1: \documentclass[aps,prl,preprint,showpacs]{revtex4}
2: %\documentclass[twocolumn,showpacs,preprintnumbers,amsmath]{revtex4}
3: \usepackage{epsfig}
4: \usepackage{amssymb}
5:
6: \begin{document}
7:
8: \title{Motion of a sphere in an oscillatory boundary layer: an optical tweezer based study }
9: \author{Prerna Sharma \footnote{email : prerna@tifr.res.in}, Shankar Ghosh and S. Bhattacharya}
10: \date{\today }
11: %\pacs{}
12:
13: \begin{abstract}
14: The drag forces acting on a single polystyrene sphere in the
15: vicinity of an oscillating glass plate have been measured using an
16: optical tweezer. The phase of the sphere is found to be a sensitive
17: probe of the dynamics of the sphere. The evolution of the phase from
18: an inertially-coupled regime to a purely velocity-coupled regime is
19: explored. Moreover, the frequency dependent response is found to be
20: characteristic of a damped oscillator with an effective inertia
21: which is several orders of magnitude greater than that of the
22: particle.
23: \end{abstract}
24:
25: \affiliation{Department of Condensed Matter Physics and Materials Science\\
26: Tata Institute of Fundamental Research\\
27: Homi Bhabha Road, Mumbai 400-005, India\\
28: }
29: \maketitle
30:
31: \section{Introduction}
32:
33: Study of the forces between two sliding bodies in the presence of an
34: intervening fluid is an important problem in fluid mechanics
35: \cite{landau}. In recent years, a variety of experimental and
36: theoretical/ computational tools have been brought to bear of this
37: central issue. Experimental studies of the ``simple'' case of
38: hydrodynamic drag forces acting on a single particle suspended in a
39: liquid near a plate executing sinusoidal oscillations are rare,
40: although there have been measurements on the lift forces acting on
41: spheres in oscillatory flows \cite{sleath}. Such measurements have
42: shown a highly non-monochromatic response of the lift force as a
43: function of the oscillating frequency. This problem is important in
44: understanding the role played by colloidal particles in a
45: lubricating liquid as well as the nature of interactions between the
46: colloidal particles and the solid interface formed say by the walls
47: of narrow channels, e.g., blood vessels.
48:
49: In this paper we present a new technique based on an optical tweezer
50: that addresses an important aspect of this problem: how does one
51: probe the dynamical response of a solid particle in ``contact'' with
52: a substrate in the presence of a lubricating liquid, which requires
53: an elucidation of the notion of the ``contact'' itself. The study
54: reveals the importance of the phase response of the particle which
55: provides insight into the nature of the dynamics.
56:
57:
58: A schematic of the experimental setup is shown in figure 1 . A very
59: dilute monodisperse colloidal solution of polystyrene spheres of
60: radius(a) of 1$\mu $m (Alfa Aesar) was placed between two parallel
61: glass plates separated by a distance of 2mm. The top glass plate was
62: kept stationary. The bottom glass plate was attached to a $xyz$
63: piezo stage subjected to an oscillatory motion in the $y$ direction
64: with an amplitude $y_{po}$ and frequency $\omega $, i.e.,
65: $y_{p}=y_{po}\sin (\omega t)$, where $y_{p}$ is the instantaneous
66: displacement of the glass plate . A colloid particle was trapped
67: near the glass plate using an IR laser of wavelength 1064 nm focused
68: using a 100X oil immersion objective lens. A He-Ne laser of
69: wavelength 632.8 nm was used to track the position of the sphere.
70: The position of the sphere was detected using a quadrant photo diode
71: (UDT instruments). The positional response of the sphere was then
72: locked on to the drive signal of the piezo plate using a SRS 830
73: lock-in amplifier and both phase and amplitude of the locked in
74: signal were analyzed. The separation between the particle and the
75: oscillating glass plate was varied by moving the piezo stage in the
76: $z$ direction. The trap stiffness and the corner frequency of the
77: trap were analyzed using the power spectrum of the equilibrium
78: positional fluctuations of the trapped particle.
