Control of magnetic field distribution by using nickel powder@PDMS pillars in microchannels

Abstract

A simple and robust approach to control the magnetic field distribution by nickel powder@PDMS pillars was established. The nickel powder@PDMS pillars were fabricated in several simple steps, using a simple and robust method, and no training in techniques or expensive equipment is necessary compared to other methods. The localized magnetic field distributions in microchannels can be tailored by the nickel powder@PDMS pillars with automatic generation of high magnetic field gradients around them due to the high relative magnetic permeability of the nickel powder@PDMS. The numerical simulation and red fluorescent magnetic nanoparticle capture experiment results convinced us that our approach could effectively control the localized magnetic field distribution in the microchannels. Two kinds of tailoring events were studied at the powder@PDMS pillars in the microchannels underneath two different external magnetic fields. To the best of our knowledge, this is the first time different localized magnetic field distributions have been obtained in microchannels by nickel powder@PDMS pillars due to different external magnetic fields. This approach was used to capture fluorescent magnetic nanoparticles and magnetic bead–yeast cell complexes. We believe that this approach has great potential applications in chemistry, biology, biomedicine and tissue engineering.

---

Introduction

Magnetic nanoparticles, owing to many advantages such as high magnetization values, flexible surface functionalization, simple manipulation by magnets and good biocompatibility, have been widely used in therapy, diagnosis, tissue engineering, and other biomedical applications.^1,2 Microfluidic chips can integrate multiple functional units into one chip with benefits such as saving time, low cost and reducing the amount of reagents needed. A strategy that combines magnetic nanoparticles and microfluidic chips can remove complicated washing processes,^3,4 enhance their advantages, and maximize their applicability.^5,6 However, it is a big challenge to capture the magnetic nanoparticles at a high flow velocity (typically at dozens of microliters per minute) due to low magnetic field gradients in the microchannels.^7,8 The magnetic nanoparticles captured at low flow velocities in microfluidic chips (less than 1–2 microliter per minute) have low throughput and are not a good choice for many applications. Also, it is not efficient to remove waste and non-specific adsorptions⁹ on the magnetic nanoparticles. It is important and significant to control the localized magnetic field distributions in order to capture the magnetic nanoparticles at high flow velocities. Nickel and Ni–Fe alloys were applied to increase the localized magnetic field gradients in the microchannels,^10,11 owing to the significantly higher magnetic permeability of nickel relative to the buffer solution.^9,12,13 In our previous study, we used poly(dimethylsiloxane) (PDMS) encapsulated nickel patterns to control the magnetic field distribution and pattern magnetic bead arrays for on-chip detection of multiple cancer biomarkers.^3,14 The nickel patterns were fabricated by an electroplating process which needs electroplating instruments, an electroplating bath composition, intensive labour and a lot of waste generation.

Here, a new approach to tailor the localized magnetic field distribution in microchannels, by using the nickel powder@PDMS pillars, is reported. The pillars were composed mostly of the nanosized nickel powders and a little PDMS. Two kinds of different localized magnetic field distributions were studied in the microchannels at two different external magnetic fields. The red fluorescent-magnetic nanoparticles^2,15–17 were captured in the microchannels in order to observe the local magnetic field distributions. We also used COMSOL Multiphysics 3.5a software to simulate the localized magnetic field distributions in the microchannels. The simulation results agreed with the experiment results and confirmed that our approach could effectively control the localized magnetic field distribution in the microchannels. Our approach has many advantages: (1) the localized magnetic field distribution in the microchannels could be tailored by the nickel powder@PDMS pillars. By using nickel powder@PDMS pillars, there were higher magnetic field gradients around the nickel powder@PDMS pillars in the microchannels. (2) The nickel powder@PDMS pillars were generated in several simple steps, using a simple and robust method, and no training in electroplating techniques was necessary. (3) To our knowledge, other materials which have high relative permeability such as iron powder, Fe₃O₄ nanoparticles or nickel–iron alloy particles etc. can be used for the same purpose, to control the local magnetic field in microchannels. (4) This approach could be used to capture magnetic nanoparticles and generate magnetic nanoparticle arrays which could be used for cancer cell sorting, immunoassays, drug delivery, drug release and so on. To the best of our knowledge, this is the first time that different localized magnetic field distributions have been obtained in microchannels by nickel powder@PDMS pillars due to the different external magnetic fields.

