A new setup for giant soap films characterization
Sandrine Mariot, Marina Pasquet, Vincent Klein, Frédéric Restagno, Emmanuelle Rio
AA new setup for giant soap films characterization
Sandrine Mariot ∗ , Marina Pasquet, Vincent Klein, Frédéric Restagno, Emmanuelle Rio Université Paris-Saclay, CNRS, Laboratoire de physique des solides, 91405 Orsay
February 16, 2021
Abstract
Artists, using an empirical knowledge, manage to generate and play with giant soap films and bubbles. Untilnow, scientific studies of soap films generated at a controlled velocity and without any feeding from the top, studiedfilms of a few square centimeters. The present work aims to present a new setup to generate and characterize giantsoap films (2 m × Artists combine formulation and practice to generate gi-ant soap films and play with their shapes and bright col-ors [1]. They manage thereby to generate huge bubbles,up to several meters high: the world record for the tallestfree-standing soap bubble is 7.004 meters high [2]. Suchbig and fragile objects are not only fascinating, they alsoraise specific scientific questions important for industrialaspects, as well as for the global comprehension of thephenomenon: what is the role of physical chemistry onthe stability of the films? What are the role of inertiaand gravity on the foam film generation and stability?Up to now, most experimental studies have beenfocused on soap films generation at small scales, upto less than 5 cm high, as we can see in Fig. 1[5, 10, 26, 25, 8, 3, 27, 3]. These soap films were generatedat low velocities, less than a few centimetres per second.Lionti-Addad et al. [18], Cohen-Adad et al. [10], Adelizzi et al. [5] and Berg et al. [8] have studied the withdrawalof entrained films with different surfactants and hydro-soluble polymers. They have measured film thicknesses h by light reflectivity, in order to investigate the area ofvalidity of Frankel’s law. For a solution of viscosity η andsurface tension γ , the entrained film thickness predictedby the Frankel’s law [20] is: h F = 1 . κ − Ca / (1)where κ − = ( γ/ρg ) / is the capillary length and Ca = ηV /γ the capillary number. Van Nierop et al. [29] havecompiled experimental data from the literature on thethickness of soap films as a function of their entrained ve-locity. They have shown that the Frankel’s law in Ca / corresponds well to the experimental data, neverthelessdeviations from this law appear at “large” capillary num-ber Ca (cid:38) − . In this article we will show that we canreach much larger values of Ca.More recently, Saulnier et al. [26, 25] have studiedentrained small films stabilized either by C E (hex-aethylene glycol monododecyl ether) or by SDS (sodiumdodecylsulfate). They have studied the rupture of thefilms created by a frame pulled out of a liquid bath [25].They obtained the same lifetime for the different sur-factants, and conclude that the films maximum lengthdoes not depend on the physico-chemistry. On the otherhand, the artists who make giant bubbles pay attentionto the composition of the solutions they use: with thisnew setup, we will be able to work under controlled con-ditions and see whether or not the solutions play a rolein the stability of the giant films.The aim of this article is to present a new setup togenerate giant soap films at controlled velocities, com-parable to those apply by the artists when they cre-ate bubbles by hand. We thus need to increase by afactor of 300 the pulling velocity compared to previousexperiments [5, 10, 26, 25]. Until now, no study hasbeen carried out during films generation at such veloc-ities (Fig. 1). We also need to increase the size of thesoap films by more than a factor of 40 compared to themaximum size reported in the literature. Note that thequestion that we want to arise here is the stability of gi-ant soap films during and after their generation, whichis a problem different than the stability of films fed fromthe top. The later are called “curtains” in the literature,and many studies have been carried out to characterizethem [4, 15, 24], to use them as a support to study two-1 a r X i v : . [ c ond - m a t . s o f t ] F e b igure 1: Schematic diagram of the range of maximum lengths and velocities studied in the literature for soap filmsgenerated by pulling a frame out of a liquid bath.dimensional problems such as turbulence [22, 14, 13] orto investigate three-dimensional hydrodynamic problems(impact of liquid jets on the films [7], formation of bub-bles [23], ...). The largest curtain made in a laboratoryhad a height of 20 m by a width of 4 m [22], much longerthan any laboratory non fed foam film.We present a new setup allowing to create 2 m heightsoap