Gravity-driven fluid oscillations in a drinking straw

Ryan P. Smith, and Eric H. Matlis

Citation: American Journal of Physics 87, 433 (2019); doi: 10.1119/1.5095945

View online: https://doi.org/10.1119/1.5095945
View Table of Contents: https://aapt.scitation.org/toc/ajp/87/6
Published by the American Association of Physics Teachers

ARTICLES YOU MAY BE INTERESTED IN

Electromagnetic surface wave propagation in a metallic wire and the Lambert W function
American Journal of Physics 87, 476 (2019); https://doi.org/10.1119/1.5100943

Superconductivity, broken gauge symmetry, and the Higgs mechanism

American Journal of Physics 87, 436 (2019); https://doi.org/10.1119/1.5093291

The candle seesaw

American Journal of Physics 87, 370 (2019); https://doi.org/10.1119/1.5096886

The sliding ladder problem revisited in phase space

American Journal of Physics 87, 444 (2019); https://doi.org/10.1119/1.5096278

2018 Millikan Medal Award Lecture: Breaking out of the physics silo
American Journal of Physics 87, 415 (2019); https://doi.org/10.1119/1.5096623

Orbits through polytropes

American Journal of Physics 87, 452 (2019); https://doi.org/10.1119/1.5093295
Gravity-driven fluid oscillations in a drinking straw
Ryan P. Smith
Department of Physics, California State University - East Bay, Hayward, California 94542
Eric H. Matlis
Department of Aerospace and Mechanical Engineering, University of Notre Dame, Notre Dame,
Indiana 46556
(Received 23 September 2018; accepted 12 March 2019)
We describe a simple experiment observing oscillations of fluid in a drinking straw immersed in a
bath of water. The motion of this oscillator system with changing mass is matched remarkably well
by a model derived from Newton’s laws with the inclusion of only a phenomenological damping
coefficient. We compare the frequency of the oscillations to Hooke’s law for small displacements
and find 0.25% discrepancy. This experiment can be shown as a demonstration of oscillations of
fluids, or it can be performed by students as a laboratory activity in upper-division undergraduate
physics and engineering courses. VC 2019 American Association of Physics Teachers.
https://doi.org/10.1119/1.5095945

I. STATEMENT OF THE PROBLEM changing as the liquid level changes in time. Thus, we can
write down the mass of the liquid in the straw as
Students are often familiar with the phenomenon of
capping a straw while submerged in water, then lifting the m ¼ qzA: (2)
still-capped straw; water is held inside by a balance of atmo-
spheric pressure and the vacuum created toward the top of Equation (1) can be rearranged as a second-order ordinary
the straw. Another interesting concept involves turning differential equation where q and A have both canceled out
around this idea, as illustrated in Fig. 1: what happens when
1 2

we submerge a capped straw into a liquid bath and suddenly €z ¼  z_ þ gz  gh þ bz_ : (3)
release the capping? The fluid rushes into the straw, driven z
by the hydrostatic pressure from below, causing the fluid
level in the straw to rise rapidly above the surface of the Here, the phenomenological damping coefficient is redefined
bath, then undergo several oscillations before settling to the b0 ¼ bqA to simplify the final equation. This equation can be
level of the bath. While an involved analysis of this problem solved numerically, e.g., using odeint from the scipy
has previously been performed,1 we proceed to understand scientific python library (python notebook linked below7),
this behavior by connecting a simple Newtonian model with using the experimental conditions (g ¼ 9:8 m/s2, h ¼ 10:2
experimental observation. Such a model is accessible to cm) and the initial conditions that the fluid level starts near
junior-level undergraduates, and it provides exposure to the edge of the straw (z0 ¼ 0:2 cm) with initial velocity
numerical solutions to ordinary differential equations. We z_ 0 ¼ 0. A fluid statics analysis confirms that fluid level is
also note some related educational works that engage stu- pushed into the bottom of the straw by 2% the length of
dents in modeling and exploring fluid dynamics2–4 and oscil- the tube, as observed. Since this distance is small compared
latory motion.5,6 with the levels measured, the solution to the differential
equation zðtÞ is robust to a range in z0.
II. MODELING THE DYNAMICS USING NEWTON’S
III. RECORDING DATA
SECOND LAW
Data were recorded using a smartphone video camera (24
Our model follows Newton’s second law on the mass of frames per second), and the video was imported using
fluid in the cylindrical straw column. Our model includes the Tracker software4 (equivalently, PASCO Capstone), allow-
hydrostatic pressure from the bottom of the straw and the ing actual position data of the fluid level in the straw tube to
force of gravity, and energy losses. Following this approach, be extracted frame by frame. The equipment needed for this
an equation of motion for the liquid level above the bottom experiment is as follows: a clear glass or plastic container,
of the straw, z, as shown in Fig. 1, can be expressed as water, a straw (10 mm diameter works best), food coloring
dp (optional, to improve optical contrast), and a video camera.
¼ Fg þ Fpressure þ Fdamping The length of the straw itself (200 mm long) can provide a
dt scale in the video, and the fluid level is measured above the
z ¼ mg þ qghA  b0 z;
m_ z_ þ m€ _ (1) bottom of the straw. The initial time t ¼ 0 was fixed as the
moment that the finger was lifted from the top of the straw.
where p ¼ mz_ is the momentum of the liquid in the straw, q
is the density of the liquid, h is the depth of the bottom of the IV. DISCUSSION OF RESULTS
straw below the bath surface, A ¼ pr 2 is the cross-sectional
area of the straw, and b0 is a phenomenological damping After the data were processed into a file, we compared the
coefficient. An important difference with most introductory measurement with the solution to Eq. (3). By inputting the
physics problems is that the mass of liquid in the tube is initial conditions ðz0 ¼ 0:2 cm; z_ 0 ¼ 0 cm=sÞ and modifying