Experimental section

Chemical reagents

SU-8 2050 photoresists and developer were purchased from MicroChem (MicroChem Corp., USA). AZ50XT photoresists and developer AZ400K were purchased from AZ Electronic Materials (AZ Electronic Materials USA Corp., USA). Poly(dimethylsiloxane) (PDMS) and curing agent were purchased from GE (GE Toshiba Silicones Co., Ltd., Japan). Nanonickel powder (50 nm, 99.9%) was obtained from DeKeDaoKing (Beijing, China). Yeast cells (Saccharomyces cerevisiae AY, BY4742) were purchased from the China Center for Type Culture Collection (Wuhan University, China). Red fluorescent magnetic nanoparticles were synthesized as in our previous reports.^2,15,17–20 MyOne carboxylic acid magnetic beads (MyOne, 1.05 μm diameter) were obtained from Dynabeads (Invitrogen Dynal, Norway). Dextrose was obtained from JingKeHongDa Bio-Tech Co., (Beijing, China). Peptone and yeast extract were purchased from BD (Bectone, Dickinson and company, USA). N-Ethyl-N′-(3-dimethyl-aminopropyl) carbodiimide hydrochloride (EDC), N-hydroxysulfosuccinimide sodium salt (sulfo-NHS) and 2-(N-morpholino) ethane sulfonic acid hydrate (MES) were obtained from Sigma-Aldrich. Bovine serum albumin (BSA) and concanavalin A (Con A) were obtained from Biosharp. Trimethylchlorosilane (TMCS), Na₂HPO₄·12H₂O, NaCl, NaH₂PO₄·2H₂O, Tween-20 and NaN₃ were purchased from GuoYao (Shanghai, China). All solutions were prepared with 18.2 MΩ cm ultrapure water obtained from a Millipore water-purification system (Millipore, USA).

Fabrication of nickel powder@PDMS pillars

The fabrication of the nickel powder@PDMS pillars and the microfluidic channel is shown in Fig. 1. The microfluidic chip with nickel powder@PDMS pillars was fabricated as per the following steps. First, standard soft lithography technology was used to fabricate the SU-8 mould. Briefly, the SU-8 2050 photoresist was spin coated on a clean silicon wafer to fabricate a ∼40 μm SU-8 mould with 50–100 μm holes by using designed masks of different shapes. Then the silicon SU-8 mould was exposed to TMCS vapour for 2 min in order to facilitate the mould release. Then the nickel powders (50 nm) were filled into the holes on the silicon SU-8 mould with the aid of a small brush. Next the degassed RTV615A and RTV615B components were mixed at a ratio of 10 : 1 by weight and cast on the silicon SU-8 mould for a thickness of about 500 μm and then kept at 75 °C for 1 hour. After that, the solidified PDMS pattern was peeled off from the silicon SU-8 mould with the nickel powders. Some of the pre-polydimethylsiloxane flew into the holes containing nickel powder and were cured to form the nickel powder@PDMS pillars. The PDMS pattern with nickel powder@PDMS pillars was bound to a clean slide glass to compose the microfluidic chip (Fig. 1(A)). This method for the fabrication of nickel powder@PDMS pillars was simple and robust and the silicon moulds can be reused. By designing and fabrication of different shaped holes in the moulds, we can obtain different shaped nickel powder@PDMS pillars, which could be used for the generation of different magnetic nanoparticle patterns in microfluidic chips.

Fig. 1 Fabrication process of the microfluidic chip with integrated nickel powder@PDMS pillars. (A) Fabrication of the nickel powder@PDMS pillars. (B) Fabrication of the microfluidic channel. (C) Side view of the integrated microfluidic chip with nickel powder@PDMS pillars.

Fabrication of microfluidic channels

The microfluidic channel layer was fabricated using a standard soft lithography method.²¹ An AZ50XT photoresist was spin coated on a silicon wafer to fabricate a positive mould about 40 μm thick. The PDMS mixture (RTV615A : RTV615B = 10 : 1 (w/w)) was poured onto the silicon mould and baked at 75 °C for 4 hours. The solid PDMS was peeled off from the positive mould and punched with a blunt needle for inlets and outlets, as shown in Fig. 1(B). Afterwards, the microfluidic channel layer was irreversibly bonded to the PDMS pattern with nickel powder@PDMS pillars (Fig. 1(C)).

