Institut des
NanoSciences de Paris
Accueil > Pages personnelles > Borensztein Yves > Scientific activities

Scientific activities

Master Internship proposal 2020-2021 / Proposition de stage de M2 2020-2021


Current Research

Ultra-sensitive plasmonic sensors (click for details)


1. Localized plasmon surface resonance in metal nanoparticles : principle of plasmonic sensors

Nanoparticles (NPs) of different metals (gold, silver, palladium, etc.) have particular optical properties, related to the Localized Surface Plasmon Resonances (LSPR), which are collective oscillations of the conduction electrons confined within the NPs, as illustrated in Fig.1. For example, the LSPR for Au is located in the visible optical range, and gives a red or purple color to Au NPs instead of the usual yellow color (Fig.2).



Fig.1 Localized surface plasmon resonance of a metal nanoparticle : collective oscillation of conductions electrons induced by the electric field of light

Fig.2 Solution of gold nanoparticles in suspension



This LSPR is very sensitive to the immediate environment of the NPs and can be strongly affected when the NPs interact with molecules or ions. Thanks to this very high sensitivity, gas or biological plasmonic sensors based on gold NPs, gold-based alloys or other metals [1] are being developed (see Fig. 3). They commonly use two methods of measurements. The first most common mode, which investigates the shift in wavelength of the LSPR induced by the analyte, requires the use of high-resolution monochromators. The second mode, based on self-reference or perfect absorbing systems, measures the changes of the signal intensity, usually requires the use of sophisticated samples elaborated by lithography techniques.

In our current research, we have been developing an original method, which combines the very sensitive reflectance anisotropy spectroscopy technique (RAS), previously used for investigating crystal surfaces (see RAS), with anisotropic plasmonic gold films, formed of slightly elongated nanoparticles, easily elaborated by grazing deposition on microscopy glass slides (see next paragraph).

This allows us to achieve a much higher sensitivity than conventional plasmonic sensors. We have been using it to study the reaction of hydrogen with gold nanoparticles [2], to make a prototype hydrogen sensor [3], and we recently demonstrated the possibility to determine the adsorption of a couple of biomolecules on a micrometer range sample [4].








Fig. 3 Shift of the plasmon resonance induced by the adsorption of molecules






2. Anisotropic samples prepared by grazing incidence deposition on glass slides

The metal films are elaborated by vacuum deposition at a grazing incidence of the order of 10° on microscope glass slides, as shown in the left scheme of figure 4.

The metal NP are in average preferentially elongated normally to the direction of deposition, which is demonstrated by Scanning Electron Microscopy (SEM) (Figure 4.b). This counterintuitive observation is explained by the shadowing effect (see Fig. 5) : the metal nuclei create exclusion areas at the rear, that the impinging atoms cannot reach, leading to a larger NP growth in the lateral direction to the atom flux direction, than in the parallel direction. The use of a kinetic Monte Carlo simulation at the atomic scale, where the various moves that an atom can experience on the film (Fig. 6) allows us to reproduce correctly the average shape and size of the NPs, observed by Scanning Electron Microscopy (SEM), as illustrated in Fig.4.c.

These anisotropic samples display corresponding anisotropic LSPR, leading to different transmissions for light polarization parallel and perpendicular to the deposition direction (Fig. 7) : the LSPR is red shifted along the long size of the NPs with respect to the LSPR along the short size, which is very well reproduced by use of the dipolar approximation, based on an area of metal ellipsoids in electromagnetic interaction.

Figure 7. Transmission spectra for polarized light parallel (blue) and perpendicular (red) to the deposition direction, and scheme of the anisotropic film. Continuous lines : experiment ; dotted lines : calculation based on the dipolar approximation.

