Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Hydroelastic wake on a thin elastic sheet floating on water

Hydroelastic wake on a thin elastic sheet floating on water Hydroelastic wake on a thin elastic sheet floating on water 1 2 1 Jean-Christophe Ono-dit-Biot, Miguel Trejo, Elsie Loukiantcheko, 1 2 1, 2 3, 4, ∗ Max Lauch, Elie Rapha¨el, Kari Dalnoki-Veress, and Thomas Salez Department of Physics and Astronomy, McMaster University, 1280 Main Street West, Hamilton, Ontario, L8S 4M1, Canada. Laboratoire de Physico-Chimie Th´eorique, UMR CNRS Gulliver 7083, ESPCI Paris, PSL Research University, 75005 Paris, France. Univ. Bordeaux, CNRS, LOMA, UMR 5798, F-33405 Talence, France. Global Station for Soft Matter, Global Institution for Collaborative Research and Education, Hokkaido University, Sapporo, Hokkaido 060-0808, Japan. (Dated: June 21, 2018) We investigate the hydroelastic wake created by a perturbation moving at constant speed along a thin elastic sheet floating at the surface of deep water. Using a high-resolution cross-correlation imaging technique, we characterize the waves as a function of the perturbation speed, for different sheet thicknesses. The general theoretical expression for the dispersion relation of hydroelastic waves includes three components: gravity, bending and stretching. The bending modulus and the tension in the sheet are independently measured. Excellent agreement is found between the experimental data and the theoretical expression. In 1963, Richard Feynman described water waves in inertia [32–38]. The dispersion relation in the inertial his famous Lectures on Physics [1] as “the worst possi- case was analytically derived and found to depend on ble example [of waves], because they are in no respect three components: gravity, bending and stretching [39]. like sound and light; they have all the complications that A few experimental studies developed in different con- waves can have”. Several decades later, some questions texts have studied the limiting cases where only bending remain unanswered and the study of these waves contin- and stretching [40, 41], or gravity and bending [20, 21], ues to be an area of great interest. For example, Kelvin’s contribute. gravity wake behind a ship [2] still stimulates fundamen- In this Letter, we study the hydroelastic wake cre- tal questions [3, 4]. Surface tension of the liquid-air in- ated by a perturbation moving at constant speed along terface also influences the wave propagation, resulting an elastic sheet floating on deep water. The waves are in gravito-capillary waves and wake [5–8]. Unlike the gravity wake, the capillary wake appears ahead of the perturbation [6]. This is particularly relevant for the lo- (v) comotion of insects [9–13], as well as for nanorheologi- cal applications involving e.g. atomic-force microscopy (i) probes moving along thin viscous samples [14–16]. (iii) Other waves of interest are the ones that propagate on elastic plates and membranes. Their properties are dic- tated by both the bending and stretching rigidities of the material [17]. Floating such an elastic sheet on a liquid (ii) further leads to the coupling of the elastic waves to hydro- dynamics. The resulting hydroelastic waves are of par- ticular interest, as elastic sheets surrounded by fluids are ubiquitous in nature. Examples can be found in fluid me- (iv) chanics [18, 19], geophysics [20–23], and biophysics [24]. Hydroelastic waves are also relevant to practical appli- cations in civil engineering [25, 26], as well as in energy FIG. 1. 3D schematics of the experimental setup. (i) Ro- harvesting through piezoelectric flags [27] and control of tating transparent annular tank (outer radius R = 50 cm, out energy radiation by trucks moving on ice sheets [28]. In- inner radius R = 30 cm) filled with water to a depth of in terestingly, the propagation of such waves can be finely about 16 cm, and with an elastic sheet floating atop. (ii) controlled in an optical-like fashion by using model thin Infrared beam-breaking setup to measure the angular speed sheets with heterogeneous elastic properties [29]. Dif- Ω of the elastic sheet. (iii) A pipette perturbs the surface ferent properties of these waves, such as the wave re- by blowing air, causing a wake to form. (iv) Light sheet and dot pattern used to characterize the waves using the Schlieren sistance or non-linear effects, have been further studied method [42]. (v) A camera is placed ∼ 2 m above the tank to theoretically [30, 31], including the overdamped limit of image the dot pattern. lubrication settings where viscosity dominates over fluid arXiv:1806.07472v1 [physics.flu-dyn] 19 Jun 2018 2 imaged using a high-resolution optical method. By us- ing elastic sheets with different thicknesses, the bending modulus of the sheet is varied over more than two or- ders of magnitude. We find excellent agreement between experimental data and the general theoretical dispersion relation, accounting for the three different contributions: gravity, bending, and stretching. A transparent annular tank is filled with water, as shown in Fig. 1. Thin elastic sheets of Elastosil (Wacker Chemie AG) with nominal thicknesses h of 50, 100, 200, 250 and 350 μm, and lateral dimensions of 20 cm × 16 cm are floated onto the surface of water. A thin rigid plas- tic support (18 cm × 1 cm × 0.1 cm) is placed atop the leading and trailing edge of the elastic sheet to ensure the sheets do not crumple. We experimentally verify that adding the supports does not introduce an anisotropic tension in the sheet, by ensuring that the deformation induced by ball bearings placed atop the sheet is axially symmetric (see SI). The tank is rotated at constant an- −1 gular speed Ω, ranging from 0 to 2.5 rad.s , causing the water to flow and the sheet to move. We take advan- tage of the opaque plastic supports to measure the angu- lar speed of the sheet, using an infrared beam-breaking technique. Because of inertia, both the sheet and water do not follow the tank’s speed instantaneously. Hence, FIG. 2. (a) Raw 2D dot-displacement data measured with all experiments are performed only once the speed of the Ncorr [44], for an elastic sheet of nominal thickness h ≈ −1 200 μm moved at speed v = 0.9 m.s . The displacement sheet is constant and equal to the speed of the tank. vectors are only shown every 10 pixels for clarity. The bot- A glass capillary (World Precision Instruments, USA) tom arrow indicates the speed of the sheet, with respect to the is pulled to a diameter of about 100 μm at one end stationary air jet whose position is indicated by the red ar- with a pipette puller (Narishige, Japan), and used to row. (b) Zoom around the perturbation, corresponding to the blow air at the surface of the sheet (Fig. 1(iii)). The black box in (a). (c) The y-component d of the displacement pipette is placed in the middle of the tank, i.e. at a ra- field. Warm colours (green to orange) correspond to positive dius R = 40 cm from the center. The air jet acts as a displacements, while cold colours (green to blue) correspond to negative ones. (d) Zoom around the perturbation, corre- perturbation moving at speed v = ΩR in the reference sponding to the black box in (c). All scale bars correspond to frame of the elastic sheet, which generates an hydroelas- 1 cm. tic wake. The latter is imaged using a synthetic Schlieren method [42] involving a random dot pattern refracted by the surface topography. The dot pattern is generated us- ing Matlab [42, 43] and printed onto a transparency