79:
80: \section{Forces acting on the sphere due to oscillatory flow}
81:
82:
83:
84:
85:
86: The two extreme scenarios possible in the problem are (i) when the
87: sphere is in physical contact with the bottom plate and (ii) when
88: the sphere is far away from the glass plate. In the first scenario
89: the coupling between the sphere and the bottom plate is mainly
90: inertial and this is reflected in an ``in-phase'' response of the
91: sphere with respect to the motion of the bottom plate. When the
92: sphere is far away from the plate then the forces acting on the
93: sphere will be dominantly due to viscosity mediated shear stresses
94: and drag forces arising from the motion of the liquid. Therefore, in
95: this experiment we have two control parameters, i.e., the height of
96: the sphere from the plate and the velocity of the plate.
97:
98:
99:
100:
101: The motion of the liquid in presence of oscillatory shear is
102: governed by the Naiver Stokes equation $\rho\left( \frac{\partial
103: v}{\partial t}+(v.\nabla )v\right) =-\nabla p+\eta \nabla ^{2}v$,
104: where $v$ is the velocity, $\eta $ is viscosity ,$\nabla p$ is the
105: pressure gradient and $\rho $ is the density of the liquid. For the
106: present problem there is no pressure gradient, i.e., $\nabla p=0$,
107: and equation of continuity demands $\mathbf{\nabla \cdot \ v}=0$,
108: thus $(v.\nabla )v=0$. With the above assumptions the Naiver Stokes
109: equation can be written as $ \rho \frac{{\partial v_y }}{{\partial
110: t}} = \eta \frac{{\partial ^2 v_y }}{{\partial z^2 }} $. This has
111: traveling wave solutions of the form, $ v_y (z,t) = v_0 e^{i(kz -
112: \omega t)} $ with Stokes boundary layer, $\delta =\sqrt{\frac{2\eta
113: }{\omega\rho }}$ and $k=\pm \frac{(1+\imath )}{\delta }$ \footnote{
114: We have ignored the reflected travelling waves considering the top
115: plate to be at infinity. This is because the distance between the
116: plates is much larger compared to the boundary layer.} . The
117: frictional force acting on the sphere due to the velocity gradient
118: is in the direction of the liquid motion. The force per unit area is
119: given by
120: \begin{eqnarray}
121: S &=& \eta \frac{{\partial v}}{{\partial z}}=- \frac{{\sqrt 2 }}{\delta }\eta v_0 e^{ - i\left( {\omega t + \frac{\pi }{4}} \right)} e^{\left( {i - 1} \right)\frac{z}{\delta
122: }}.
123: \end{eqnarray}
124: We obtain the total frictional force ($F_s$) acting on the sphere
125: by computing the surface integral.
126: \begin{equation}
127: F_s = 2\pi \eta av_0 e^{ - \left( {\frac{{h + a}}{\delta }}
128: \right)} (1 - e^{\frac{{2a}}{\delta }} )e^{ - i\omega t},
129: \end{equation}
130: where $h$ is the height of the sphere from the glass plate.
131:
132: The forces experienced by the trapped particle \ at a height $h$
133: from the glass plate are (i) the \ spring like force $-k_{op}x$,
134: exerted by the tweezer; (ii) the viscous drag force $-6\pi \eta
135: a(\dot x-u)$,where $\dot x$ is the velocity of the sphere and $u$
136: the velocity of the liquid in the vicinity of the
137: sphere.(ii)hydrodynamic force $F_h$ and (iv) a frictional force
138: ($F_{plate}$) due to the interaction with the plate.
139:
140: Thus, the equation of motion of the sphere is
141: \begin{equation}
142: m_{eff} \ddot{x}+6\pi \eta a \dot x + k_{op} x = F_{h} +
143: F_{plate}+6\pi \eta a u,
144: \end {equation}
145: where $m_{eff}$ is the effective mass of the particle. We introduce
146: $m_{eff}$ in place of actual mass of the particle to account for the
147: effects of frequency, hydrodynamic interactions, and presence of
148: wall to the Stoke's drag.
