Transport of cargo by catalytic Janus micro-motors

Abstract

Catalytically active Janus micro-spheres are capable of autonomous motion and can potentially act as carriers for transportation of cargo at the micron-scale. Focusing on the cases in which a single or a pair of Janus micro-motors is used as carrier, we investigate the complex dynamics exhibited by various active carrier–cargo composites.

---

Autonomous motion at the micron- and nano-scale is indispensable for the development of new generations of devices for modern bio-analytical or chemical assays, performed either on a chip,¹ or in a biological environment. Several components of the cellular metabolism, e.g., mitosis² and the internal protein transport,³ are supported by various types of molecular motors.^1,2,4–6 In many cases the motion of these biological machines emerges from a complex interplay of random fluctuations with deterministic driving forces.⁴

Man-made synthetic micro- and nano-motors capable of moving cargo loads within a liquid medium are of significant interest for applications such as targeted drug delivery,⁷ biosensing, or shuttle-transport of the emulsion droplets⁸ and living cells.⁹ The fast progress in the fields of micro-fabrication,¹⁰ colloidal chemistry,¹¹ and catalysis¹² has led to a variety of microscopic man-made objects capable of self-motility, which is an essential step in the quest for building synthetic nano- and micro-machines.^13–18 One promising approach is the use of catalytically active objects such as bi-metallic nanorods,^13,19,20 metallic and dielectric particles,^21,22 and tubular catalytic jet-engines.^23–26 As a general principle, these artificial motors use an asymmetric decoration of their surface with a catalyst (usually platinum,^13,14,21nickel,²⁷ or enzymes²⁴) which promotes a specific chemical reaction in the surrounding liquid²⁸ and leads to concentration gradients along the surface of the particle. Depending on the systems, various propulsion mechanisms emerge, such as interfacial tension gradients,²⁹ bubble propulsion,²⁴ self-electrophoresis,²⁹ or self-diffusiophoresis.³⁰ For such chemically active particles directed motion at short timescales has been successfully achieved. But at medium and long timescales the directionality is lost because of the thermal rotational diffusion processes, intrinsic to the motion of such small-scale objects.^21,31,32 In order to achieve long time directionality, various strategies of employing “guiding” external fields have been proposed.¹⁷

While there is a relatively good understanding of the motion of single active particles, studies of cargo transport by such active particle carriers are scarce.^{17,26,33–35} Here we use catalytically active spherical Janus micro-motors in order to address the dynamics of the transportation of cargoes at the micron-scale. We investigate experimentally the motion (trajectories and velocities) of carrier(s)–cargo(s) systems in which either a single or a pair of active Janus particles^21,36 upload and transport catalytically inert cargoes. We further show that the rich dynamic behaviour observed in these systems can be qualitatively understood within the framework of self-diffusiophoresis.^30,37–39

The active Janus spheres which we employ are carboxylate modified polystyrene colloids (Duke Scientific, diameter d of approximately 5 μm and average polydispersity below 15%) with a platinum cap. These are fabricated as follows. A self-assembled monolayer of colloidal particles is formed by coating previously cleaned glass slides with a droplet of the colloidal suspension. The monolayer is dried via slow evaporation of the solvent at room temperature. Thereafter, a thin bilayer of Pt (5 nm)/Ti (2 nm) is deposited on top of the spherical colloids using electron beam evaporation.^40,41 The Ti layer is deposited in order to facilitate the good adhesion of the Pt film to the surface of the polystyrene particles. The metal films form the hemispherical caps on the surface of the particles. These caps are responsible for the optical contrast within a single sphere, where the dark region coincides with the catalytic area of particles. In the following, this contrast helps to distinguish the Janus spheres from the cargo particles.

After fabrication, the Janus particles are detached from the substrate viasonication and are mixed with an aqueous solution of hydrogen peroxide. The suspension is placed onto a glass substrate where the particles settle due to gravity. At last, the non-catalytic plain polystyrene particles of the same size, which represent the cargoes, are added to the suspension. The resulting density of the particles at the substrate is rather low (the average distance L between the particles is more than 10 particle diameters). The Janus particles, the inert polystyrene particles, and the glass substrate in contact with the aqueous solution are all negatively charged. Therefore the suspension is stable and the particles do not stick to the glass wall after sedimentation.

