Article pubs.acs.org/JPCC
Bubble-Propelled Microjets: Model and Experiment Manoj Manjare,† Bo Yang,‡ and Y.-P. Zhao*,† †
Nanoscale Science and Engineering Center, Department of Physics and Astronomy, The University of Georgia, Athens, GA 30605, United States ‡ Department of Mechanical and Aerospace Engineering, University of Texas, Arlington, TX 76019, United States ABSTRACT: Using basic principles of diffusion and reaction, a onedimensional mass transport model inside a catalytic microjet has been proposed. The effect of microjet geometry on mass transport has been investigated, and its effect on bubble growth and microjet motion has been predicted. Oxygen generation and its flux to one end of the microjet induce nucleation and growth of bubbles. The bubble growth and ejection/ burst cause the microjet to move forward. Numerical investigations of the motion-related parameters, such as O2 flux, bubble generation rate and frequency, and average speed of the microjet motors during bubble growth, are found to depend closely on the length and opening radius of the microjet and concentration of H2O2 in the surrounding environment. The predicted results are compared to experimental data obtained from graphene oxide-based microjets, and they show good qualitative agreement. These results demonstrate that this simple model could be used to optimize catalytic microjet design.
bubbles to propel the nanomotors.16 They have also used acidic solutions such as HCl to oxidize Zn and produce H2 bubbles to power the motors.17 With the addition of magnetic components such as Ni or Fe, the moving direction of microjets can be controlled by external magnetic fields. This allows for maneuverability to be controlled in three dimensions, which increases the number of applications.18 In fact, several applications of catalytic microjets have already been demonstrated. By modification of microjets with self-assembled monolayers, microjets are able to adsorb oil, which could be used in water−oil separation.19 Microjet surfaces can be functionalized with nucleic acid probes to selectively detect and target nucleic acids.20 Microjets can also be used in diverse cargo towing applications with proper functionalization.21,22 In spite of the recent attention given to microjets, little is known regarding the detailed propulsion mechanisms of these microjets. Mei et al. have recently proposed a simple model, the body deformation model, which is based on the experimental observation that the average speed of a microjet is approximately equal to the product of the bubble radius and the bubble ejection frequency.18,23,24 The model assumes that the system consisting of the microjet and the bubble undergoes a deformation from “bubble inside the tube” to “bubble detached from the tube”. Only the state of the deformation, in which the bubble starts to expel out of the microjet until its detachment from the microjet,
1. INTRODUCTION Recently, using catalytic chemical reactions to propel micro- or nano-objects has attracted a great deal of attention.1−5 The most commonly used reaction is the decomposition of H2O2 by metallic catalysts, such as Pt or Ag.2,5 Many different shapes of micromotors have been produced, for example, Janus particles,6 rod-shaped nanomotors,7 microjets,4,8 and motors with other shapes.9−12 Among all these motors, microjets show the highest moving speed and have been studied extensively.13 There are primarily two methods to fabricate tubular microstructures, stress engineering to roll up materials into microjets4 and casting materials into molds of microtubes.14 For rolled-up micromotors, multilayers of materials are deposited in the form of thin films, which introduces an engineered strain gradient and causes the sheets to roll up into nano/microtubes after a sacrificial layer has been etched away and the films are released.4 Microjets can also be fabricated without using sacrificial layers. For example, recently Yao et al. used graphene oxide as a template to make microjets.15 The casting method is employed in Wang’s group by growing material layers inside tube templates.8 The shape of the tubes is usually cylindrical or conical. All the methods used to make the microjets incorporate a layer of catalyst on the inner surface of the microjet. When aqueous fuel like H2O2 is introduced, it breaks down at the catalyst surface into H2O and O2. The O2 is released in the form of bubbles and is expelled out of the microjet from one end. This provides the primary mechanism for microjets to move. Other fuels have also been used to produce the bubble-propelled motion. Gao et al. have used Ga/Al in water where Al reacts with H2O to generate H2 © 2013 American Chemical Society
Received: December 5, 2012 Revised: February 3, 2013 Published: February 8, 2013 4657
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
Figure 1. Geometry of the cylindrical microjet, an attached bubble, and definitions of related geometric parameters. Horizontal forces acting on the microjet and the bubble during the bubble growth are also illustrated.