149: The force acting on the sphere due to the
150: glass plate is mainly electrostatic in origin and hence strongly
151: dependent on the sphere-plate separation. When the sphere is far
152: away from the plate and frequency of oscillations are low, $F_h$ is
153: the frictional shear force $F_s$. However, in cases where the sphere
154: is within the stokes boundary layer and the length scale of the
155: vortex shedding is greater than the plate-sphere separation one
156: expects flow field around the particle to be altered by the presence
157: of bottom plate and the inertial effects to differ from those for a
158: particle in an unbounded flow \cite{FISCHER_oscillatory,Sumer}.Thus,
159: for the sphere near the bottom plate the hydrodynamic force will
160: differ from the simple minded frictional shear force.
161:
162:
163:
164:
165:
166:
167:
168: \section{Experimental Results}
169:
170:
171:
172:
173:
174:
175: We will first discuss the results which show transition of the
176: motion of the sphere dominated by inertial contact with the plate to
177: the regime where the motion of the sphere is mainly driven by the
178: liquid velocity. Top panel of Fig. \ref{fig:fig2} shows the motion
179: of the sphere (solid line) suspended in water and the corresponding
180: drive signal to the $y$-direction of the piezo (dotted line). The
181: data plotted is for a constant driving frequency (1Hz) and constant
182: drive amplitude(1$\mu m$ ) for various sphere-plate separation ($h$)
183: \footnote{The heights have been calculated from the $z$ displacement
184: of the piezo.}. The drive signal shown in the top panel is only a
185: guide to the eye and the value on the y axis is not a reflection of
186: its absolute magnitude. When $h$ ($\sim$ 0.1 $\mu$ m) is small, the
187: response of the sphere is in phase with the drive signal. As the
188: plate-sphere separation is increased the motion of the sphere
189: develops a phase lag with respect to the plate. This is shown in the
190: bottom panel of Fig. \ref{fig:fig2} where the histogram of the phase
191: of the motion of the sphere(averaged over 3 seconds) is plotted as a
192: function of the sphere-plate separation.`` Phase '' of the motion of
193: the sphere is defined by the phase of the locked in signal with
194: respect to the driving signal. The phase of the sphere's motion
195: confirms the two regimes-inertia coupled (in-phase) and velocity
196: coupled (finite phase lag). In between the two extremes of in phase
197: and $\sim$ $\pi /2$ out of phase motion, the histogram of the phase
198: of the motion of the sphere shows a broad distribution \footnote {On
199: carefully inspecting the displacement of the sphere one finds the
200: sphere to show slip-stick behavior.}.
201:
202:
203: We now present our results which measures the effect of large drive
204: frequencies on the motion of the sphere for appreciably large values
205: of $h$. In this case the sphere executes a $\pi /2$ out of phase
206: motion with respect to the motion of the bottom plate for low drive
207: frequencies. This confirms that the motion of the sphere is
208: predominantly velocity coupled. We have used two aqueous mediums of
209: suspension, namely water($\eta = 1mPas$) and glycerol($\eta=
210: 760mPas$) for this study. Left panel of Fig. \ref{fig:fig3} shows
211: response of the sphere(solid spheres joined by line) suspended in
212: glycerol as the frequency of the oscillation of the plate is varied
213: when the sphere-plate separation is constant at about $\sim$ 1.3
214: $\mu$m (sphere is far away from the plate).The drive amplitude of
215: the plate was kept constant at $0.1 \mu m$. The trace of the
216: velocity of the liquid is shown by a solid line. Note that the phase
217: of the velocity of the liquid is $\pi /2$ phase shifted with respect
218: to that of the bottom plate.\footnote{$y_{p}= y_{p0}sin(\omega t)$,
219: $\dot y_{p}=y_{p0}\omega sin(\omega t+\pi /2)$, where $\dot y_p$ is
220: the velocity of the plate and assuming no-slip boundary condition is
221: also the velocity of the liquid.} It can be seen that at 2Hz the
222: sphere moves almost in phase with the velocity of the liquid.
