CALIBRATION SOFTWARE. H2020 Grant Agreement N° 761036 is commonly considered the most accurate one in the measurements of permittivity and loss factors (especially of low-loss materials). It is an industrial (and a formal IEC) standard. What restricts the application of DRs to materials containing nanoparticles or thin-film stacks such as solar cells, is their limited resolution. Each DR provides complex permittivity of the material averaged under the DR’s “head”. At a nominal frequency of 10 GHz the averaging area should be considered as ca. 16 mm in diameter (or at least 8 mm with ca. 90% accuracy). While theoretically the spatial resolution increases linearly with the increase of frequency, such an approach is impractical due to high complexity (and hence prohibitive costs) of manufacturing and calibrating the resonator. The family of SPDRs readily available at an acceptable price of a few kEUR terminates at 15 GHz with ca. 12 mm effective resolution. In Task 4.4 we have now proposed and evaluated the possibility to downscale the DR measurements without reducing the geometrical size of the DR’s “head” or increasing the measurement frequency. To begin with, we took advantage of the micro-scale insight into the electromagnetic quantities and phenomena within the DR, as facilitated by QuickWave EM software and exploited within WP3. We observed that the measured microwave signal is due to a geometrically weighted contribution of the material parameters, where the 2D weighting function is the squared electric field, further referred to as the DR’s template. In principle, the template is obtainable with any state-of-the-art EM simulator, but QuickWave features an exceptional variety of the relevant co- and post-processing functionalities. Our state-of-the-practice is as follows. Within WP3, we generate the DR’s template by ultra- fast BOR FDTD (Bodies of Revolution Finite-Difference Time-Domain) simulations implemented in the QW-V2D version of QuickWave. Further, in WP5 we have designed and manufactured a prototype of the 2D scanner, where a sample of up to 80 mm x 80 mm size is initially imaged in 1 mm steps (thought the applied commercial motors allow us to go down to 5µm steps). Then in this Task 4.4, the matrix of raw coarse measurements is transformed into the refined image, using the template matrix. In this Deliverable, we apply our own downscaling technique based on direct matrix inversion. The inversion is shown to be ill-conditioned, and thus we need to apply numerical t...
CALIBRATION SOFTWARE. H2020 Grant Agreement N° 761036 The linear matric algebra technique for downscaling the DR measurements becomes Foreground Knowledge of QWED. The concept has been proposed for presentation at MMAMA workshops (IMS and EuMW 2019) and its electromagnetic bases are planned to be explained in a regular submission to the IMS 2019 SC-33. QWED aims to incorporate the technique into its commercial products. Publication in full is not intended. Patenting does not appear worthwhile at this stage. Relation to third party IPRs: Our resulting new linear matrix algebra technique is initially compared in this Deliverable to the deconvolution in the Fourier domain. The Fouried domain approach was initiated by QWED team and pursued in a Ph.D. by X. Xxxxxx of the Warsaw University of Technology (also in “Deconvolution-based spatial resolution improvement technique for resistivity scans acquired with split-post dielectric resonator”, MIKON 2018). Xx. Xxxxxx holds IPRs for the Fourier domain method. He will be involved in the MMAMA activities when the two methods will further be compared and validated throughout the project. Document ID: D4.1 CALIBRATION SOFTWARE H2020 Grant Agreement N° 761036
CALIBRATION SOFTWARE. H2020 Grant Agreement N° 761036 assuming there a constant value of εd of the dielectric sample. We will get the energy increase caused by insertion of the dielectric sample: K L d W = a2h(ε − ε0 ) ∑ ∑E2 (k,l) = h(ε T d k =− K l =− L − ε0 )
CALIBRATION SOFTWARE. H2020 Grant Agreement N° 761036 We aim at a relation: [X] [V]’=[W]’ where [X] and [W]’ are known and we need to obtain [V]’ (3) Matrix [X] is generated in a way that the element in the chosen position (row, column) denoted x(r,s), for each r and s is equal to : T
CALIBRATION SOFTWARE. H2020 Grant Agreement N° 761036
CALIBRATION SOFTWARE. The software shall have a calibration utility to guide the technician though the calibration process. This utility will be accessible to qualified calibration personnel via a password protected, unique calibration login.