433 Am. J. Phys. 87 (6), June 2019 http://aapt.org/ajp C 2019 American Association of Physics Teachers
V 433
 phenomenological damping factor is needed: the solution to
the differential equation using the chosen initial conditions
matches the data. While the damping coefficient does depend
on the straw diameter and fluid viscosity, we consider the
calculation of this damping coefficient beyond the scope of
these educational goals of this work. We decided that the
Newton’s-law model is more transparent in the derivation,
requires little background in fluid mechanics, and convinc-
ingly models the data.
We note that stronger experimental damping compared with
both the Newtonian and Lorenceau models is noticeable at later
times. This discrepancy was also not fully explained by
Lorenceau et al. In the Newtonian model, we simply chose the
best damping parameter to fit the data but did not modify the
initial conditions. Internal friction effects of the tube (or straw),
surface roughness, development length effects, transition to tur-
bulence, or capillary effects were not taken into account in the
Fig. 1. (a) Straw submerged in the bath, capped by a finger; (b) moments after model, which may explain the discrepancies.
cap is released, oscillations of the fluid level in the straw follow. Relevant
dimensions labeled. Finger can remain attached to the experimentalist.
Additionally, a discrete Fourier transform of this oscilla-
tion yields more detailed information about the nature of the
oscillations, shown as a power spectrum in Fig. 3. The posi-
only the phenomenological damping coefficient, b, we are tion of the peak in the spectrum as fit by a Lorentzian repre-
able to match the data using our Newtonian model with sents the natural oscillation frequency. The full-width at
remarkable agreement, shown in Fig. 2. The anharmonic half-max (FWHM) of this spectrum, Df ¼ 0:31 Hz, relates to
nature of the damped oscillatory motion can be seen clearly the decay time of the oscillation as Tdecay ¼ 1=Df , which in
by setting the damping coefficient b ¼ 0 in the model, result- this case evaluates to 3 s, setting the decay timescale for
ing in a sharp turnaround toward z ¼ 0. This behavior can be
the damping.
understood since the lower mass at fluid levels near the straw
bottom (see Eq. (2)) means less inertia, yielding larger accel-
eration of the liquid level than when the straw is full of V. COMPARISON OF OSCILLATION FREQUENCY
liquid. WITH HOOKE’S LAW FOR SMALL
We also show in this figure the solution to the more DISPLACEMENTS
sophisticated model of Lorenceau et al., which does not
Another exciting extension of this measurement can be
require any phenomenological factors.1 This model considers
found by simplifying Eq. (1) to explore behavior in the limit of
the laminar-flow regime, neglects capillary forces, and calcu-
small displacements about the surface of the bath. For conve-
lates the damping coefficient by considering the viscous
nience, we define y ¼ z  h to be the fluid level in the straw
Poiseuille friction as well as the singular pressure loss at the
referenced to the surface level of the bath, and then take y  h
tube opening at the tube-reservoir interface, giving rise to
(and thus y_ is also small). Thus, Eq. (1) simplifies to
eddies; both effects dissipate energy.
By solving the differential equation from the more sophis- g
ticated model created by Lorenceau et al., using the same y€ ¼  y; (4)
h
numerical strategy, we do find slightly better agreement with
our experimental data, as shown in Fig. 2. We input the same which is can be identified as Hooke’s law,8 and the natural
initial conditions, the straw diameter, the density of water, frequency f0 of oscillation can be readily read off as
the fluid viscosity, and atmospheric pressure, yielding the
curve shown in Fig. 2. A benefit for this approach is that no

Fig. 2. Matching the recorded damped oscillatory motion of the fluid level Fig. 3. Fourier transform of oscillatory data fit by a Lorentzian curve yields
with the solution to the simple Newtonian model (solid line), as well as the the natural frequency of oscillation which compares well with the analytic
more sophisticated model of Lorenceau et al. (dashed line). estimate in the limit of small displacements (shown as a vertical line).