film. wave pattern of the hydroelastic wake appears clearly, Light shines through the dot pattern (Fig. 1(iv)) and the with a dominant – centimetric – wavelength λ. As the wake, before being collected by a camera above the tank hydroelastic waves propagate along the y-direction, the (Fig. 1(v)). The Schlieren method consists in measuring projection of the displacement along that direction pro- the apparent displacement of the dots due to light re- vides the strongest signal for analysis. Figures 2(c,d) fraction by the wake. The displacement is measured rel- thus focus only on the y-component d of the displacement ative to a reference image of the unperturbed surface (i.e. field. To characterize experimentally the dispersion rela- no air jet) moving at angular speed Ω – which ensures tion of the hydroelastic wake, the wavelength λ is mea- that the collected information is only due to the wake. sured as a function of the speed v. Figures 3(a-c) show This measurement is performed using an open-source dig- the y-component d of the displacement field, for various ital 2D image-correlation algorithm (Ncorr, Matlab) [44]. speeds. The wake is slightly tilted and not symmetric From the displacement of the dots, one can access the about the y-axis, because of the centrifugal force and the slope of the surface and thus the surface topography [42]. surface of water assuming a parabolic profile when ro- Figures 2(a,b) show a typical vectorial displacement tated. This distorsion is avoided in the wavelength mea- field. The air jet creates a localized perturbation in the surement by analyzing the displacement field normal to sheet, as evidence by the large magnitude of the dis- the wave front, as shown in Fig. 3(a,d). We observe both placement field therein. Ahead of the perturbation, in the wavelength and the displacement to decrease as the the reference frame of the elastic sheet, the upstream speed increases. 3 vanishes in the far field and that satisfies Laplace’s equa- 2 2 2 tion [45]: ∇ ϕ = (∇ + ∂ )ϕ = 0. At lowest order in r z the flow (i.e. for small-amplitude hydroelastic waves), the linearized Bernoulli equation for unsteady potential flows provides the excess hydrodynamic pressure exerted on the sheet: P = −ρ ∂ϕ/∂t| − ρgζ, with ρ the liquid z=0 density and g the acceleration due to gravity. To obtain the dispersion relation, one substitutes the expression for P into Eq. (1) in the absence of driv- ing (P = 0), and invokes the kinematic condition ext ∂ϕ/∂z| = ∂ζ/∂t at the water-sheet interface. Con- z=0 sidering a plane wave ϕ ∝ exp[i(ky − ωt − ikz)] satisfy- 0.5 ing Laplace’s equation, with angular wavenumber k and angular frequency ω(k), yields [39]: Bk σk ω = + + gk . (2) ρ ρ -0.5 We now consider the wake created by the driving per- -2 0 2 4 6 8 turbation P traveling at constant speed v along y. ext In the comoving frame of the perturbation, the angular frequency ω of a plane-wave component of the wake is FIG. 3. (a) - (c) y-component d of the displacement field (see shifted by the Doppler effect, and thus reads ω = ω−kv. −1 Fig. 2), for speeds v = 0.4, 0.6, and 0.9 m.s , respectively. Furthermore, since in that comoving frame the wake is All scale bars correspond to 1 cm. (d) y-component d of the stationary, ω = 0 is a necessary condition. Using Eq. (2), displacement field normal to the wave front (see red line in one thus obtains the central relation connecting the an- (a)), as a function of the distance ∆ x from the perturbation, gular wavenumber k and the perturbation speed v, for a for the three speeds as indicated (shifted vertically for clarity). The position of the air-jet perturbation is indicated by the hydroelastic wake on deep water: arrow. s Bk σk g v = + + . (3) ρ ρ k In order to quantify further and rationalize these obser- vations, we now introduce the relevant theoretical frame- An extensive analysis of this relation, similar to the one work. The mechanical system we consider is the thin performed for the gravito-capillary case [7], reveals the elastic sheet in its reference frame. Neglecting the solid main features of the present wake (see also Fig. 4). First, inertia owing to the slenderness of the sheet, the out- below a certain minimal speed v wave propagation is im- of-plane displacement field z = ζ(r, t) with respect to its possible. Secondly, at a given speed v > v there are two flat horizontal state z = 0 satisfies the F¨oppl-von K´arm´an possible values for the observed wavelength: i) the small- equation [17]: est value corresponds to a group velocity that is higher 4 2 than the perturbation speed v, and therefore the waves B∇ ζ − σ∇ ζ = P + P , (1) ext r r propagate upstream of the perturbation. This is the sit- along the 2D horizontal space coordinate r = (x, y) and uation studied in the present work (see Figs. 2 and 3), time t, where ∇ is the nabla operator in 2D, B is the which is dominated by bending and stretching at suffi- bending stiffness of the sheet, and σ represents the ten- ciently large speed; ii) the largest value corresponds to a sion in the sheet. The first and second terms respec- group velocity that is lower than the perturbation speed tively account for bending and stretching. The system v, and therefore the waves propagate downstream of is further subjected to two external forces: the excess the perturbation. This situation corresponds to Kelvin’s hydrodynamic pressure P (r, t) (with respect to the at- classical wake [2], which is dominated by gravity at suf- mospheric one) exerted on the sheet by the water flow ficiently large speed. under gravity, and the driving pressure P (x, y − vt) The values of the bending modulus B and the tension ext modelling the perturbation by the air jet translating at σ are measured independently. The bending modulus 3 2 constant speed v along y. B = Eh /[12(1 − ν )] depends on three parameters: i) The water contribution, P (r, t), is calculated by as- the Young’s modulus E = 1.11±0.06 MPa of Elastosil , suming an incompressible and irrotational flow of an measured using the stress-strain curve; ii) the sheet thick- inviscid fluid, in a semi-infinite half space located at ness h, depending on the sample and measured through z < ζ(r, t). In this context, the fluid velocity field can optical microscopy; iii) the Poisson’s ratio ν = 0.5, as- be written as ∇ϕ, where ϕ(r, z, t) is a potential that suming that Elastosil is an incompressible elastomer. 