We observe the carriers and cargoes located close to the glass substrate (“wall”) and their motion is confined to an (x–y) plane parallel to the wall. The distance to the wall is roughly estimated to be 100 nm, set by the balance between the apparent weight of a colloid and the electrostatic repulsion (across an aqueous film) from the wall. When placed into the solution of hydrogen peroxide, the Janus particles move in response to the development of O₂ gradients along their surface because of the Pt-catalyzed decomposition of H₂O₂. The dynamics is determined by the interplay between thermodynamic forces, including particle interactions with the O₂ gradient, and the hydrodynamic flow response of the solution.^30,42 We observe experimentally that the motion of these particles is always directed away from the catalyst site, which is in agreement with previous reports.²¹ The motion of singlets, doublets, and other aggregated configurations of active particles as well as that of carrier–cargo composites is observed using a videomicroscopy setup⁴¹ and recorded by a high speed camera (Zeiss AxioCam HSm). The corresponding trajectories are tracked and analyzed using the ImageJ software.⁴³

The investigated systems are illustrated in Fig. 1. Although more complex configurations are possible (see also the examples of self-assembled Janus particles^34,44), here we focus on two cases paradigmatic for the phenomenology observed in active particle–cargo composites: (i) a “single carrier”, with a single active colloid acting as a carrier for cargo transportation (Fig. 1, top row), and (ii) a “double carrier” corresponding to the scenario in which two active colloids are involved in the transport (Fig. 1, bottom row). The single or double carriers can load one or more colloidal cargoes.‡

Fig. 1 Janus micro-motors with Pt caps (carriers) loading the inert polystyrene beads (cargo) for transportation. From left to right, top row: single free carrier, loaded with one and two cargo particles, respectively. Bottom row: double carrier in similar configurations. Dark regions on the spheres (marked by the yellow arrows) correspond to the Pt catalytic cap. The black arrows indicate the momentary direction of motion.

The motion of the composite depends significantly on the number of particles involved, their position, and the relative orientation of the catalytic caps³¹ (see Movie S1 in the ESI†). In Fig. 2 we show representative trajectories of free single and double carrier configurations, as well as those transporting a single cargo or a double cargo, moving in an aqueous solution of H₂O₂ of 10% v/v (volume/volume) concentration. It can be easily inferred that for all objects the directed motion is maintained only over time intervals of a few seconds, while during the observation time of 40 s the directionality is significantly affected. In particular, the single carrier (Fig. 2(a), left) prominently reveals this tendency. In the case of one or two loaded cargoes, a circular-like motion of the whole composites can appear (Fig. 2(a), right). For the double carrier (Fig. 2(b), left), circular trajectories appear even without any loaded cargo. Such features cannot be attributed just to a thermal rotational diffusion, but are most likely due to a misalignment of the catalytic cap orientation with respect to the symmetry axis of the colloidal composite. These will be discussed below in detail.

Fig. 2 Recorded trajectories of Janus carriers transporting cargoes according to the configurations shown in Fig. 1. Fuel: 10% v/v H₂O₂. The red areas in the schematic drawing of the particles depict the Pt covered caps of the carriers. A slight misalignment of the carriers is indicated in (b) (see Fig. 4 and discussion in the main text).

Fig. 3 presents the H₂O₂ concentration dependence of the absolute value of the velocity V corresponding to the various carrier–cargo composites presented in Fig. 2. In this graph V denotes a mean velocity averaged over several experimental realisations.§ The velocities of free and cargo-loaded active carriers show a quasi-linear dependence on the H₂O₂ concentration up to around 5% v/v; beyond this value they gradually saturate to a concentration-independent value. It is intuitively expected that the velocity of cargo-loaded carriers is lower than the one corresponding to free carriers because the same source of energy has to account for more particles moving.

Fig. 3 Velocities of the carrier–cargo composites versus the concentration n of H₂O₂. Panel (a) displays the dependences of the velocities of the free single carrier (black curve) and carriers loaded with one or two cargoes (red and blue curves, respectively). In turn, panel (b) describes the corresponding velocities of the free double carriers (black curve), with one cargo (red curve), and with two cargoes loaded (blue curve). The insets in the panels (a) and (b) display ideal, perfectly aligned configurations of the carrier–cargo composites (see Fig. 4 and discussion in the main text).