The external magnetic fields acting on the microfluidic chip

The localized magnetic field distribution in the microchannels was studied under two different kinds of external magnetic fields: (1) two permanent magnets were fixed on a glass slide to generate an “NS” pole and put under each side of the microchannel (the gap between the permanent magnets was 6 mm) to provide a relatively uniform magnetic field with a magnetic flux density about 105 mT.¹⁴ (2) Only one permanent magnet was put under the microchannel to provide an upright magnetic field throughout the microchannel.

Results and discussions

SEM and optical imaging of the nickel powder@PDMS pillar arrays

Scanning electron microscopy (SEM) provided a straightforward way to characterize the nickel powder@PDMS pillars. Fig. 2 shows the SEM results of the nickel powder@PDMS pillars. The nickel powders on the surfaces of the PDMS pillars, demonstrated that the nickel powders filled the holes and were doped onto the PDMS pillars. The optical images of the nickel powder@PDMS pillars are shown in the ESI (Fig. S1†). The optical images demonstrate that the nickel powders (black) were filled out onto the PDMS pillars. Moreover, by designing and fabrication of different shaped holes in the photoresist moulds, we can obtain correspondingly shaped nickel powder@PDMS pillars. As a result, a large area of regular nickel powder@PDMS pillar patterns, without any defects, can be fabricated in the microchannels by using this robust and simple method.

Fig. 2 SEM image of the nickel powder@PDMS pillars. (A) 90 μm square shaped nickel powder@PDMS pillars. (B) 100 μm circular shaped nickel powder@PDMS pillars. (C) 65 μm “∼” shaped nickel powder@PDMS pillars. (D) 70 μm hexagonal shaped nickel powder@PDMS pillars. Inserts are amplifications of single nickel powder@PDMS pillars. The scale bars are 200 μm.

The stability of the nickel powder@PDMS pillars in microchannels

The nickel powder@PDMS pillars are very stable under high velocities, such as 50 μL min⁻¹ or higher. Because the nickel powders were encapsulated in the cured elastic PDMS, they cannot leak out from the pillars. Moreover, the nickel powder@PDMS pillars are fixed on the thin PDMS which is bonded to a clean glass slide. We could use an ultrasound cleaner to clean the microfluidic channels and reuse the chip. The channels weren't destroyed even under these extreme conditions.

Simulation of the external magnetic field distribution

The magnetic field is commonly described by the magnetic flux density B, which can be calculated by using eqn (1),^14,22

B = μ₀(H + M) = μ₀(H + χH) = μ₀(1 + χ)H = μ₀μ_rH (1)

μ₀ is the permeability of a vacuum (4π × 10⁻⁷, H m⁻¹), H is the magnetic field intensity (A m⁻¹), M is the magnetic moment per unit volume (A m⁻¹), χ is the volumetric magnetic susceptibility (dimensionless), μ_r is the relative permeability (dimensionless).

For the permanent magnets, the distribution of the magnetic fields can be described by eqn (2),^3,14

B = μ₀μ_rH + B_r (2)

B_r is the remnant magnetic flux density, which was one where no magnetic field was present. In our model, we assumed that the relative permeability of the magnets (μ_r) was 1.05 and the remnant magnetic flux density of the permanent magnets (B_r) was 1.17 T, as in our previous reports.^3,14 In our research, we have studied two kinds of the magnetic field distributions in microchannels under two different external magnetic fields. First, two permanent magnets were positioned at each side with opposite poles under the microchannel. The gap between the two permanent magnets was 6 mm. We called this external magnetic field magnetic field-1. The magnetic field distribution at magnetic field-1 is shown in the ESI (Fig. S2(A) and 2(B)†). In this situation there was an almost uniform magnetic field distribution in the microchannels and the magnetic flux density was about 105 mT.¹⁴ From Fig. S2(B),† it is clearly shown that the direction of the external magnetic field through the microchannel was in the Y-axis direction. However, if only one magnet was put underneath the microfluidic chip (we called this magnetic field-2), as shown in Fig. S2(C),† the magnetic field distribution was different from that of magnetic field-1 in the microchannel. As shown in Fig. S2(D),† the direction of the external magnetic field through the microchannel was in the Z-axis direction (the direction of the arrows represents the direction of the magnetic field distribution). The tailoring events of the localized magnetic field distributions in the microchannels by the nickel powder@PDMS pillars might be influenced by the different external magnetic fields.

