Capteurs autonomes de substances chimiques dans le sous-sol des infrastructures pétrolières System for monitoring sub-surface pollution risks in oil industry infrastructures using passive sensors

Fig 2: Schematic illustration of a SAW delay line

Measurement in a liquid medium requires the use of a shear wave weakly coupling with the surrounding fluid. We have chosen lithium tantalate as piezoelectric substrate due to its properties, which avoid packaging issues thanks to its high permittivity. This property allows covering interdigitated transducers with water without inducing capacitive short-circuit. Nevertheless, this material possesses a major disadvantage because it only propagates a pseudo-shear wave. Therefore, we have to prevent this wave from radiating in the bulk.

The pseudo-shear wave can remain confined to the surface by the addition of a conducting (metallic) layer over the sensing area (see Fig 2). An organic layer with the dual purpose of acoustic wave guiding and sensing capability can be also implemented over the substrate/metallic layer. Tuning these wave properties by varying the guiding layer thickness, piezoelectric substrate orientation and electrode geometry induces varying dependences of the acoustic velocity with temperature, stress and environmental conditions, which can only be properly predicted by an accurate modeling of the acoustic delay line response with these parameters.

Acoustic sensors exhibit complex behavior since they are exposed to a multitude of varying parameters. Before manufacturing transducers, accurate simulation is needed to minimize temperature and stress dependence, while maximizing gravimetric sensitivity – defined as velocity variation due to mass absorption on the sensing area. FEMTO-ST has developed all software needed to model acoustic propagation characteristics as a function of piezoelectric substrate orientation, guiding layer properties (thickness, density, Young modulus) and electrode geometry. Then, after having identified an appropriate propagation mode (interdigitated transducer design), the impact of the surrounding fluid, both on the insertion losses due to capacitive short circuit as well as variation of the static capacitance and mirror efficiency, must be considered in simulation. Furthermore, the impact of the organic guiding layer on the acoustic propagation properties and especially the mirror reflection coefficient needs to be considered in this design stage, accounting for possibly varying polymer physical properties (stiffness, density) against which the design must be robust. While water has been shown to hardly affect the interdigitated transducer electromechanical conversion capability, the influence of the high permittivity fluid over the mirrors has remained so far elusive.

Then, after simulation, we will be able to propose a complete design of transducers for impedance matching and operating in a frequency band compatible with the Ground Penetrating RADAR antenna. This design will include all the geometrical parameters (i. e. the electrode thickness, number of interdigitated transducer for optimal coupling and minimal pulse duration, also allowing for impedance matching by tuning the acoustic aperture) for the fabrication of the chips. This step of fabrication will be achieved by using the skills and competencies of FEMTO-ST and SENSeOR and the facilities of the technological platform MIMENTO.

Finally, an experimental assessment of the gravimetric sensitivity of a Love mode lithium tantalate-based device will be completed using a reference model (electrochemical deposition of copper or biochemical with well known antigen-antibody reactions) in liquid phase.

Here, we will synthesize a class of resists which (i) exhibit dedicated sensing capability towards targeted analytes (H₂S, BTEX, or HAP) and (ii) can be integrated in collective processes of fabrication of acoustics sensors.

In the case of H₂S, we will develop organometallic monomers including metallic cations (Pb²⁺, Zn²⁺ etc.) which can specifically react with H₂S. In the case of organic analytes (BTEX or PAH), we base our strategy on the hypothesis that there is no organic solvent in an unpolluted soil. Therefore, we assume that selectivity towards sub-species is not needed since any organic contamination of soils has to be detected. As a consequence, we will develop hydrophobic monomers, which can adsorb or absorb organic solvents without selectivity. In these two cases, reaction will not be reversible under the experimental conditions of use of the sensor, which is compatible with the requirements of the long-term monitoring of the pollution risk of soils (WP3). All monomers will also bear UV-active functional groups (C=C doubles bonds, acrylate, etc.) in order to be polymerized by UV-light exposure (wavelengths between 365 and 436 nm) used during the fabrication of sensors.