films with a width of 0.7 m. We reach high en-trained velocity, from 0.02 m/s up to 2.5 m/s by combin-ing efficient motors, meaning that the capillary numberCa goes up to 10 − for typical aqueous solutions. Thiswill be fully described in the section 2. In section 3, wewill see that we can work in a controlled environmentwhere the humidity is regulated and measured. Indeed,knowing the ambient humidity during the generation isimportant and few experiments have been done in thisway [17, 16, 9, 19], whereas Champougny et al. [9] re-ported in a recent work that the evaporation has a hugeimpact on the film rupture. In section 4, we will describeall the measurements allowed by this setup. In partic-ular, we will show that the automatization allows us tohave a robust statistical analysis and accuracy, which isnecessary as shown by Tobin et al. [28]. The overall setup that we have designed, shown in Fig. 2,is composed of 30 ×
30 mm aluminium square profiles(Bosh Rexroth purchased from Radiospare) assembledtogether as a parallelepiped with overall dimensions of2.2 m × × × d (see Fig. 4 b)). Threecombined light-sensitive resistors (NSL-19M51 purchasedfrom Radiospare) collect the reflected light and assessthe presence of a film. During a run, we continuouslysave the reflected intensity collected by the light-sensitiveresistors and compare it to a reference intensity, whichcorresponds to the room lighting before generating a soapfilm. As soon as a film is generated and has reached thelight-sensitive resistors position, the collected intensityincreases and a timer starts until its rupture. We canthen obtain the lifetime as well as the maximum lengthof the soap films (see section 4.1). This also allows toautomatically start the generation of a new film afterthe rupture.A UV-VIS spectrometer (Nanocalc 2000 with a 400 µ mdiameter fiber, purchased from Ocean Optics), shown inFig. 4 (G), is placed on an aluminium profile along theheight and we can vary its position (denoted H in Fig.4 b)). This allows to measure the film thickness, as it isdescribed in section 4.2.Figure 3: The stepping motor (a) entrains a first pulley(b), diameter 25.06 mm, that is coupled to a second pul-ley (c), diameter 70 mm, thanks to a band (d). It allowsto multiply the speed. The bands (e) strain leading thesoap film is set by a round 3D printed piece (f).When a soap film is entrained, the pulleys rotate inthe tank containing the soapy solution: this results in thecreation of foam, which changes the conditions of the gen-eration and can even inhibit the formation of soap films.To overcome this problem and to be able to have statisti-cal measurements, we set up an automatic ethanol spraythat destroys the foam. To do this, we connected twoethanol sprayers to compressed air using a programmablesolenoid valve. We can therefore choose the duration ofspraying, as well as its frequency, allowing us to generatefilms under similar conditions.3igure 4: (a) 3-D view of the bottom of the frame. Aglassy tank (A) contains the soapy solution. A whitelight spot coupled to 3 photodiodes (B) is used to de-tect the film presence and a UV-VIS spectrometer (C)to study the thickness. The fishing line is entrained bythe movement of the pulleys (D). (b)Side drawing of thetrough. Half of the pulley is immersed in the soapy solu-tion (A). (B) shows the position where the film is gener-ated. d represents the fixed distance between the lighting(B) and the beginning of the film at the surface. H repre-sents the variable distance of the spectrometer (D) alongthe height of the film. l c , the circular length, is fixed intothe program. Most of the components are connected to an electronicbox that hosts programmable microcontroller boards(NXP LPC1768 MCU, purchased from Radiospare),which is depicted in Fig. 5. The first board is relatedto the motors, the reflected light device and the ethanolsprayers, and the second one to the humidity sensors sothat there is no conflict with the MCU timers. Using acode editor and a C/C++ compiler, we then create theuser interface with LabVIEW. The UV-VIS spectrom-eter device is a portable device, controlled by anothercomputer.Figure 5: Diagram of the electronic control system. (A)is an electronic box with 2 MBED microcontrollers. Thefirst one controls the variators (B) of the motors (C) andthe photoconductive cells (E), the second one controlsthe humidity captors (D), placed respectively at the topand at the bottom of the film. (F) is the spectrometer,(G) is the tank, (H) represents a soap film. This schemeis not to scale.At the beginning of an experiment, the user interfaceallows to redefine the position of the origin (correspond-ing to the surface of the liquid