The experimental observations indeed confirm this expectation. For example, in solutions with 10% v/v H₂O₂ the velocity of a free active particle is V_f = 6.5 μm s⁻¹, while when transporting one and two cargoes it decreases to V₁ = 4.8 μm s⁻¹ and V₂ = 3.8 μm s⁻¹, respectively (Fig. 3).

It has already been mentioned that the motion of the Pt capped polystyrene Janus spheres in H₂O₂ solutions can be attributed to a self-diffusiophoretic mechanism.²⁸ Furthermore, it can be well described by a simplified model accounting for O₂ concentration gradients along the surface of the particle and neglecting the depletion of reactant (H₂O₂).^30,37,42 Therefore, our discussion below will focus on the experimental results at the larger concentrations of H₂O₂, for which the latter assumption is more likely to hold. By alluding to the simplified model of self-diffusiophoresis, we shall qualitatively explain the particular features of the dynamics of the catalytic carrier–cargo composites as observed experimentally.

The concentration gradients of O₂ (solute), occurring due to the non-uniform production over the surface of the particle, and the solute–particle effective interactions lead to a solvent flow within a very thin surface layer. This is accounted for by a so-called phoretic slip V_s(r) = −b∇_‖ρ(r) around the carrier where ρ(r) is the solute concentration along the particle surface, ∇_‖ denotes the projection of the gradient operator onto the plane locally tangent to the surface of the particle, and b is an effective mobility coefficient.^30,37–39 For repulsive (attractive) effective surface interactions, b is negative (positive).³⁸ The motion of the particles will be away from the high concentration gradients for repulsive solute–particle interactions, and towards it for attractive interactions.^37–39 According to the experimental observation that the Janus particles always move away from the catalyst side, we conclude that for the presented system the solute–particle effective interaction is repulsive.

For our approach we assume small Peclet and Reynolds numbers for the motion of the micro-particles, and fast diffusion of O₂ in the H₂O₂ aqueous solution.⁴² Therefore a steady distribution of solute quickly establishes around the moving particles. In the co-moving reference system, i.e., the system of coordinates moving with the active particle, the steady-state solute concentration ρ(r) obeys the Laplace equation (∇²ρ = 0) subject to the boundary conditions of a source current at the catalytic part of the particles (−D∂_nρ = P, where D is the diffusion coefficient of the O₂ product, P is the rate of O₂ production per unit surface, and n denotes the outer normal to the surface of the particle) and zero normal current at the inert part of the particles (∂_nρ = 0) and at the surface of the confining glass wall (∂_zρ = 0), because both are impermeable to the solute. Due to the complex structure of the carrier–cargo composites considered here, this equation has to be solved numerically. The solute concentration profiles allow the computation of the phoretic slip V_s(r) around the particle. The translational and rotational velocities of the object are determined as integrals, over the surface of the particles, of the phoretic slip weighted by complex hydrodynamic resistance matrices (which depend only on the geometry of the system^30,45). Consequently, the orientation of the motion and the magnitude of the phoretic velocity can be qualitatively understood from the distribution of the catalytic reaction product around the carrier–cargo composites.

