betaFeedback

WorksPhysics and Astronomy

Fluid-structure coupling governs a dynamic transition in foam scraping

Abstract

Foam scraping exhibits a dynamic transition between a slip state, in which the foam moves beneath a plate, and a scraping state, in which the foam is expelled from the confined region. Although this transition has been associated with the propagation of local T1 rearrangements, the physical parameter controlling their propagation remains unclear. Here, we investigate the dependence of the critical scraping velocity on the liquid fraction, bulk viscosity, and surfactant system. For all examined conditions, the critical capillary number follows Cac∝ϕ−1\mathrm{Ca}_c\proptoϕ^{-1}, apart from a solution-dependent prefactor. We propose a local fluid–structure-coupling model in which viscous work transmitted through the Plateau-border network competes with the effective energetic cost required for one T1 event to trigger the next. The observed scaling implies an effective energetic cost governed by the Laplace pressure and is consistent with a small local compressive component of the bubble deformation. These results identify local coupling between interstitial flow and bubble deformation as a mechanism controlling the macroscopic slip–scraping transition.

Introduction

Foams are soft jammed materials in which gas bubbles are densely packed within a small amount of liquid. Because the dispersed elements are deformable, foams belong to a broader class of soft jammed systems that also includes emulsions and biological tissues. Their combination of structural rigidity and local deformability gives rise to a characteristic set of properties, including elasticity, liquid uptake, thermal insulation, and suppression of mass transport [1, 2]. These properties underlie the widespread use of foams in foods, fire-extinguishing agents, cosmetics, detergents, and insulating materials.

In many practical uses of foams, external forces are applied to induce large deformation [3, 4]. Examples include controlling food texture, ejecting fire-extinguishing foams, and spreading cosmetic or detergent foams over a surface. For such two-component systems, pioneering studies by Tanaka and co-workers have demonstrated that the rheological response can be governed by the dynamical coupling between the dispersed structure and the surrounding fluid [5, 6, 7, 8]. The mechanical properties of foams, including their yield stress, elasticity, and viscoelastic response, have been studied extensively [9, 10, 11, 12, 13]. Nevertheless, much less is understood about how the dynamical coupling between the bubble structure and the interstitial liquid selects a macroscopic deformation mode under large-scale forcing.

Recent experiments on foam scraping have revealed non-monotonic deformation patterns that cannot be explained solely by the mean shear rate or the macroscopic yield stress [14, 15]. On a hydrophilic substrate, the foam slips beneath the plate at low scraping velocities, whereas at higher velocities it is expelled from beneath the plate, resulting in a scraping state. The critical velocity UcU_{\mathrm{c}} decreases as the gap distance increases, showing that the transition does not occur at a constant shear rate. Furthermore, UcU_{\mathrm{c}} also depends on the lateral width of the foam, providing further evidence that the transition cannot be described by a simple local shear-rate criterion [15].

The boundary between the slip and scraping states was suggested to exhibit a dynamic critical phenomenon belonging to the directed-percolation universality class [16, 17, 18]. In the slip state, a local rearrangement (T1 event) can trigger subsequent rearrangements in the surrounding foam, leading to spatial propagation of T1 activity. However, the physical parameter that determines whether a T1 event propagates, and therefore controls the transition from slip to scraping, remains unknown.

In this study, we show that the critical capillary number for the slip–scraping transition follows

Cac∼ϕ−1,(1)\mathrm{Ca}_{\mathrm{c}} \sim\phi^{-1}, \tag*{(1)}

where ϕ\phi is the liquid fraction. To explain this scaling, we propose a local fluid–structure-coupling model in which the propagation of a T1 event is determined by the balance between viscous work in the Plateau-border network and the energy required for a small local compression of the bubbles near the rearranging region. This result identifies the interstitial liquid as an active component that selects the macroscopic deformation mode of the foam. The present framework may also provide insight into how local fluid–structure coupling controls macroscopic deformation in other soft jammed systems, including emulsions and biological tissues.

