Sensitivity of PS/CoPd Janus particles to an external magnetic field

The dual nature of Janus particles confers fascinating properties such as a response to multiple stimuli. In this communication, we systematically study the sensitivity to a uniform external magnetic field of isolated Janus rod-shaped and spherical particles in water confined to two dimensions. The Janus asymmetry of the particles is given by magnetic [Co(0.28 nm)/Pd(0.90 nm)]8 multilayer films deposited onto monodisperse polystyrene (PS) nanorods and microspheres, respectively. It is shown that the particles dispersed in water respond to weak magnetic field applied in in-plane direction. Here we demonstrate that a precise control of the in-plane particle orientation can be obtained for magnetic field strengths higher than 0.1 mT for microspheres and 0.4 mT for nanorods.


Introduction
The very sensitive response of so matter to external stimuli, like temperature, pH, electric and magnetic elds, causes strong research interest. The small-scale motion control has attracted wide-spread attention in various elds including chemistry, material science, biology and medicine. [1][2][3][4] Exposing colloids to external elds, for example, electric, 5 magnetic, 6 optical 7 and acoustic 8 ones, gives the fascinating opportunity to study so matter in a controlled way. Magnetostatic interaction is widely independent from temperature uctuations, surface charges and the pH-value of the environment. Moreover, the magnetic eld can penetrate into non-magnetic or weakly magnetic materials that are used for fabricating microuidic devices or biological matter. Therefore, micro-and nano-objects in closed environments can be effectively manipulated without being directly contacted. Within the past decade, researchers have developed a variety of techniques for synthesizing colloidal magnetic particles with a wide range of reproducible shapes and sizes that differ in the distribution of the magnetic material within each particle. Magnetic nanoparticles with isotropic magnetization distribution have been fabricated via chemical synthesis, [9][10][11] electrodeposition 12 and lithographic techniques. 13 Single particles with distinct magnetic properties can be used as basic building blocks for complex structures and applications. Magnetic nanoparticles have also been embedded in hydrogels or polymers via microuidic channels or by encapsulation. [14][15][16] Narrow size distribution of the magnetic nanoparticles is important, since the blocking temperature of nanomagnets depends critically on the size of the particles. A polydisperse sample results in broad blocking temperature unwanted in applications. Another synthesis method relies on the surface modication of polymer or silica particles via metal deposition. Janus particles, named aer the two-faced Roman god Janus, are characterized by surface areas with different polarities or functionalities. By using this approach, one part of the particle surface is covered with a composition containing Co, 17,18 Ni 19 or Fe 20,21 metals providing ferromagnetic properties. Magnetic properties of Janus particles depend strongly on their size and shape in addition to their intrinsic magnetic characteristics, such as magnetic moment and anisotropy. The dynamics of a magnetic active Janus particle undergoing Brownian motion under the inuence of a uniform magnetic eld has been analytically and numerically examined by several groups. 22,23 The interaction of a magnetic eld with the magnetic dipole of a particle induces a torque that competes with the randomization of its orientation due to rotational diffusion. Most of the experimental studies dealing with magnetic control of spherical active Brownian particles have been focused on monitoring their spatial trajectory. In this work, we focus on the monitoring of the particle orientation in dependency of the magnetic eld strength. To our knowledge, there is no systematic experimental study on the sensitivity of the orientation of magnetic Janus particles to low external magnetic elds.
Here we introduce spherical and rodlike magnetic Janus particles. The Janus particles are based on polystyrene (PS) microspheres with a 4.2 mm in diameter or nanorods with a diameter of 0.4 mm and a length of 5 mm. The hemispherical ferromagnetic coating by Co/Pd multilayers possesses a perpendicular magnetic anisotropy. For the spherical Janus particles, one half of their surface was covered by magnetic material, while for Janus nanorods, the metal lm was deposited onto one tip only and covered 4% of the total rod surface. The sensitivity of articially designed magnetic particles and clusters dispersed in water has been studied via video microscopy. The magnetic-eld sensitivity is dened by the uctuations of the particle orientation in low magnetic elds. The reorientation time of Janus particles was studied as a function of magnetic eld strength. Both types of particles demonstrated a high sensitivity to low magnetic elds. The PS/CoPd microspheres follow exactly the direction of the magnetic eld starting from 0.1 mT. A reasonable response of the nanorods was observed for elds starting from 0.4 mT.