4 The values of h(μm) and B(N.m) for our five different 2.5 samples are found to be: (h, B) = {[51 ± 1, (1.6 ± 0.1) × −8 −7 10 ], [104 ± 2, (1.4 ± 0.1) × 10 ], [213 ± 7, (1.2 ± 0.1) × −6 −6 10 ], [258 ± 2, (2.1 ± 0.1) × 10 ], [362 ± 3, (5.9 ± 0.3) × −6 10 ]}. We note that the 362 μm film was obtained by stacking two films with nominal thicknesses of 250 μm 1.5 and 100 μm. Besides, as the sheet is freely floating on water, the tension in the sheet is equal to the air-water surface tension: σ = γ. The latter is measured to be −1 γ = 50 ± 10 mN.m , as in [29], from two different techniques (see SI): i) using a Wilhelmy-plate setup; ii) 0.5 characterizing the dispersion relation of gravito-capillary waves on water. The low value of γ and the large un- certainty are attributed to the fact that the tank is filled with an important volume of tap water, and thus subject 0 0.2 0.4 0.6 0.8 1 to contamination. As shown in Fig. 4 (dashed lines), using the above mea- sured values of B and σ, one can predict the evolution of FIG. 4. Angular wavenumber k = 2π/λ as a function of the angular wavenumber k as a function of the perturba- perturbation speed v, for five different sheet thicknesses h as tion speed v. The uncertainties in B and σ are taken into indicated in the legend. Each data point was obtained using account through two limiting predictions and an interval the procedure detailed in Fig. 3 and an average over four in between. Note that the uncertainty on σ accounts different experimental displacement profiles. The error bars correspond to the standard deviation, which is comparable for most of the spread between the two limiting predic- to the marker size. For each thickness h, the blue and red tions. Interestingly, for the values of σ and B considered dashed lines separated by a grey region indicate the upper and here, all three terms in the right-hand side of Eq. (3) are lower theoretical predictions obtained from Eq. (3), using the of the same order of magnitude, especially at low speed independently-measured values of B and σ (see main text) −1 v < 0.4 m.s (see SI). This highlights the counterin- and their uncertainties. tuitive role of gravity in the wavelength selection of the upstream hydroelastic wake. Finally, as expected from Eq. (3), all the theoretical curves for different h (and thus formed by moving a thin elastic sheet, floating on water, B) collapse onto Kelvin’s gravity-dominated branch [2], past a stationary air jet. Specifically, we experimentally at both large v and small k. measured the wavelength of the wake as a function of the Using the experimental procedure detailed above, we perturbation speed, for sheets with bending moduli vary- measure the wavelength λ (see Fig. 3), or equivalently ing over two orders of magnitude. For thin elastic sheets the angular wavenumber k = 2π/λ, as a function of v. (thickness smaller than 100 μm), stretching plays a sig- The results for the five different sheet thicknesses h are nificant role in the propagation of the waves. For thicker shown in Fig. 4 (data points). We find excellent agree- elastic sheets (thickness larger than 100 μm), the bending ment between the experimental data and the theoretical contribution becomes dominant – a regime that is par- predictions, with no adjustable parameter. The experi- ticularly relevant for floating ice [20, 21, 28]. The results mental data points for the two thinnest sheets seem to are found to be in excellent agreement with theoretical be in slightly better agreement with the upper predic- predictions, based on the elasticity of slender structures tion at low speed, and with the lower prediction at high coupled to the hydrodynamics of inviscid incompressible speed. This observation could perhaps be related to a flows, with no adjustable parameter. Interestingly, for slight, but not quantifiable, increase in the sheet tension thin elastic sheets, bending, stretching and gravity all due to the increase in curvature of the air-water interface. contribute to the hydroelastic wake – a result with prac- Another interesting feature of Fig. 4 is that the difference tical consequences in geophysics, biophysics and civil en- gineering. between the upper and lower predictions decreases as the thickness h of the sheet is increased. Indeed, the relative The financial support by the Natural Science and Engi- contribution of bending to Eq. (3) increases, and the dif- neering Research Council of Canada and the Joliot chair ference between both predictions, which is mainly due to of ESPCI Paris is gratefully acknowledged. The authors the uncertainty in tension, decreases. Note that Kelvin’s thank the Global Station for Soft Matter, a project of classical gravity-dominated branch [2] corresponds to: i) Global Institution for Collaborative Research and Edu- a wake propagating behind the perturbation; ii) a wave- cation at Hokkaido University. They are also grateful to length that would almost reach the meter range in our ex- Andreas Koellnberger and Wacker Chemie AG for tech- periments, which is not attainable with the current setup. nical information and the donation of Elastosil films, as In this Letter, we have studied the hydroelastic wake well as to Antonin Eddi, Lucie Domino, Andreas Carlson 5 and Yacine Amarouchene for stimulating discussions. [30] F. Milinazzo, M. Shinbrot, and N. Evans, J. Fluid Mech. 287, 173 (1995). [31] E. I. Pa ˘r˘ au and J.-M. Vanden-Broeck, Phil. Trans. R. Soc. A 369, 2973 (2011). [32] A. Hosoi and L. Mahadevan, Physical review letters 93, 137802 (2004). thomas.salez@u-bordeaux.fr [33] H. Vandeparre, S. Gabriele, F. Brau, C. Gay, K. K. [1] R. P. Feynman, R. B. Leighton, and M. Sands, The Parker, and P. Damman, Soft Matter 6, 5751 (2010). Feynman lectures on physics, Vol. I (Basic books, 2011). [34] J. R. Lister, G. G. Peng, and J. A. Neufeld, Physical [2] L. Kelvin, Proc. Inst. Mech. Eng 3, 409 (1887). review letters 111, 154501 (2013). [3] M. Rabaud and F. Moisy, Phys. Rev. Lett. 110, 214503 [35] T. T. Al-Housseiny, I. C. Christov, and H. A. Stone, (2013). Physical review letters 111, 034502 (2013). [4] A. Darmon, M. Benzaquen, and E. Rapha¨ el, J. Fluid [36] A. Carlson and L. Mahadevan, Mech. Rapids 738 (2014). Physics of Fluids 28, 011702 (2016). [5] L. D. Landau and E. M. Lifshitz, Fluid Mechanics, 2nd [37] M. Arutkin, R. Ledesma-Alonso, T. Salez, and ed. (Pergamon, 1987). E. Rapha¨ el, J. Fluid Mech. 829, 538 (2017). [6] D. J. Acheson, Elementary fluid dynamics (Oxford Uni- [38] O. Kodio, I. M. Griffiths, and D. Vella, versity Press, 1990). Phys. Rev. Fluids 2, 014202 (2017). [7] E. Rapha¨ el and P.-G. de Gennes, Phys. Rev. E 53, 3448 [39] R. M. S. M. Schulkes, R. Hosking, and A. Sneyd, Journal (1996). of Fluid Mechanics 180, 297 (1987). [8] F. Dias and C. Kharif, Annu. Rev. Fluid Mech. 31, 301 [40] L. Deike, J.-C. Bacri, and E. Falcon, J. Fluid Mech. 733, (1999). 394 (2013). [9] D. L. Hu, B. Chan, and J. W. Bush, Nature 424, 663 [41] L. Deike, M. Berhanu, and E. Falcon, Phys. Rev. Fluids (2003). 2, 064803 (2017). [10] A. Chepelianskii, F. Chevy, and E. Rapha¨ el, Phys. Rev. [42] F. Moisy, M. Rabaud, and K. Salsac, Exp. Fluids 46, Lett. 100, 074504 (2008). 1021 (2009). [11] F. Closa, A. Chepelianskii, and E. Rapha¨ el, Phys. Fluids [43] F. Moisy and M. Rabaud, 22, 052107 (2010). “Free-surface synthetic schlieren (fs-ss): A tutorial,” . [12] J. Voise and J. Casas, J. Royal Soc. Interface 7, 343 [44] J. Blaber, B. Adair, and A. Antoniou, Exp. Mech. 55, (2010). 1105 (2015). [13] T. Steinmann, M. Arutkin, P. Cochard, [45] G. K. Batchelor, An Introduction to Fluid Dynam- E. Rapha¨ el, J. Casas, and M. Benzaquen, ics (Cambridge University Press, Cambridge, England, Journal of Fluid Mechanics 848, 370 (2018). 1967). [14] K. Wedo lowski and M. Napi´ orkowski, Soft Matt. 11, 2639 (2015). [15] R. Ledesma-Alonso, M. Benzaquen, T. Salez, and E. Rapha¨ el, J Fluid Mech. 792, 829 (2016). [16] R. Ledesma-Alonso, E. Raphael, T. Salez, P. Tordjeman, and D. Legendre, Soft Matter 13, 3822 (2017). [17] L. D. Landau and E. M. Lifshitz, Theory of Elasticity, Vol. 7, 3rd ed. (Elsevier, New York, 198). [18] M. J. Shelley and J. Zhang, Annu. Rev. Fluid Mech. 43, 449 (2011). [19] E. Virot, X. Amandolese, and P. H´ emon, J. Fluid. Struct. 