In a second step, the monomers will then formulate to achieve the fabrication of resists. These resists have to be compatible with collective fabrication (i. e. at the wafer scale) of acoustic sensors. The resists require a good processability for thin film deposition (thickness ranging from 50 to 2500 nm depending on the acoustic wave polarization and film guiding properties) and patterning (lateral precision around 10 um) in order to separate the temperature effect and define a reference path for getting rid of the RADAR to sensor distance delay. In addition, pad opening requires the removal of the guiding layer from the electrical contacts. To resume, the resists will be formulated by the mixing of three main components:

1. 
   a monomer bearing the adsorption site of the analyte (i. e. a metallic cation - H₂S or hydrophobic functional group – PAH/BTEX) and a UV-active polymerizable function,
   
2. 
   a reticulating agent for the tuning of mechanical properties of thin films,
   
3. 
   a solvent for the deposition of thin layers by spin-coating onto the acoustics components at the wafer scale
   

Having completed tasks 1.1 and 1.2, we have to manufacture enough chips to investigate packaging and to assess long term stability when the sensing element is exposed to realistic, outdoor environments (See Figure 3).

The packaging of a sensor has to satisfactory two competing parameters: 1) the packaging should protect the sensing area from physical damage and yet 2) allow the analyte contained in the fluid to permeate towards the sensing area. In our case, the packaging should prevent the deposition of small soil particles (i. e. dust). To achieve this protection, we propose to use porous polymer membranes, which allow the diffusion of organic targets (H₂S, BTEX, PAHs) but stop the particles. We start the packaging by using Nafion, which is widely used in packaging processes, and compensate for the poor mechanical stability of the membrane by stacking it between two metallic grids to reinforce the mechanical stability of the package. If Nafion is not efficient, we will consider inorganic membranes (porous silicon, silica etc.).

The chips will provide the basic input for finding a packaging scheme that will not deterioriate the sensing capability, before being used by all other workpackages, namely wireless interrogation using GPR and sub-surface sensing in relevant sites. Reaching this goal will require wafer scale fabrication of functionalized chips and radiofrequency characterization before dicing to obtain individual chips. Packaging solutions might involve wafer scale packaging allowing for the sensing area to be exposed to the analyte while preventing particles from reaching the chip surface. If no wafer scale solution is identified, commercially available packages will be modified for such a result at the individual chip level. By using lithium tantalate, we have already avoided the major issue of depositing microfluidic structures on the acoustic path: the package is solely attached to the die outside the acoustic paths.

Finally, the last step is dedicated to the long-term stability of the packaged device under various environmental conditions simulated by accelerated aging in a climatic chamber including temperature and moisture variations as available in the FEMTO-ST facility. This investigation will also be relevnt to monitor the long-term stability of the sensing layer by introducing a sufficient number of transducers in the climatic chamber and periodically sampling a few sensors to assess the evolution of their response under a probe station (wired configuration).

Major Risks: Packaging is a major risk due to the challenging conditions of use in soil. In addition, radiofrequency compatibility is needed, meaning the package is compatible with high frequency signals.

Fallback solution: If we fail to develop a dedicated packaging solution, commercial packaging solutions exist for gas sensing in dust-saturated environments. They are based on 200 mesh screen (75 um) grid capping a TO5 metallic package. Adapting such solutions to our needs requires identifying a supplier of such packages, or manufacturing ourselves the cap by milling commercially available TO5 packages (Sigma Aldrich Screens for CD-1).

Work package 2. Interrogation unit based on a Ground Penetrating RADAR concept, T₀ - T₀ +42

Leader: D. Rabus (SENSeOR)

Personel involved: J.-M. Friedt, and J. Jeannoutot (FEMTO-ST) and A PhD’s student (to be hired) – D. Rabus, F. Gegot and L. Arapan (SENSeOR) - M. Baqué, I. Betremieux and M.F. Benassy (TOTAL SA)