Materials and methods
2.1 Synthesis of PS/CoPd Janus particles 2.1.1 Preparation of nanorods with hemispherical tip shape. The PS nanorods with hemispherical tips were fabricated in two steps following the procedure reported earlier. 24 Briey, as a rst step, a 1 : 10 solution of PS (M w ¼ 24 700 g mol À1 M w /M n ¼ 1.03, Polymer Source Inc., Canada) in toluene was inltrated into the nanopores of a homemade anodic aluminum oxide (AAO) membrane template 25 with a pore diameter of 0.4 mm and a pore depth of 5 mm. To complete the inltration, AAO membranes with PS bulk lm were annealed at 160 C for 6 hours under Ar while applying a load of 0.3 bar. Then, the samples were cooled to room temperature at a rate of À5 C min À1 for homogeneity and thermal relaxation of the PS lm. In the second step, the separation of nanorods from the PS lm with the subsequent formation of the hemispherical tips, was performed by modication of procedures reported earlier. 26,27 The samples were immersed in a polyethylene glycol (PEG) 400 bath at 80 C for 1 h to induce dewetting of PS. To complete dewetting process, the samples were annealed at 200 C for approximately 4 h in a PEG 400 bath. Finally, the AAO templates with inltrated PS nanorods (Fig. 1a) were washed in an ethanol/ water mixture, at least 5 times followed by drying at 40 C for 18 h.
2.1.2 Fabrication of PS microsphere arrays. The Janus spherical particles were prepared according to a previous study. 28 A 2.5 wt% aqueous solution of PS microspheres with diameter of 4.2 mm was purchased from Bangs Laboratories, Inc. and used as received. A monolayer of PS microspheres was prepared by drop-casting method using pre-cleaned n-doped Siwafers with areas of 2 cm 2 (Fig. 1d). Then, the Co/Pd multilayers were deposited on the top of microsphere arrays.

Metal deposition.
A magnetic thin lm was sputtered normal to the samples represented by either nanorod arrays or monolayers of PS microspheres via magnetron sputter deposition at a base pressure of 2 Â 10 À8 mbar and the Ar gas pressure of 3.5 Â 10 À3 mbar at room temperature. The deposition rates were kept at 0.03 nm s À1 for Ta, 0.05 nm s À1 for Pd and 0.03 nm s À1 for Co, respectively. As a rst step, to improve the adhesion and the perpendicular magnetic anisotropy of thin metal lms, the samples were covered by a seed-layer consisting of Ta(3.0 nm)/ Pd(3.0 nm). 29 Then, a magnetic multilayer sequence consisting of [Co(0.28 nm)/Pd(0.9 nm)] 8 was deposited. As a nal step, the lm was capped with an additional Pd(2.1 nm) layer to prevent oxidation. The resulting deposited composition was Ta(3 nm)/Pd(3 nm)/[Co(0.28 nm)/Pd(0.9 nm)] 8 /Pd(2.1 nm). The resulting lm thickness was 17.54 nm. Finally, the magnetically capped particles were saturated at 1 T for 5 min at room temperature to align the cap moments in the out-of-plane direction.
2.1.4 Compositional characterization. The morphological and compositional characterization of the Janus particles were performed by Scanning Electron Microscopy (ZEISS Auriga, Germany) equipped with an Energy Dispersive X-ray Spectrometer (Oxford Instruments, UK). The samples were prepared by dropping 2 mL Janus particle solutions onto glass substrates previously cleaned with oxygen plasma. The measurements were performed with in-lens and X-ray detectors, acceleration voltage of 14 kV and 5 mm working distance. Prior to the imaging tests, the samples were sputter coated with Pt/Ir at 20 mA for 15 s three times in a K575X Emitech device.
2.1.5 X-ray diffraction measurements. The X-ray diffraction was performed on a diffractometer system XPERT-PRO (PANalytical, the Netherlands) having sample stage MRD cradle conguration. The Cu Ka radiation source, l ¼ 0.15418 nm was used. The measurement was conducted from start 2Q angle of 10 till end angle of 90 values with a step size of 0.02. The total time for measurements were about approximately 9 hours.  2.2.2 Detaching of microspheres. The microspheres were separated from the Si/SiO 2 wafer in ultrasonic water bath within 20-30 min at room temperature. Then, the particles were collected and resuspended in DI water.
2.2.3 Sample cell. Few droplets of the suspended Janus particles were placed on a pre-cleaned glass slide and le for particles sedimentation. Once particles were settled down, the external magnetic eld was applied.