Because the axis of the Janus particle remains, on average, parallel to the wall (x–y plane), the overall behaviour (translation, rotation, or both) can be inferred from the structure of the concentration distributions in this plane. Fig. 4 displays the calculated concentration distribution of the catalytic reaction products formed around the Janus spheres of radius R and various carrier–cargo composites. For the numerical calculation it has been assumed that the particles are half-coated by the catalyst and they perform motion in close vicinity to the wall, without being in mechanical contact with it. If the symmetry axis of the carrier–cargo composite coincides with that of the in plane concentration distribution, the motion involves only translational motion along that axis and no rotational component. This type of motion holds for perfectly aligned and rigidly “glued” spheres, as shown in Fig. 4(a) and (b), and in their corresponding insets. However, if the spheres are not “glued”, as it is the case in the experiment, the misalignment of the catalytic particles with respect to the axis of the colloidal cluster is likely to happen. For instance, the concentration distribution in Fig. 4(b) implies that the catalytic spheres actually experience a tendency to turn such that the directions of the catalytic parts are facing away from each other. The larger concentration gradient near the contact point (red coloured area in Fig. 4(b)) imposes a misalignment of the particles, as shown in Fig. 4(c). The same conclusion can be drawn for all the configurations in the insets of Fig. 4(b). Therefore, for any composite comprising more than one active particle or one cargo, the perfectly aligned configurations with a symmetry axis in the plane of motion are unstable so that purely translational motion cannot occur. The configuration in Fig. 4(c) shows that the overall distribution of the reaction product exhibits broken axial symmetry and will necessarily lead to a phoretic rotation of the whole composite. Such a rotation can emerge as clockwise- or anti-clockwise with equal probability, but once it emerges its direction is maintained. This explains the rotations observed experimentally (see the trajectories in Fig. 2), and in particular the systematic clockwise and anti-clockwise rotations exhibited by the double carrier configurations (Fig. 2(b)) and the anti-clockwise rotation exhibited by the single carrier with two cargoes (Fig. 2(a), right). Further geometrical misalignments, as shown in Fig. 1 (bottom right), additionally contribute to such rotations. These scenarios are qualitatively different from the commonly invoked diffusional rotation, which involves random changes in the instantaneous direction of motion due to thermal fluctuations.^21,32

Fig. 4 Colour coded steady-state distribution ρ(r) of reaction products around single (a) and double (b) free carriers. The insets show the corresponding distributions in the presence of one or two cargoes. (c) Misalignment of a free double carrier, driven by the large concentration gradient at the “contact” point. For the same configurations as in (a)–(c), the bottom panels show ρ(r) in the plane perpendicular to the wall and the symmetry axis and passing through the point of contact of the two active spheres (or through the centre of the free single carrier, respectively). The densities ρ(r) are represented in units of the total rate of solute production × the characteristic diffusion time R²/D.

Additionally, one should note that the presence of the glass substrate (“wall”) modifies the structure of the steady state distributions of the reaction products around the particles. Fig. 4(d)–(f) clearly demonstrate the “deformations” of the concentration distributions in the y–z plane with higher solute (O₂ molecules) concentrations close to the wall. As already mentioned before, such an asymmetry in conjunction with the solute–particle repulsive effective interactions in the system implies a tendency for the carrier to move in the direction away from the wall. However, this competes with the stronger opposite tendency of sedimentation, which explains why the active motion remains confined near and parallel to the wall.

Finally, we note that Fig. 3(a) shows that at large H₂O₂ concentrations (about 15% v/v) the ratio of velocities of the [one carrier, one cargo] and of the [one carrier] configurations is approximately V₁/V_f = (5.3/8.6) ≈ 0.62. This value agrees well with the recently reported⁴⁶ theoretically one calculated for the same configuration in an unbounded solution.¶ Furthermore, a comparison of the velocities of the single and double micro-motor(s) carrying two cargoes (the blue lines in Fig. 3(a) and (b)) reveals that at all fuel concentrations studied the single micro-motor performs as good, or even better, than a double carrier; for example, at 15% v/v of H₂O₂ the velocities are approximately 4.25 and 4.21 μm s⁻¹, respectively. This can be explained qualitatively from visual inspection of the corresponding theoretical O₂ concentration distributions ρ(r) shown in the insets of Fig. 4(a) and (b). The single carrier, loaded by one or two cargoes, exhibits carrier–cargo and cargo–cargo contact points (Fig. 4(a), inset). Due to the symmetry of the composite, the concentration of the reaction product is low in the vicinity of these contact points. Therefore, for the single carrier the concentration gradient along the surface of the composite runs mainly along the Ox direction, with negligibly weak variations at the carrier–cargo contacts. For the double carrier composite there is an additional type of contact points, between the two carriers. In contrast to the other points of contact, this one is associated with a large concentration of the reaction product (red area in Fig. 4(b), and insets). The resulting gradient of the concentration, oriented towards this point, induces a significant flow of the solvent along the surface of the particles and towards the carrier–carrier contact point (along the −Oy direction). This component of the flow does not contribute to the translational motion of composites (along Ox). Near the carrier–carrier contact, this flow is eventually directed towards the contact point, and thus contributes in the negative Ox direction, which slows down the translational motion of the whole double carrier composite. This results in an effectively larger “driving force” for the single motor configuration, while the hydrodynamic resistances are expected to be similar (because in a first approximation they are determined by the cross-sections exposed in the Ox direction of motion). These effects of the carrier–carrier contact point on the solution flow are superposed to the already identified tendency of the double micro-motor to break the alignment of the catalytic caps (leading to configurations shown in Fig. 4(c)), which additionally decreases the efficiency of the double carrier composite with two cargoes.