3. Enhanced sensitivity and figure of merit

4. Plasmonic di-hydrogen sensor

An illustration of the sensing sensitivity down to a few ppm of H2 is shown in Fig. 5. The top panel shows the modification of the RAS signal measured in real time, when switching the gas from pure Ar to H2 diluted in Ar and back, for decreasing concentration of H2, from pure (100%) to highly diluted (0.0002%, i.e 2 part per million). On the left part, from 100% to 4%, the change of signal is large, and corresponds to the formation of the dense ’beta’ phase of Pd hydride. On the right part, at and below 0.25%, the change of signal is much smaller, and corresponds to the formation of the dilute ’alpha’ phase of Pd hydride. At intermediate concentration of H2 (1%), there is coexistence of the two phases. The zoom shown in the bottom panel shows that, except for the intermediate case, the kinetics for the detection of H2 is fast, even at very low concentration, showing that this method provides a very competitive H2 sensor. More details can be found in 2018 PhD defense by William Watkins


Fig.5 Effect of H2 diluted in Ar on the RAS signal measured on an anisotropic plasmonic Pd sample. Up : change of the RAS signal measured as a function of switching between pure H2 in Ar and pure Ar, for decreasing concentrations of H2, from pure H2 (100%) to 2 ppm. Middle : zoom on three cycles (16%, 1% and 0.015%), the first and last ones showing a fast saturation Bottom : scheme of PdHx for dense, dilute and intermediate phases







Formation and properties of monolayers of Si and of silicene (click for details)


2D-Xenes (where X = Si, Ge, Sn ...) are a class of two-dimensional materials from the group IVA that possess electronic properties different from those of the corresponding 3D bulk. Calculations have shown that silicene and germanene, i.e. Si or Ge atoms arranged in a honeycomb structure, could possibly exist and display electronic, transport or magnetic properties similar to graphene’s one. With the prospect of device engineering, silicene and germanene present a better compatibility with silicon-based microelectronics. Moreover, the spin-orbit coupling leads to the opening of a small band gap (a few meV) for these materials. However, contrary to graphene, there is no layered material from which 2D-Xenes could be exfoliated. Thus, intensive research has been made in order to synthesize these 2D materials while trying to maintain their intrinsic properties.


At the Institute of NanoSciences of Paris, we study the structure and growth mechanisms of silicene and germanene obtained by material evaporation under UHV on various substrates (Ag, Al, layered materials). We use in-situ scanning tunneling microscopy (STM), grazing incidence X-ray diffraction (GIXD) and optical investigations, together with density functional theory (DFT) calculations (in collaboration).

A complete overview of our results can be found in the page :Structure, optical properties, and growth mechanisms of 2D-Xenes


  • Optical investigation of so-called « silicene » monolayers

Theoretical calculations have shown that the optical properties of free-standing silicene are similar to the ones of graphene, and are dominated by the presence of the Dirac cone at the Fermi level, around the K point, and by interband transitions around the M point.

However, the Si monolayers when grown on Ag(110) or Ag(111) are in strong interaction with the sustrate, and their electronic properties are completely modified with respect to what is expected for the free-standing silicene. Consequently, the optical response is also different, and appears to be rather similar to amorphous or disordered silicon, which indicates that the Si layers are sp3 hybridized and not sp2.

Fig.1 shows the Surface Differential Reflectance Spectrum, measured for 0.5 and 1 Si ML grown on Ag(111), defined by DR= (RSi-RAg)/RAg, where RAg and RSi are the reflectances of the pristine Ag surface and the same surface covered by Si, respectively. The spectra are completely different to what would be expected if the Si film has got the same optical response as free-standing silicene (red curve). The experiment is correctly reproduced by DFT calculation (right). From these data, the effective dielectric function of the silicon layer can be isolated from the optical response of Ag, and is shown in Fig.2.a. The imaginary part (red) corresponds to the optical absorption. It is very well reproduced by the DFT calculation (b). On the contrary, it is different from the dielectric function calculated for silicene (black curve in (c), and close to that of amorphous Si (red curve in (c)).



Fig. 1 SDRS of Si/Ag(111) : (a) Experimental measurements at 515K for 0.5 ML (dashed black line) and 1.0 ML (solid black line) and at 600K for 1.0 ML (dot-dashed red line). (b) Computed spectrum for the (4 × 4) model (corresponding to 1.0 ML coverage).