Video recording and analysis
The trajectories and orientations of particles were studied by a Leica DMI3000B transmission microscope integrated with a Leica DFC395 CCD video camera and a 63X objective with a numerical apperture (NA) of 0.70. Two pairs of Helmholtz coils with diameter of 2 cm 2 provided homogeneous magnetic elds in x and y directions. All videos were captured at a frame rate of 20 frames per second (fps). The particle movements were subsequently analyzed by homemade Python and LabView scripts. The video processing algorithm was used to track the positions and the orientations of individual nanorods and microspheres with pixel resolution. 1 pixel corresponds to a size of 0.127 mm 2 .

z-potential measurements
The z-potential measurements were performed using a Malvern Zetasizer Nano ZSP instrument equipped with a 632 nm HeNe laser operating at a 173 detector angle. Prior to the experiment, the particles were dispersed in water. Three and ve tests were performed for each sampleset, respectively.

Results and discussion
The geometric shape of the PS/CoPd particles changes negligibly aer the deposition of 17.54 nm thick magnetic lm ( Fig. 1b and e). The hemispherical metallic caps follow the curvature of the particles. Due to directed deposition, the lm thickness varies across the cap following a cosine behavior with the maximum at the center of the cap. It was shown that below a critical Co layer thickness of about 0.8 nm, the magnetization points perpendicular to the lm plane due to interface anisotropy. 30 Transition of easy axis from perpendicular (out-of-plane) to a parallel (in-plane) direction occurs at a Co thickness of about 0.8 nm. 31 For a Co lm thickness of 0.28 nm as deposited here, a planar lm exhibits perpendicular anisotropy. The total magnetic moment depends weakly on the thickness of the Pd layers in the multilayer. 32 Magnetic force microscopy (MFM) measurements at a plane of Co/Pd multilayer deposited on Si conrmed the perpendicular magnetic anisotropy. The magnetization orientation also follows the curvature of the particle. The magnetization decays from the top to the rim of the cap and hence there is a preferred magnetic directionalong the symmetry axis of the cap.
In order to nd the magnetic saturation point, 10 mL suspension were placed onto Si-wafer with area of about 10 mm 2 . Aer the evaporation of water, the sample was measured at temperatures of 5 K and 300 K, respectively. The SQUID measurements at a temperature of 300 K are represented in Fig. 1c and f. For the outplane orientation of the external magnetic eld, the saturation for nanorods was achieved at 0.5 T. In turn, the Janus microspheres reached magnetic saturation at 0.2 T. For Janus particles easy axis of magnetization was found in out-of-plane.
To identify the elemental composition of Janus particles, a spatially resolved EDX characterization was performed. The EDX spectra directly revealed the presence of the atomic elements in the sample as shown in Fig. 2a and b. The system was le to automatically detect the elements and the following elements were determined in this way Si, O and C. Manually we have checked if one of the following is also present: Pd and Co. Pd and Co were registered in small amounts. Additionally, we performed X-ray diffraction to analyze the structure of the metal thin lm. The Janus nanorods embedded in AAO membrane with the capped side facing to the air were used for analysis. Empty AAO template with a pore diameter of 400 nm and a pore depth of 5 mm was used as a reference. The XRD pattern of empty alumina is presented in Fig. 2c. The diffraction peak position was recorded as the detector angle, 2Q. The XRD pattern shows the hexagonal structure of AAO with a preferential orientation along the (200) direction. For PS/CoPd nanorod arrays embedded in AAO template, two peaks were identied as shown in Fig. 2d. Here, the peak at 44.6 corresponds to (200) plane reection of AAO template. 33 The double peak at 69 corresponds to (220) planes reection of CoPd alloy. 34,35 As the next step, the particles were suspended in deionized water. The z-potential measurements on the rodlike particles showed À35.1 AE 7.3 mV (5 measurements), while the average value of z-potential on the microspheres was À45.6 AE 2.7 mV (3 measurements). Hence, the Janus particles in contact with the aqueous solution were negatively charged. To encapsulate the particle dispersion for long time from the environment, a droplet of about 3-5 mL of particle suspension was enclosed between two glass slides separated by a distance of several ten micrometers. Then, the glass slides were sealed with UV glue to prevent evaporation of the solvent. Due to the density mismatch and the electrostatic repulsion, the particles settled within a few nanometers above the bottom glass surface. When the Janus nanorods were sedimented in water, the Co/Pd caps pointed in random directions, while the Co/Pd caps of microspheres were oriented downwards to the glass slide. Applying sufficient magnetic eld, Co/Pd caps reoriented and aligned to the magnetic eld. Fig. 3a represents a schematic illustration of spherical and rodlike Janus particles with one magnetic cap. The orientation of the cap is dened by a polar angle q formed by the magnetic eld direction and the macrospin orientation. In other words, q iden-ties the rotational displacement of Janus particle in a polar direction. Depending on the eld strength, the orientation of the macrospin of the Co/Pd cap uctuates around the direction of the external eld B. For B applied along y axis, the uctuations around the externally dened direction are shown in Fig. 3b-d. We measured the orientation of Co/Pd caps for each value of the magnetic eld strength B, at least, over 60 s. Prior to setting a different value of B, we applied B y ¼ 0 for, at least, 30 s to orient the caps randomly. As can be seen in Fig. 3b for  The reorientation times T of particles in weak (T rw ) and strong (T rs ) magnetic elds are shown in Fig. 4. Here, the particle was rotated by 90 by switching the external magnetic eld at constant amplitude from x to y direction. The reorientation time is dened between the crossing of the linear t of the transition region of the trajectory with the 0 and 90 levels. The intervals T rs and T rw might have a different le border. If the orientation of particles for certain eld reached values B w AE DB w or B s AE DB s and then remained in this range, at least, for 2 s, we considered the response of particle as a reorientation. The reorientation times for single Janus nanorods and microspheres are summarized in Fig. 5.
A 90-degrees reorientation time of particle decreased upon the increase of B. We found that the reorientation time of Janus nanorods changes from 12.9 AE 8.3 s for B ¼ 0.4 mT to 6.2 AE 0.9 s Fig. 3 (a) Schematic representation of Janus particles with one side coated by a metallic cap. Sensitivities of single Janus nanorods (b), cluster consisting of two rods (c) and microspheres (d) to externally defined applied magnetic field. The data represent average measurements for 4-6 different single particles and 1 cluster. All observations were taken over 240 s.  for B ¼ 1 mT, correspondingly. In the presence of strong magnetic eld of B ¼ 1 mT, the reorientation of rodlike particle occurred about two times faster than that for B ¼ 0.4 mT. The reorientation time of PS/CoPd spherical particles was one order of magnitude smaller than that measured for nanorods. The microspheres were rotated by 90 within 1.3 AE 0.13 s in a weak magnetic eld of B ¼ 0.1 mT, while in a strong eld of B ¼ 0.7 mT the reorientation was observed in 0.69 AE 0.08 s.