Materials and Methods

Foams were prepared using aqueous solutions of three surfactant systems: 5.0 wt% tetradecyltrimethylammonium bromide (TTAB), sodium dodecyl sulfate (SDS), both purchased from FUJIFILM Wako Pure Chemical Corporation, Japan, and a commercial household detergent (Charmy, Lion Corp., Japan). The TTAB concentration was well above its critical micelle concentration. (∼0.1\sim0.1 wt%) [19]. The surface tensions of the TTAB, SDS, and Charmy solutions were measured using a surface tensiometer (DY300, Kyowa Interface Science Co., Ltd., Japan) and were 37, 34, and 25 mN/m, respectively. Previous measurements showed that the drainage times of foams prepared with these three surfactant systems were comparable and that their Boussinesq numbers were approximately Bo=ηs/ηr≃2\mathrm{Bo} = \eta_s/\eta_r \simeq2, where ηs\eta_s is an interfacial viscosity, rr is the Plateau-border radius, and η\eta is a bulk viscosity [20]. Thus, their interfaces have comparable mechanical properties. The bulk viscosity was varied by adding glycerol to the surfactant solution. Within the experimental uncertainty of approximately 0.1 mN/m, the surface tension was independent of the glycerol concentration. The equilibrium contact angle on an acrylic substrate was θE=23±2∘\theta_E = 23 \pm2^{\circ} and was also independent of the glycerol concentration.

Foams were generated using a foam dispenser manufactured by Awahour (Torigoe Plastic Industry Co., Ltd., Japan). For each experiment, a portion of the prepared foam was sampled, and its liquid fraction ϕ\phi was determined from ϕ=m/ρVsample\phi= m/\rho V_{\mathrm{sample}}, where mm is the mass of the foam sample, ρ\rho is the density of the liquid phase, and VsampleV_{\mathrm{sample}} is the total volume of the foam sample. The uncertainty in the measured liquid fraction was approximately ±1%\pm1\%. The mean bubble radius was R=0.22R = 0.22 mm, with a standard deviation of 0.082 mm. No appreciable drainage was observed during the experimental period, which was at most 20 s.

The experimental setup is shown in Figure 1. An acrylic plate was positioned perpendicular to the substrate, forming a gap of width b=1.5b = 1.5 mm. The confinement length was set to L=5.0L = 5.0 mm. To construct the foam-scraping state diagram, a hemispherical foam sample with a diameter W=20W = 20 mm was placed on the substrate. The substrate was translated horizontally at a constant velocity in the range U=1.0U = 1.0–35.035.0 mm/s using an electric slider (LTS300/M, Thorlabs, Inc., USA). The scraping dynamics were recorded from above using a video camera (EOS R, Canon Inc., Japan).

Schematic of the experimental setup

Figure 1. Schematic of the experimental setup for foaming. An acrylic plate was positioned perpendicular to the substrate, forming a gap of width b=1.5b = 1.5 mm and a confinement length L=5.0L = 5.0 mm. A hemispherical foam sample with a diameter W=20W = 20 mm was placed on the substrate, which was translated horizontally at a constant velocity UU.

Results and Discussion

Liquid-fraction dependence of the transition

The transition from slip to scraping can be described in terms of a directed percolation process, in which an isolated T1 event propagates to neighboring regions and develops into a continuous sequence of rearrangements through the foam [14, 15]. Within this framework, the critical point is determined by the probability that one T1 event triggers the next.

In foams, the liquid fraction is known to be a key parameter controlling the occurrence of T1 events [10, 21]. We therefore examined the liquid-fraction dependence of the critical scraping velocity UcU_c. The results for foams prepared with SDS, Charmy, and TTAB are shown in

Image not extracted

Figure 2. For all three surfactant systems, the transition occurred at velocities of the same order of magnitude, and the data were well described by Uc∝ϕ−αU_c \propto\phi^{-\alpha}. The fitted exponents were α=1.17±0.22\alpha= 1.17 \pm0.22 for SDS, α=0.93±0.11\alpha= 0.93 \pm0.11 for Charmy, and α=1.16±0.18\alpha= 1.16 \pm0.18 for TTAB. The uncertainties mainly originate from variations in the liquid fraction among independently prepared foam samples. In all three systems, the fitted exponent is consistent with α=1\alpha= 1, indicating

Uc∝ϕ−1.(2)U_c \propto\phi^{-1}. \tag*{(2)}

Capillary-number scaling

The capillary number, Ca=ηU/γ\mathrm{Ca} = \eta U/\gamma is a fundamental dimensionless parameter for describing foam spreading. To examine the viscosity dependence of the slip–scraping transition, we used the three surfactant systems and additionally prepared TTAB solutions containing glycerol. Figure 3 shows the liquid-fraction dependence of the critical capillary number normalized by the solution-dependent prefactor kk. After normalization by a solution-dependent prefactor kk, all data collapse onto the relation

Critical capillary number normalized by the solution-dependent prefactor as a function of liquid fraction

Figure 3. Liquid-fraction dependence of the critical capillary number normalized by the solution-dependent prefactor kk. Results are shown for SDS, Charmy, TTAB, and TTAB solutions containing 40 and 50 wt% glycerol. The solid line represents Cac/k=ϕ−1Ca_c/k = \phi^{-1}. Horizontal error bars represent the uncertainty in the measured liquid fraction, whereas vertical error bars are calculated from the uncertainty in the critical velocity.

Cac=kϕ−1,(3)\mathrm{Ca}_c = k\phi^{-1}, \tag*{(3)}

including those obtained at different bulk viscosities. The values of the prefactor kk are summarized in Figure 4. They range from 4.6×10−54.6 \times10^{-5} to 7.0×10−57.0 \times10^{-5} and are therefore of the same order of magnitude for all solutions. Nevertheless, kk varies with glycerol concentration even within the TTAB system. The possible origin of this variation is discussed in § 2 C.

Prefactor $k$ obtained for each solution

Figure 4. Prefactor kk obtained from Cac=kϕ−1Ca_c = k\phi^{-1} for each solution. The values of kk remain of the same order for all surfactant systems, while a systematic variation is observed with glycerol concentration in the TTAB solutions. Error bars represent the uncertainty obtained from the fitting.

Critical scraping velocity as a function of liquid fraction for foams prepared with SDS, Charmy, and TTAB

Figure 2 (PDF p. 3). Liquid-fraction dependence of the critical scraping velocity UcU_c for foams prepared with SDS, Charmy, and TTAB. The solid lines represent fits to Uc∝ϕ−αU_c \propto\phi^{-\alpha}. The fitted exponents are α=1.17±0.22\alpha= 1.17 \pm0.22 for SDS, α=0.93±0.11\alpha= 0.93 \pm0.11 for Charmy, and α=1.16±0.18\alpha= 1.16 \pm0.18 for TTAB. The background colors indicate the slip and scraping regimes. Horizontal error bars represent the uncertainty in the measured liquid fraction, whereas vertical error bars indicate the velocity increment of 0.5 mm/s used to determine the transition.

Model

We next consider the physical mechanism that determines the transition from the slip regime to the scraping regime. During scraping, the relative motion between the bubble structure and the liquid in the Plateau border network generates a viscous stress. The characteristic velocity gradient is estimated as U/rU/r, where UU is the scraping velocity and rr is the characteristic radius of the Plateau border. The corresponding viscous stress is therefore

σvis∼ηUr,(4)\sigma_{\mathrm{vis}} \sim\eta\frac{U}{r}, \tag*{(4)}

where η\eta is the liquid viscosity.

We assume that this stress acts over a cross-sectional area of order r2r^2. The force generated through the Plateau-border network is then transmitted over a bubble-scale distance of order aa, where aa is the characteristic bubble radius. The viscous work available for inducing a neighboring rearrangement is consequently estimated as

Wvis∼CW(ηUr)r2a=CWηUra.(5)W_{\mathrm{vis}} \sim C_W \left(\eta\frac{U}{r}\right) r^2 a = C_W \eta U r a. \tag*{(5)}

Here, CWC_W is a dimensionless coefficient describing the efficiency with which viscous stresses in the Plateau border network are transmitted to the bubble structure. Its value is expected to depend on the mobility of the interface characterized by the Boussinesq number Bo=ηs/ηr\mathrm{Bo} = \eta_s/\eta_r [2]. A large Bo\mathrm{Bo} corresponds to a relatively immobile interface, whereas a smaller Bo\mathrm{Bo} corresponds to a more mobile, or softer, interface.

