arXiv subjects

Christophe Blanc

Publications and source records attributed to Christophe Blanc.

13 recordsLinked to original sources

arXiv2022

Microparticle Brownian Motion near an Air-Water Interface Governed by Direction-Dependent Boundary Conditions

Although the dynamics of colloids in the vicinity of a solid interface has been widely characterized in the past, experimental studies of Brownian diffusion close to an air-water interface are rare and limited to particle-interface gap distances larger than the particle size. At the still unexplored lower distances, the dynamics is expected to be extremely sensitive to boundary conditions at the air-water interface. There, ad hoc experiments would provide a quantitative validation of predictions. Using a specially designed dual wave interferometric setup, the 3D dynamics of 9 micrometers diameter particles at a few hundreds of nanometers from an air-water interface is here measured in thermal equilibrium. Intriguingly, while the measured dynamics parallel to the interface approaches expected predictions for slip boundary conditions, the Brownian motion normal to the interface is very close to the predictions for no-slip boundary conditions. These puzzling results are rationalized considering current models of incompressible interfacial flow and deepened developing an ad hoc model which considers the contribution of tiny concentrations of surface active particles at the interface. We argue that such condition governs the particle dynamics in a large spectrum of systems ranging from biofilm formation to flotation process.

cond-mat.soft↗

arXiv2021

Helfrich-Hurault elastic instabilities driven by geometrical frustration

The Helfrich-Hurault (HH) elastic instability is a well-known mechanism behind patterns that form as a result of strain upon liquid crystal systems with periodic ground states. In the HH model, layered structures undulate and buckle in response to local, geometric incompatibilities, in order to maintain the preferred layer spacing. Classic HH systems include cholesteric liquid crystals under electromagnetic field distortions and smectic liquid crystals under mechanical strains, where both materials are confined between rigid substrates. However, richer phenomena are observed when undulation instabilities occur in the presence of deformable interfaces and variable boundary conditions. Understanding how the HH instability is affected by deformable surfaces is imperative for applying the instability to a broader range of materials. In this review, we re-examine the HH instability and give special focus to how the boundary conditions influence the mechanical response of lamellar systems to geometrical frustration. We use lamellar liquid crystals confined within a spherical shell geometry as our model system. Made possible by the relatively recent advances in microfluidics within the past 15 years, liquid crystal shells are composed entirely of fluid interfaces and have boundary conditions that can be dynamically controlled at will. We examine past and recent work that exemplifies how topological constraints, molecular anchoring conditions, and boundary curvature can trigger the HH instability in liquid crystals with periodic ground states. We then end by identifying similar phenomena across a wide variety of materials, both biological and synthetic. With this review, we aim to highlight that the HH instability is a generic and often overlooked response of periodic materials to geometrical frustration.

cond-mat.soft↗

arXiv2020

Learning Sparse Filters in Deep Convolutional Neural Networks with a l1/l2 Pseudo-Norm

While deep neural networks (DNNs) have proven to be efficient for numerous tasks, they come at a high memory and computation cost, thus making them impractical on resource-limited devices. However, these networks are known to contain a large number of parameters. Recent research has shown that their structure can be more compact without compromising their performance. In this paper, we present a sparsity-inducing regularization term based on the ratio l1/l2 pseudo-norm defined on the filter coefficients. By defining this pseudo-norm appropriately for the different filter kernels, and removing irrelevant filters, the number of kernels in each layer can be drastically reduced leading to very compact Deep Convolutional Neural Networks (DCNN) structures. Unlike numerous existing methods, our approach does not require an iterative retraining process and, using this regularization term, directly produces a sparse model during the training process. Furthermore, our approach is also much easier and simpler to implement than existing methods. Experimental results on MNIST and CIFAR-10 show that our approach significantly reduces the number of filters of classical models such as LeNet and VGG while reaching the same or even better accuracy than the baseline models. Moreover, the trade-off between the sparsity and the accuracy is compared to other loss regularization terms based on the l1 or l2 norm as well as the SSL, NISP and GAL methods and shows that our approach is outperforming them.

cs.NE↗

arXiv2020

2D Wasserstein Loss for Robust Facial Landmark Detection

The recent performance of facial landmark detection has been significantly improved by using deep Convolutional Neural Networks (CNNs), especially the Heatmap Regression Models (HRMs). Although their performance on common benchmark datasets has reached a high level, the robustness of these models still remains a challenging problem in the practical use under noisy conditions of realistic environments. Contrary to most existing work focusing on the design of new models, we argue that improving the robustness requires rethinking many other aspects, including the use of datasets, the format of landmark annotation, the evaluation metric as well as the training and detection algorithm itself. In this paper, we propose a novel method for robust facial landmark detection, using a loss function based on the 2D Wasserstein distance combined with a new landmark coordinate sampling relying on the barycenter of the individual probability distributions. Our method can be plugged-and-play on most state-of-the-art HRMs with neither additional complexity nor structural modifications of the models. Further, with the large performance increase, we found that current evaluation metrics can no longer fully reflect the robustness of these models. Therefore, we propose several improvements to the standard evaluation protocol. Extensive experimental results on both traditional evaluation metrics and our evaluation metrics demonstrate that our approach significantly improves the robustness of state-of-the-art facial landmark detection models.

cs.CV↗

arXiv2020

Facial Landmark Correlation Analysis

We present a facial landmark position correlation analysis as well as its applications. Although numerous facial landmark detection methods have been presented in the literature, few of them explicitly take into account the inherent relationship among landmarks. To reveal and interpret this relationship, we propose to analyze landmark correlation by using Canonical Correlation Analysis~(CCA). We experimentally show that the dense facial landmark annotations in current benchmarks are strongly correlated. We propose two applications based on this analysis. First, by analyzing the landmark correlation, we gain some interesting insights into the predictions of different landmark detection models (including random forests model and CNN models). We also demonstrate how CNNs progressively learn to predict facial landmarks. Second, we propose a few-shot learning method that allows to considerably reduce the manual effort for dense landmark annotation.

cs.CV↗

arXiv2017

The Universality of Thermal Transport in Amorphous Nanowires at Low Temperatures

Thermal transport properties of amorphous materials at low temperatures are governed by the interaction between phonons and localized excitations referred to as tunneling two level systems (TLS). The temperature variation of the thermal conductivity of these amorphous materials is considered as universal and is characterized by a quadratic power law. This is well described by the phenomenological TLS model even though its microscopic explanation is still elusive. Here, by scaling down to the nanometer scale amorphous systems much below the bulk phonon-TLS mean free path, we probed the robustness of that model in restricted geometry systems. Using very sensitive thermal conductance measurements, we demonstrate that the temperature dependence of the thermal conductance of silicon nitride nanostructures remains mostly quadratic independently of the nanowire section. It is not following the cubic power law in temperature as expected in a Casimir-Ziman regime of boundary limited thermal transport. This shows a thermal transport counter intuitively dominated by phonon-TLS interactions and not by phonon-boundary scattering in the nanowires. This could be ascribed to an unexpected high density of TLS on the surfaces which still dominates the phonon diffusion processes at low temperatures and explains why the universal quadratic temperature dependence of thermal conductance still holds for amorphous nanowires.

cond-mat.mes-hall↗

arXiv2016

Reduction of phonon mean free path: from low temperature physics to room temperature applications in thermoelectricity

It has been proposed for a long time now that the reduction of the thermal conductivity by reducing the phonon mean free path is one of the best way to improve the current performance of thermoelectrics. By measuring the thermal conductance and thermal conductivity of nanowires and thin films, we show different ways of increasing the phonon scattering from low temperature up to room temperature experiments. It is shown that playing with the geometry (constriction, periodic structures, nano-inclusions), from the ballistic to the diffusive limit, the phonon thermal transport can be severely altered in single crystalline semiconducting structures; the phonon mean free path being in consequence reduced. The diverse implications on thermoelectric properties will be eventually discussed.

cond-mat.mes-hall↗

arXiv2015

Nanoscale phase-engineering of thermal transport with a Josephson heat modulator

Macroscopic quantum phase coherence has one of its pivotal expressions in the Josephson effect [1], which manifests itself both in charge [2] and energy transport [3-5]. The ability to master the amount of heat transferred through two tunnel-coupled superconductors by tuning their phase difference is the core of coherent caloritronics [4-6], and is expected to be a key tool in a number of nanoscience fields, including solid state cooling [7], thermal isolation [8, 9], radiation detection [7], quantum information [10, 11] and thermal logic [12]. Here we show the realization of the first balanced Josephson heat modulator [13] designed to offer full control at the nanoscale over the phase-coherent component of thermal currents. Our device provides magnetic-flux-dependent temperature modulations up to 40 mK in amplitude with a maximum of the flux-to-temperature transfer coefficient reaching 200 mK per flux quantum at a bath temperature of 25 mK. Foremost, it demonstrates the exact correspondence in the phase-engineering of charge and heat currents, breaking ground for advanced caloritronic nanodevices such as thermal splitters [14], heat pumps [15] and time-dependent electronic engines [16-19].

cond-mat.mes-hall↗

arXiv2015

The thermal conductivity of silicon nitride membranes is not sensitive to stress

We have measured the thermal properties of suspended membranes from 10 K to 300 K for two amplitudes of internal stress (about 0.1 GPa and 1 GPa) and for two different thicknesses (50 nm and 100 nm). The use of the original 3 ω-Volklein method has allowed the extraction of both the specific heat and the thermal conductivity of each SiN membrane over a wide temperature range. The mechanical properties of the same substrates have been measured at helium temperatures using nanomechanical techniques. Our measurements show that the thermal transport in freestanding SiN membranes is not affected by the presence of internal stress. Consistently, mechanical dissipation is also unaffected even though Qs increase with increasing tensile stress. We thus demonstrate that the theory developed by Wu and Yu [Phys. Rev. B 84, 174109 (2011)] does not apply to this amorphous material in this stress range. On the other hand, our results can be viewed as a natural consequence of the "dissipation dilution" argument [Y. L. Huang and P. R. Saulson, Rev. Sci. Instrum. 69, 544 (1998)] which has been introduced in the context of mechanical damping.

cond-mat.mes-hall↗

arXiv2013

Modal Decomposition in Goalpost Micro/nano Electro-mechanical Devices

We have studied the first three symmetric out-of-plane flexural resonance modes of a goalpost silicon micro-mechanical device. Measurements have been performed at 4.2K in vacuum, demonstrating high Qs and good linear properties. Numerical simulations have been realized to fit the resonance frequencies and produce the mode shapes. These mode shapes are complex, since they involve distortions of two coupled orthogonal bars. Nonetheless, analytic expressions have been developed to reproduce these numerical results, with no free parameters. Owing to their generality they are extremely helpful, in particular to identify the parameters which may limit the performances of the device. The overall agreement is very good, and has been verified on our nano-mechanical version of the device.

cond-mat.mes-hall↗

arXiv2013

Nematic droplets in aqueous dispersions of carbon nanotubes

Aqueous dispersions of exfoliated, bile-salt stabilized single-wall carbon nanotubes exhibit a first order transition to a nematic liquid-crystalline phase. The nematic phase presents itself in the form of micron-sized nematic droplets also known as tactoids, freely floating in the isotropic host dispersion. The nematic droplets are spindle shaped and have an aspect ratio of about four, irrespective of their size. We attribute this to a director field that is uniform rather than bipolar, which is confirmed by polarization microscopy. It follows that the ratio of the anchoring strength and the surface tension must be about four, which is quite larger than predicted theoretically but in line with earlier observations of bipolar tactoids. From the scatter in the data we deduce that the surface tension of the coexisting isotropic and nematic phases must be extremely low, that is, of the order of nN/m.

cond-mat.soft↗

arXiv2013

Orientational order of carbon nanotube guests in a nematic host suspension of colloidal viral rods

In order to investigate the coupling between the degrees of alignment of elongated particles in binary nematic dispersions, surfactant stabilized single-wall carbon nanotubes (CNTs) have been added to nematic suspensions of colloidal rodlike viruses in aqueous solution.We have independently measured the orientational order parameter of both components of the guest-host system by means of polarized Raman spectroscopy and by optical birefringence, respectively. Our system allows us therefore to probe the regime where the guest particles (CNTs) are shorter and thinner than the fd virus host particles. We show that the degree of order of the CNTs is systematically smaller than that of the fd virus particles for the whole nematic range. These measurements are in good agreement with predictions of an Onsager-type second-viral theory, which explicitly includes the flexibility of the virus particles, and the polydispersity of the CNTs.

cond-mat.soft↗

arXiv2013

Phonon Heat Conduction in Corrugated Silicon Nanowires Below the Casimir Limit

The thermal conductance of straight and corrugated monocrystalline silicon nanowires has been measured between 0.3 K and 5 K. The difference in the thermal transport between corrugated nanowires and straight ones demonstrates a strong reduction in the mean free path of the phonons. This averaged mean free path is remarkably smaller than the smaller diameter of the nanowire, evidencing a phonon thermal transport reduced below the Casimir limit. Monte Carlo simulations highlight that this effect can be attributed to significant multiple scattering of ballistic phonons occuring on the corrugated surfaces. This result suggests an original approach to transforming a monocrystalline material into a phonon glass.

cond-mat.mes-hall↗