434 Am. J. Phys., Vol. 87, No. 6, June 2019 R. P. Smith and E. H. Matlis 434
  1 data, estimating the decay time and the period of oscillations.
x0 1 g 2
f0 ¼ ¼ : (5) Instructors can guide more advanced students through the
2p 2p h
derivation of the Newtonian model, yielding a differential
For our experimental parameters (g ¼ 9:8 m/s2, h ¼ 10:2 equation which they can then solve numerically. Performing
cm), Eq. (5) evaluates to f0 ¼ 1:560 Hz. For comparison, we a discrete Fourier transform of the resulting output exposes
fit a Lorentzian to the power spectrum of our data as shown students to an important analysis technique with readily
in Fig. 3, yielding f0 ¼ 1:563 ð60:1%Þ Hz, having a 0.25% interpretable results. As some suggestions for exploration,
discrepancy with the analytic result, smaller than the spacing students could repeat this experiment for different fluids,
of points in the power spectrum. The error in the power spec- straw diameters, or straw shapes, to discover the effects of
trum fit is determined through the covariance matrix in a changing each parameter on the oscillatory behavior.
least-squares optimization and is significantly smaller than
the spacing of data points. ACKNOWLEDGMENTS
Because of limitations in the precision of video equip-
The authors gratefully acknowledge CSUEB colleagues
ment, measuring very small displacement oscillations is
Dr. Erik Helgren and Dr. Jason Singley, who helped support
challenging. However, we attribute the agreement to the fact
that after releasing the cap, the liquid level spends most of the implementation of this into the Physics curriculum; Dr.
the time near the equilibrium. With more sensitive observa- Justin Kuczynski, Dr. Candace Gilet, Dr. Derek Jackson
tion equipment, such as monitoring the electrical resistance Kimball, Dr. Amy Furniss, Dr. Katy Grimm, and Dr. Marlin
of a wire in the straw as a proxy for fluid height, one could Halim for helpful discussions; and CSUEB students for their
experimentally verify this small-displacement oscillation fre- participation and insights.
quency more precisely.
1
E. Lorenceau, D. Quere, J. Y. Ollitrault, and C. Clanet, “Gravitational oscil-
VI. OUTLOOK
lations of a liquid column in a pipe,” Phys. Fluids 14, 1985–1992 (2002).
2
This simple experiment in fluid oscillations provides an S. Shamim et al., “Investigating viscous damping using a webcam Investigating
viscous damping using a webcam,” Am. J. Phys. 78, 433–436 (2010).
alternative perspective on Newton’s second law and fluids, 3
F. B. Morinigo, “Fluid Oscillations in a U Tube,” Am. J. Phys. 40, 350–351
and the topic also connects naturally to other topics involv- (1972).
ing complex oscillatory motion. Our analysis yields agree- 4
V. K. Gupta, G. Shanker, and N. K. Sharma, “Experiment on fluid drag and
ment between a simple Newtonian model and experimental viscosity with an oscillating sphere,” Am. J. Phys. 54, 619–622 (1986).
5
data, and the observed natural oscillation frequency is S. Theodorakis and K. Paridi, “Oscillations of a candle burning at both
matched within 0.25% by a simplified analytic equation (Eq. ends,” Am. J. Phys. 77, 1049–1054 (2009).
6
(4)) in the limit of small displacements. R. Hauko, D. Andreevski, D. Paul, and R. Repnik, “Experiments teaching
of the harmonic oscillator damped by a constant force: The use of analogy
Undergraduate students can try this experiment by first
and experiments,” Am. J. Phys. 86, 657–662 (2018).
sketching a prediction of the fluid level in the straw as a 7
R. P. Smith, <https://github.com/r31415smith/AdvLab/tree/master/fluid_
function of time, recording a short video of the experiment, osc>, last accessed January 2, 2019.
and then extracting data using the software mentioned ear- 8
J. R. Taylor, Classical Mechanics (University Science Books, Sausalito,
lier. Students can then compare their sketch with the actual CA 2005).

435 Am. J. Phys., Vol. 87, No. 6, June 2019 R. P. Smith and E. H. Matlis 435