Brownian motion in a presence of magnetic eld
Brownian motion is the result of random collisions between a microscopic particle and the molecules of the surrounding uid. The presence of external constant magnetic elds inuences the random, uncontrolled movement of ferromagnetic Janus particles in a uid due to the interplay between Brownian diffusion and magnetic torques. In turn, the magnetic eld causes a conning potential of the angular motion and the Brownian particle is captured in the local minimum (lowest energy state) of a potential well. In the most simplest case, the Brownian particle behaves like a one-dimensional stochastically damped harmonic oscillator. 36,37 The width of the potential energy function depends on the strength of the magnetic eld and the magnetic moment of the cap. Moreover, the width of the potential function decreases with the increase in the magnetic eld strength. Classically, the Brownian motion at temperature T is restricted to the region where the connement potential is less than k B T, with k B the Boltzmann constant. At a constant temperature, if the magnetic eld is weak, the particle moves in a broad region while for strong elds the region becomes more narrow. Therefore, the orientation of magnetic particle directed parallel to strong eld B s deviates less from the eld direction than that for weak eld B w . Fig. 6 shows the eld-dependent standard deviation functions, when the magnetic eld points vertically (B along y direction) or horizontally (B along x direction). These results have been obtained for Janus rod-like particles. As can be seen from the graph, the two functions have nearly the same prole indicating the equivalence of the responses of Janus nanorods to vertically or horizontally applied magnetic elds. As expected, Dq decreases with the eld strength. For a weak magnetic eld, the Brownian diffusion dominates over magnetic interaction. For a strong magnetic eld, the thermal uctuations of the orientational motion of the particle was nearly suppressed, and the magnetic dipole was aligned along the eld direction. Due to the larger magnetic moment, the alignment is much stronger for the microspheres with respect to the nanorods.