in the bath) and to gen-erate unique soap films to perform some tests. For sta-tistical analysis, cycles can be conducted, i.e. successivefilm generation can be initiated automatically. The usercan choose the height target, the entrained velocity, thenumber of cycles and the frequency of use of ethanolsprayers. In each cycle, the lifetime of the soap films,their height as well as the humidity and the temperaturein the enclosed box are automatically recorded.As rotational motors need an acceleration time, accel-eration is made before the fishing line reaches the surfaceof the soapy solution. The acceleration phase occurs inthe bath, and when the film gets out the entrained veloc-ity is constant. The circular length l c in Fig. 4 representsthe added length ran by the fishing line when acceler-ating. To check that the velocity is actually constant,we have measured for 6 different velocities, the heightreached by the fishing line over time using a high-speedcamera (Photron Fastcam SA3) at 125 to 250 frames persecond. The videos obtained allow to locate the fishingline relatively to the surface of the solution, and to com-4are this to the expected height, corresponding to theproduct of the velocity multiplied by time. As can beseen in Fig. 6, the measured and expected height are inagreement, which allows us to validate the system devel-oped for controlling the entrained velocity.Figure 6: Altitude reached by the fishing line as a func-tion of time. The dots represent measurements madewith a high-speed camera with a frame rate equal to125 for velocities between 20 and 140 cm.s − , and aframe rate equal to 250 for the velocities 180 and 250cm.s − . The dotted lines correspond to the theoreticalheight V × t . The inset is a zoom of this graph at shorttimes, for t < 0.5 s. Evaporation has an important impact on the stabilityof soap films [9]. It is thus essential to control it in theenvironment where the films are generated and to be ableto measure it over time.For this purpose, and in order to protect the filmsfrom "draughts", the setup is enclosed with PVC filmsand PMMA doors. Two room humidifiers, one at thetop and one at the bottom of the frame, are used. Hu-midity and temperature sensors (SHT25 purchased fromRadiospare) are present in the chamber to measure thehumidity and temperature over time. The humidifierat the bottom (Okoia AH450) allows us to select thetarget humidity and the humidifier at the top (BionaireBU1300W-I), which has an adjustable flow rate, allowsus to have more homogeneous humidity in the enclosure.This one is set manually at the beginning of an experi-ment.If we want to work for example at 60 % humidity, wecan keep it during more than 30 hours (this is longer than the classical experience times) with an absolute accuracyof 4 % as can be seen in Fig. 7. The temperature ismeasured throughout the experiments with an accuracyof 0.1 ◦ C, and corresponds to the room temperaturewhich is about 22 ◦ C. Night Night
Figure 7: Humidity and temperature recorded over timein the enclosed box for 30 hours, for an experiment per-formed at RH = 60 ± % . In this controlled environment, the automatic detectionof the films, thanks to the lighting system presented insection 2.1, allows us to measure the films lifetime andto ensure their statistical significance. In this paper, allsoap films are made with a 4 % solution of dishwashingliquid (Fairy Original from Procter & Gamble) dissolvedin ultrapure water (resistivity > 18.2 M Ω .cm). Fig. 8 (a)shows the lifetime successive measurements of 100 soapfilms generated continuously at a velocity of 20 cm.s − .The relative humidity is fixed at 42 % ± % (in orange)and 60 % ± % (in red) humidity. The bottom axis indi-cates the number of the successive films generated duringa cycle. For these experiments, soap films break duringtheir generation. The obtained histograms of these mea-surements are shown in Fig.8 (b). For a given humidity,we measured well-defined distributions. The lines repre-sent the corresponding normal distribution. As expected[9, 19], the average lifetime increases with humidity. Forsoap films generated at 20 cm.s − , we measure an in-crease of the lifetime from 4.55 s at 42 % to 6.81 s at 60 % . To study film stability in more detail and to character-ize them in the regime of high entrained velocities, it5 a)(b)