A further direct comparison between the velocities attained in various configurations is difficult at the qualitative level because of the single-particle spinning and reorientation inside the composite as discussed above; a complete numerical analysis is required for elucidating the effects of the cargoes.

In summary, we have investigated the motion (i.e., the trajectories and the velocities) of carrier(s)–cargo(es) systems by using catalytically active Janus micro-spheres. Specific attention has been paid to the cases in which a single or a pair of active carriers upload and transport colloidal cargoes. In order to gain insight into the complex dynamics of these systems, we have calculated the distributions of the catalytic reaction products formed at the catalytic parts of the micro-motors. These findings allow us to qualitatively explain and understand the systematic rotations observed for the various carrier–cargo composites. We have found that multi-particle configurations with perfectly aligned carriers are intrinsically unstable against spinning and reorientation of the catalytic caps of the carriers, leading to rotations of the carrier–cargo composites. Therefore rigid inter-particle linkers, or external fields, are necessary for achieving unidirectional motion. These studies on the transport of cargoes by Janus micro-motors are expected to contribute to the development of targeted load, transport, and delivery of micro-objects for lab-on-a-chip and drug delivery applications.

Acknowledgements

This work was supported by the Volkswagen Foundation (I/84 072). We thank Dr. D. Makarov for support in preparing the samples and for fruitful discussions.