Simulation of |(B·∇)B| under two kinds of magnetic field

The characterization of the magnetic field in the microchannels on a micrometer scale is tricky by common methods. We used a numerical simulation method and magnetic bead capture experiments to demonstrate the effective control of the magnetic field distribution in the microchannels. The magnetic force acting on a magnetic nanoparticle in a magnetic field can be described by eqn (3),^5,14Δχ is the difference in magnetic susceptibility between the magnetic nanoparticle and the surrounding buffer medium (dimensionless), V_m is the volume of the magnetic nanoparticle (m³), μ₀ is the permeability of a vacuum (4π 10⁻⁷, H m⁻¹), B and ∇B are the magnetic flux density (T) and magnetic field gradient (T m⁻¹), respectively. From this equation, it is easy to conclude that the magnetic force is laid on the (B·∇)B, if Δχ and V_m are constant. (B·∇)B can be calculated by using eqn (4),²²

We simulated the magnetic field distributions in two different external magnetic fields (magnetic field-1 (Fig. 2(A)) and magnetic field-2 (Fig. 2(C))). The directions of the external magnetic fields acting in the microchannels were different (ESI, Fig. S2†).

In the 2D model, the absolute value of (B·∇)B could be calculated using eqn (5),¹⁴

As shown in Fig. 3(A), |(B·∇)B| increased significantly as the Y-axis direction when the external magnetic field was in the same direction. However, in the X-axis direction, there was no obvious increase generated around the nickel powder@PDMS pillars. In the middle of the model, there was an area without any nickel powder@PDMS pillars, where no obvious increase of |(B·∇)B| along the direction of external magnetic field-1 was found. As shown in Fig. 3(B), there were eight peaks of values for |(B·∇)B| along the X-axis direction which corresponded to eight nickel powder@PDMS pillars. Meanwhile, in the middle nickel powder@PDMS-deficient area, there was no peak value for |(B·∇)B|. This demonstrated that our approach can tailor the magnetic field distribution in the microchannels by using the nickel powder@PDMS pillars.

Fig. 3 Simulation of |(B·∇)B| in the microchannel with the nickel powder@PDMS pillars in the external magnetic field-1 (Y-axis direction, X–Y plane). (A) Simulation result of |(B·∇)B| in the microchannel in the X–Y plane. (B) Simulation result of |(B·∇)B| vs. the distances along the X-axis direction, as marked by the white dashed line shown in (A). The scale bar in (A) is 100 μm.

|(B·∇)B| was simulated at magnetic field-2 (ESI, Fig. S3†). As shown in Fig. S3,† there was a significant increase of the absolute value of |(B·∇)B| around the nickel powder@PDMS pillars. It was also found that high |(B·∇)B| was generated around the whole nickel powder@PDMS pillars area, which was different from the simulation result at magnetic field-1 (Fig. 3). This was attributed to the difference of the external magnetic fields.

Control of the magnetic field distribution by nickel powder@PDMS pillars

The red fluorescent magnetic nanoparticles were used to carry on the capture experiments in order to display the magnetic field distribution tailored by the nickel powder@PDMS pillars clearly. Fig. 4(C) shows that the red fluorescent magnetic nanoparticles were captured in the microchannel with nickel powder@PDMS pillars under magnetic field-1. In this situation, the red fluorescent nanoparticles were captured at both sides around the nickel powder@PDMS pillars along the direction of the external magnetic field (Y-axis direction, Fig. S2(B)†). This result agreed with the simulation results of |(B·∇)B| under magnetic field-1 (Fig. 3). Fig. S4† also shows that the red fluorescent magnetic nanoparticles were captured in the microchannel with nickel powder@PDMS pillars under magnetic field-1. From the results, we found that the red fluorescent magnetic nanoparticles were captured only at the areas of the pillars along the direction of external magnetic field (Y-axis direction), which corresponded to the simulation results of |(B·∇)B| (Fig. 3) under magnetic field-1. In addition, if there was a defect unit in the nickel powder@PDMS pillars, no red fluorescent magnetic nanoparticles were captured at the area of the defect unit, but red fluorescent magnetic nanoparticles were captured at other areas in the same microchannel (ESI, Fig. S5†).