the microjet and mass transport of O2 out of the microjet are restricted and are only possible from the two open ends of the microjet. H2O2/O2 is transported in/out of the microjet by means of diffusion. Thus, the mass transport of H2O2 and O2 depends strongly on the physical dimension of the microjet. When immersed in an aqueous solution of H2O2, both ends of the microjet are open for H2O2 intake immediately before the reaction starts to happen. Assuming that the diffusion of H2O2 in and O2 out of the microjet follows Fick’s first law,26 Jp = −Dp dCp/dx,where Jp is the flux (number of moles of matter passing through a unit area in unit time), Dp is the diffusion constant, Cp is the concentration (number of moles per unit volume), and subscript “p” denotes species, H2O2 or O2. Consider a cylindrical microjet of radius R and length L as shown in Figure 1, which has a cross section of area A = πR2 and perimeter B = 2πR. At the inner Pt-coated surface of the microjet, reaction 2H2O2 → 2H2O + O2 occurs, consuming H2O2 while releasing O2. Considering the processes, Fick’s second law (i.e., the law of mass conservation) requires
contributes to the forward thrust of the microjet. This phenomenological model is sufficient to explain only the average speed of a microjet in certain cases. However, the detailed theoretical considerations of mass transport and its effect inside the microjet, along with the geometrical parameters on the motion of the microjets, have not been considered. The nucleation and growth of the bubble, along with how the bubbles eject from the microjet, have not been included in the model, which could be very important for bubble-governed dynamics. Recently we have shown that these parameters are of key importance to understanding the quasi-oscillatory motion of microbead-based catalytic motors.25 Therefore, we believe that for microjet motors, these factors will be helpful in understanding the fundamental physical and chemical processes occurring at the site of the catalyst, and they could prove important principles for future designs of microjet-based motors. In this article, we develop a simplified one-dimensional mass transport model to describe the motion dynamics of selfpropelling microjets. We investigate the effects of geometrical parameters such as microjet length and opening diameter on the mass transport of H2O2 and O2 across the microjet. O2 generation and the corresponding flux of O2 to one end of the microjet induce nucleation and growth of bubbles. The bubble growth and ejection/burst then in turn cause the microjet to move forward. Numerical predictions of the motion-related parameters, such as O2 flux, bubble generation rate and frequency, and average speed of the microjet motors during bubble growth, are presented. The bubble ejection and burst mechanisms are discussed briefly. The predicted results are compared to the experimental data obtained from graphene oxide-based microjets, and they are in good qualitative agreement with each other.
∂C H2O2 ∂t ∂CO2 ∂t
= DH2O2
= DO2
∂ 2C H2O2
∂ 2CO2 ∂x 2
∂x 2 +
−
BK C H2O2 A
BK CH O 2A 2 2
(1)
(2)
where t is the time and K is the reaction rate constant of H 2O 2 per unit area at a Pt surface. It is assumed that the reaction rate is proportional to local H2O 2 concentration. We do not consider the effect of H 2O generated through the reaction for two reasons. First, we are using the waterbased solvent, which means the entire reaction takes place in water. The H2O 2 concentrations used in the experiments are usually small; particularly in our experiments, we use 5% H 2O 2. Compared to the solvent, the amount of water generated through the catalytic reaction is almost negligible and will not affect the local concentration of H 2O2 and O 2. Second, the small amount of water generated in the reaction may cause convection, but this effect should be negligible compared to the bubble-driven hydrodynamic flow. If the reaction is fast, the concentration distribution inside the microjet will rapidly approach a steady state, i.e., ∂Cp/∂t = 0.
2. MASS TRANSPORT IN MICROJET Microjets are cylindrical or conical in shape and have a layer of a catalyst, such as Pt, on their inner surface. The motion of microjets is a direct consequence of bubble growth and bubble burst/ejection at one end of the microjet, as the microjet is pushed away from the bubble burst/ejection site. Bubble generation results from the catalytic decomposition of H2O2 into O2 along the inner wall of the microjet, which then diffuses to a bubble nucleation center at one end of the microjet. Due to the geometric confinement of the microjet, the flux of H2O2 into 4658
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
that no bubbles will nucleate in the center of the microjet. The position of the O2 concentration maximum is given by
Thus, eq 1 and eq 2 can be simplified into DH2O2 DO2
∂ 2C H2O2 ∂x
∂ 2CO2 ∂x 2
2
−
BK C H2O2 = 0 A
(3)
BK + CH O = 0 2A 2 2
xmax =
CO2
(7)
Figure 3 shows variations of normalized CH2O2 and CO2 with x/L for various microjet lengths with fixed R = 6 μm. It shows
(4)
According to most of our experimental observations, we can assume that there is no bubble at one end of the microjet, say, the left end (x = −L/2), while a bubble grows at the other, right end (x = L/2), as shown in Figure 1. Thus, we have the following boundary conditions, CH2O2|x=−L/2 = CH2O2∞, (∂CH2O2)/(∂x)|x=+L/2 = 0, and CO2|x=±L/2 = 0. Here we assume that any O2 produced is expelled into gaseous phase at the right end of the microjet, which contributes to the growth of the bubble. The bubble seals the right end of the microjet. The concentrations of O2 and H2O2 at the microjet ends are those of the open environments, except H2O2 at the bubble-sealed end. Solving eqs 3 and 4 along with the boundary conditions, one has C H2O2 =
⎛ 1 ⎞ 1 L sinh−1⎜ (1 − cosh(βL))⎟ + β 2 ⎝ βL ⎠
C H∞2O2
⎡ ⎛ L ⎞⎤ cosh⎢β ⎜x − ⎟⎥ ⎝ ⎣ cosh(βL) 2 ⎠⎦
(5)
⎛ ⎞ DH2O2C H∞2O2 ⎜ 1 ⎛ 1 1 = − 1⎟x + ⎜ ⎜ 2DO2 ⎜ L ⎝ cosh(βL) 2cosh(βL) ⎠ ⎝
((
L
cosh β x − 2 1 + − 2 cosh(βL)
Figure 3. Normalized CH2O2 and CO2 by C∞ H2O2 versus normalized location x/L inside the microjet for different microjet lengths L at R = 6 μm.
)) ⎞⎟ ⎟ ⎟ ⎠
that CH2O2 decreases more rapidly along the microjet as the microjet gets longer. This could be due to the fact that while the supply of H2O2 is restricted by the microjet opening, there is more Pt available for decomposition of H2O2 inside the tube. This results in greater consumption of H2O2 and generation of O2. As the length of microjet increases, the maximum O2 concentration becomes larger. This directly supports the fact that more H2O2 is consumed for longer microjets resulting in greater O2 production. 2.1. Bubble Growth at the End of the Microjet. The O2 produced in the microjet accumulates and moves toward the right end. It is expelled into the bubble, as shown in Figure 1. The oxygen generation, i.e., the flux of O 2 going into the bubble, can be obtained using Fick’s first law, JO2 = −DO2((dCO2)/(dx)):
(6)
where β = ((2K)/(DH2O2R)) . Figure 2 shows a few representative plots of normalized CH2O2 and CO2 by C∞ H2O2 along the 1/2
JO |x = L /2 2
Figure 2. Normalized CH2O2 and CO2 by C∞ H2O2 versus normalized location x/L inside the microjet for different radii R at fixed L = 60 μm.
⎞ DH2O2C H∞2O2 ⎛ 1 =− − 1⎟ ⎜ 2L ⎝ cosh(βL) ⎠
(8)
The total mass of O2 released at the end of the microjet, the bubble growth rate GO2, can be written as GO2 = JO2AMO2, where MO2 is the molar mass of O2. From a good number of bubble growth studies, we know that the bubble volume grows linearly with time, Vb = ((4π)/3)γ3t, so that the bubble radius Rb follows a power law with respect to the growth time t, Rb = γt1/3.28,30 Since GO2 is the amount of oxygen blown into bubbles per unit time, one has γ = ((3GO2)/(4πρO2))1/3, where ρO2 is the mass density of oxygen; thus
microjet axis (x/L) for various microjet opening radii R with a fixed microjet length L = 60 μm. Here, we use K = 6.83 × 10−7 m/s (derived from Paxton et al.7), DH2O2 = 1.43 × 10−9 m2/s, and DO2 = 2.06 × 10−9 m2/s.27 The concentration of H2O2 decreases monotonically with x, i.e., from the open end to the bubble end. With the increase of R, the overall CH2O2 value increases, which is the result of the decreasing effective reaction rate constant BK/A scaled inversely with R in eq 1. In contrast, CO2 exhibits a maximum value near the center and hence the gradient for release of O2 out of the microjet. Here we assume
GO2 4659
⎞ DH2O2C H∞2O2πR2MO2 ⎛ 1 =− − 1⎟ ⎜ 2L ⎝ cosh(βL) ⎠
(9)
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
1/3 ∞ 2 ⎡ 3D ⎞⎤ 1 H 2O2 C H 2O2R MO2 ⎛ ⎢ ⎥ − 1⎟ γ= − ⎜ ⎢⎣ 8ρO L ⎝ cosh(βL) ⎠⎥⎦ 2
where Fgrowth = ρwπRb2((3/2)CsṘ b2 + RbR̈ b), with ρw the mass density of water and Cs an empirical constant, and the drag force, Fdrag = (2πηLv)/(ln(L/R) − 0.72) for a tubular microobject. The above drag force formula was originally derived for a solid cylinder of length L and radius R.29 We numerically verified its applicability to a thin-walled tube as well by using a boundary element method. Since Rb = γt1/3, the instantaneous speed v of the microjet during bubble growth can be expressed as
(10)
Figure 4 shows JO2 and GO2 as a function of microjet length L for three different values of R. JO2 is always greater for smaller
v = (β
∫0
t
t ′−2/3 ek t′dt ′)e−kt
(14)
where β = γ (ρπ/m)[(3/2)Csn + n(n − 1)], k = (2πηLv)/ (m[ln(L/R) − 0.72]), and m is the mass of the microjet. The average speed of the microjet during the bubble growth can be expressed as 4
vavg = f
∫0
2
1/ f
(β
∫0
t
t ′−2/3 ek t′dt ′)e−kt dt
(15)
Equation 15 can be solved numerically.
3. EXPERIMENTAL RESULTS AND DISCUSSION In order to evaluate the predictions given by the proposed model, we have designed systematic experiments on microjets based on rolled-up graphene oxide (GO). The fabrication of these microjets is described in detail in ref 15. The microjets are made from GO nanosheets which are spontaneously rolled up into microjetlets after thin Ti and Pt layers are deposited onto them. The resulting microjets are long and thin with small openings and multilayered walls. Figure 5 shows a representative scanning electron microscope (SEM) image of one GO microjet. As one
Figure 4. Predicted JO2 and GO2 versus microjet length L for different microjet radii R.
openings at any fixed length. However, GO2 is always smaller for smaller openings. It is clear that the openings of the microjet play an important role in bubble growth. Figure 4 shows that for a microjet with R = 2 μm, there is a value of L for which JO2 reaches a maximum. From eq 8, one can find the condition for this JO2 maximum L = 1.506
DH2O2 2K
R1/2
(11) 7
For the specific K values observed in Paxton et al., we have L = 4.87 × 10−2R1/2. This gives a criterion for optimizing microjet design for maximizing the O2 generation and could potentially maximize the speed of microjets. GO2 is more relevant to the growth of large bubbles, while JO2 is critical for bubble nucleation and early stage growth. Figure 2 shows that microjets with smaller openings have a higher concentration of O2 at any point in the microjet than microjets of the same length but with bigger openings; this translates directly into flux, JO2. However, since the overall Pt area available for reaction is greater in microjets with larger openings, the total oxygen produced per unit time is greater, which directly translates into larger GO2. The amount of oxygen expelled from the microjet, GO2, can be used to calculate the frequency f for bubble ejection or burst given the maximum radius of the bubble RB f=
Figure 5. Representative SEM image showing the morphology of a GO microjet.
3GO2 4ρO πRB3 2
can see, the microjet is not perfectly cylindrical, but rather tapered. The length L and mean opening radius R of the microjet are 30 and 2.5 μm, respectively. Most of the microjets studied in these experiments have similar dimensions, with L varying from 15 to 40 μm and R varying from 1.5 to 3 μm. In order to study the motion behavior, 5 μL of an aqueous dispersion containing the fabricated microjets was pipetted onto a clean Si substrate, followed by the introduction of 5 μL of 10% H2O2 to activate the motion. After a steady reaction rate was reached and observed (less than 1 min), pictures and videos of microjet motion were captured by a Mitituya FS110 optical microscope with a fast charge-coupled device (CCD) camera (Phantom v9.1) mounted onto the trinocular head, using 10× magnification objective lenses. The location of an observed microjet was marked using a copper mesh by recording its relative position to the mesh after
(12)
2.2. Bubble Growth-Induced Motion. During the bubble growth, there are different forces exerted on the microjet which could make the microjet move as discussed in a recent paper.25 In particular, the bubble growth will exert a growth force on the object to which the bubble is attached, and this force pushes the object forward as shown in Figure 1 (here we have neglected all the vertical forces since all observed motion is horizontal). The growth force, which arises from the formation and growth of bubble surface against the fluid environment, is the only driving force on the microjet. While moving in a liquid environment, a microjet also experiences a drag force. The net force on the microjet is given by Fnet = Fgrowth − Fdrag
(13) 4660
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
Table 1. Geometrical Parameters of the Individual Tubes and the Experimentally Derived Quantities for Each Tube microjet
length L (um) (±0.5 um)
radius R (um) (±0.25 um)
maximum bubble radius RB (um) (±0.25 um)
bubble frequency f (/s)
bubble growth prefactor γ (m/s1/3)
growth rate Gexp = ρf Vb (kg/s)
exptl reaction rate K (m/s)
avg microjet speed vavg (μm/s)
bubble ejection/ burst mode
1 2 3 4 5 6 7 8 9 10
23.4 15.5 17 16 25.8 21.6 23 24 23 20
2.8 1.8 1.8 1.8 2.5 2.3 2.5 2.9 2 2.5
2.0 2.5 2.3 2.5 2 2 2 2 2 2
580 ± 2 300 ± 5 310 ± 3 308 ± 1 380 ± 4 206 ± 5 569 ± 5 740 ± 5 268 ± 4 66 ± 3
2.0 × 10−5 1.9 × 10−5 2.1 × 10−5 1.7 × 10−5 1.7 × 10−5 1.2 × 10−5 1.7 × 10−5 1.8 × 10−5 1.3 × 10−5 8.0 × 10−6
2.7 × 10−14 2.8 × 10−14 2.1 × 10−14 2.9 × 10−14 1.8 × 10−14 9.8 × 10−15 2.7 × 10−14 3.5 × 10−14 1.3 × 10−14 3.2 × 10−15
3.06 × 10−5 5.41 × 10−5 1.80 × 10−5 4.41 × 10−5 1.65 × 10−5 1.04 × 10−5 1.85 × 10−5 1.65 × 10−5 2.1 × 10−5 1.1 × 10−5
580 590 510 960 470 360 580 630 400 87
burst burst burst burst burst burst burst burst ejection ejection
time t for three different microjets. The accumulative distance is calculated by s = Σ((xi+1 − xi)2 + (yi+1 − yi)2)1/2, where xi and yi are the coordinates of the microjet in the ith frame. The s−t plots show a linear relationship, indicating that all the microjets are moving at a constant average speed, even though the instantaneous speed of the microjet changes rapidly during the growth-burst/ ejection cycle of the bubbles. The average speeds of the microjets are in the range of 100−1000 μm/s. However, the instantaneous speed of the microjet is not constant due to bubble growth and burst/ ejection. The instantaneous speed is calculated using vinst = ((xi+1 − xi)2 + (yi+1 − yi)2)1/2/Δt, where Δt is the time interval between each adjacent frame. Figure 7b shows plots of the instantaneous speed versus time for three of the microjets. The plot exhibits a pattern of low speed and high speed jumps. The black-filled symbols represent the microjet speed during bubble growth, while the gray open symbols represent speed right after the burst or ejection events. The high speed jumps that are marked by circles are due to the bubble ejection/burst events. The motional behaviors depicted in Figure 7b are discretized since the instantaneous speeds are obtained from frame to frame and we are limited by the spatial resolution of the CCD. For example, at 100 μs temporal resolutions the movement of the microjet may not be discernible in the given spatial resolution in certain subsequent frames; hence, the instantaneous speeds are calculated as zero for these instances. From this plot, and the observation of bubble ejection/burst, we can determine the bubble ejection/burst frequency f, which is summarized in Table 1. Using the high speed CCD camera to capture both bubble growth radius Rb and the microjet displacement, we can verify whether the assumption in the model proposed by Li et al. is valid.18,23,24 According to the model proposed by Li et al., the center of the bubble is stationary during bubble growth, so that the motion of the microjet is caused by the expansion of the bubble. Therefore, the microjet displacement should be the same as the radius Rb of the bubble. Figure 8a plots the bubble radius Rb as function of time t for five cycles of bubble growth-burst/ ejection for one of the GO microjets along with the accumulated microjet moving distance s during the bubble growth. Clearly, the microjet moving distance s is significantly smaller than the bubble radius Rb at any given time. This is inconsistent with the assumption by Li et al. In addition, eq 15 shows that the relation between frequency f and average speed vavg is not as simple and linear as depicted in ref 23. Our experimentally observed vavg versus RB × f relationship is only partially consistent with the model proposed by Li et al. In Figure 8b, we compare the predicted and observed average speed in ref 23 with the average speed versus the product RB × f observed in our experiments. We see that the average speed observed in our experiments does not match well with the theory predicted in ref 23.
the liquid had dried. This microjet was then observed by SEM, and the geometric parameters were measured. Using this process, the motion of a particular motor could be directly linked to its morphological parameters. We observe that most of the microjets exhibit circular motion with bubbles coming from only one particular end of the microjet in the high speed and high resolution videos. Most of the microjet motions result from bubble burst, and only a few are due to bubble ejection. We can directly obtain the following information from the videos: the length L and radius R of the microjet, the bubble ejection or burst events, the dynamic change of the radius Rb(t) of the bubble, and coordinates x(t), y(t) of the microjet. From these measurements we can determine the motion trajectory, the instantaneous speed v, average speed vavg of the microjet, the maximum bubble size RB, the flux of O2 at the end of the microjet JO2, oxygen generation rate GO2, the bubble ejection/ burst frequency f, and the prefactor for bubble growth γ. All of the experimentally derived parameters are summarized in Table 1. After introducing them in aqueous H2O2 solution, the reaction takes place inside the microjets, and we observe that the bubble is almost always ejected from or bursts at the end with greater radius. The bubble growth and burst/ejection cause the microjet to move forward. Figure 6 plots representative trajectories of
Figure 6. Representative 2D trajectories of five different GO microjets.
several different microjets for movies taken at 10 000 fps. Since the ends of the microjets are not always symmetric, the bubbles eject/burst at an angle rather than parallel to the long axis of the microjet. This results in a circular trajectory, as shown in Figure 6. Figure 7a shows the plots of accumulative distances s versus 4661
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
Figure 7. (a) Representative plots of accumulated distance s traveled versus time t by three different microjets. The linear fitting (red line) represents a constant average speed; (b) instantaneous speed v for three different microjets with gray circles representing bubble burst/ejection events.
Figure 8. (a) Bubble radius Rb and microjet displacement s versus bubble growth time t during bubble growth. The dotted lines represent the fitting usingRb = γtn, which can be used to extract γ experimentally. (b) Microjet average speed vavg versus RB × f. Predicted (dotted line) and experimental (○) values from ref 23 with experimental values of average speed of microjets with burst mechanism (■) and that with ejection mechanism (▲).
In fact, from the model proposed in section 2, all the parameters associated with microjet motion are related to GO2, i.e., the prefactor γ for the bubble growth. This γ can be extracted experimentally and compared to the theoretical prediction. From Figure 8, we find that Rb−t follows the relation Rb = γtn, and according to eq 10, the parameter γ is directly related to the shape of the microjet and the reaction rate, K
can compare the measured O2 rate Gexp with the predicted GO2 calculated using eq 9. The values of physical dimensions of the microjets and reaction rates K are obtained from Table 1. Figure 9a shows the comparison of Gexp and GO2 for microjets with different aspect ratios, ξ = L/2R. Although the exact values of Gexp and GO2 are different for different ξ, these two parameters follow the same trend. The discrepancy in the values could be due to several reasons: (1) the accuracy in determining γ values, (2) the assumption that ρO2 is constant within the bubble could be inaccurate, and (3) the accuracy in determining the bubble ejection/frequency f. Figure 9b plots the comparison of predicted frequency of bubble ejection/ burst with that of the measured values for different ξ. The predicted frequencies are calculated by eq 12 using the physical parameters such as L, R, K, and RB, obtained from Table 1. Again the theoretical prediction follows the experimental trend very well. The speed of the microjet depends closely on the dimensions of the microjet as well as the bubble’s growing radius Rb, the maximum bubble radius RB, and the frequency f as seen from eq 15. Note that the average speed in eq 15 is the speed of the microjet during bubble growth; however, it does not consider the effect of thrust induced by bubble burst or bubble ejection. It is observed that the process of burst or ejection provides a much greater thrust to the microjets than the thrust provided during bubble growth, as is seen in Figure 7b. Figure 10 shows the comparison of the predicted average speed vp with the measured average
2 ⎛ ⎞⎞ DH2O2R ⎛⎜ 3DH2O2C H∞2O2R2MO2 −1⎜ ⎟ ⎟ K= cosh ⎜ 2 3 ∞ ⎟⎟ 2L2 ⎜⎝ ⎝ 3DH2O2C H2O2R MO2 − 8γ ρO2 L ⎠⎠
(16)
By fitting the bubble growth curves shown in Figure 8 and using eq 16, we can estimate the reaction constant K. Table 1 summarizes the γ obtained by fitting the K values obtained from the experimental data along with the physical dimensions of each microjet for which the K was calculated. All the K values we obtained are on the order of 10−5 m/s. Experimentally, Paxton et al. have determined K to be 6.83 × 10−7 m/s7 while Li et al. obtained K to be 8.4 × 10−4 m/s.24 Our K values are within the range of these two values and can be considered as reasonable values. In addition, one can also measure the O2 generation rate Gexp using Gexp = fρO2Vb, where Vb is the volume of a single bubble, Vb = (4/3)πR3B, RB is the maximum radius of the bubble, and f is the measured bubble burst/ejection frequency. Here we assume that the O2 gas density is the same as that at the atmospheric pressure. We 4662
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
Figure 9. Comparison between experimental value (■), theoretical prediction (●) of (a) bubble growth rate GO2 and (b) bubble ejection/burst frequency f.
speed vavg of the microjets. The measured speeds are obtained from slopes of plots in Figure 7a for each microjet, while the estimated average speeds are calculated using eq 15 and the parameters from Table 1. The frequency f and parameter γ are obtained from eqs 12 and 10, respectively. Figure 10 shows that both vp and vavg follow the same trend, although the predictions significantly underestimate the magnitude of the speed since the theory does not take into account the effect of burst or ejection. Figure 10 also shows average experimental speed vgrowth during
of O2 at the end of the microjet. The maximum size of the bubble and how it leaves the end of the microjet are influenced by a number of factors. Below we discuss how bubble ejection or burst could occur. According to Zeng et al.,30 the maximum diameter of the bubble before it detaches from a microjet depends on the rate of bubble growth, interfacial tension, liquid pressure, buoyancy, etc. Since the growth and departure of bubbles is a dynamic process, the momentum exchange between the bubble and the liquid must be considered. The force equation for a growing bubble in one dimension can be expressed as Fnet = Fσ + Fgrowth + Fbuoy + Fexcess + Fwake
(17)
2
where Fσ = ((2σ)/Rb)πR is the force due to surface tension (σ is the interfacial tension), Fbuoy is the buoyant force on the bubble, Fexcess = 2πRσ is the contact pressure force due to surrounding liquid, and Fwake is the force created by the preceding bubble. We can neglect the effect of buoyant force since the motion of microjets is only on a horizontal plane. In addition, the magnitude of Fwake is estimated to be Fwake = 10−4Fgrowth; thus, we can also neglect Fwake. Therefore, the total force acting on the bubble can be written as ⎛3 ⎞ ⎛ 2σ ⎞ Fnet = 2πRσ − ρw πR b2⎜ CsṘ b2 + R bR̈ b⎟ + ⎜ ⎟πR2 ⎝2 ⎠ ⎝ Rb ⎠
Figure 10. Plot of the average speed of the microjet vavg (■), the average speed of the microjet during bubble growth vgrowth (▲), and the predicted average speed vp (●) versus the aspect ratio ξ of the microjets.
(18)
In quasi-equilibrium condition, Ftotal = 0. Therefore, the condition for bubble detaching from the microjet is
bubble growth obtained from measuring the motion of the microjet without the impact caused by bubble burst/ejection. The average speed without the impact of bubble burst/ejection is calculated by using vgrowth = s/t, where s is obtained from Figure 8. We see that the vgrowth values are much closer to the predicted values vp. The discrepancy in values could be attributed to same reasons given in the paragraph above.
⎛ ⎛3 ⎞ R⎞ 2σR ⎜1 + ⎟ ≤ ρw R b2⎜ CsṘ b2 + R bR̈ b⎟ ⎝ ⎠ Rb ⎠ 2 ⎝
(19)
Using the relation RB = γtB , where tB is the time when Rb = RB, one has 1/3
⎛ ⎛1 R ⎞ 2⎞ ⎟ = ρw γ 4t B−2/3⎜ Cs − ⎟ 2σR ⎜⎜1 + 1/3 ⎟ ⎝ 6 9⎠ γt B ⎠ ⎝
4. BUBBLE DETACHMENT: EJECTION OR BURST During the motion of the microjets, it is observed that the bubble grown at one end of the microjet either ejects from the microjet or bursts (disappears instantly) when it reaches a maximum size RB and imposes a large impulse to make the microjet move in the opposite direction. We cannot predict the behavior of the bubble after it reaches RB with the current one-dimensional mass transport model. In a quasi-steady state, there is a constant supply
⎛ R + t B = ⎜⎜ − 2 ⎝ γ
(20)
⎞3 ρw γ 4 ⎛ 1 R 2 2⎞⎟ ⎜ C − ⎟ + s 2σR ⎝ 6 9 ⎠ ⎟⎠ 4γ 2 −6
−6
(21) −3
Using R = 2 × 10 m, L = 25 × 10 m, σ = 72 × 10 N/m, ρw = 1000 kg/m3, and Cs = 6.67 from Zeng et al.,30 we get tB = 20 ms, 4663
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
■
which is close to the values (∼1−5 ms) obtained in our experiments. The process of bubble burst could be explained by the pressure difference created by the greater gap in O2 leak rate out of the bubble and O2 flux flowing into the bubble. In our onedimensional model, we did not consider the effect of O2 leak from the bubble to the solution. There are at least two reasons for the burst. First, the bubble growth is a dynamic process; it starts from its critical nucleation size R*; at this point it cannot seal one end of the microjet entirely. This means that during the bubble growth, there is still H2O2 flux coming from the bubble nucleation end. This could lead to higher production rate of O2 than that predicted by the current model. When the bubble becomes larger and covers more area of the microjet opening, the O2 flux becomes smaller. Meanwhile, the O2 in the bubble could have a tendency to dissolve back into the solution during bubble growth. This “leakage” effect can rapidly intensify as the bubble’s surface area becomes larger and the O2 concentration becomes smaller at the far side during the bubble growth. After the bubble is grown so large as to cover the microjet opening and hence prevent H2O2 intake, the O2 production would soon stop. At this point, the bubble would only lose O2 and the bubble can collapse abruptly, i.e., burst.25 The pressure inside the growing bubble is described by a general, Rayleigh−Plesset equation31 Pb = P∞ +
⎛ 4η ̇ 2σ 3 ⎞ + R b + ρ⎜R bR̈ b + Ṙ b2⎟ ⎝ Rb Rb 2 ⎠
Article
AUTHOR INFORMATION
Corresponding Author
*Fax: +1 706 542 2492. Tel.: +1 706 542 2485. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We acknowledge Dr. Christopher Barret and Prof. Tina Salguero for providing us with graphene oxide sheet suspension. We thank Brett Granberg and Derrik Mccoy for helping in some of the movie analyses. We also thank George Larsen for proofreading this manuscript. This research is funded by National Science Foundation under contract No. ECCS-0901141.
■
REFERENCES
(1) Ozin, G. A.; Manners, I.; Fournier-Bidoz, S.; Arsenault, A. Dream Nanomachines. Adv. Mater. 2005, 17, 3011−3018. (2) Paxton, W. F.; Sundararajan, S.; Mallouk, T. E.; Sen, A. Chemical Locomotion. Angew. Chem., Int. Ed. 2006, 45, 5420−5429. (3) Wang, J.; Manesh, K. M. Motion Control at the Nanoscale. Small 2010, 6, 338−345. (4) Mei, Y. F.; Solovev, A. A.; Sanchez, S.; Schmidt, O. G. Rolled-up Nanotech on Polymers: From Basic Perception to Self-Propelled Catalytic Microengines. Chem. Soc. Rev. 2011, 40, 2109−2119. (5) Gibbs, J.; Zhao, Y. Catalytic Nanomotors: Fabrication, Mechanism, and Applications. Front. Mater. Sci. 2011, 5, 25−39. (6) Gibbs, J. G.; Zhao, Y. P. Autonomously Motile Catalytic Nanomotors by Bubble Propulsion. Appl. Phys. Lett. 2009, 94. (7) Paxton, W. F.; Kistler, K. C.; Olmeda, C. C.; Sen, A.; St. Angelo, S. K.; Cao, Y. Y.; Mallouk, T. E.; Lammert, P. E.; Crespi, V. H. Catalytic Nanomotors: Autonomous Movement of Striped Nanorods. J. Am. Chem. Soc. 2004, 126, 13424−13431. (8) Manesh, K. M.; Cardona, M.; Yuan, R.; Clark, M.; Kagan, D.; Balasubramanian, S.; Wang, J. Template-Assisted Fabrication of SaltIndependent Catalytic Tubular Microengines. ACS Nano 2010, 4, 1799−1804. (9) Campuzano, S.; Kagan, D.; Orozco, J.; Wang, J. Motion-Driven Sensing and Biosensing Using Electrochemically Propelled Nanomotors. Analyst 2011, 136, 4621−4630. (10) Gibbs, J. G.; Zhao, Y. P. Design and Characterization of Rotational Multicomponent Catalytic Nanomotors. Small 2009, 5, 2304−2308. (11) Gibbs, J. G.; Kothari, S.; Saintillan, D.; Zhao, Y. P. Geometrically Designing the Kinematic Behavior of Catalytic Nanomotors. Nano Lett. 2011, 11, 2543−2550. (12) He, Y. P.; Wu, J. S.; Zhao, Y. P. Designing Catalytic Nanomotors by Dynamic Shadowing Growth. Nano Lett. 2007, 7, 1369−1375. (13) Gao, W.; Sattayasamitsathit, S.; Wang, J. Catalytically Propelled Micro-/Nanomotors: How Fast Can They Move? Chem. Rec. 2012, 12, 224−231. (14) Gao, W.; Sattayasamitsathit, S.; Orozco, J.; Wang, J. Highly Efficient Catalytic Microengines: Template Electrosynthesis of Polyaniline/Platinum Microtubes. J. Am. Chem .Soc. 2011, 133, 11862−11864. (15) Yao, K.; Manjare, M.; Barrett, C. A.; Yang, B.; Salguero, T. T.; Zhao, Y. Nanostructured Scrolls from Graphene Oxide for Microjet Engines. J. Phys. Chem. Lett. 2012, 3, 2204−2208. (16) Gao, W.; Pei, A.; Wang, J. Water-Driven Micromotors. ACS Nano 2012, 6, 8432−8438. (17) Gao, W.; Uygun, A.; Wang, J. Hydrogen-Bubble-Propelled ZincBased Microrockets in Strongly Acidic Media. J. Am. Chem. Soc. 2012, 134, 897−900. (18) Huang, G. S.; Wang, J. S.; Mei, Y. F. Material Considerations and Locomotive Capability in Catalytic Tubular Microengines. J. Mater. Chem. 2012, 22, 6519−6525. (19) Guix, M.; Orozco, J.; Garcia, M.; Gao, W.; Sattayasamitsathit, S.; Merkoci, A.; Escarpa, A.; Wang, J. Superhydrophobic Alkanethiol-Coated Microsubmarines for Effective Removal of Oil. ACS Nano 2012, 6, 4445−4451.
(22)
where Pb, P∞, Rb, σ, and ρ are the pressure of gas inside the bubble, the pressure of the liquid at remote distances (≈101 kPa, the atmospheric pressure), the radius of the bubble, the surface tension, and the mass density of water, respectively. The bubble bursts when the O2 flowing into the bubble supplied by the catalytic reaction is much smaller than the O2 flux from the bubble into the liquid.25 The solution to this question can be obtained by combining eq 21 with mass transport equation and dynamic boundary conditions.
5. CONCLUSION In summary, we have formulated one-dimensional reaction diffusion equations to describe the mass transport and reaction in a cylindrical microjet. With this model, we can predict the H2O2 consumption rate and distribution, O2 generation, transport and flux at one end of the microjet, the O2 bubble growth, ejection/ burst frequency given the maximum radius, and the average speed of the microjet during bubble growth. The predicted O2 growth rate, bubble ejection/burst frequency, and the microjet speed during the bubble growth are compared with the experimental values. They agree qualitatively well and follow the same trends versus microjet geometry. The current model cannot predict the overall microjet speed due to the ejection or burst of the microjet, but the physical processes that could induce either ejection or burst have been discussed. The conditions for those processes could be solved numerically by combining mass transport equations, Rayleigh−Plesset equation, and dynamic boundary conditions. Obviously, the location of the bubble nucleation and how it seals one end of the microjet add complexity to the problem. The nonuniformity of the microjet shape, i.e., the conic shape of the microjet, should also be considered in future models. 4664
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665
The Journal of Physical Chemistry C
Article
(20) Kagan, D.; Campuzano, S.; Balasubramanian, S.; Kuralay, F.; Flechsig, G. U.; Wang, J. Functionalized Micromachines for Selective and Rapid Isolation of Nucleic Acid Targets from Complex Samples. Nano Lett. 2011, 11, 2083−2087. (21) Wang, J. Cargo-Towing Synthetic Nanomachines: Towards Active Transport in Microchip Devices. Lab Chip 2012, 12, 1944−1950. (22) Solovev, A. A.; Xi, W.; Gracias, D. H.; Harazim, S. M.; Deneke, C.; Sanchez, S.; Schmidt, O. G. Self-Propelled Nanotools. ACS Nano 2012, 6, 1751−1756. (23) Solovev, A. A.; Mei, Y. F.; Urena, E. B.; Huang, G. S.; Schmidt, O. G. Catalytic Microtubular Jet Engines Self-Propelled by Accumulated Gas Bubbles. Small 2009, 5, 1688−1692. (24) Li, J. X.; Huang, G. S.; Ye, M. M.; Li, M. L.; Liu, R.; Mei, Y. F. Dynamics of Catalytic Tubular Microjet Engines: Dependence on Geometry and Chemical Environment. Nanoscale 2011, 3, 5083−5089. (25) Manjare, M.; Yang, B.; Zhao, Y. P. Bubble Driven Quasioscillatory Translational Motion of Catalytic Micromotors. Phys. Rev. Lett. 2012, 109, 128305. (26) Fick, A. Ueber Diffusion. Ann. Phys. 1855, 170, 59−86. (27) Vanstroebiezen, S. A. M.; Everaerts, F. M.; Janssen, L. J. J.; Tacken, R. A. Diffusion-Coefficients of Oxygen, Hydrogen-Peroxide and Glucose in a Hydrogel. Anal. Chim. Acta 1993, 273, 553−560. (28) Thorncroft, G. E.; Klausner, J. F.; Mei, R. An Experimental Investigation of Bubble Growth and Detachment in Vertical Upflow and Downflow Boiling. Int. J. Heat Mass Transfer 1998, 41, 3857−3871. (29) Happel, J. Low Reynolds Number Hydrodynamics with Special Applications to Particulate Media; Noordhoff International Publishing: Leyden, The Netherlands, 1973. (30) Zeng, L. Z.; Klausner, J. F.; Mei, R. A Unified Model for the Prediction of Bubble Detachment Diameters in Boiling Systems.1. Pool Boiling. Int. J. Heat Mass Transfer 1993, 36, 2261−2270. (31) Franc, J.-P.; Michel, J.-M. Fundamentals of Cavitation: Kluwer Academic Publishers: Dordrecht; Boston, MA, 2004.
4665
dx.doi.org/10.1021/jp311977d | J. Phys. Chem. C 2013, 117, 4657−4665