223: However, as the drive frequency is increased the sphere develops a
224: $\sim \pi /2$ phase difference for drive frequency of 11Hz and $\sim
225: \pi$ phase difference for a drive frequency of 40 Hz. The
226: corresponding average phase is shown by solid circles joined by line
227: in top panel of Fig.\ref{fig:fig4}. The inset of the top panel in
228: fig. \ref{fig:fig4} shows the amplitude of the locked in signal.
229:
230: The right panel of Fig. \ref{fig:fig3} shows response of the
231: sphere(solid spheres joined by line) suspended in water as the
232: frequency of the oscillation of the plate is varied when the
233: sphere-plate separation is constant at about $\sim$ 1.3 $\mu$m
234: (sphere is far away from the plate).The drive amplitude of the plate
235: was kept constant at $1 \mu m$. The sphere's motion at the driving
236: frequency of 2Hz is purely sinusoidal and monochromatic, with a
237: dominant frequency of 2Hz. At this frequency the motion of the
238: sphere is almost in phase with the velocity of the liquid. However
239: as the frequency is increased, the motion of the sphere develops a
240: phase lag with respect to the velocity of the liquid. This is also
241: accompained by a large distortion of the waveform of the motion of
242: the sphere. It is interesting to note that such distortions are
243: absent for the case of glycerol.The corresponding average phase is
244: shown by solid circles joined by line in bottom panel of
245: Fig.\ref{fig:fig4}. The inset of the bottom panel in fig.
246: \ref{fig:fig4} shows the amplitude of the locked in signal.
247:
248:
249:
250: In the analysis of the phase as shown in fig.\ref{fig:fig4} we have
251: shifted the phase of the drive signal by $\pi/2$ to account for the
252: velocity contribution. That is, the phase is plotted with respect to
253: the velocity of the liquid. As the frequency is increased from 2Hz
254: to 60Hz the phase decreases monotonically from zero to $-\pi$.
255: Since the sphere is far away from the plate we ignore $F_{plate}$.
256: We can then consider this system as a forced damped harmonic
257: oscillator. The equation of motion of a forced damped harmonic
258: oscillator is $ \ddot x + \gamma \dot x + \omega _0^2 x = \frac{{F_0
259: }}{m}\cos \omega t $ which has a steady state solution $x= A \cos
260: \omega t$ where $A = \frac{{F_0 }}{m}\frac{1}{{\left[ {(\omega _0^2
261: - \omega ^2 )^2 + (\omega \gamma )^2 } \right]^{1/2} }} $ and $ \phi
262: = \arctan \left( {\frac{{\gamma \omega }}{{\omega ^2 - \omega _0^2
263: }}} \right) $. Comparing it with eq no 3 we see that in our case, $
264: \gamma = \frac {6\pi\eta a} {m_{eff}},
265: \omega_0^2=\frac{k_{op}}{m_{eff}}$.
266:
267:
268:
269:
270:
271: Figure \ref{fig:fig4} also shows the fit (open squares joined by
272: line) to Eqn., $\phi = \arctan \left( {\frac{{\gamma \omega
273: }}{{\omega ^2 - \omega _0^2 }}} \right) $ to the experimental data.
274: From the fit we obtain $\gamma=150$ and $\omega_0=11Hz$ for glycerol
275: and $ \gamma=220$ and $\omega_0=20Hz$ for water. This implies
276: implies $m_{eff}=0.095$x$10^{-6}$Kg for the case of glycerol and
277: $m_{eff}=0.085$x$10^{-9}$Kg for water. The value of $k_{op}=m_{eff}
278: \omega_0^2$ obtained from the fit parameters for glycerol is
279: 5.96x$10^{-4}$N/m which agrees well with that obtained from the
280: equilibrium positional fluctuations of the
281: sphere(2.7x$10^{-4}$N/m).The value of $k_{op}=m_{eff} \omega_0^2$
282: obtained from the fit parameters for water is 13.4x$10^{-7}$N/m
283: which agrees well with that obtained from the equilibrium positional
284: fluctuations of the sphere(11.4x$10^{-7}$N/m). The quality of the
285: fit improves if $m_{eff}$ is considered as function of effective
286: mass. But in that case the equation of motion should be solved using
287: perturbation theory.