Figure 8: (a) Lifetime measurement of 100 soap filmsalong an experiment performed at a velocity of 20 cm.s − with controlled humidities of 42 % and 60 % . The soapfilms break during their generation. (b) Histograms ob-tained for these two experiments. The lines represent thenormal distribution. is particularly relevant to be able to measure also theirthickness. For that, we used the UV-visible spectrom-eter described in section 2.1. We are thus able to mapthe thickness of the soap films as a function of the en-trained velocity and humidity over time. An example ofa spectrum measured for a film thickness h = 1559 nmis given in the inset of Fig. 9. The measured spectra areanalysed with a Python script. The maxima and minimaare detected, which allows to determine the best thick-ness that fit the data. The spectra are recorded every100 ms.By positioning the spectrometer in the center of thesoap film, 13 cm above the soapy solution bath, werecorded its thickness over time during its generation ata velocity of 60 cm.s − . As can be seen in Fig. 9, thesoap film thickens over time at this position. This is thefirst time that this behaviour of films thickening is ob-served during the generation. This setup will thus allowus to study in more detail the impact of inertia on thestability of soap films.Figure 9: Evolution of a film thickness over time duringits generation at a velocity of 60 cm.s − , 13 cm above thesoapy solution bath. The inset represents the spectrumof the intensity reflected by a given soap film as a func-tion of the wavelength (blue line: signal; orange squares:minimum detected; green circles: maximum detected).For this spectrum, the film thickness measured is h = The inhomogeneity of the soap films thickness gives riseto beautiful inhomogeneities of colour on the surface ofthese films. These colours are due to light interferencebetween the two air/liquid interfaces of the soap film.Each colour is characteristic of a thickness and the moreintense are the colours, the thinner is the film. To visu-alize this, we installed the setup shown in Fig. 10 (a),which allows us to get photographs from the films.6o illuminate uniformly the soap films, we use a largewhite photo studio background paper, which will serveas a reflector. We made a hole in it in order to pass thelarge angle Nikon lens (AF-P DX 10-20 mm) of a NikonD7200 camera. We also use 4 halogen lamps, attached tothe frame of the setup. The light emitted by these lampswill then be reflected by the large white paper towardsthe soap films: the colours then appear on their surface,without any parasitic reflection (see Fig. 10 (b)).
In this paper, we presented a new automatized setup forthe generation of unfed giant soap films. With it, we cancontrol the height (up to 2 meters) and the entrainedvelocity (up to 2.5 m/s). This allows us to work in anunexplored regime for the generation of soap films with-drawn by pulling out a frame of a soapy solution, whichcorresponds to the regime used by the bubbling artists.In parallel, we can also control the ambient humidity andthe physico-chemistry of the solutions. Implemented re-flective setup and UV-VIS spectrometer gives informa-tion about the lifetime and the thickness of the films.Thanks to automatic repetitions of the experiments, wecan access to robust statistical information on the rup-ture process. These numerous diagnostics allowed by thispromising setup will help us to study the stability of gi-ant soap films in a controlled environment.
We acknowledge funding from ESA (MAP Soft MatterDynamics and contract 4000115113) and CNES (throughthe GDR MFA).Moteurs couplés
References [1] Slash bubbles artist. .[2] Tallest free-standing soap bubble (guin-ness world records). .[3] N. Adami and H. Caps. Capillary-driven two-dimensional buoyancy in vertical soap films.
EPL(Europhysics Letters) , 106(4):46001, May 2014.[4] N. Adami and H. Caps. Surface tension profiles invertical soap films.
Physical Review E , 91(1):013007,2015. halogen lampsoap film Camera (a)(b) hole
Figure 10: (a) Experimental setup used to visualize soapfilms. (b) Photograph obtained by using this set-up witha Nikon D7200 camera and a large angle objective lens.The film is generated at a velocity of 50 cm.s − at aheight of 120 cm with a 4 % Fairy solution.75] Eric A. Adelizzi and Sandra M. Troian. InterfacialSlip in Entrained Soap Films Containing Associat-ing Hydrosoluble Polymer.
Langmuir , 20(18):7482–7492, August 2004.[6] P Ballet and F Graner. Giant soap curtains forpublic presentations.
European Journal of Physics ,27(4):951–967, July 2006.[7] Saikat Basu, Ali Yawar, Andres Concha, and M MBandi. On angled bounce-off impact of a drop im-pinging on a flowing soap film.
Fluid Dynamics Re-search , 49(6):065509, nov 2017.[8] Steffen Berg, Eric A Adelizzi, and Sandra M Troian.Experimental Study of Entrainment and DrainageFlows in Microscale Soap Films. page 10.[9] Lorène Champougny, Jonas Miguet, Robin Henaff,Frédéric Restagno, François Boulogne, and Em-manuelle Rio. Influence of evaporation on soap filmrupture.