43, 385 (2013). [20] T. Takizawa, Cold regions Science and Technology 11, 171 (1985). [21] V. Squire, W. Robinson, P. Langhorne, and T. Haskell, Nature 333, 159 (1988). [22] V. A. Squire, J. P. Dugan, P. Wadhams, P. J. Rottier, and A. K. Liu, Annu. Rev. Fluid Mech. 27, 115 (1995). [23] E. Pa ˘r˘ au and F. Dias, J. Fluid Mech. 460, 281 (2002). [24] J. B. Grotberg and O. E. Jensen, Annu. Rev. Fluid Mech. 36 (2004). [25] M. Murai, H. Kagemoto, and M. Fujino, Journal of ma- rine science and technology 4, 123 (1999). [26] E. Watanabe, T. Utsunomiya, and C. Wang, Engineering structures 26, 245 (2004). [27] D. T. Akcabay and Y. L. Young, Physics of Fluids 24, 054106 (2012). [28] J. W. Davys, R. J. Hosking, and A. D. Sneyd, Journal of Fluid Mechanics 158, 269?287 (1985). [29] L. Domino, M. Fermigier, E. Fort, and A. Eddi, EPL (Europhysics Letters) 121, 14001 (2018). Supplemental Information for: “Hydroelastic wake on a thin elastic sheet floating on water” 1 2 1 Jean-Christophe Ono-dit-Biot, Miguel Trejo, Elsie Loukiantcheko, 1 2 1, 2 3, 4, ∗ Max Lauch, Elie Rapha¨el, Kari Dalnoki-Veress, and Thomas Salez Department of Physics and Astronomy, McMaster University, 1280 Main Street West, Hamilton, Ontario, L8S 4M1, Canada. Laboratoire de Physico-Chimie Th´eorique, UMR CNRS Gulliver 7083, ESPCI Paris, PSL Research University, 75005 Paris, France. Univ. Bordeaux, CNRS, LOMA, UMR 5798, F-33405 Talence, France. Global Station for Soft Matter, Global Institution for Collaborative Research and Education, Hokkaido University, Sapporo, Japan. (Dated: June 21, 2018) TENSION ISOTROPY (a) yy (b) σ σ xx xx yy FIG. S1. (a) Top-view schematic of the elastic sheet floating on water. Plastic beams are placed at the leading and trailing edges of the elastic sheet to ensure the sheet does not crumple. Ball bearings are dropped atop the sheet to verify that no anisotropy is introduced when placing the beams. (b) Picture of the reference dot pattern, seen through the water and the sheet, around a ball bearing. The solid lines are sample isodisplacement lines. The dashed lines are the best fits of the isodisplacement lines to ellipses. The best-fit ellipticities are found to be equal to 1 for all cases. A schematic of the elastic sheet is shown in Fig. S1(a). The tensions along the x- and y-axes are denoted σ xx and σ , respectively. We place ball bearings directly on the sheet floating on water, and we image the resulting yy deformation using the optical Schlieren method [1]. We then calculate the magnitude of the displacement vector field, which is directly linked to the deformation of the elastic sheet. Sample isodisplacement lines are shown in Fig. S1(b). We quantify the anisotropy of the deformation by fitting the isodisplacement lines to ellipses. The best-fit ellipticities for the four cases shown in Fig. S1(b) are all found to be equal to 1, meaning that the isodisplacement lines are circles, which thus indicates that the tension in the sheet is isotropic. Indeed, if σ or σ was larger than the other, the xx yy deformation would be elongated along the low-tension direction, leading to an ellipticity larger than 1. TENSION MEASUREMENT The tension σ along the x-axis is set by the water-air surface tension γ, as both the left and right edges are free (see xx Fig. S1(a)). Since all the experiments presented in the study are performed on sheets where the tension is isotropic, one can safely assume that σ = σ = σ = γ. The tabulated value for the pure water-air surface tension under yy xx −1 ambient conditions is γ = 72 mN.m , but it is extremely sensitive to contamination by all kinds of surfactants. The arXiv:1806.07472v1 [physics.flu-dyn] 19 Jun 2018 2 experiments being conducted in an open tank containing ∼ 200 L of water, contamination is unavoidable. Therefore, γ was measured from two independent methods. −1 First, using a Wilhelmy-plate setup, γ was found to be between 40 and 55 mN.m , for water from three different sources: water from the tank after one day, tap water, and deionized water. The largest value of γ was obtained for deionized water, and the smallest one for the water from the tank – which is consistent with tank contamination over time. 0 0.2 0.4 0.6 0.8 1 FIG. S2. Angular wavenumber k as a function of perturbation speed v, for the gravito-capillary wake at the surface of deep water. The dashed line shows the best fit of Eq. (S1) to the experimental data points. The water-air surface tension −1 γ = 47 mN.m is obtained as the only adjustable parameter. Another approach to measure γ is to invoke the gravito-capillary wake formed at the surface of deep water by a perturbation moving at constant speed v. In such a case, the analogue of Eq. (3) is [2]: γk g v = + . (S1) ρ k Therefore, γ can be evaluated by fitting Eq. (S1) to the experimental evolution of the angular wavenumber k as a function of speed v for a gravito-capillary wake. In fact, as the elastic sheet only covers a small portion of the water in the tank, the hydroelastic wake is only observed once a lap when the sheet moves across the stationary perturbation at speed v. Otherwise, water flowing at speed v is directly exposed to the perturbation, and a gravito-capillary wake is instead formed at the surface. Using the Schlieren method, the wavelength λ = 2π/k of the upstream gravito-capillary wake is measured as a function of speed v. The measurements of the wavelengths for both the hydroelastic and the gravito-capillary wakes are thus performed simultaneously. Figure S2 shows the evolution of the angular wavenumber k as a function the speed v, for the gravito-capillary wake. By fitting the experimental data to Eq. (S1), one finds −1 γ = 47 mN.m . Note that the last measurement was performed during the characterization of the hydroelastic wake on a sheet of thickness h = 50 μm. Similar measurements were also performed during the characterization of the hydroelastic wake on sheets with larger thicknesses h = {100; 200; 250} μm. However, in those cases, the wavelength for the −1 gravito-capillary wake was measurable only for the lowest speed, v ≈ 0.2 m.s . For all those three measurements, −1 we get γ = 47 mN.m . Considering all the measured values from both methods, we reach the conclusion that −1 γ = 50 ± 10 mN.m . 3 CONTRIBUTIONS OF BENDING, STRETCHING AND GRAVITY Let us consider Eq. (3). The first term in the square root corresponds to bending, the second one to stretching, and the third one to gravity. Using all the measured values for B, and the value of σ, the respective contributions of those three terms, as well as their sum, are computed from Eq. (3) and plotted in Fig. S3. We experimentally measure −1 angular wavenumbers ranging from 400 to 2000 m (see Fig. 4). For the thinnest film h = 50 μm (Fig. S3(a)), all three terms do contribute in that range. The elastic sheet with h = 100 μm shows an interesting behaviour −1 (Fig. S3(b)): at low angular wavenumbers (k < 1000 m ), all three terms are relevant, while bending becomes predominant at larger angular wavenumbers. Finally, for the three largest thicknesses, h = 200, 250, and 350 μm ((Fig. S3(c-e))), bending clearly dominates. 0.4 0.3 0.2 0.1 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 FIG. S3. Contributions of bending, stretching and gravity in Eq. (3). The five panels correspond to different thicknesses (and thus bending moduli B) – from left to right: h = 50, 100, 200, 250, and 350 μm. The blue line corresponds to bending (Bk /ρ), the red line to stretching (σk/ρ), and the green line to gravity (g/k); while the dashed line represents the sum of the three contributions. thomas.salez@u-bordeaux.fr [1] F. Moisy, M. Rabaud, and K. Salsac, Exp. Fluids 46, 1021 (2009). [2] E. Rapha¨el and P.-G. De Gennes, Phys. Rev. E 53, 3448 (1996). http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Condensed Matter arXiv (Cornell University)

Hydroelastic wake on a thin elastic sheet floating on water

Loading next page...