At the transition, the viscous work is assumed to become comparable to the effective energetic cost required for the propagation of a T1 event,

Wvis∼ET1eff.(6)W_{\mathrm{vis}} \sim E_{\mathrm{T1}}^{\mathrm{eff}}. \tag*{(6)}

The superscript “eff” emphasizes that this quantity is not the equilibrium energy barrier of an isolated T1 event. It represents the energetic cost for one rearrangement to induce a subsequent rearrangement under the externally driven and spatially asymmetric conditions produced by scraping.

Using the definition Ca=ηU/γ\mathrm{Ca} = \eta U/\gamma, where γ\gamma is the surface tension, Eqs. (5) and (6) give, apart from a dimensionless prefactor,

ET1eff∼γCacra,(7)E_{\mathrm{T1}}^{\mathrm{eff}} \sim\gamma\mathrm{Ca}_{c} r a, \tag*{(7)}

where Cac\mathrm{Ca}_{c} is the critical capillary number. For a dry foam, the characteristic Plateau-border radius scales as

r∼aϕ1/2,(8)r \sim a\phi^{1/2}, \tag*{(8)}

where ϕ\phi is the liquid fraction. Combining Eq. (8) with the experimentally observed relation Cac∼ϕ−1\mathrm{Ca}_{c} \sim\phi^{-1} yields

ET1eff∼γa2ϕ−1/2.(9)E_{\mathrm{T1}}^{\mathrm{eff}} \sim\gamma a^{2}\phi^{-1/2}. \tag*{(9)}

yields

ET1eff∼γa2ϕ−1/2.(10)E_{\mathrm{T1}}^{\mathrm{eff}} \sim\gamma a^{2}\phi^{-1/2}. \tag*{(10)}

Thus, the experimental scaling directly determines the liquid-fraction dependence of the effective energetic cost for T1 propagation. In particular, the energy required to propagate rearrangement activity increases as the foam becomes drier.

Using r∼aϕ1/2r \sim a\phi^{1/2}, Eq. (10) can equivalently be written as

ET1eff∼γra3.(11)E_{\mathrm{T1}}^{\mathrm{eff}} \sim\frac{\gamma}{r}a^{3}. \tag*{(11)}

Because γ/r\gamma/r represents the characteristic Laplace pressure, this scaling has the form of pressure–volume work. We therefore write

ET1eff∼PLΔVeff,PL∼γr,ΔVeff=CVa3,(12)E_{\mathrm{T1}}^{\mathrm{eff}} \sim P_{\mathrm{L}}\Delta V_{\mathrm{eff}}, \qquad P_{\mathrm{L}} \sim\frac{\gamma}{r}, \qquad\Delta V_{\mathrm{eff}} = C_{\mathrm{V}}a^{3}, \tag*{(12)}

where CVC_{\mathrm{V}} is a dimensionless measure of the effective volumetric deformation associated with T1 propagation, and the experimental prefactor kk gives k=CV/CWk = C_{\mathrm{V}}/C_{\mathrm{W}}. Thus, the relation

ET1eff∝γra3,(13)E_{\mathrm{T1}}^{\mathrm{eff}} \propto\frac{\gamma}{r}a^{3}, \tag*{(13)}

suggests that a small local compression of the bubbles is a governing parameter for a T1 event under scraping. The prefactor kk remains of the same order for TTAB, SDS, and Charmy. This is consistent with the similar Boussinesq numbers reported for these surfactant systems, suggesting that their efficiencies of viscous-stress transmission, represented by CWC_{\mathrm{W}}, are also comparable. A systematic change in kk is nevertheless observed when glycerol is added to the TTAB solution. The bulk viscosity η\eta then increases substantially, whereas the interfacial viscosity ηs\eta_s is not expected to increase by the same factor. The resulting decrease in Bo\mathrm{Bo} makes the interface more mobile and is expected to reduce CWC_{\mathrm{W}}. This interpretation is consistent with the observed increase in kk with glycerol concentration.