Langmuir , 34(10):3221–3227, 2018. PMID:29498527.[10] Sylvie Cohen-Addad and Jean-Marc di Meglio. Sta-bilization of aqueous foam by hydrosoluble poly-mers. 2. Role of polymer/surfactant interactions.
Langmuir , 10(3):773–778, March 1994.[11] Sylvie Cohen-Addad and Jean-Marc di Meglio. Sta-bilization of aqueous foam by hydrosoluble poly-mers. 2. Role of polymer/surfactant interactions.
Langmuir , 10(3):773–778, March 1994.[12] Stephen Frazier, Xinyi Jiang, and Justin C. Burton.How to make a giant bubble.
Physical Review Fluids ,5(1):013304, 2020.[13] H. Kellay. Hydrodynamics experiments with soapfilms and soap bubbles: A short review of re-cent experiments.
Physics of Fluids , 29(11):111113,November 2017.[14] Hamid Kellay and Walter I Goldburg. Two-dimensional turbulence: a review of some recentexperiments.
Reports on Progress in Physics ,65(5):845–894, apr 2002.[15] Ildoo Kim and Shreyas Mandre. Marangoni elas-ticity of flowing soap films.
Phys. Rev. Fluids ,2:082001, Aug 2017.[16] Xueliang Li, Stoyan I. Karakashev, Geoffrey M.Evans, and Paul Stevenson. Effect of Environmen-tal Humidity on Static Foam Stability.
Langmuir ,28(9):4060–4068, March 2012.[17] Xueliang Li, Ryan Shaw, and Paul Stevenson. Ef-fect of humidity on dynamic foam stability.
Interna-tional Journal of Mineral Processing , 94(1-2):14–19,February 2010. [18] Sylvie Lionti-Addad and Jean Marc Di Meglio. Sta-bilization of aqueous foam by hydrosoluble poly-mers. 1. Sodium dodecyl sulfate-poly(ethylene ox-ide) system.
Langmuir , 8(1):324–327, January 1992.[19] Jonas Miguet, Marina Pasquet, Florence Rouyer,Yuan Fang, and Emmanuelle Rio. Stability of bigsurface bubbles: impact of evaporation and bubblesize.
Soft Matter , 16(4):1082–1090, 2020.[20] KJ Mysels, K. Shinoda, and S.Frankel.
Soap Films:Studies of their Thinning and a Bibliography . Perg-amon, New York, 1959.[21] P. Petit, J. Seiwert, I. Cantat, and A.-L. Biance.On the generation of a foam film during a topo-logical rearrangement.
Journal of Fluid Mechanics ,763:286–301, January 2015.[22] M. A. Rutgers, X. L. Wu, and W. B. Daniel. Con-ducting fluid dynamics experiments with verticallyfalling soap films.
Review of Scientific Instruments ,72(7):3025–3037, 2001.[23] Louis Salkin, Alexandre Schmit, Pascal Panizza,and Laurent Courbin. Generating soap bubbles byblowing on soap films.
Phys. Rev. Lett. , 116:077801,Feb 2016.[24] Aakash Sane, Shreyas Mandre, and Ildoo Kim. Sur-face tension of flowing soap films.
Journal of FluidMechanics , 841:R2, 2018.[25] Laurie Saulnier, Lorène Champougny, Gaël Bastien,Frédéric Restagno, Dominique Langevin, and Em-manuelle Rio. A study of generation and rupture ofsoap films.
Soft Matter , 10:2899–2906, 2014.[26] Laurie Saulnier, Frédéric Restagno, Jérôme Dela-cotte, Dominique Langevin, and Emmanuelle Rio.What is the mechanism of soap film entrainment?
Langmuir , 27(22):13406–13409, 2011.[27] Jacopo Seiwert, Ronan Kervil, Soniraks Nou, andIsabelle Cantat. Velocity Field in a Vertical FoamFilm.
Physical Review Letters , 118(4):048001, Jan-uary 2017.[28] S. T. Tobin, A. J. Meagher, B. Bulfin, M. Möbius,and S. Hutzler. A public study of the lifetime distri-bution of soap films.
American Journal of Physics ,79(8):819–824, August 2011.[29] Ernst A. van Nierop, Benoit Scheid, and Howard A.Stone. On the thickness of soap films: an alterna-tive to frankel’s law.