 
/lp/arxiv-cornell-university/hydroelastic-wake-on-a-thin-elastic-sheet-floating-on-water-9H0PX9R8b0
ISSN
2469-990X
eISSN
ARCH-3331
DOI
10.1103/PhysRevFluids.4.014808
Publisher site
See Article on Publisher Site

Abstract

Hydroelastic wake on a thin elastic sheet floating on water 1 2 1 Jean-Christophe Ono-dit-Biot, Miguel Trejo, Elsie Loukiantcheko, 1 2 1, 2 3, 4, ∗ Max Lauch, Elie Rapha¨el, Kari Dalnoki-Veress, and Thomas Salez Department of Physics and Astronomy, McMaster University, 1280 Main Street West, Hamilton, Ontario, L8S 4M1, Canada. Laboratoire de Physico-Chimie Th´eorique, UMR CNRS Gulliver 7083, ESPCI Paris, PSL Research University, 75005 Paris, France. Univ. Bordeaux, CNRS, LOMA, UMR 5798, F-33405 Talence, France. Global Station for Soft Matter, Global Institution for Collaborative Research and Education, Hokkaido University, Sapporo, Hokkaido 060-0808, Japan. (Dated: June 21, 2018) We investigate the hydroelastic wake created by a perturbation moving at constant speed along a thin elastic sheet floating at the surface of deep water. Using a high-resolution cross-correlation imaging technique, we characterize the waves as a function of the perturbation speed, for different sheet thicknesses. The general theoretical expression for the dispersion relation of hydroelastic waves includes three components: gravity, bending and stretching. The bending modulus and the tension in the sheet are independently measured. Excellent agreement is found between the experimental data and the theoretical expression. In 1963, Richard Feynman described water waves in inertia [32–38]. The dispersion relation in the inertial his famous Lectures on Physics [1] as “the worst possi- case was analytically derived and found to depend on ble example [of waves], because they are in no respect three components: gravity, bending and stretching [39]. like sound and light; they have all the complications that A few experimental studies developed in different con- waves can have”. Several decades later, some questions texts have studied the limiting cases where only bending remain unanswered and the study of these waves contin- and stretching [40, 41], or gravity and bending [20, 21], ues to be an area of great interest. For example, Kelvin’s contribute. gravity wake behind a ship [2] still stimulates fundamen- In this Letter, we study the hydroelastic wake cre- tal questions [3, 4]. Surface tension of the liquid-air in- ated by a perturbation moving at constant speed along terface also influences the wave propagation, resulting an elastic sheet floating on deep water. The waves are in gravito-capillary waves and wake [5–8]. Unlike the gravity wake, the capillary wake appears ahead of the perturbation [6]. This is particularly relevant for the lo- (v) comotion of insects [9–13], as well as for nanorheologi- cal applications involving e.g. atomic-force microscopy (i) probes moving along thin viscous samples [14–16]. (iii) Other waves of interest are the ones that propagate on elastic plates and membranes. Their properties are dic- tated by both the bending and stretching rigidities of the material [17]. Floating such an elastic sheet on a liquid (ii) further leads to the coupling of the elastic waves to hydro- dynamics. The resulting hydroelastic waves are of par- ticular interest, as elastic sheets surrounded by fluids are ubiquitous in nature. Examples can be found in fluid me- (iv) chanics [18, 19], geophysics [20–23], and biophysics [24]. Hydroelastic waves are also relevant to practical appli- cations in civil engineering [25, 26], as well as in energy FIG. 1. 3D schematics of the experimental setup. (i) Ro- harvesting through piezoelectric flags [27] and control of tating transparent annular tank (outer radius R = 50 cm, out energy radiation by trucks moving on ice sheets [28]. In- inner radius R = 30 cm) filled with water to a depth of in terestingly, the propagation of such waves can be finely about 16 cm, and with an elastic sheet floating atop. (ii) controlled in an optical-like fashion by using model thin Infrared beam-breaking setup to measure the angular speed sheets with heterogeneous elastic properties [29]. Dif- Ω of the elastic sheet. (iii) A pipette perturbs the surface ferent properties of these waves, such as the wave re- by blowing air, causing a wake to form. (iv) Light sheet and dot pattern used to characterize the waves using the Schlieren sistance or non-linear effects, have been further studied method [42]. (v) A camera is placed ∼ 2 m above the tank to theoretically [30, 31], including the overdamped limit of image the dot pattern. lubrication settings where viscosity dominates over fluid arXiv:1806.07472v1 [physics.flu-dyn] 19 Jun 2018 2 imaged using a high-resolution optical method. By us- ing elastic sheets with different thicknesses, the bending modulus of the sheet is varied over more than two or- ders of magnitude. We find excellent agreement between experimental data and the general theoretical dispersion relation, accounting for the three different contributions: gravity, bending, and stretching. A transparent annular tank is filled with water, as shown in Fig. 1. Thin elastic sheets of Elastosil (Wacker Chemie AG) with nominal thicknesses h of 50, 100, 200, 250 and 350 μm, and lateral dimensions of 20 cm × 16 cm are floated onto the surface of water. A thin rigid plas- tic support (18 cm × 1 cm × 0.1 cm) is placed atop the leading and trailing edge of the elastic sheet to ensure the sheets do not crumple. We experimentally verify that adding the supports does not introduce an anisotropic tension in the sheet, by ensuring that the deformation induced by ball bearings placed atop the sheet is axially symmetric (see SI). The tank is rotated at constant an- −1 gular speed Ω, ranging from 0 to 2.5 rad.s , causing the water to flow and the sheet to move. We take advan- tage of the opaque plastic supports to measure the angu- lar speed of the sheet, using an infrared beam-breaking technique. Because of inertia, both the sheet and water do not follow the tank’s speed instantaneously. Hence, FIG. 2. (a) Raw 2D dot-displacement data measured with all experiments are performed only once the speed of the Ncorr [44], for an elastic sheet of nominal thickness h ≈ −1 200 μm moved at speed v = 0.9 m.s . The displacement sheet is constant and equal to the speed of the tank. vectors are only shown every 10 pixels for clarity. The bot- A glass capillary (World Precision Instruments, USA) tom arrow indicates the speed of the sheet, with respect to the is pulled to a diameter of about 100 μm at one end stationary air jet whose position is indicated by the red ar- with a pipette puller (Narishige, Japan), and used to row. (b) Zoom around the perturbation, corresponding to the blow air at the surface of the sheet (Fig. 1(iii)). The black box in (a). (c) The y-component d of the displacement pipette is placed in the middle of the tank, i.e. at a ra- field. Warm colours (green to orange) correspond to positive dius R = 40 cm from the center. The air jet acts as a displacements, while cold colours (green to blue) correspond to negative ones. (d) Zoom around the perturbation, corre- perturbation moving at speed v = ΩR in the reference sponding to the black box in (c). All scale bars correspond to frame of the elastic sheet, which generates an hydroelas- 1 cm. tic wake. The latter is imaged using a synthetic Schlieren method [42] involving a random dot pattern refracted by the surface topography. The dot pattern is generated us- ing Matlab [42, 43] and printed onto a transparency film. wave pattern of the hydroelastic wake appears clearly, Light shines through the dot pattern (Fig. 1(iv)) and the with a dominant – centimetric – wavelength λ. As the wake, before being collected by a camera above the tank hydroelastic waves propagate along the y-direction, the (Fig. 1(v)). The Schlieren method consists in measuring projection of the displacement along that direction pro- the apparent displacement of the dots due to light re- vides the strongest signal for analysis. Figures 2(c,d) fraction by the wake. The displacement is measured rel- thus focus only on the y-component d of the displacement ative to a reference image of the unperturbed surface (i.e. field. To characterize experimentally the dispersion rela- no air jet) moving at angular speed Ω – which ensures tion of the hydroelastic wake, the wavelength λ is mea- that the collected information is only due to the wake. sured as a function of the speed v. Figures 3(a-c) show This measurement is performed using an open-source dig- the y-component d of the displacement field, for various ital 2D image-correlation algorithm (Ncorr, Matlab) [44]. speeds. The wake is slightly tilted and not symmetric From the displacement of the dots, one can access the about the y-axis, because of the centrifugal force and the slope of the surface and thus the surface topography [42]. surface of water assuming a parabolic profile when ro- Figures 2(a,b) show a typical vectorial displacement tated. This distorsion is avoided in the wavelength mea- field. The air jet creates a localized perturbation in the surement by analyzing the displacement field normal to sheet, as evidence by the large magnitude of the dis- the wave front, as shown in Fig. 3(a,d). We observe both placement field therein. Ahead of the perturbation, in the wavelength and the displacement to decrease as the the reference frame of the elastic sheet, the upstream speed increases. 3 vanishes in the far field and that satisfies Laplace’s equa- 2 2 2 tion [45]: ∇ ϕ = (∇ + ∂ )ϕ = 0. At lowest order in r z the flow (i.e. for small-amplitude hydroelastic waves), the linearized Bernoulli equation for unsteady potential flows provides the excess hydrodynamic pressure exerted on the sheet: P = −ρ ∂ϕ/∂t| − ρgζ, with ρ the liquid z=0 density and g the acceleration due to gravity. To obtain the dispersion relation, one substitutes the expression for P into Eq. (1) in the absence of driv- ing (P = 0), and invokes the kinematic condition ext ∂ϕ/∂z| = ∂ζ/∂t at the water-sheet interface. Con- z=0 sidering a plane wave ϕ ∝ exp[i(ky − ωt − ikz)] satisfy- 0.5 ing Laplace’s equation, with angular wavenumber k and angular frequency ω(k), yields [39]: Bk σk ω = + + gk . (2) ρ ρ -0.5 We now consider the wake created by the driving per- -2 0 2 4 6 8 turbation P traveling at constant speed v along y. ext In the comoving frame of the perturbation, the angular frequency ω of a plane-wave component of the wake is FIG. 3. (a) - (c) y-component d of the displacement field (see shifted by the Doppler effect, and thus reads ω = ω−kv. −1 Fig. 2), for speeds v = 0.4, 0.6, and 0.9 m.s , respectively. Furthermore, since in that comoving frame the wake is All scale bars correspond to 1 cm. (d) y-component d of the stationary, ω = 0 is a necessary condition. Using Eq. (2), displacement field normal to the wave front (see red line in one thus obtains the central relation connecting the an- (a)), as a function of the distance ∆ x from the perturbation, gular wavenumber k and the perturbation speed v, for a for the three speeds as indicated (shifted vertically for clarity). The position of the air-jet perturbation is indicated by the hydroelastic wake on deep water: arrow. s Bk σk g v = + + . (3) ρ ρ k In order to quantify further and rationalize these obser- vations, we now introduce the relevant theoretical frame- An extensive analysis of this relation, similar to the one work. The mechanical system we consider is the thin performed for the gravito-capillary case [7], reveals the elastic sheet in its reference frame. Neglecting the solid main features of the present wake (see also Fig. 4). First, inertia owing to the slenderness of the sheet, the out- below a certain minimal speed v wave propagation is im- of-plane displacement field z = ζ(r, t) with respect to its possible. Secondly, at a given speed v > v there are two flat horizontal state z = 0 satisfies the F¨oppl-von K´arm´an possible values for the observed wavelength: i) the small- equation [17]: est value corresponds to a group velocity that is higher 4 2 than the perturbation speed v, and therefore the waves B∇ ζ − σ∇ ζ = P + P , (1) ext r r propagate upstream of the perturbation. This is the sit- along the 2D horizontal space coordinate r = (x, y) and uation studied in the present work (see Figs. 2 and 3), time t, where ∇ is the nabla operator in 2D, B is the which is dominated by bending and stretching at suffi- bending stiffness of the sheet, and σ represents the ten- ciently large speed; ii) the largest value corresponds to a sion in the sheet. The first and second terms respec- group velocity that is lower than the perturbation speed tively account for bending and stretching. The system v, and therefore the waves propagate downstream of is further subjected to two external forces: the excess the perturbation. This situation corresponds to Kelvin’s hydrodynamic pressure P (r, t) (with respect to the at- classical wake [2], which is dominated by gravity at suf- mospheric one) exerted on the sheet by the water flow ficiently large speed. under gravity, and the driving pressure P (x, y − vt) The values of the bending modulus B and the tension ext modelling the perturbation by the air jet translating at σ are measured independently. The bending modulus 3 2 constant speed v along y. B = Eh /[12(1 − ν )] depends on three parameters: i) The water contribution, P (r, t), is calculated by as- the Young’s modulus E = 1.11±0.06 MPa of Elastosil , suming an incompressible and irrotational flow of an measured using the stress-strain curve; ii) the sheet thick- inviscid fluid, in a semi-infinite half space located at ness h, depending on the sample and measured through z < ζ(r, t). In this context, the fluid velocity field can optical microscopy; iii) the Poisson’s ratio ν = 0.5, as- be written as ∇ϕ, where ϕ(r, z, t) is a potential that suming that Elastosil is an incompressible elastomer. 4 The values of h(μm) and B(N.m) for our five different 2.5 samples are found to be: (h, B) = {[51 ± 1, (1.6 ± 0.1) × −8 −7 10 ], [104 ± 2, (1.4 ± 0.1) × 10 ], [213 ± 7, (1.2 ± 0.1) × −6 −6 10 ], [258 ± 2, (2.1 ± 0.1) × 10 ], [362 ± 3, (5.9 ± 0.3) × −6 10 ]}. We note that the 362 μm film was obtained by stacking two films with nominal thicknesses of 250 μm 1.5 and 100 μm. Besides, as the sheet is freely floating on water, the tension in the sheet is equal to the air-water surface tension: σ = γ. The latter is measured to be −1 γ = 50 ± 10 mN.m , as in [29], from two different techniques (see SI): i) using a Wilhelmy-plate setup; ii) 0.5 characterizing the dispersion relation of gravito-capillary waves on water. The low value of γ and the large un- certainty are attributed to the fact that the tank is filled with an important volume of tap water, and thus subject 0 0.2 0.4 0.6 0.8 1 to contamination. As shown in Fig. 4 (dashed lines), using the above mea- sured values of B and σ, one can predict the evolution of FIG. 4. Angular wavenumber k = 2π/λ as a function of the angular wavenumber k as a function of the perturba- perturbation speed v, for five different sheet thicknesses h as tion speed v. The uncertainties in B and σ are taken into indicated in the legend. Each data point was obtained using account through two limiting predictions and an interval the procedure detailed in Fig. 3 and an average over four in between. Note that the uncertainty on σ accounts different experimental displacement profiles. The error bars correspond to the standard deviation, which is comparable for most of the spread between the two limiting predic- to the marker size. For each thickness h, the blue and red tions. Interestingly, for the values of σ and B considered dashed lines separated by a grey region indicate the upper and here, all three terms in the right-hand side of Eq. (3) are lower theoretical predictions obtained from Eq. (3), using the of the same order of magnitude, especially at low speed independently-measured values of B and σ (see main text) −1 v < 0.4 m.s (see SI). This highlights the counterin- and their uncertainties. tuitive role of gravity in the wavelength selection of the upstream hydroelastic wake. Finally, as expected from Eq. (3), all the theoretical curves for different h (and thus formed by moving a thin elastic sheet, floating on water, B) collapse onto Kelvin’s gravity-dominated branch [2], past a stationary air jet. Specifically, we experimentally at both large v and small k. measured the wavelength of the wake as a function of the Using the experimental procedure detailed above, we perturbation speed, for sheets with bending moduli vary- measure the wavelength λ (see Fig. 3), or equivalently ing over two orders of magnitude. For thin elastic sheets the angular wavenumber k = 2π/λ, as a function of v. (thickness smaller than 100 μm), stretching plays a sig- The results for the five different sheet thicknesses h are nificant role in the propagation of the waves. For thicker shown in Fig. 4 (data points). We find excellent agree- elastic sheets (thickness larger than 100 μm), the bending ment between the experimental data and the theoretical contribution becomes dominant – a regime that is par- predictions, with no adjustable parameter. The experi- ticularly relevant for floating ice [20, 21, 28]. The results mental data points for the two thinnest sheets seem to are found to be in excellent agreement with theoretical be in slightly better agreement with the upper predic- predictions, based on the elasticity of slender structures tion at low speed, and with the lower prediction at high coupled to the hydrodynamics of inviscid incompressible speed. This observation could perhaps be related to a flows, with no adjustable parameter. Interestingly, for slight, but not quantifiable, increase in the sheet tension thin elastic sheets, bending, stretching and gravity all due to the increase in curvature of the air-water interface. contribute to the hydroelastic wake – a result with prac- Another interesting feature of Fig. 4 is that the difference tical consequences in geophysics, biophysics and civil en- gineering. between the upper and lower predictions decreases as the thickness h of the sheet is increased. Indeed, the relative The financial support by the Natural Science and Engi- contribution of bending to Eq. (3) increases, and the dif- neering Research Council of Canada and the Joliot chair ference between both predictions, which is mainly due to of ESPCI Paris is gratefully acknowledged. The authors the uncertainty in tension, decreases. Note that Kelvin’s thank the Global Station for Soft Matter, a project of classical gravity-dominated branch [2] corresponds to: i) Global Institution for Collaborative Research and Edu- a wake propagating behind the perturbation; ii) a wave- cation at Hokkaido University. They are also grateful to length that would almost reach the meter range in our ex- Andreas Koellnberger and Wacker Chemie AG for tech- periments, which is not attainable with the current setup. nical information and the donation of Elastosil films, as In this Letter, we have studied the hydroelastic wake well as to Antonin Eddi, Lucie Domino, Andreas Carlson 5 and Yacine Amarouchene for stimulating discussions. [30] F. Milinazzo, M. Shinbrot, and N. Evans, J. Fluid Mech. 287, 173 (1995). [31] E. I. Pa ˘r˘ au and J.-M. Vanden-Broeck, Phil. Trans. R. Soc. A 369, 2973 (2011). [32] A. Hosoi and L. Mahadevan, Physical review letters 93, 137802 (2004). thomas.salez@u-bordeaux.fr [33] H. Vandeparre, S. Gabriele, F. Brau, C. Gay, K. K. [1] R. P. Feynman, R. B. Leighton, and M. Sands, The Parker, and P. Damman, Soft Matter 6, 5751 (2010). Feynman lectures on physics, Vol. I (Basic books, 2011). [34] J. R. Lister, G. G. Peng, and J. A. Neufeld, Physical [2] L. Kelvin, Proc. Inst. Mech. Eng 3, 409 (1887). review letters 111, 154501 (2013). [3] M. Rabaud and F. Moisy, Phys. Rev. Lett. 110, 214503 [35] T. T. Al-Housseiny, I. C. Christov, and H. A. Stone, (2013). Physical review letters 111, 034502 (2013). [4] A. Darmon, M. Benzaquen, and E. Rapha¨ el, J. Fluid [36] A. Carlson and L. Mahadevan, Mech. Rapids 738 (2014). Physics of Fluids 28, 011702 (2016). [5] L. D. Landau and E. M. Lifshitz, Fluid Mechanics, 2nd [37] M. Arutkin, R. Ledesma-Alonso, T. Salez, and ed. (Pergamon, 1987). E. Rapha¨ el, J. Fluid Mech. 829, 538 (2017). [6] D. J. Acheson, Elementary fluid dynamics (Oxford Uni- [38] O. Kodio, I. M. Griffiths, and D. Vella, versity Press, 1990). Phys. Rev. Fluids 2, 014202 (2017). [7] E. Rapha¨ el and P.-G. de Gennes, Phys. Rev. E 53, 3448 [39] R. M. S. M. Schulkes, R. Hosking, and A. Sneyd, Journal (1996). of Fluid Mechanics 180, 297 (1987). [8] F. Dias and C. Kharif, Annu. Rev. Fluid Mech. 31, 301 [40] L. Deike, J.-C. Bacri, and E. Falcon, J. Fluid Mech. 733, (1999). 394 (2013). [9] D. L. Hu, B. Chan, and J. W. Bush, Nature 424, 663 [41] L. Deike, M. Berhanu, and E. Falcon, Phys. Rev. Fluids (2003). 2, 064803 (2017). [10] A. Chepelianskii, F. Chevy, and E. Rapha¨ el, Phys. Rev. [42] F. Moisy, M. Rabaud, and K. Salsac, Exp. Fluids 46, Lett. 100, 074504 (2008). 1021 (2009). [11] F. Closa, A. Chepelianskii, and E. Rapha¨ el, Phys. Fluids [43] F. Moisy and M. Rabaud, 22, 052107 (2010). “Free-surface synthetic schlieren (fs-ss): A tutorial,” . [12] J. Voise and J. Casas, J. Royal Soc. Interface 7, 343 [44] J. Blaber, B. Adair, and A. Antoniou, Exp. Mech. 55, (2010). 1105 (2015). [13] T. Steinmann, M. Arutkin, P. Cochard, [45] G. K. Batchelor, An Introduction to Fluid Dynam- E. Rapha¨ el, J. Casas, and M. Benzaquen, ics (Cambridge University Press, Cambridge, England, Journal of Fluid Mechanics 848, 370 (2018). 1967). [14] K. Wedo lowski and M. Napi´ orkowski, Soft Matt. 11, 2639 (2015). [15] R. Ledesma-Alonso, M. Benzaquen, T. Salez, and E. Rapha¨ el, J Fluid Mech. 792, 829 (2016). [16] R. Ledesma-Alonso, E. Raphael, T. Salez, P. Tordjeman, and D. Legendre, Soft Matter 13, 3822 (2017). [17] L. D. Landau and E. M. Lifshitz, Theory of Elasticity, Vol. 7, 3rd ed. (Elsevier, New York, 198). [18] M. J. Shelley and J. Zhang, Annu. Rev. Fluid Mech. 43, 449 (2011). [19] E. Virot, X. Amandolese, and P. H´ emon, J. Fluid. Struct. 43, 385 (2013). [20] T. Takizawa, Cold regions Science and Technology 11, 171 (1985). [21] V. Squire, W. Robinson, P. Langhorne, and T. Haskell, Nature 333, 159 (1988). [22] V. A. Squire, J. P. Dugan, P. Wadhams, P. J. Rottier, and A. K. Liu, Annu. Rev. Fluid Mech. 27, 115 (1995). [23] E. Pa ˘r˘ au and F. Dias, J. Fluid Mech. 460, 281 (2002). [24] J. B. Grotberg and O. E. Jensen, Annu. Rev. Fluid Mech. 36 (2004). [25] M. Murai, H. Kagemoto, and M. Fujino, Journal of ma- rine science and technology 4, 123 (1999). [26] E. Watanabe, T. Utsunomiya, and C. Wang, Engineering structures 26, 245 (2004). [27] D. T. Akcabay and Y. L. Young, Physics of Fluids 24, 054106 (2012). [28] J. W. Davys, R. J. Hosking, and A. D. Sneyd, Journal of Fluid Mechanics 158, 269?287 (1985). [29] L. Domino, M. Fermigier, E. Fort, and A. Eddi, EPL (Europhysics Letters) 121, 14001 (2018). Supplemental Information for: “Hydroelastic wake on a thin elastic sheet floating on water” 1 2 1 Jean-Christophe Ono-dit-Biot, Miguel Trejo, Elsie Loukiantcheko, 1 2 1, 2 3, 4, ∗ Max Lauch, Elie Rapha¨el, Kari Dalnoki-Veress, and Thomas Salez Department of Physics and Astronomy, McMaster University, 1280 Main Street West, Hamilton, Ontario, L8S 4M1, Canada. Laboratoire de Physico-Chimie Th´eorique, UMR CNRS Gulliver 7083, ESPCI Paris, PSL Research University, 75005 Paris, France. Univ. Bordeaux, CNRS, LOMA, UMR 5798, F-33405 Talence, France. Global Station for Soft Matter, Global Institution for Collaborative Research and Education, Hokkaido University, Sapporo, Japan. (Dated: June 21, 2018) TENSION ISOTROPY (a) yy (b) σ σ xx xx yy FIG. S1. (a) Top-view schematic of the elastic sheet floating on water. Plastic beams are placed at the leading and trailing edges of the elastic sheet to ensure the sheet does not crumple. Ball bearings are dropped atop the sheet to verify that no anisotropy is introduced when placing the beams. (b) Picture of the reference dot pattern, seen through the water and the sheet, around a ball bearing. The solid lines are sample isodisplacement lines. The dashed lines are the best fits of the isodisplacement lines to ellipses. The best-fit ellipticities are found to be equal to 1 for all cases. A schematic of the elastic sheet is shown in Fig. S1(a). The tensions along the x- and y-axes are denoted σ xx and σ , respectively. We place ball bearings directly on the sheet floating on water, and we image the resulting yy deformation using the optical Schlieren method [1]. We then calculate the magnitude of the displacement vector field, which is directly linked to the deformation of the elastic sheet. Sample isodisplacement lines are shown in Fig. S1(b). We quantify the anisotropy of the deformation by fitting the isodisplacement lines to ellipses. The best-fit ellipticities for the four cases shown in Fig. S1(b) are all found to be equal to 1, meaning that the isodisplacement lines are circles, which thus indicates that the tension in the sheet is isotropic. Indeed, if σ or σ was larger than the other, the xx yy deformation would be elongated along the low-tension direction, leading to an ellipticity larger than 1. TENSION MEASUREMENT The tension σ along the x-axis is set by the water-air surface tension γ, as both the left and right edges are free (see xx Fig. S1(a)). Since all the experiments presented in the study are performed on sheets where the tension is isotropic, one can safely assume that σ = σ = σ = γ. The tabulated value for the pure water-air surface tension under yy xx −1 ambient conditions is γ = 72 mN.m , but it is extremely sensitive to contamination by all kinds of surfactants. The arXiv:1806.07472v1 [physics.flu-dyn] 19 Jun 2018 2 experiments being conducted in an open tank containing ∼ 200 L of water, contamination is unavoidable. Therefore, γ was measured from two independent methods. −1 First, using a Wilhelmy-plate setup, γ was found to be between 40 and 55 mN.m , for water from three different sources: water from the tank after one day, tap water, and deionized water. The largest value of γ was obtained for deionized water, and the smallest one for the water from the tank – which is consistent with tank contamination over time. 0 0.2 0.4 0.6 0.8 1 FIG. S2. Angular wavenumber k as a function of perturbation speed v, for the gravito-capillary wake at the surface of deep water. The dashed line shows the best fit of Eq. (S1) to the experimental data points. The water-air surface tension −1 γ = 47 mN.m is obtained as the only adjustable parameter. Another approach to measure γ is to invoke the gravito-capillary wake formed at the surface of deep water by a perturbation moving at constant speed v. In such a case, the analogue of Eq. (3) is [2]: γk g v = + . (S1) ρ k Therefore, γ can be evaluated by fitting Eq. (S1) to the experimental evolution of the angular wavenumber k as a function of speed v for a gravito-capillary wake. In fact, as the elastic sheet only covers a small portion of the water in the tank, the hydroelastic wake is only observed once a lap when the sheet moves across the stationary perturbation at speed v. Otherwise, water flowing at speed v is directly exposed to the perturbation, and a gravito-capillary wake is instead formed at the surface. Using the Schlieren method, the wavelength λ = 2π/k of the upstream gravito-capillary wake is measured as a function of speed v. The measurements of the wavelengths for both the hydroelastic and the gravito-capillary wakes are thus performed simultaneously. Figure S2 shows the evolution of the angular wavenumber k as a function the speed v, for the gravito-capillary wake. By fitting the experimental data to Eq. (S1), one finds −1 γ = 47 mN.m . Note that the last measurement was performed during the characterization of the hydroelastic wake on a sheet of thickness h = 50 μm. Similar measurements were also performed during the characterization of the hydroelastic wake on sheets with larger thicknesses h = {100; 200; 250} μm. However, in those cases, the wavelength for the −1 gravito-capillary wake was measurable only for the lowest speed, v ≈ 0.2 m.s . For all those three measurements, −1 we get γ = 47 mN.m . Considering all the measured values from both methods, we reach the conclusion that −1 γ = 50 ± 10 mN.m . 3 CONTRIBUTIONS OF BENDING, STRETCHING AND GRAVITY Let us consider Eq. (3). The first term in the square root corresponds to bending, the second one to stretching, and the third one to gravity. Using all the measured values for B, and the value of σ, the respective contributions of those three terms, as well as their sum, are computed from Eq. (3) and plotted in Fig. S3. We experimentally measure −1 angular wavenumbers ranging from 400 to 2000 m (see Fig. 4). For the thinnest film h = 50 μm (Fig. S3(a)), all three terms do contribute in that range. The elastic sheet with h = 100 μm shows an interesting behaviour −1 (Fig. S3(b)): at low angular wavenumbers (k < 1000 m ), all three terms are relevant, while bending becomes predominant at larger angular wavenumbers. Finally, for the three largest thicknesses, h = 200, 250, and 350 μm ((Fig. S3(c-e))), bending clearly dominates. 0.4 0.3 0.2 0.1 0 1 2 0 1 2 0 1 2 0 1 2 0 1 2 FIG. S3. Contributions of bending, stretching and gravity in Eq. (3). The five panels correspond to different thicknesses (and thus bending moduli B) – from left to right: h = 50, 100, 200, 250, and 350 μm. The blue line corresponds to bending (Bk /ρ), the red line to stretching (σk/ρ), and the green line to gravity (g/k); while the dashed line represents the sum of the three contributions. thomas.salez@u-bordeaux.fr [1] F. Moisy, M. Rabaud, and K. Salsac, Exp. Fluids 46, 1021 (2009). [2] E. Rapha¨el and P.-G. De Gennes, Phys. Rev. E 53, 3448 (1996).

Journal

Condensed MatterarXiv (Cornell University)

Published: Jun 19, 2018

There are no references for this article.