Fig. 2 Surface dielectric difference (SDD) spectra of Si/Ag(111). (a) Experimental data at 515 and 600K (imaginary part only, blue dot-dashed line). (b) Computed SDD spectra for the 4 × 4 structure (solid red and black dashed curves). Also shown are the imaginary parts of the slab dielectric function for the clean (Ag : dot-dashed magenta line) and Si-covered surface (Si/Ag : green dots). (c). Imaginary part of the dielectric function of free-standing silicene, of 1 crystalline Si monolayer and of 1 amorphous Si monolayer.



  • « Silicene » multilayers

Optical measurements, combined with Auger electron spectroscopy and X-ray diffraction show that Si multilayers grown on Ag(111) have actually the bulk-Si (diamond-like) structure, and are covered by a thin layer of Ag which gives the well-known root(3)xroot(3) R30° reconstruction. Surface Differential Reflectance Spectroscopy and Thermo-Reflectance Spectroscopy (TRS) have both been used and showed that a film of 2.1 nm grown at high temperature on Ag(111) display optical properties intermediate between bulk crystalline and amorphous Si, which is typical of poorly crystallized diamond-like Si.

Fig.3 shows the experimental TR spectrum for 2.1 nm Si / Ag (red). It displays transitions which are typical of the transitions measured on bulk crystalline Si and on Ag. Comparison with calculation shows that the spectrum cannot be reproduced by free-standing silicene (1) or by “silicite” (2). The features are reproduced correctly by bulk crystalline Si (4), and a mixture half amorphous (3), half crystalline Si (4) permits to reproduce almost perfectly the experimental (5).




Fig. 3 Thermoreflectance spectra. Circles : experiments ; red : 2.1-nm Si/Ag ; black : bulk Si ; blue : bulk Ag. Continuous lines : calculations for different dielectric functions of the 2.1-nm silicon film on Ag. (1) : free-standing silicene ; (2) : “silicite” ; (3) : amorphous Si ; (4) : crystalline Si (2) ; (5) : half amorphous/crystalline Si. For spectra (1) and (2), the dashed lines are the spectra amplified by a factor of 7. The curves have been shifted vertically for a better visualization.







See details in page : 3. Silicene thin films : are they silicene ?.  




Previous Research

Catalytic properties of Au and Au-alloys nanoparticles
  • Catalytic Properties
  • Oxidation of CO
  • Transmission Electron Microscopy
  • Real-time investigation of plasmon resonance in Au/oxide catalysts during reaction
  • Environmental Scanning Tunneling Microscopy of Au and AuCu Nanoparticles
Optical response of silicon surfaces
  • Intrinsic optical response of Si(100) and Si(111) surfaces.

Adsorption and molecular layers on Silicon
  • Dissociative / non-dissociative adsorption mode of molecules (oxygen, ethylene, benzene...) on Si(100).
  • Monitoring the adsorption kinetics
  • Monte-Carlo Simulation



Surface Optical Spectroscopies
Surface Differential Reflectance Spectroscopy (SDRS)
The Reflectance Anisotropy Spectrometer (RAS)
The Reflectance Anisotropy Spectrometer (RAS) is a linear optical method, which permits one to measure minute anisotropies of the surface reflectance of a sample, as low as 10-5. It delivers the spectral dependence of the complex quantity : delta r / r = (rx - r y) / r, where rx and ry are the reflectances of light for polarization parallel to the two main axes of the surface. PNG

This home-made technique, similar to a normal incidence ellipsometer, is schematized in the following figure. After being polarized normally to the working plane, the light beam is reflected by the anisotropic sample, which is oriented at 45° with respect to the polarization of light. The reflected light is analyzed by means of a Photo-Elastic Modulator, working at 50 kHz and coupled with another polarizer, playing the role of an analyzer. It is then dispersed by a monochromator, and the signal measured by the photomultiplier located at the exit slit of the monochromator is analyzed by use of a lock-in amplifier. PNG .

Differential Diffuse Reflectance Spectroscopy (DDRS)
Environnemental Scanning Tunnelling Microscopy (E-STM)
Ultra-High Vacuum techniques


Selected talks and posters


Publications from 1995 click hereclick here