Rolling Ribbons
P. S. Raux,1 P.M. Reis,1 J.W.M. Bush,1 and C. Clanet2
1Department of Mathematics, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA
2LadHyX, E´cole Polytechnique, 91128 Palaiseau, France
(Received 23 March 2010; published 23 July 2010)
We present the results of a combined experimental and theoretical investigation of rolling elastic
ribbons. Particular attention is given to characterizing the steady shapes that arise in static and dynamic
rolling configurations. In both cases, above a critical value of the forcing (either gravitational or
centrifugal), the ribbon assumes a two-lobed, peanut shape similar to that assumed by rolling droplets.
Our theoretical model allows us to rationalize the observed shapes through consideration of the ribbon’s
bending and stretching in response to the applied forcing.
DOI: 10.1103/PhysRevLett.105.044301 PACS numbers: 46.70.De, 46.25.�y, 46.35.+z
Galileo’s study of rigid spheres rolling down an inclined
ramp [1] is often considered as the starting point of modern
physics, since it involves both theory and experiment [2,3].
The influence of ramp flexibility on the dynamics was
recently considered by Aristoff et al. [4]. We here consider
another variant of Galileo’s problem in which the ramp is
rigid but the rolling body, an elastic cylindrical shell, is
deformable. As will be shown, this dynamical elastic prob-
lem presents some common features with the rolling of a
liquid drop on a hydrophobic surface [5–7] or a lubricated
ramp [8]. We first present our experimental observations
and then develop a supporting theoretical model.
The ribbons are cast out of three different types of
vinylpolysiloxane that produce three elastic polymers
with Young’s moduli E ¼ 0:26 MPa, 0.56 MPa and
1.2 MPa, and respective densities � ¼ 1050 kg=m3,
1100 kg=m3, and 1200 kg=m3. The associated Poisson
ratio is measured to be � ¼ 0:5. The geometrical character-
istics of the ribbons are radius R0 ¼ ð23:2� :4Þ mm,
length L0 ¼ 2�R0, thickness ð0:8< h0 < 3:25Þ mm,
width ð18 4, the
circular shape is only weakly affected by gravity while
strong deformations are observed for �g � 1. Touchdown
is achieved at �?g ¼ 0:19. The observed dependence of
static ribbon shape on �g is rationalized in our subsequent
theoretical developments.
We observe that, as the speed increases, the rolling
ribbon develops a two-lobed shape. A similar behavior
has been reported for rotating liquid drops [7,9]. As the
drop rotation speed increases, its shape evolves through a
series of axisymmetric forms, from a sphere to an oblate
ellipsoid to a torus. However, above a critical speed, the
axisymmetric equilibrium shape is no longer adopted
[10,11], and the drop assumes a two-lobed, peanut shape.
The critical rotation speed�L abovewhich a drop of radius
RL loses its axisymmetry is found to scale as
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
�L=�LR
3
L
q
,
where �L and �L are the surface tension and density,
respectively. This scaling emerges from balancing the
destabilizing rotational energy �LR
5
L�
2
L with the stabiliz-
ing surface energy �LR
2
L. Drawing an analogy between the
liquid drop and our elastic ribbon suggests that the latter
will lose its axisymmetric form when its rotational energy
�S0R
3
0�
2
c greatly exceeds its elastic energy EI0=R0. Thus,
even in the absence of the flattening influence of gravity,
one anticipates deformed ribbons for velocities larger thanffiffiffiffiffiffiffiffiffi
E=�
p
h0=R0. For the ribbon presented in Fig. 1, this critical
velocity is of the order of 1 m=s, which is substantially less
than that observed experimentally (Uc � 7:5 m=s). This
discrepancy has motivated the more precise analysis de-
tailed below.
Guided by the scaling analysis, we characterize the
ribbon deformation in terms of the parameter �i ¼
Eh2=�R20U
2 to study the ribbon deformation. We note
that stretching is important in our experiments; thus, we
define �i in terms of h instead of h0. The observed varia-
tion of � with �i is reported in Fig. 3. In the zero velocity
limit (�i ! 1), the aspect ratio tends towards the static
value imposed by �g. Since the results presented in Fig. 3
correspond to four different ribbons with four different
values of �g, we observe four different asymptotic values.