Local compression as a trigger for T1 propagation

The experimentally observed scaling suggests that the propagation of a T1 event involves a small compressive component. If CWC_{\mathrm{W}} is of order unity, the estimated volume change is

ΔVeffa3∼6.0×10−5.(14)\frac{\Delta V_{\mathrm{eff}}}{a^{3}} \sim6.0 \times10^{-5}. \tag*{(14)}

This corresponds to a characteristic linear scale

δla∼(ΔVeffa3)1/3∼3.9×10−2,(15)\frac{\delta l}{a} \sim\left(\frac{\Delta V_{\mathrm{eff}}}{a^{3}}\right)^{1/3} \sim3.9 \times10^{-2}, \tag*{(15)}

so that the inferred local deformation is of the order of a few percent of the bubble size. To obtain an order-of-magnitude comparison, we estimate the transient increase in surface-energy cost as

EA∼γΔA∼γδl2∼1.5×10−3γa2.(16)E_{\mathrm{A}} \sim\gamma\Delta A \sim\gamma\delta l^{2} \sim1.5 \times10^{-3}\gamma a^{2}. \tag*{(16)}

By contrast, the compressive work is

EV∼ΔPLΔVeff∼γrΔVeff∼6.0×10−5arγa2.(17)E_{\mathrm{V}} \sim\Delta P_{\mathrm{L}}\Delta V_{\mathrm{eff}} \sim\frac{\gamma}{r}\Delta V_{\mathrm{eff}} \sim6.0 \times10^{-5}\frac{a}{r}\gamma a^{2}. \tag*{(17)}

Using r/a∼1.7ϕ1/2r/a \sim1.7\phi^{1/2} [1, 2] and ϕ≃0.1\phi\simeq0.1, we obtain

EV∼1.1×10−4γa2,(18)E_{\mathrm{V}} \sim1.1 \times10^{-4}\gamma a^{2}, \tag*{(18)}

which is approximately one order of magnitude smaller than EAE_{\mathrm{A}}. This order-of-magnitude estimate indicates that the compressive pathway can be energetically more favorable than a volume-preserving deformation accompanied by a comparable transient increase in interfacial area. This interpretation is also plausible for a jammed foam. Even a small local volume change can generate an elastic heterogeneity in the otherwise rigid bubble network and thereby facilitate localized bubble motion. The local compression may therefore act as a trigger for the propagation of the next T1 event.

Conclusion

Foam scraping exhibits a dynamic transition between a slip state, in which the foam passes beneath the plate, and a scraping state, in which it is expelled from the confined region [14]. Previous studies have shown that this transition cannot be explained solely by the mean shear rate or the macroscopic yield stress and have suggested that it is controlled by the propagation of local T1 rearrangements [15]. However, the physical parameter governing the propagation of these rearrangements has remained unclear. To address this problem, we examined the effects of liquid fraction, bulk viscosity, and surfactant system on the critical scraping velocity. For all conditions examined, the critical capillary number followed Cac=kϕ−1\mathrm{Ca}_{\mathrm{c}}=k\phi^{-1}, where kk is a solution-dependent prefactor. This scaling demonstrates that the interstitial liquid plays an active role in determining whether a local rearrangement propagates through the foam.

We proposed a local fluid–structure-coupling model in which viscous work generated in the Plateau-border network is transmitted to the bubble structure and competes with the effective energetic cost required for one T1 event to trigger the next. The experimentally observed scaling implies that this cost is governed by the Laplace pressure and is consistent with a small local compressive deformation near the rearranging region. Such local compression may generate an elastic heterogeneity in the jammed bubble network and facilitate the propagation of subsequent T1 events.

These results identify local coupling between interstitial flow and bubble deformation as a mechanism controlling the macroscopic slip–scraping transition. Because deformation in emulsions, biological tissues, and other soft jammed materials is also mediated by localized rearrangements, the present framework may provide a broader basis for understanding how fluid–structure coupling selects macroscopic deformation modes in multicomponent soft matter systems.

Acknowledgments

R. K. was supported by a Grant-in-Aid for Scientific Research (B) (Grant No. 26K00676) from the JSPS.

Author Contributions

RK conceived the project, and ME performed the experiments. RK developed the model for the transition mechanism and wrote the manuscript.

Data Availability

All data supporting the findings of this study are included within the article.

Competing Interests

The authors declare that they have no competing financial interests.

Correspondence

Correspondence and requests for materials should be addressed to R.K. ([email protected]).