The inuence of anisotropy of PS/CoPd Janus particles on the rotational dynamics
The cap signicantly impacts the rotational dynamics of the particle due to gravitational torque and the electrostatic torque resulting from the asymmetry in electrostatic forces. The gravitational torque on a Janus particles is caused by the density mismatch between the cap and native particle. The gravitational torque depends on the weight of the cap, which in turn depends on the thickness, total size and cap material. In our experiments, the thickness of Co/Pd caps was xed at 17.54 nm for both types of Janus particles. The estimation of weight of non-coated PS spherical and rodlike particles gives 40.9 Â 10 À15 kg and 0.64 Â 10 À15 kg, respectively. The spherical particle without coating is approximately 64 times heavier than nanorods. The mass of the CoPd caps amounts to approximately 2.8 Â 10 À15 kg for microspheres and 0.026 Â 10 À15 kg for nanorods, respectively. Hence, the mass of the cap amounts to about 7% for the spheres and about 4% for the rods. Aer the deposition of the Co/Pd multilayer, the caps of spherical particles in an aqueous solution were mostly oriented downwards similar to Rashidi et al. 38 The authors investigated the impact of particle size and coating thickness on the magnitude of rotation for Au-coated PS spherical Janus particles with radii ranging from 1 to 6 mm and a cap thicknesses up to 20 nm. It was shown that the magnitude of downward rotation decreased with increasing particle diameter at a xed cap thickness. We estimated the hydrodynamic sizes (d h ) of Janus nanorods and microspheres. d h of spherical particles equals to 4.2 mm while d h for nanorods is 1.8 mm. The calculation of d h for nanorods was based on the equivalent sphere theory 39 where T is temperature, h is the viscosity of water. The translational diffusion coefficient for rodlike particles D t ¼ 0.24 (mm) 2 s À1 was estimated based on the Tirado and Garcia de la Torre theory. 40 In line with the Brownian dynamics simulation of Rashidi et al., in our experiments, we also observed that the amplitude of uctuation of q for spherical particles was much smaller than that for nanorods at xed magnetic elds.

The impact of Co/Pd cap size on the sensitivity to magnetic elds
To investigate the inuence of cap size on the sensitivity, we calculated the ratio between the Co/Pd cap surfaces and the PS parts for spherical and rodlike particles based on the SEM images. For Janus microspheres, the ratio is estimated to be 0.5 (half of particle is coated with Co/Pd). For nanorods, only one of the hemispherical tips with radius of 0.2 mm was coated with Co/Pd. The total surface of the nanorod can be represented by the side of the cylinder, which when unrolled is a rectangle with sides of 5 mm and 1.256 mm, and the surface of the sphere with radius of 0.2 mm. The ratio of the metalized surface of the rodlike particle to the total surface area constitutes was determined as small as 0.04. The Co/Pd surface area Fig. 6 The standard deviation of polar angle q for single Janus nanorods and microspheres (shown as an inclusion).
of the cluster pair increased by a factor two as compared to a single particle. At the same time, the polymeric surface area is 1.7 times higher compared to a single nanorod. The ratio between metalcoated and nonreactive surface areas is approximately 0.05, 20% higher than that of a single particle. Despite of the tiny amount of magnetic materials, Janus nanorods were observed to be highly sensitive to application of weak magnetic elds. Co/Pd magnetic capped particles revealed that Janus microspheres were more sensitive to magnetic elds as compared to single nanorods. The reason might be the larger amount of ferromagnetic material on the microspheres, and the larger size of the nanorods with respect to the magnetically active caps.

Conclusion
We synthesized CoPd-capped PS Janus particles, which respond very sensitively to application of low external magnetic elds.
The results reveal that a uniform magnetic eld affects the dynamic behavior of magnetic active particles. The precise control of the orientation of the particles was achieved at 0.1 mT for microspheres and at 0.4 mT for nanorods, respectively. The rotation time of 90 degrees of Janus nanorods decreases by a factor of 2 upon increasing the magnetic eld strength from 12.9 AE 8.3 s at 0.4 mT to 6.2 AE 0.9 s at 1 mT. The reorientation time values of microspheres were one order of magnitude smaller than that for nanorods and amounted to 1.3 AE 0.13 s at 0.1 mT and 0.69 AE 0.08 s at 0.7 mT. Owing to the magnetic properties, the PS/CoPd particles may be used, for example, to sense the presence of weak magnetic elds as so-called micromagnetometers.

Author contributions
A

Conflicts of interest
There are no conicts to declare.