We observe in Fig. 3 that the aspect ratio differs substan-
tially from its static limit only for �i � 1. In the large
velocity limit (�i ! 0), the deformation increases (�! 0)
up until touchdown at �?i , the value of which depends on
�g. Typically, �
?
i � 0:02. One way to show that �?i indeed
depends on �g is to consider the limit of a ribbon for which
�g ¼ �?g . In this limit, touchdown is achieved without any
rotation, that is for U ¼ 0 or �?i ¼ 1. Apart from this
FIG. 2 (color online). Aspect ratio of static ribbons, � ¼ H2R0 ,
as a function of their normalized stiffness �g ¼ EI0�gS0R30 . The
line is deduced by integrating Eq. (7) with Fr ¼ 0. For the letters
on the curve, we present in the inset the corresponding picture.
The theoretical shape obtained through numerical integration of
Eq. (7) is superposed as a thin white dashed line. Scale bars,
10 mm.
FIG. 3 (color online). Aspect ratio, � ¼ H2R0 , of different roll-
ing ribbons as a function of the speed parameter �i ¼ Eh2�R2
0
U2
.
Curves correspond to those predicted by Eq. (7). Circles,
squares, diamonds, and triangles correspond to four different
ribbons; specifically, �g ¼ 0:278, 0.372, 0.429, 0.488. E ¼
0:26 MPa for the first three ribbons and E ¼ 0:56 MPa for the
last. Inset: Shape comparison for a single ribbon (R0 ¼
22:8 mm, �g ¼ 0:488, E ¼ 0:56 MPa) at different rotation
speeds: (a) Cy ¼ 0:09, (b) Cy ¼ 0:16, (c) Cy ¼ 0:20, (d) Cy ¼
0:25, where Cy ¼ �U2=E. Dashed curves are computed by
integration of Eq. (7). Scale bar: 10 mm.
PRL 105, 044301 (2010) P HY S I CA L R EV I EW LE T T E R S
week ending
23 JULY 2010
044301-2
dependence of the critical velocity on the initial gravita-
tional deformation of the ribbon, we observe that, qualita-
tively, the shape transition prompted by centripetal forces
is similar to that caused by gravity and presented in Fig. 2.
We proceed with a quantitative study, by developing a
theoretical model that predicts the steady shapes of both
static and rolling ribbons.
A schematic diagram of the ribbon system is presented
in Fig. 1(f). The ribbon shape is described in terms of the
arc length s along its center line. Let n be the outward
pointing unit vector normal to the ribbon, � the angle
between the local tangent vector t and the horizontal unit
vector ex. Considering the ribbon as a slender elastic
structure, the local force and torque balances can be ex-
pressed as [12]:
F 0 ¼ �K; (1)
and
EI�00 ¼ F� n; (2)
where F represents the internal forces andKds the external
ones. Primes denote derivatives with respect to s. To
integrate this system, we divide the ribbon into two zones
where the external constraints K are different: a contact
zone and a free zone. In the free zone (�Lf=2< s < Lf=2,
where Lf is the free length), there is no friction, but
gravitational and centripetal forces act. In the contact
zone (jsj> Lf=2), friction pins the ribbon to the drum.
Consequently, the ribbon has a small but finite curvature
prescribed by that of the drum: its center line is in trans-
lation at a constant speed Uð R
Rþh2
Þ.
In the free zone, the external force acting on an infini-
tesimal element of ribbon has both gravitational and cen-
tripetal contributions:
K ¼ �Sgþ �SU2�0n: (3)
In order to determine boundary conditions for the vertical
and horizontal components of the internal force, Fy and Fx
respectively, we assume that the shape is symmetric with
respect to the vertical midplane, an assumption consistent
with experimental observations prior to touchdown of the
ribbon. It is thus sufficient to consider one half of the
ribbon. The symmetry also sets the boundary conditions
at the top of the ribbon where the shape must be perpen-
dicular to the axis of symmetry (�js¼0 ¼ 0) and where
there is no variation of curvature (�00js¼0 ¼ Fyjs¼0 ¼ 0).