References

  1. [1]D. L. Weaire and S. Hutzler, The physics of foams (Oxford University Press, 2001).DOI
  2. [2]I. Cantat, S. Cohen-Addad, F. Elias, F. Graner, R. Höhler, O. Pitois, F. Rouyer, and A. Saint-Jalmes, Foams: structure and dynamics (OUP Oxford, 2013).DOI
  3. [3]M. Marchand, F. Restagno, E. Rio, and F. Boulogne, Roughness-induced friction on liquid foams, Phys. Rev. Lett. 124, 118003 (2020).DOI
  4. [4]A. Deblais, R. Harich, D. Bonn, A. Colin, and H. Kellay, Spreading of an oil-in-water emulsion on a glass plate: Phase inversion and pattern formation, Langmuir 31, 5971 (2015).DOI
  5. [5]T. Araki and H. Tanaka, Dynamics of colloidal particles in soft matters, Progress of Theoretical Physics Supplement 175, 37 (2008).DOI
  6. [6]A. Furukawa and H. Tanaka, Violation of the incompressibility of liquid by simple shear flow, Nature 443, 434 (2006).
  7. [7]T. Yanagishima, Y. Liu, H. Tanaka, and R. P. A. Dullens, Particle-level visualization of hydrodynamic and frictional couplings in dense suspensions of spherical colloids, Phys. Rev. X 11, 021056 (2021).DOI
  8. [8]M. Tateno, J. Yuan, and H. Tanaka, The impact of colloid-solvent dynamic coupling on the coarsening rate of colloidal phase separation, Journal of Colloid and Interface Science 684, 21 (2025).DOI
  9. [9]M. L. Merrer, S. Cohen-Addad, and R. Höhler, Bubble rearrangement duration in foams near the jamming point, Phys. Rev. Lett. 108, 188301 (2012).
  10. [10]Y. Furuta, N. Oikawa, and R. Kurita, Close relationship between a dry-wet transition and a bubble rearrangement in two-dimensional foam, Sci. Rep. 6, 37506 (2016).DOI
  11. [11]G. Katgert, B. P. Tighe, and M. van Hecke, The jamming perspective on wet foams, Soft Matter 9, 9739 (2013).DOI
  12. [12]A. O. Siemens and M. Van Hecke, Jamming: A simple introduction, Physica A: Statistical Mechanics and its Applications 389, 4255 (2010).
  13. [13]S. Cohen-Addad, R. Höhler, and O. Pitois, Flow in foams and flowing foams, Annual Review of Fluid Mechanics 45, 241 (2013).
  14. [14]M. Endo, M. Tani, and R. Kurita, Scraping of foam on a substrate, J. Colloid and Interface Sci. 650, 1612 (2023).DOI
  15. [15]M. Endo and R. Kurita, Critical-like behavior in foam dynamics: Transition from slip to scraping, Phys. Rev. Res. 7, 023013 (2025).DOI
  16. [16]M. Sano and K. Tamai, A universal transition to turbulence in channel flow, Nat. Phys. 12, 249 (2016).DOI
  17. [17]H. Hinrichsen, Non-equilibrium critical phenomena and phase transitions into absorbing states, Adv. in Phys. 49, 815 (2000).
  18. [18]K. A. Takeuchi, M. Kuroda, H. Chaté, and M. Sano, Experimental realization of directed percolation criticality in turbulent liquid crystals, Phys. Rev. E 80, 051116 (2009).DOI
  19. [19]K. D. Danov, P. A. Kralchevsky, and K. P. Ananthapadmanabhan, Micelle–monomer equilibria in solutions of ionic surfactants and in ionic–nonionic mixtures: A generalized phase separation model, Adv. Colloid Interface Sci. 206, 17 (2014).DOI
  20. [20]A. Kaneda and R. Kurita, Absorptive limits of foams governed by kinematic coupling between solution and bubbles, J. Colloid and Interface Sci. 695, 137746 (2025).DOI
  21. [21]R. Kurita, Y. Furuta, N. Yanagisawa, and N. Oikawa, Dynamical transition in a jammed state of a quasi-two-dimensional foam, Physical Review E 95, 062613 (2017).DOI

Paper details

vlnc:W_yh9nrkxb
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents