stable magnetic chemical locomotive with pd nanoparticle incorporated ferromagnetic oxide

Published: June 15, 2011

r 2011 American Chemical Society 12708 dx.doi.org/10.1021/jp111890d | J. Phys. Chem. C 2011, 115, 12708–12715

ARTICLE

pubs.acs.org/JPCC

Stable Magnetic Chemical Locomotive with Pd NanoparticleIncorporated Ferromagnetic OxideKrishna Kanti Dey,† Kula Kamal Senapati,‡ Prodeep Phukan,^ Saurabh Basu,§ and Arun Chattopadhyay*,†,||

†Centre for Nanotechnology, ‡Central Instruments Facility, §Department of Physics, and )Department of Chemistry, Indian Instituteof Technology Guwahati, Guwahati 781039, India^Department of Chemistry, Gauhati University, Guwahati 781014, India

bS Supporting Information

1. INTRODUCTION

Controlled autonomous movement of submicrometer scaleobjects inside a liquid holds promise toward the development ofsmart micro/nanobots for efficient transportation of materials atthis length scale. Manipulation of dynamics of these objectsinside a liquid not only put forward scopes for effective orga-nization of useful biomolecules1,2 but also opens the possibilitiesfor minimally invasive surgeries.3 Attainment of guided self-propulsion of small-scale objects might prove to be significanttoward the development of “lab-on-a-chip” heterogeneous micro-systems4 and could also be useful for self-assembly of molecularstructures,5 energy conversions,6 and micropump designs.7 Thevery first attempt toward realizing controlled autonomous trans-port was to harness the self-propelling ability of various biomo-lecular proteins,8 which spontaneously carry chemical payloadswithin a cell with an efficiency that is difficult to achieve with aninorganic analogue. The attempt was to couple motor proteinswith inorganic structures and then to let them move along themicrotubular network in a directedmanner. Applications of thesebiointegrated motors were, nevertheless, found to be limited notonly for the need of defined environments for protein activationbut also for the quick natural degradation of these molecules.9

The next significant approach toward small-scale autonomoustransport was the development of artificial microscale swimmersusing chemical reaction catalyzed on their surface,10 followed byfabrication of a variety of microstructures, which moved sponta-neously inside dilute hydrogen peroxide (H2O2) solution bycatalytically decomposing the liquid.11,12 Some of those objectswere incisively steered inside the liquid using external magnetic

fields that marked their development as potential attemptstoward targeted delivery.13 Magnetic manipulation of self-pro-pelled structures, while they move autonomously inside a liquid,has always been of importance since the magnetic field canpermeate through most of the fluids and also offers greater rangeof selectivity in attaining such controls.3 Recently, using Au/Ni/Au/Pt-CNT nanowires, Burdick et al. demonstrated controlledpropulsion of micrometer scale polystyrene beads coated withiron oxide particles along predetermined trajectories within amicrochannel network containing solutions of 5 wt % H2O2.

14 Aweak external magnetic field was used to direct the objects at thejunctions of the network, without affecting their speed. Inanother interesting report, Dreyfus et al. have demonstratedthe motion of micrometer scale magnetic structures, where anexternal oscillating magnetic field was used to propel as well as todirect these structures during their movement. The structuresconsisted of linear chains of colloidal magnetic particles linked byDNA strands, attached with red blood cells which behaved likeflexible artificial flagella, in the presence of the external magneticfield. Although such a magnetically actuated colloidal devicepromised to be useful in precise and selective positioning ofmicro objects or in realizing controlledmotion at small scales, themotion, however, was not autonomous.15

In most of the reports on autonomously moving objectsrealizing directional manipulation, the metal that has been used

Received: December 15, 2010Revised: April 26, 2011

ABSTRACT: In this article, we present results on the development of anewmagnetic chemical locomotivemade of ferromagnetic cobalt ferrite(CoFe2O4), doped with Pd nanoparticles (NPs). The compositeparticles were found to be chemically stable in realizing magneticchemical propulsion. Further, these particles strongly catalyzed thedecomposition of hydrogen peroxide (H2O2) solution even when thestrength of the solution was as low as 0.3%, making it efficient ininducing autonomous chemical locomotion. In addition, the velocity ofthe particles could be controlled by using dimethyl sulfoxide (DMSO),which is known to quench hydroxyl radicals in the solution. We alsoshowed that these ferromagnetic structures possessed appreciablemagnetization and were capable of being guided magnetically, whileroving autonomously inside the liquid.

12709 dx.doi.org/10.1021/jp111890d |J. Phys. Chem. C 2011, 115, 12708–12715

The Journal of Physical Chemistry C ARTICLE

to impart the magnetic guidance over the movement is primarilythe bulk or nanoparticle (NP) form of Ni.13,14,16 Apart frombeing ferromagnetic, Ni is also known to catalyze the decom-position of H2O2 effectively that apparently projects it to be apromising choice for fabricating chemically driven magneticlocomotives. However, even in the least concentrated H2O2,bulk Ni or Ni NPs are prone to get degraded rapidly.17 Thisresults in the formation of the bulk or NP form of NiO, which ineither form is antiferromagnetic in nature18�20 and responds toan external magnetic field rather feebly. As an alternative, devel-opment of a composite structure consisting of individual mag-netic and catalytic components would be ideal, where eachcomponent could be used for the designated purpose—one forcatalytic decomposition of H2O2 to provide propulsion to thedevice and the other for externally guiding its motion.

In this article, we present results on the development ofa chemically stable magnetic locomotive using ferromagneticcobalt ferrite (CoFe2O4) doped with Pd NPs. The averagedimension of the CoFe2O4 particles was found to be around40 nm, and when agglomerated, the bigger particles formed wereof a mean dimension of 150 μm. The larger particles helped inobserving the motion using an optical microscope. The particlesshowed maximum chemical stability, so far as its oxidation inaqueous H2O2 was concerned, in comparison to materials usedearlier in this regard.16,21 The extent of oxidation of theseparticles in different strengths of H2O2 was negligible comparedto that observed in the case of commonly used metals like Au, Fe,or Ni.16,21,22 It is worthwhile to mention here that CoFe2O4 itselfis known to catalyze the decomposition of H2O2,

23 the rate ofwhich was found to be much less than that doped with Pd NPs(refer to Supporting Information, (SI)). Further, these particlesstrongly catalyzed the decomposition of hydrogen peroxidesolution even when the strength of the solution was 0.3%. Wealso showed that these ferromagnetic structures possessed appre-ciable magnetization and hence were capable of being guidedmagnetically while they moved spontaneously inside the liquid.Interestingly, the presence of dimethyl sulfoxide (DMSO), aknown scavenger for the •OH radical, reduced the velocity ofthe particle indicating quencher controllable motion. We de-monstrate that a simple model based on catalytic decompositionof H2O2 in the presence of a scavenger (DMSO herein)could account for the variable velocity of the microstructure inthe liquid.

2. EXPERIMENTAL SECTION

2.1. Preparation of Pd NPs. Pd NPs with narrow size dis-tribution were prepared following the method reported byWanget al.24 A mixture of 0.05 g (2.2 � 10�4 M) palladium acetate(obtained from Sigma Aldrich Chemical Co.) and 4.0 g (2.0 �10�3 M) poly(ethyleneglycol) (Merck, India), PEG (MW2000), was stirred using a magnetic stirrer at 80 �C for 1 h.The resulting light yellow homogeneous solution was furtherstirred for 2 h at the same temperature during which the color ofthe solution slowly turned dark gray from light yellow, indicatingthe formation of Pd NPs. The mixture of PEG and Pd NPs wasthen solidified by cooling at room temperature. Pd NPs obtainedin this process were dispersed in ethanol (obtained from Merck,India) by ultrasonication. The dispersion obtained was thencentrifuged, and the precipitate collected was washed withethanol. The particles thus synthesized were air-dried and werepreserved for further use.

2.2. Preparation of Pd NP Incorporated CoFe2O4 Particles.CoFe2O4 particles were synthesized following a method re-ported by Ceylan et al.25 Briefly, 50 mL of 9.3 � 10�3 Maqueous solution of FeCl3 (Merck) was mixed with 50 mL of4.2� 10�3 M aqueous solution of CoCl2 3 6H2O (Merck, India).To the salt solution, 25 mL of 3.0 M aqueous KOH (Merck,India) was added dropwise. Before mixing, all three solutionswere ultrasonicated for a period of 30 min to remove the dis-solved O2. The pH of the solution was regularly monitored as theKOH solution was added, and the reaction mixture was con-stantly stirred using a magnetic stirrer, until a pH level of 11�12was reached. A specified amount of oleic acid (Merck, India) wasthen added to the solution as a surfactant and coating material.To the reaction mixture, the dispersion of Pd NPs (in ethanol)was then added dropwise, and the reaction temperature wasslowly brought to a value of 80 �C. The mixture was stirredcontinuously for 1 h in a magnetic stirrer and then slowly cooledto room temperature. The black precipitate was then separatedby centrifugation at 15 000 rpm for 15 min, washed several timeswith both distilled water and ethanol (Merck), and kept over-night in an incubator at 60 �C. The precipitate was then furtherdried in an oven at 100 �C for nearly 1 h and subsequently keptin an evacuated environment (10�2 bar) for another 1 h. Theresidual water contained in the sample was then removed byannealing at 600 �C for 6 h.2.3. Characterization of Pd NP Incorporated CoFe2O4

Particles. Pd NP doped CoFe2O4 was first characterized usinga LEO 1430 VP scanning electron microscope (SEM), operatingat a maximum voltage of 15 kV. For transmission electron micro-scopy (TEM), a dilute dispersion of the particles was prepared inwater and was ultrasonicated for 30min. The dispersion was thencentrifuged at 8000 rpm for 5 min. The supernatant solution wascollected, and the whole process was repeated three times. Thefinal dispersion was drop-cast on a carbon-coated Cu grid andwas then left overnight for air-drying. The grid was then analyzedusing a JEOL 2100 TEM, operated at a maximum voltage of200 kV. Magnetization of Pd doped CoFe2O4 was investigated ina vibrating sample magnetometer (VSM, model 7410, LakeShore Cryotronics Inc., USA). X-ray diffraction (XRD) studiesof the sample before and after the treatment with H2O2 wererecorded using a Bruker AXS, Advance D8 diffractometer, with aCu KR source of X-ray wavelength 1.54 Å.2.4. Kinetic Measurements. Two sets of experiments were

performed for the determination of rate constants for decom-position of H2O2 in the presence of CoFe2O4 and Pd-CoFe2O4.For each set of experiments, there was an ensemble of eightsamples containing 5 mL of 0.4% (w/v) H2O2 solution. To eachof the samples in the first group, 0.1 g of CoFe2O4 was added, andthe same amount of Pd-CoFe2O4 was used for the second group.An aliquot of 1mL volume from each of the collections was takenafter a definite interval of time and titrated against standardizedKMnO4. The strength of the KMnO4 solution used for titrationwas measured using 0.2 N oxalic acid as the primary standard.To prevent further decomposition of H2O2 (refer to SI), 0.2 gof boric acid (Merck, India) was added to each 1.0 mL aliquot,immediately after withdrawal, from the reaction mixture.26

The strength of the H2O2 solution thus determined was plottedin logarithmic scale against the time of titration. The graphswere found to be straight lines which indicated that the decom-position of H2O2 solution followed first-order kinetics. Thevalues of the rate constants were obtained as the slopes of thesestraight lines.



3. RESULTS AND DISCUSSION

The presence of Pd NPs in the CoFe2O4 particles wasconfirmed by selected area electron diffraction (SAED)measure-ment in transmission electron microscopy (TEM). The d-spa-cings corresponding to the diffraction pattern obtained matchedclosely with that of metallic Pd (refer to SI). Pd NP incorporatedCoFe2O4 particles, when analyzed in SEM, revealed the overallparticle sizes to be within the range 40�50 nm, which was furtherconfirmed in TEM characterization (see later). The dimension ofthe bigger agglomerated particle was found to be nearly 150 μm.Figure 1(A) shows a typical SEM image recorded for such asample. Figure 1(B) shows the magnetization curve of PdNP doped CoFe2O4 particles, recorded within the field range�2 to 2 T using 0.02 g of the sample. The sample was found tobe strongly ferromagnetic with a maximum value of satura-tion magnetization of 85.6 emu g�1. After the withdrawal ofthe external magnetic field, the sample was found to retain amagnetization of 46.4 emu g�1. This indicated that theseparticles could potentially be guided magnetically inside a liquideven with a low external magnetic field.

The stability of Pd-CoFe2O4 particles in aqueous H2O2

solution was tested using a solution of 1.0% H2O2. An amountof 0.2 g of PdNP doped CoFe2O4 was first characterized by XRDand thenwas kept immersed in 1.0% aqueousH2O2 solution for aperiod of 6 h. The sample was then washed and air-dried. Whenthese particles were again characterized in XRD, the positions ofall the peaks corresponding to their crystal structure were foundto be present. Further, when characterized by TEM, the particlesof Pd-CoFe2O4 were found to remain discernibly unchanged inshape and composition even after treating with 1.0% H2O2 for 6h with an average size that remained the same as before. Thedetails of XRD data and TEM images are shown in Figure 2.

When Pd NP incorporated CoFe2O4 particles were placed inan aqueous solution of H2O2, rapid formation of O2 bubbles,around their surfaces, was observed. The bubbles grew continu-ously in size and upon their detachment provided a recoil thrustto the particle. The resultant thrust of all the detaching bubbles ata particular instant impartedmotion to the particle. The videos ofsuch movements, in a drop of solution placed on a glass slide,were recorded with a Nikon E200 optical microscope equippedwith a Nikon E8400 digital camera. A sample video is shown in

the SI (Video1). The direction of motion at each instant wasdecided by the direction of the resultant recoil momentumimparted at that instant. However, each particle, in a definitetime interval, was observed to suffer a net displacement, whichwas found to depend on the number of bubble generating sites,except for the cases when the bubbles were forming symmetri-cally around the particle and the particle hardly was displacedfrom its position. Further, smaller particles were seen to movemuch faster than the bigger ones possibly owing to the higherinertia of the latter. A typical trajectory of movement of relativelylarge particles is shown in Figure S5 (SI). It is important here tomention that the present system is characterized by a lowReynold’s number (∼10�3) (for calculation refer to SI), andthus it may be reasonable to conclude that hydrodynamic forcesmay be responsible for the motion of the particles. However, theorigin of the driving force behind the particle propulsion herein isthe force arising out of the detachment of the bubble from itssurface, whose magnitude was found to be much higher than theviscous force associated with the motion of the particle (calcula-tion to follow). The hydrodynamic forces which would arise outof detachment of the bubble from the particle would be smallerthan the original force due to detachment, as additional energywould be required for the forward motion of the particle. Acalculation of the velocity of the particle considering the mo-mentum conservation would provide a simple way even thoughthe exact velocity would be overestimated. This would at leastenable us to calculate the trend in particle velocity as a function ofexperimental parameters without resorting to complications thatwould involve the exact estimation of the hydrodynamic forces.Further, occasionally, the particles were found to get stuck to thebottom of the glass slides; however, O2 bubbles continued togrow over them until they detached from the surface imparting alarge recoil momentum—making the particles free from thesurface of the slides. Interestingly, Pd incorporated CoFe2O4

particles were found to move quite rapidly even when thestrength of H2O2 solution used was 0.3% (although the resultsreported in this article were obtained using 0.4% H2O2). Thisis, to the best of our knowledge, the lowest concentration ofperoxide fuel used so far in prototypes for autonomously movingmagnetic objects. When the strength was kept at 0.4%, Pdincorporated CoFe2O4 particles were seen to move with a

Figure 1. (A) Typical SEM image recorded for Pd-CoFe2O4 microparticles. (B) Magnetization curve of Pd doped CoFe2O4 microparticles, recordedwithin the field range of�2 to 2 T using 0.02 g of the sample. Here B is the external magnetic field in Tesla andM is the corresponding magnetization inemu g�1.



maximum velocity of 0.04( 0.02mm s�1, the average dimensionof the particles being 150.0 μm. It was observed that differentparticles of approximately the same dimension of 150 μm, in asolution of 0.4% H2O2, moved with different velocities depend-ing on the number of catalytic sites and other factors like defects,interfacial tension, and bubble detachment rate. Velocities of anumber (typically 8�10) of such particles (of same size) werestudied, and the highest value recorded was called the maximumvelocity, denoting it withVmax.With an increase in the strength ofH2O2 solution, Vmax was found to increase linearly. The increasein maximum velocity of a particle of average dimension 150.0 μmwith the increase in strength of H2O2 solution is shown inFigure 3.

We were also interested in finding ways to chemically controlthe motion of these particles. This could possibly be achieved bycontrolling the formation of O2 bubbles during the decomposi-tion of H2O2. To accomplish this, we used dimethyl sulfoxide(DMSO), which is known to chemically quench radical species,especially •OH radicals generated upon decomposition ofH2O2.

27 Now, decomposition of H2O2 on metal surfaces, asproposed by Weiss, occurs via a mechanism as follows28

SþH2O2 fk1S

Sþ þOH� þ •OH ð1Þ

H2O2 þ •OHfk2S

H2Oþ •HO2 ð2Þ

Sþ þO2� f

k3SSþO2 ð3Þ

Sþ •HO2 fk4S

Sþ þHO2� ð4Þ

Sþ þHO2� f

k5SSþ •HO2 ð5Þ

Here, S and Sþ stand for the uncharged and charged part of themetal surface, respectively. The hydroxyl radical •OH formed instep 1 is key to the formation of O2 bubbles in the decompositionH2O2 f H2O þ O2. Thus, control over the formation orremoval of this radical would possibly lead to the controlledproduction of O2 bubbles during H2O2 decomposition. When

Figure 2. XRD patterns recorded for Pd-CoFe2O4 particles (A) before and (B) after the treatment with 1.0% H2O2 solution. The crystal planescorresponding to CoFe2O4 have been identified. TEM images obtained for Pd-CoFe2O4 particles (C) before and (D) after the treatment with 1.0%H2O2 solution. The corresponding SAED patterns are shown in the insets.

Figure 3. Variation of maximum velocity Vmax of Pd incorporatedCoFe2O4 microparticles with the increase in concentration of H2O2

solution.



DMSO was added to the medium, the formation of O2 bubbleswas found to decrease systematically with increasing concentra-tion of DMSO, and the maximum velocity of the catalyticparticles was also found to decrease consequently. Figure 4 showsthe change in the maximum speed of a Pd NP doped CoFe2O4

particle (of dimension 150.0 μm) with the increase in DMSOconcentration in 0.4% aqueous H2O2 solution. It is evident fromthe figure that the velocity decreased linearly with increasingconcentration of DMSO. Further, it is interesting to note that theconcentration of DMSO required to stop the motion of theparticle completely was sufficiently low (0.056M) in comparisonto the concentration of H2O2 present in the medium (0.12 M).

To account for the trend in velocity of the particle in thepresence of DMSO, the involvement of •OH radicals in thereactions may be considered. It may be recalled from eqs 1 and 2that there are two steps in the overall reactions of catalyticdecomposition of H2O2 which involve production or consump-tion of •OH. On the other hand, in the presence of DMSO theconsumption of •OH by DMSO may be considered to followoverall second-order kinetics as written below

•OHþDMSOfkDOH� CH3S ¼ Oþ •CH3 ð6Þ

Here, kD is the rate constant for the reaction between •OHradical and DMSO. Further, if one considers the steady stateconcentration of •OH, then from eqs 1, 2, and 6, we get

½•OH� ¼ k1S½S�½H2O2�k2S½H2O2� þ kD½DMSO� ð7Þ

Now, the rate of production of O2 in the absence of DMSO,following Weiss’ theory, is given by

d½O2�dt

¼ k1S½S�½H2O2� ð8Þ

In the presence of DMSO, eq 8 may be rewritten as

d½O2�dt

¼ k1S½S�½H2O2� � kD½•OH�½DMSO� ð9Þ

Using eq 7, we get

Or,d½O2�dt

¼ k1S½S�½H2O2� 1� kD½DMSO�k2S½H2O2� þ kD½DMSO�

� �ð10Þ

Let rΔt1 be the instantaneous radius of an O2 bubble when it getsdetached from the particle, after a timeΔt1, from the initiation ofthe reaction. We consider VO2

and FO2to be the volume and

density of the oxygen bubble getting detached from the particlesurface and ubi(= 0) and ubf to be the initial and final velocitiesof the bubble attained immediately following its detachment.If Δt2 is the duration for which the recoil thrust arising out ofbubble detachment acts on the particle, the force which sets theparticle in motion is then given by

F ¼ VO2FO2ubf

Δt2ð11Þ

If m is the mass of the particle and v is its velocity after a smalltime intervalΔt3 from the start of motion, then the force balanceequation (assuming the particle initially to be at rest) becomes

F� f ¼ mv� 0Δt3

� �ð12Þ

where f denotes the effective drag force acting on the particle. Tohave a quantitative estimate, we assume the catalytic particle to bea sphere accelerating uniformly through the liquid. The dragforce acting on such a sphere is given by29

f ¼ 6πRηvþ 12

43πR3

� �Faþ 6R2ðπηFÞ1=2

Z t

0

aðt0Þðt � t0Þ1=2

dt0

¼ F1 þ F2 þ F3

Here, F1 is equal to the steady-state viscous drag on the sphereacting in the direction opposite to its motion; F2 has the samemagnitude as the resistance of an accelerating sphere in irrota-tional motion; and F3 is the term signifying the history of theacceleration. Further, in the expression of f, R signifies the effec-tive radius of the particle; η and F are the coefficient of viscosityof 0.4% H2O2 solution and density of CoFe2O4 particles respec-tively; v and a are the velocity and acceleration (assumed uniform)of the particle; and t denotes the time interval for which themotionis being observed. Force on a particle arising out of bubble detach-ment is given by F = [(VO2

FO2ubf)/(Δt2)] = [((4/3)πrΔt1

3 FO2ubf)/

(Δt2)], where Δt2 is the duration for which the recoil force actson the particle. Now, the final velocity of a detached bubble wascalculated from measuring its displacement in a particular timeinterval, by analyzing a recorded video of locomotion frame byframe. The capture rate of such a video was 30 fps, and ourmeasurement of velocity will thus be limited by the least count ofthe measuring device, i.e., 0.033 s. This means that the estimationof the time interval for which the displacement is measuredcannot be smaller than this. Taking into consideration micro-scope magnification, we observed the displacement of an O2

bubble (assuming its motion to be uniform) to be 1.6� 10�4 min a time interval 0.2 ((0.033) s (for details refer to SI). The finalbubble velocity thus was found to be ubf = 8.0 � 10�4 m s�1.We also measured the approximate radius of the bubble to berΔt1 = 0.001 m from the video. Since it is not possible to measurethe duration Δt2 for which the recoil force acts on the particle at

Figure 4. Profile of maximum velocity attained by Pd NP dopedCoFe2O4 particles, in 0.4% aqueous H2O2 solution, with the concentra-tion of DMSO.



the moment of bubble detachment, we arbitrarily assumed it tobe equal to 0.001 s, which is closer to the observable time intervalof 0.033 s. Assuming the O2 formed to be an ideal gas at STP, thedensity is given by FO2

= 1.43 kg m�3. Using all these values, themagnitude of the driving force acting on the particle as a con-sequence of detachment of an O2 bubble is F = 4.8 � 10�9 N.Assumption of smaller values of bubble detachment time Δt2would make the magnitude of the force even larger. Further,considering the following values for the various parametersη = 0.001 N s m�2, R = 75 � 10�6 m, v = 4 � 10�5 m s�1,F = 2.8� 103 kg m�3, a = 2.0� 10�4 m s�2, and t = 0.2 s, we getf = 7.5� 10�11 N (refer to SI for calculation). Thus, the viscousdrag is small compared to the total driving force due to bubbledetachment. However, with increased acceleration and viscosityof the medium, the opposing force may be significant. On theother hand, one could still argue from the standpoint of animaginary situation where the bubbles are generated from thesame point of the sphere—one after another—thus constantlyincreasing the velocity to infinity. Hence, there is a need for aforce opposing the acceleration to explain the observations. Wewould like to emphasize here that in addition to the drag forcementioned above there are two other forces which would opposethe motion in question. First of all, owing to depositions of (Pd)NPs on all sides, the bubbles would continuously be generated(and thus detached) from all directions. This would prevent anyuniform increase in velocity in one direction. In addition, even inthe case of coalescence of smaller bubbles it is difficult to imaginethat the larger bubble would grow keeping the center of thehemisphere the same all the time, as smaller bubbles wouldcoalesce from all sides randomly. In other words, there may be aconstant shift of the center of the bubble hemisphere as new largebubbles would form out of coalescence of smaller ones. In thatcase also, the direction of motion would shift regularly. Further,our calculations indicated that the weight of the magnetic particlewas about 3 times the buoyancy force in water (refer to SI).Hence, gravitational force would also contribute to the opposi-tion of the force due to bubble detachment, although at an angledifferent from 180�. For simplicity, we neglect the gravitationalpull and consider the case of single bubble formation and detach-ment and thus include the influence of viscous drag only.

The motion of a catalytic particle, taking into considerationthe viscous drag and quenching of the •OH radical by DMSO,can then be described following eq 12 as

VO2FO2ubf

Δt2� 6πRηvþ 1

243πR3

� �Fa

�

þ 6R2ðπηFÞ1=2Z Δt3

0

aðt0Þðt � t0Þ1=2

dt0!

¼ mvΔt3

ð13Þ

Or

v ¼½ð43πr3Δt1FO2

ubf Þ=Δt2� � 23πR3Fa� 12R2a

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiπηFΔt3

pmΔt3

þ 6πηR

� �ð14Þ

If [nO2]Δt1 is the number of moles of O2 produced in an interval

Δt1, then we have

r3Δt1 ¼ 34π

22:4� 10�3½nO2 �Δt1 ð15Þ

Therefore, the velocity after time Δt3 is given by

v ¼½ð22:4� 10�3½nO2 �Δt1FO2

ubf Þ=Δt2�� 23πR3Fa� 12R2a


pmΔt3

þ 6πηR

� �ð16Þ

Now, using eq 10 and assuming the rate of O2 production to beuniform for a fixed concentration of DMSO, the total number ofmoles of O2 produced in time Δt1 is given by

½nO2 �Δt1 ¼ k1S½S�½H2O2� 1� kD½DMSO�k2S½H2O2� þ kD½DMSO�

� ��Δt1 ð17Þ

Therefore,

v¼ 22:4� 10�3k1S½S�½H2O2�Δt1Δt2

1� kD½DMSO�k2S½H2O2� þ kD½DMSO�

� �(

�FO2ubf � 2

3πR3Fa� 12R2a


p o� mΔt3

þ 6πηR

� �� ð18Þ

Or eq 18 can be rewritten as

v ¼ A� B0kD½DMSO�

k2S½H2O2� þ kD½DMSO� ð19Þ

where

A ¼22:4� 10�3k1S½S�½H2O2�Δt1

Δt2FO2

ubf �23πR3Fa�12R2a


pmΔt3

þ 6πηR

� �

ð20Þand

B0 ¼ 22:4� 10�3k1S½S�½H2O2�Δt1FO2ubf

Δt2mΔt3

þ 6πηR

� � ð21Þ

In the present case, [H2O2] = 0.4% (w/v) = 0.12 M and[DMSO]max = 0.056 M. Hence, putting [H2O2] . [DMSO]in eq 19 and assuming [H2O2] = constant, we get

v ¼ A� B½DMSO� ð22Þ

B ¼ B0kDk2S½H2O2�

Thus, the variation of v as a function of [DMSO] is a straight linewith a slope �(B0kD)/(k2S[H2O2]), when the concentration ofDMSO is sufficiently low (which is the case herein). This is whathas been observed in experiments and as shown in Figure 4. Onemay also consider the effect of viscosity, due to the presence ofDMSO, on the velocity of the particle. For example, a binarymixture of DMSO and 0.4% H2O2 solution possesses a slightlyhigher value of viscosity than that of 0.4% H2O2 solution only atroom temperature.30 However, the effect, if at all, would be smallin comparison to that of quenching of the •OH radical by DMSO(refer SI). In addition, as opposed to coalescence leading to



formation of a larger bubble in the absence of DMSO, itspresence led to the generation of smaller bubbles all aroundthe particle. The detachment of smaller bubbles from all sides didnot generate enough net force for the motion of the compositeparticle. The presence of DMSO possibly lowered the interfacialtensions between the bubble, liquid, and the particle and thusprevented formation of larger bubbles through coalescence.31

However, accounting it quantitatively is rather difficult and out ofbound of the present work. Considering the low concentration ofDMSO, it may be the most appropriate if we attribute thedecrease in maximum speed of the particles to be due to effectivequenching of •OH radicals by DMSO.Wewere also interested tostudy the effect of DMSO on the catalytic decomposition ofH2O2. Unfortunately, the efforts to titrate H2O2 in the presenceof the catalyst and DMSO failed as there was precipitationobserved upon addition of KMnO4 to the mixture.

We were finally interested in achieving a directional controlover the movement of these microparticles in H2O2 solution.Room-temperature VSM-basedmeasurement revealed that the par-ticles possessed a net saturation magnetization of 85.6 emu g�1.We observed that a weak external magnetic field of even3.0 � 10�3 T, with a positive field gradient, was sufficient todirect the course of these particles inside the liquid. Letm be themagnetic moment of a Pd-CoFe2O4 microparticle, induced inthe presence of an external magnetic field B. If the direction ofinduced magnetization is parallel to the direction of motion ofthe particle, it will experience a translatory force of magnitudeFm = (mr)B,11 which will make the particle accelerate towardthe nearest pole of the external magnet. We consider i to be theunit vector along the þX-axis in which both the externalmagnetic field and the magnetization of the particle are assumedto be directed. Taking themagnetic momentm to be irrotational,B to be solenoidal, and the external field to be large enough tocompletely magnetize the particle, the acceleration experiencedby the particle is given by a = Mγi, where M is equal tomagnetization of the particle per unit mass and γ denotesthe linear gradient in the external field. For a particle of dimen-sion 150 μm, when kept in an external field with a gradient0.366 T m�1, the theoretical estimate of the acceleration ayielded 3.49 m s�2. When experimentally measured, the valueof acceleration was found to be 6.67 m s�2 (refer to SI), whichfairly agreed with the theoretical estimation. It may bementionedhere that in the absence of H2O2, i.e., O2 not being producedfrom its catalytic decomposition, there was no motion of theparticle as the force generated by the weak magnetic field was notsufficient to overcome the forces due to gravity and friction of thecontainer surface. On the other hand, in the presence of anO2 gasbubble the particle was lifted from the bottom of the containerdue to buoyancy and could be moved with much ease when theexternal magnetic field was applied. If we now consider the direc-tion of the external magnetic field to be perpendicular to themotion of the particle, the net force experienced by the particlewill be zero. The particle will, however, experience a torque givenby τ = m � B, which will be directed perpendicular to the planecontaining bothm andB. If we consider the magnetic moment ofthe particle to be directed along the þX-axis as earlier and thedirection of the external magnetic field to be along the þY-axis,the magnitude of the torque is given by τ = m � B (directedalong theþZ-axis). The torque will try to align the particle alongthe direction of the applied field. The magnitude of the torqueincreases with the increase in the value of the external magneticfield, and the particle will try to align itself more quickly along

the external field lines. This is exactly what we had observed inthe experiments, where by applying a small magnetic field of3.0 � 10�3 T it was possible to manipulate the course of themagnetic particles inside the liquid (see SI, Sample Video 2).While most of the smaller particles responded to the externalmagnetic field as expected and aligned themselves along the fieldlines, often for the bigger particles (of dimension more than150 μm), it was observed that the rotation was not complete andfinally that the particle was found not lying along the direction ofB. The particle often remained aligned at an angle θc (made withits initial direction of motion) even at moderately high values ofthe external field. Figure 5 shows how θc varied with the appliedmagnetic field, for a typical particle of dimension nearly 200 μm.This could be due to the fact that, in the presence of the externalmagnetic field, the particle was magnetized in a direction that isnot parallel to its principal axis. Thus, even when its magnetiza-tion was actually parallel to the direction of the external magneticfield, geometrically the particle made an angle θc with itsinitial direction of motion, which in an ideal case should havebeen 90�. Other than these occasions, directional control for thePd-CoFe2O4 particles could be achieved by carefully applying a lowexternal magnetic field. At large field values, the rotation vanished,and all the particles congregated at the nearest pole of the magnet.

To summarize, we have demonstrated that Pd NP dopedCoFe2O4 microparticles can be used to develop a modelmagnetic chemical locomotive. The structures were found tobe chemically stable and capable of moving effectively even in0.3% aqueous H2O2 solution. This is, to our knowledge, thelowest reported concentration of peroxide fuel used so far, inrealizing self- propelled magnetic chemical locomotives. Themaximum velocity attained by such particles as a function ofH2O2 concentration was studied, and the increase in velocitywith concentration was attributed to the increased O2 bubbleformation on the surface of the particles, which eventuallyincreased the resultant recoil momentum. To achieve a controlon the velocity of the particles by quenching the formation of theradicals in the solution, DMSOwas mixed with the peroxide fuel,the former being well-known as a quencher of the •OH radical.The speed of the ferrite particles was found to be effectivelylowered with the increase in the concentration of DMSO in theperoxide fuel. Finally, with an aim to achieve a directional controlover the movement, we used a weak external magnetic field tonavigate the particle through the liquid.

Figure 5. Increase in the angle of rotation of a Pd-CoFe2O4 micro-particle of dimension about 200 μm with the increase in the magnitudeof the external field. Increased rotation of a typical particle at increasingvalues of the field is shown sucessively from a to f in the inset.



’ASSOCIATED CONTENT

bS Supporting Information. Figures S1�S5; calculation ofReynolds number; determination of final velocity of an oxygenbubble after it gets detached from the particle surface; change invelocity of a CoFe2O4 particle with the concentration of glyceroland DMSO; calculation of magnetization (m) of a single Pd-CoFe2O4 particle; calculation of forces acting on a ferrite particleduring its movement; acceleration in the presence of an externalmagnetic field, applied parallel to the direction of motion;theoretically estimated acceleration; experimentally observedacceleration; torque in the presence of an external magneticfield, applied perpendicular to the direction of its motion; andvideos. This material is available free of charge via the Internet athttp://pubs.acs.org.

’AUTHOR INFORMATION

Corresponding Author*E-mail: [email protected].

’ACKNOWLEDGMENT

We thank CIF, IITG, for SEM analysis and CSIR, India(01(2172)/07/EMR-II), 03(1097)/07/EMR-II), for funds.The authors also thank DST, India, for financial support (SR/S5/NM-108/2006, 2/2/2005-S.F.; SR/S2/CMP-23/2009; andSR/S1/RFPC-07/2006). K.K.D. acknowledges a fellowshipfrom CSIR, India (09/731(0051)/2007-EMR-I).

’REFERENCES

(1) Ahn, C. H.; Allen, M. G.; Trimmer, W; Jun, Y. N; Erramilli, S.J. Microelectromech. Syst. 1996, 5, 151.(2) Gunnarsson, K; Roy, P. E; Felton, S; Pihl, J; Svedlindh, P; Berner,

S; Lidbaum, H; Oscarrson, S. Adv. Mater. 2005, 17, 1730.(3) Bhatt, S. K. Locomotion of Magnetic Objects in Fluids,

Ph.D. Dissertation, Drexel University, USA, 2008. http://idea.library.drexel.edu/bitstream/1860/2864/1/Bhat_Shubham.pdf (accessed onDecember 07, 2010).(4) Yellen, B. B.; Friedman, G. Adv. Mater. 2004, 16, 111.(5) Hess, H. Soft Matter 2006, 2, 669.(6) Gazeau, F.; Baravian, C.; Bacri, J.-C.; Perzynski, R.; Shliomis,

M. I. Phys. Rev. E 1997, 56, 614.(7) Kline, T. R.; Paxton, W. F.;Wang, Y.; Velegol, D.; Mallouk, T. E.;

Sen, A. J. Am. Chem. Soc. 2005, 127, 17150.(8) Soong, R. K.; Bachand, G. D.; Neves, H. P.; Olkhovets, A. G.;

Craighead, H. G.; Montemagno, C. D. Science 2000, 290, 1555.(9) Hess, H.; Vogel, V. Rev. Mol. Biotechnol 2001, 82, 67.(10) Howse, J. R.; Jones, R. A. L.; Ryan, A. J.; Gough, T.; Vafabakhsh,

R.; Golestanian, R. Phys. Rev. Lett. 2007, 99, 048102.(11) Paxton,W. F.; Kistler, K. C.; Olmeda, C. C.; Sen, A., St.; Angelo,

S. K.; Cao, Y.; Mallouk, T. E.; Lammert, P. E.; Crespi, V. H. J. Am. Chem.Soc. 2004, 126, 13424.(12) Fournier-Bidoz, S.; Arsenault, A. C.; Manners, I.; Ozin, G. A.

Chem. Commun. 2005, 441.(13) Kline, T. R.; Paxton, W. F.; Mallouk, T. E.; Sen, A. Angew.

Chem., Int. Ed. 2005, 44, 744.(14) Burdick, J.; Laocharoensuk, R.; Wheat, P. M.; Posner, J. D.;

Wang, J. J. Am. Chem. Soc. 2008, 130, 8164.(15) Dreyfus, R.; Baudry, J.; Roper, M. L.; Fermigier, M.; Stone,

H. A.; Bibette, J. Nature 2005, 437, 862.(16) Dey, K. K.; Sharma, D.; Basu, S.; Chattopadhyay, A. J. Chem.

Phys. 2008, 129, 121101.(17) Chen, W. F.; Wu, S. Y.; Ferng, Y. F.Mater. Lett. 2006, 60, 790.(18) Koillert, B.; Falicov, L. M. J. Phys. C: Solid State Phys. 1975,

8, 695.

(19) Tiwari, S. D.; Rajeev, K. P. Thin Solid Films 2006, 505, 113.(20) Smart, J. S.; Greenwald, S. Phys. Rev. 1951, 82, 113.(21) Dey, K. K.; Panda, B. R.; Paul, A.; Basu, S.; Chattopadhyay, A.

J. Colloid Interface Sci. 2010, 348, 335.(22) Zepp, R. G.; Faust, B. C.; Holgn�e, J. Environ. Sci. Technol. 1992,

26, 313.(23) Lahiri, P.; Sengupta, S. K. Can. J. Chem. 1991, 69, 33.(24) Luo, C.; Zhang, Y.; Wang, Y. J. Mol. Catal. A: Chem. 2005,

229, 7.(25) Maaz, K.; Mumtaz., A.; Hasanaina, S. K.; Ceylan, A. J. Magn.

Magn. Mater. 2007, 308, 289.(26) Mclane, C. K. J. Chem. Phys. 1949, 17, 379.(27) Yoshimura, Y.; Inomata, T.; Nakazawa, H.; Kubo, H.; Yamaguchi,

F.; Ariga, T. J. Agric. Food Chem. 1999, 47, 4653.(28) Weiss, J. Adv. Catal. 1952, 4, 343.(29) Odar, F.; Hamilton, W. S. J. Fluid Mech. 1964, 18, 302 and the

references therein.(30) Lebel, R. G.; Goring, D. A. I. J. Chem. Eng. Data 1962, 7, 100.(31) Agrawal, A.; Dey, K. K.; Paul, A.; Basu, S.; Chattopadhyay, A.

J. Phys. Chem. C 2008, 112, 2797.

stable magnetic chemical locomotive with pd nanoparticle incorporated ferromagnetic oxide

Documents