We also assume continuity with the contact zone, which is
insured by applying additional boundary conditions: a flat
contact zone for the static case (�js¼Lf=2 ¼ � and
�0js¼Lf=2 ¼ 0) and a small curvature corresponding to
that of the drum for the dynamic case (�js¼Lf=2 ¼ ��
arcsinðxjs¼Lf=2Rdrum Þ and �0js¼Lf=2 ¼ 1Rdrum ). The internal force can
be found by integrating (1) using (3),
FðsÞ ¼ F
x
Fy
� �
¼ F
xjs¼0 þ �SU2½cos�js � 1�
��gSs� �SU2 sin�js
� �
: (4)
Since the tension in the ribbon is equal to the tangential
component of the internal force, we have F� t ¼
ES ðds�ds0Þds0 where ds� ds0 is the extension of the ribbon.
Integrating this equation and combining it with (4) yields
the increase of the ribbon’s free length:
Lf � Lf0 ¼ Cy
Z �
1þ
�
�
Fr
� 1
�
cos�þ �s
Fr
sin�
�
ds0:
(5)
Here, �s ¼ sR0 denotes the nondimensional curvilinear coor-
dinate, and Lf0 the free length at rest. The Cauchy number
Cy ¼ �U
2
E indicates the relative magnitude of inertial and
stretching forces while the Froude number Fr ¼ U2gR0 ex-
presses that of inertia and gravity. Finally,� ¼ Fxjs¼0�gSR0 is the
ratio between the tension at rest and gravity. This tension
depends on the natural curvature of the ribbon. Since our
ribbons are molded, their natural curvature is close to 1=R0
and the tension is close to zero in the absence of gravity. In
our numerical integration, we take this information into
account by choosing the smallest possible value of �.
Since � varies from 0 to 2� along the free length, the
cosine and sine terms in (5) are much smaller than the
constant after integration; thus, the stretching is uniform to
leading order. In the contact zone, where there is no sliding,
we extend the assumption of uniform stretching over the
whole ribbon, so the total extended length L is given by
L ¼ ð1þ CyÞL0: (6)
The variation of thickness due to transverse stretching is
deduced from the Poisson ratio, � ¼ 0:5: since our poly-
mer conserves volume, the ribbon thickness, h ¼ h0ð1�
�CyÞ, necessarily decreases with increasing speed. In order
to determine the shape of the free length, we substitute (4)
into (2), which after nondimensionalization yields
ð1� �CyÞ2�g�00 ¼ ðFr��Þ sin�� �s cos� (7)
where �g ¼ EI0�gS0R30 again prescribes the relative magnitudes
of bending and gravity.
We use a combined fourth and fifth order explicit
Runge-Kutta method to integrate (7) numerically between
s ¼ 0 and the first value of s that satisfies the condition sþ
xjs ¼ L=2 where L is the stretched length given by (6). We
use a shooting method, and close the system by adjusting
the values of � and �0js¼0 to match the slope �js¼Lf=2 and
curvature �0js¼Lf=2 of the substrate at the edge of the
contact zone. We thus obtain the ribbon shape, �ðsÞ, by
integrating (7) with no fitting parameter.
We first compare our model and experiments in the static
(Fr ¼ 0) case. In Fig. 2, we present the shapes calculated
PRL 105, 044301 (2010) P HY S I CA L R EV I EW LE T T E R S
week ending
23 JULY 2010
044301-3
from (7) and see good agreement between theory and
experiment for the four different ribbons considered. We
proceed by verifying our assumption of constant stretching
along the rolling ribbon. In Fig. 4, we present the depen-
dence of length extension, LL0
, on the Cauchy number for
four different ribbons. Despite the uncertainty in ribbon
length introduced by the image analysis, a linear fit of the
experimental results gives a slope close to that predicted by
(6) within a 6% margin of error. The observed stretching is
thus consistent with our hypothesis of uniform extension.
For rolling ribbons, the curves in Fig. 3 represent the
prediction obtained by integrating (7) for the values of �g
corresponding to the observed deformation at rest, that is,
the asymptotic value of � at large �i. The discrepancies
observed near the critical touchdown speed are likely due
to the climbing of the ribbon along the inner wall of the
rotating cylinder, and the resulting fore-aft asymmetry of
the ribbon. In the dynamic case, we also compare observed
and predicted ribbon shapes, as presented in the inset of
Fig. 3 for four different values of Cy on the same ribbon,
which necessarily has a single �g value. As in the static
case, the agreement between observed and predicted
shapes is satisfactory.
We have considered the rolling of elastic ribbons and
shown that their shape results from a delicate coupling
between rolling, bending, and stretching. While we have
noted that the rolling ribbon has several features common
with rolling droplets [6–8], a similar family of shapes has
also been reported for tumbling blood cells [13], a tube
collapsing under uniform pressure [14], and carbon nano-
tubes deformed by van der Waals forces [15,16]. The
rationale for similar shape progressions emerging in these
disparate physical systems has yet to be carefully consid-
ered, but should be informed by our study.
[1] G. Galilei, Dialogues Concerning Two New Sciences
(Dover Publications Inc., New York, NY., 1954).
[2] H. F. Cohen, The Scientific Revolution: A Historio-
graphical Inquiry (The University of Chicago Press,
Chicago, USA, 1994).
[3] S. Drake, N.M. Swerdlow, and T. H. Levere, Essays On
Galileo And The History And Philosophy Of Science
(University of Toronto Press, Toronto, Canada, 1999).
[4] J.M. Aristoff, C. Clanet, and J.W.M. Bush, Proc. R. Soc.
A 465, 2293 (2009).
[5] L. Mahadevan and Y. Pomeau, Phys. Fluids 11, 2449
(1999).
[6] P. Aussillous and D. Que´re´, Nature (London) 411, 924
(2001).
[7] P. Aussillous and D. Que´re´, J. Fluid Mech. 512, 133
(2004).
[8] S. R. Hodges, O. E. Jensen, and J.M. Rallison, J. Fluid
Mech. 512, 95 (2004).
[9] L. Elkins, P. Ausillous, J. Bico, D. Quere, and J.W.M.
Bush, Meteorit. Planet. Sci. 38, 1331 (2003).
[10] R. A. Brown and L. E. Scriven, Proc. R. Soc. A 371, 331
(1980).
[11] Lord Rayleigh, Philos. Mag. 28, 161 (1914).
[12] L. D. Landau and E.M. Lifshitz, Theory Of Elasticity,
Course of Theoretical Physics, Vol. 7 (Pergamon Press,
Oxford, 1986).
[13] J.M. Skotheim and T.W. Secomb, Phys. Rev. Lett. 98,
078301 (2007).
[14] J. E. Flaherty, J. B. Keller, and S. I. Rubinow, SIAM J.
Appl. Math. 23, 446 (1972).
[15] T. Hertel, R. E. Walkup, and P. Avouris, Phys. Rev. B 58,
13 870 (1998).
[16] A. Pantano, D.M. Parks, and M.C. Boyce, J. Mech. Phys.
Solids 52, 789 (2004).
0 0.1 0.2 0.3 0.4
1
1.05
1.1
1.15
1.2
1.25
1.3
1.35
Cauchy number, Cy
El
on
ga
tio
n,
L
/L
0
Γg=.278, E=.26MPa
Γg=.372, E=.26MPa
Γg=.429, E=.26MPa
Γg=.488, E=.56MPa
Eqn. (6)
FIG. 4 (color online). Ribbon elongation as a function of the
Cauchy number Cy ¼ �U
2
E . The black line is the dependence
predicted by Eq. (6). Different symbols correspond to different
values of normalized stiffness �g. Characteristic error bars are
shown for only one set of experiments.
PRL 105, 044301 (2010) P HY S I CA L R EV I EW LE T T E R S
week ending
23 JULY 2010
044301-4
本文档为【【CUPT】rolling ribbons】,请使用软件OFFICE或WPS软件打开。作品中的文字与图均可以修改和编辑,
图片更改请在作品中右键图片并更换,文字修改请直接点击文字进行修改,也可以新增和删除文档中的内容。
该文档来自用户分享,如有侵权行为请发邮件ishare@vip.sina.com联系网站客服,我们会及时删除。
[版权声明] 本站所有资料为用户分享产生,若发现您的权利被侵害,请联系客服邮件isharekefu@iask.cn,我们尽快处理。
本作品所展示的图片、画像、字体、音乐的版权可能需版权方额外授权,请谨慎使用。
网站提供的党政主题相关内容(国旗、国徽、党徽..)目的在于配合国家政策宣传,仅限个人学习分享使用,禁止用于任何广告和商用目的。