Notes and references

1. A. J. DeMello, Nature, 2006, 442, 394.
2. K. T. Vaughan and E. H. Hinchliffe, Semin. Cell Dev. Biol., 2010, 21, 247.
3. M. Schliwa, Nature, 1999, 397, 204.
4. M. Schliwa and G. Woehlke, Nature, 2003, 422, 759.
5. T. D. Pollard, L. Blanchoin and R. D. Mullins, Annu. Rev. Biophys. Biomol. Struct., 2000, 29, 545.
6. T. D. Pollard, et al. , J. Biol. Chem., 1973, 248, 4682.
7. J. Kreuter, Colloidal Drug Delivery Systems, CRC Press, 1994.
8. P. Tierno, et al. , J. Phys. Chem. B, 2007, 111, 13479.
9. S. Sanchez, A. A. Solovev, S. Schulze and O. G. Schmidt, Chem. Commun., 2011, 47, 698.
10. Y. F. Mei, A. A. Solovev, S. Sanchez and O. G. Schmidt, Chem. Soc. Rev., 2011, 40, 2109.
11. C. E. Ashley, et al. , Nat. Mater., 2011, 10, 389.
12. M. Yoshida, E. Muneyuki and T. Hisabori, Nature, 2001, 2, 669.
13. W. F. Paxton, et al. , J. Am. Chem. Soc., 2004, 126, 13424.
14. R. F. Ismagilov, et al. , Angew. Chem., 2002, 114, 674.
15. G. A. Ozin, I. Manners, S. Fournier-Bidoz and A. Arsenault, Adv. Mater., 2005, 17, 3011.
16. R. Dreyfus, et al. , Nature, 2005, 437, 862.
17. T. R. Kline, et al. , Angew. Chem., Int. Ed., 2005, 44, 744.
18. M. Pumera, Nanoscale, 2010, 2, 1643.
19. P. Dhar, et al. , Nano Lett., 2006, 6, 66.
20. S. Sundararajan, et al. , Nano Lett., 2008, 8, 1271.
21. J. R. Howse, et al. , Phys. Rev. Lett., 2007, 99, 048102.
22. J. G. Gibbs, S. Kothari, D. Saintillan and Y.-P. Zhao, Nano Lett., 2011, 11, 2543.
23. A. A. Solovev, et al. , Small, 2009, 5, 1688.
24. S. Sanchez, A. A. Solovev, Y. F. Mei and O. G. Schmidt, J. Am. Chem. Soc., 2010, 132, 13144.
25. S. Balasubramanian, et al. , Angew. Chem., 2011, 123, 4109.
26. D. Kagan, et al. , Nano Lett., 2011, 11, 2083.
27. S. Fournier-Bidoz, A. Arsenault, I. Manners and G. A. Ozin, Chem. Commun., 2005, 441.
28. S. J. Ebbens and J. R. Howse, Soft Matter, 2010, 6, 726.
29. W. F. Paxton, S. Sundararajan, T. E. Mallouk and A. Sen, Angew. Chem., Int. Ed., 2006, 45, 5420.
30. M. N. Popescu, S. Dietrich, M. Tasinkevych and J. Ralston, Eur. Phys. J. E, 2010, 31, 351.
31. S. Ebbens, et al. , Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2010, 82, 015304.
32. R. Golestanian, Phys. Rev. Lett., 2009, 102, 188305.
33. L. F. Valadares, et al. , Small, 2010, 6, 565.
34. G. V. Kolmakov, V. V. Yashin, S. P. Levitan and A. C. Balazs, Soft Matter, 2011, 7, 3168.
35. A. A. Solovev, et al. , Adv. Funct. Mater., 2010, 20, 2430.
36. A. Erbe, et al. , J. Phys.: Condens. Matter, 2008, 20, 404215.
37. R. Golestanian, T. V. Liverpool and A. Ajdari, Phys. Rev. Lett., 2005, 94, 220801.
38. J. L. Anderson, Annu. Rev. Fluid Mech., 1989, 21, 61.
39. F. Juelicher and J. Prost, Eur. Phys. J. E, 2009, 29, 27.
40. M. Albrecht, et al. , Nat. Mater., 2005, 4, 203.
41. L. Baraban, et al. , Phys. Rev. E: Stat., Nonlinear, Soft Matter Phys., 2008, 77, 031407.
42. M. N. Popescu, S. Dietrich and G. Oshanin, J. Chem. Phys., 2009, 130, 194702.
43. http://rsbweb.nih.gov/ij/ .
44. Q. Chen, S. C. Bae and S. Granick, Nature, 2011, 469, 381.
45. R. Golestanian, T. V. Liverpool and A. Ajdari, New J. Phys., 2007, 9, 126.
46. M. N. Popescu, M. Tasinkevych and S. Dietrich, Europhys. Lett., 2011, 95, 28004.

---

Footnotes

† Electronic supplementary information (ESI) available: Motion of Janus particle carriers loaded by cargo (Movie S1). Video legend: the motion of single and double carriers, loaded by one or two cargoes, is displayed. Loss of the axial symmetry of the colloidal composites leads to systematic rotations of carrier–cargo systems. See DOI: 10.1039/c1sm06512b

‡ The aggregation of the carriers and cargo into composites can occur due to direct contact (as the Janus carriers move and collide with the others), which is sufficient to overcome the electrostatic barrier and to fall into the dominant (at small separations) van der Waals attraction. An alternative possibility is that of depletion forces. The effective repulsive interactions between the O₂ solute and the surface of the particles would indeed lead to an attractive depletion interaction between pairs involving at least one active particle.

§ The velocities V in Fig. 3 are determined as follows: (a) multiple trajectories of the moving particles or composites are recorded; (b) the instantaneous velocity along the trajectory is obtained as the time derivative of the arc length, and shows a fluctuating behaviour around a value 〈V〉; (c) the mean velocity V (shown in Fig. 3) is obtained by averaging the velocities 〈V〉 corresponding to the various recorded trajectories (from 6 to 10 for each carrier–cargo composite). The error bars indicate the fluctuations around this average.

¶ Fig. 2(b) in ref. 46 shows that for a “pusher” configuration (as l/R therein tends to zero, i.e., carrier and cargo being almost in contact, like in the experiment) the scaled velocity of the composite is slightly larger than 0.16. For the isolated carrier, the equally scaled velocity is 0.25. Thus the ratio of the two is ∼0.66, which is close to the present experimental result.

---