288:
289:
290:
291: Notice that $m_{eff}$ is much larger than the true mass of the
292: bead($\sim$ 4.18x$10^{-15}$Kg). We can understand this by saying
293: that it is not just the sphere but also the fluid around it which is
294: behaving like an oscillator.
295:
296: \section{Discussion}
297:
298: The relevant length scales perpendicular to direction of the flow,
299: that is, in the z-direction are the particle diameter,($2 \mu m$),
300: and the stokes boundary layer, $\delta =\sqrt{\frac{2\eta
301: }{\omega\rho }}$. In frequency range,$1Hz \ldots 60 Hz$, covered in
302: our experiments $\delta=13mm\ldots 1.6mm$ for glycerol and
303: $\delta=0.5mm\ldots 0.07mm$ for water. In the direction parallel to
304: the flow, one of the relevant length scales is again the particle
305: diameter ($2 \mu m$) and the other is associated with vortex
306: formation, shedding and potential interactions. For a sinusoidal
307: motion, the length scale associated with vortex formation is roughly
308: equal to amplitude of the plate motion \cite{FISCHER_oscillatory}
309: which in our experiments is about 0.1$\mu m$ for glycerol and 1$\mu
310: m$ for water.
311:
312:
313: The motion of the sphere comprises of two kinds of degrees of
314: freedom- translation and rotational. Translation motion arises
315: because whole fluid around it translates when the bottom plate is
316: sinusoidally driven. Since the sphere is trapped and is subjected to
317: shear stresses there will be a finite torque leading to the rotation
318: of the sphere about the $x$ axis with an angular velocity$ {\bf
319: \Omega } = \left( {\Omega,0,0} \right) $. No-slip boundary condition
320: on the sphere's surface ensures that the curl of the velocity of the
321: liquid is non zero around the surface of the sphere. The Navier
322: Stokes equation which defines the flow of the liquid about the
323: sphere is given by $ \nabla ^2 {\bf q} - \nabla p = \frac{\rho
324: }{\eta }\left( {{\bf q}.\nabla } \right){\bf q} + \frac{\rho }{\eta
325: }\frac{{\partial {\bf q}}}{{\partial t}}, $ where $\bf
326: q=(q_1,q_2,q_3)$ is the velocity of the fluid. The origin of the
327: cartesian coordinate is taken to coincide with the center of the
328: sphere. The boundary conditions are ${\bf q} \to 0$ when $z$ tends
329: to infinity, $ \bf q = \bf \Omega \times \bf r$ for the flow on the
330: sphere and ${\bf q} \to $ velocity of the plate for the flow on it,
331: here $\bf r = (x,y,z)$.This will lead to vortex formation which
332: decays with a length scale of the boundary layer thickness. It is
333: noteworthy that the stokes boundary layer is of the order of at
334: least tens of microns for both glycerol and water in the frequency
335: range covered in our experiments. Since the sphere is within the
336: boundary layer, one expects considerable potential interactions
337: between the rotational flow of the liquid and the oscillating plate.
338: This scenario could possibly result in an inertial coupling between
339: the plate and the sphere. This in turn could increase the effective
340: mass of the sphere and hence exhibit the observed phase behavior.
341:
342:
343:
344:
345:
346:
347:
348:
349:
350:
351:
352:
353: \section{Conclusion}
354: To our knowledge,there has been no experimental/computational fluid
355: dynamics simulations of a particle, rotating as well as
356: translating, in an oscillatory flow. Fischer \textit {et al} have
357: performed calculation of lift and drag forces on a stationary sphere
358: subjected to a pressure driven oscillatory flow. Our results are the
359: first direct measurements of the drag forces acting on a rotating
360: sphere subjected to an oscillatory motion. We find that the phase of
361: the motion of the sphere with respect to the drive is a sensitive
362: tool to study its dynamics. We have been able to explain our data in
363: terms of a damped harmonic oscillator. The effective mass that comes
364: out of the calculations is orders of magnitude greater than the bare
365: mass of the sphere. This highlights the importance of the role of
366: inertia in an otherwise viscosity dominated flow. We believe that
367: optical tweezers will be a effective tool to address fluid dynamics
368: problems at small length scales.
369:
370:
371: \newpage
372:
373: \begin{thebibliography}{1}
374: \bibitem{landau}
375: L.~D. Landau and E.~M. Lifshitz, \emph{Fluid mechanics}, (Second
376: edition1987).
377:
378:
379: \bibitem{FISCHER_oscillatory}
380: P.~F. Fischer, G.~K. Leaf, and J.~M. Restrepo, \emph{Forces on
381: particles in
382: oscillatory boundary layers}, Journal of Fluid Mechanics \textbf{468} (2002),
383: 327--347.
384:
385:
386:
387: \bibitem{sleath}
388: G.~N. Rosenthal and J.~F.~A. Sleath, \emph{Measurements of lift in
389: oscillatory
390: flow}, J. Fluid Mechanics \textbf{164} (1986), 449.
391:
392: \bibitem{Sumer}
393: B.~M. Sumer, B.~L. Jensen, and J.~Fredsøe, \emph{Effect of a plane
394: boundary on
395: oscillatory flow around a circular cylinder}, Journal of Fluid Mechanics
396: \textbf{225} (1991), 271--300.
397: \end{thebibliography}
398:
399:
400: \clearpage
401:
402:
403:
404: %\bibliographystyle{amsplain}
405: %\bibliography{draft}
406: \newpage
407: \begin{figure}
408: \centering
409: \centerline{\psfig{figure=fig1.eps,width=12cm} }
410:
411:
412: \caption{Schematic of the experimental setup. Liquid containing 2$\mu m$ colloidal particles is held between two glass plates separated by 2mm. The particle is optically trapped at a height $h$ from the bottom plate. The lower plate is subjected to an oscillatory motion $y_{p}=y_{po}\sin (\omega t)$, where $y_{p}$ is
413: the instantaneous displacement of the glass plate, $\omega$ is angular velocity and $y_{po}$ is the amplitude. The top plate is kept fixed. }
414: \label{fig:fig1}
415: \end{figure}
416:
417: \newpage
418: \begin{figure}
419:
420: \centering
421: \centerline{\psfig{figure=htvar_phase_xpos.eps,width=12 cm} }
422: \caption{ Top : The motion of the sphere(solid line) as a function of the sphere-plate separation: (a)$\sim$ 0.1 $\mu$m (b)$\sim$ 0.5 $\mu$m (c)$\sim$ 0.9 $\mu$m
423: (d) $\sim$ 1.3 $\mu$m. (e)$\sim$1.7 $\mu$m at a fixed driving
424: frequency of 1Hz and amplitude of 1$\mu m$. The trace showing the
425: motion of the plate (dotted line) is a guide to the eye, and the
426: corresponding $y$ axis is not a reflection of its absolute
427: amplitude.
428: Bottom : The corresponding histograms of the phase of the locked in signal as a
429: function of the sphere-plate separation. The y axis for the
430: histograms is in logarithmic scale. }
431: \label{fig:fig2}
432: \end{figure}
433:
434:
435: \newpage
436: \begin{figure}
437: \centering
438: \centerline{\epsfig{figure=time_water_glycerol.eps,width=12 cm} }
439: \caption{ Displacement of the sphere (solid circles joined by lines) in response to the velocity of the liquid (Solid line) for few select frequencies. The frequencies are shown by the side of each trace. Left: The data shown is for glycerol, Right : The data shown is for water. The trace showing the velocity of the liquid is a guide to the eye, and the corresponding $y$ axis is not a reflection of its absolute amplitude. }
440: \label{fig:fig3}
441: \end{figure}
442:
443: \newpage
444:
445:
446: \begin{figure}
447: \centering
448: \centerline{\psfig{figure=phasevsfreq.eps,width=12cm} }
449: \caption{The variation of average phase with respect to the velocity of the liquid. The inset shows the amplitude of the locked in signal. Top: The data shown is for glycerol.
450: Bottom: the data shown is for water }
451: \label{fig:fig4}
452: \end{figure}
453:
454: \newpage
455:
456:
457:
458:
459:
460:
461:
462:
463: \end{document}
464: