11
Last Modified: 23 Mar 2023 09:57
Thesis DOI: 10.5525/gla.thesis.83489
Related URLs:

Actions (login required)

View Item View Item


Downloads per month over past year