- Thin film terahertz reverse engineering to analyse vegetal tissues doi link

Auteur(s): Abautret Yannick(Corresp.), Coquillat D., Bendoula Ryad, Heran Daphne, Grezes-besset Bruno, Amra Claude, Zerrad Myriam, Buet Xavier, Soriano Gabriel, Chazallet Frederic

Conference: SPIE OPTICAL SYSTEMS DESIGN (Madrid, ES, 2021-09-13)
Actes de conférence: Proc. SPIE 11872, Advances in Optical Thin Films VII, 1187209 (12 September 2021), vol. p. (2021)

Ref HAL: hal-03413162_v1
DOI: 10.1117/12.2597056
Exporter : BibTex | endNote

Thin film techniques also find applications outside their classical sectors, and among them is the terahertz field. These waves allow to penetrate a large variety of materials opaque to optics, and find applications in security and defence, automobile and avionics, medicine … THz waves provide new imaging techniques (transverse mode) but can also be used to probe in detail the depth of samples in the form of single layers or multilayers, which is the scope of this paper. Actually we take profit from thin film design procedures (usually developed for visible and infrared ranges) in order to address reverse engineering in the THz range. We first emphasize some key differences due to the fact that most broad-band THz sources are pulsed sources (here the THz pulse duration was around 3ps). Hence conversely to optics where optical properties are intensity data issued from spectrophotometric measurements, THz pulses directly allow to record the temporal signals with and without sample interaction, which gives the modulus and phase of the spectrum in the frequency domain. The consequence is that we operate the reverse engineering procedure in the complex plane (in opposition to the real axis of photometry), which involves more data. Here the pulse duration is around 3ps, and the frequency domain with acceptable noise is limited to [0.2 THz – 3,5 THz]. A few classical (inorganic) etalon samples are first analysed and their echoes are exploited to reveal their thicknesses under the assumption of negligible absorption. Then we use reverse engineering to take account of absorption and fit all data in the THz range, which confirms the previous results but with more accuracy. The resulting thicknesses are compared with success to the provider data. In a last step we investigate vegetal tissues (sunflower leaves), which is a much more complex task3. This study falls within the context of the optimization of plant production in regard to global warming and increasing demography, a challenge which requires to analyse and control the hydric stress of plants. Actually there is a growing demand to develop non-contact techniques to analyse leaves microstructure and understand their interaction with the surrounding medium. However the vegetal leaf is highly heterogeneous and cannot be analysed with optics, due to high diffuse reflectance of transmittance (no specular beams). A solution is provided by the THz waves, due to their much larger wavelengths which reduces scattering and the weight of heterogeneities. We show that in this THz regime, the sunflower leaf indeed behaves like a homogeneous multilayer, and this allows to use reverse engineering to extract the leaf design. Results emphasize a 8-layer stack including trichomes, cuticules, epidermis and mesophyll layers4. For each layer we extract the thickness and complex index. To our knowledge this is the first time the leaf multilayer structure is extracted with accuracy with non-contact techniques.