Fig. 4 (A) External magnetic field-1. (B) External magnetic field-2. (C) Fluorescence field image of the captured fluorescent magnetic nanoparticles in a microfluidic chip with opposite cylinder-shaped nickel powder@PDMS pillars under external magnetic field-1. (D) Fluorescence field image of the captured fluorescent magnetic nanoparticles in a microfluidic chip with staggered cylinder-shaped nickel powder@PDMS pillars under external magnetic field-2. The scale bars in (A) and (B) are 1 mm, and those in (C) and (D) are 100 μm.

As shown in Fig. 4(D), red fluorescent magnetic nanoparticles were captured at the areas around the whole cycles of the nickel powder@PDMS pillars under magnetic field-2, which was also confirmed by the simulation results of |(B·∇)B| at magnetic field-2 (ESI, Fig. S2†). Moreover, we can clearly observe that no red fluorescent magnetic nanoparticles were captured in the same microchannel in the absence of nickel powder@PDMS pillars (ESI, Fig. S6†). These results demonstrated that the high magnetic field gradients are automatically generated at the areas of the nickel powder@PDMS pillars, because the relative magnetic permeability of the nickel powder@PDMS (μ_r(nickel) ≈ 200) was significantly larger than the relative magnetic permeability of the buffer medium (μ_r(buffer)) ≈ 1.^9,23 From these results, we can deduce that high magnetic field gradients were induced in the microchannels. This was proved by the numerical simulation results. Therefore, this simple method can be used to control the local magnetic field distribution in the microchannels.

Generation of fluorescent magnetic nanoparticle patterns

Under the controllable magnetic field, the magnetic nanoparticles could be greatly captured, thus generating magnetic nanoparticle patterns. The magnetic nanoparticle patterns have a lot of potential applications, such as in immunoassays, cell sorting, or drug release systems based on magnetic nanoparticles. Here we designed and fabricated two kinds of differently shaped (“∼” shaped and hexagonal shaped) nickel powder@PDMS pillars. These two nickel powder@PDMS pillars were applied to the capture of red fluorescent magnetic nanoparticles at magnetic field-1. The results are shown in Fig. 5, which also shows that the magnetic nanoparticles were captured at both sides of the nickel powder@PDMS pillars along the external magnetic field direction. The optical and fluorescence images of the captured red fluorescent magnetic nanoparticles also proved that our method could tailor the localized magnetic field distribution by a simple process.

Fig. 5 Capture of red fluorescence magnetic beads in a microfluidic chip with different shaped nickel powder@PDMS pillars (magnetic field-1). (A) and (D) Bright field images of “∼” shaped or hexagonal nickel powder@PDMS pillars. (B) and (E) Fluorescence images of the captured fluorescent magnetic beads in chips with “∼” shaped or hexagonal shaped nickel powder@PDMS pillars. (C) and (F) Merged images of the optical and fluorescence images of the captured fluorescent magnetic beads in chips with “∼” shaped or hexagonal nickel powder@PDMS pillars. The scale bars are 100 μm.

Capture of magnetic bead–yeast cell complexes and generation of yeast cell arrays

Con A is a protein, assigned to the lectin family, which was obtained from the jack bean, Canavalia ensiformis.²⁴ It can also react with carbohydrates with high specificity and affinity, especially for α-D-mannosyl and α-D-glucosyl groups,^24,25 which have been found to exist abundantly in the cell envelope of Saccharomyces cerevisiae.^24,26 In this research, Con A modified magnetic beads were used to capture yeast in the magnetic field controllable microfluidic chip, generating yeast cell arrays. Fig. 6(a) shows a two-step coupling process of conjugating the Con A to the surfaces of the magnetic beads by using the EDC and sulfo-NHS. The detailed process of conjugating Con A to the magnetic beads and yeast cell culture is shown in the ESI.† The prepared yeast cells were re-suspended in a 1.5 mL centrifuge tube and mixed with the Con A modified magnetic beads for 30 min at room temperature with a slow rotation of 150 rpm. After generation of the yeast cell–magnetic bead complexes (Fig. 6(b)-(A)), the mixed sample was passed through the magnetic field controllable microfluidic chip with nickel powder@PDMS pillars for cell sorting and the unbound yeast cells were washed out of the microfluidic chip by 1× PBS (pH 7.4).

Fig. 6 (a) Modification of the carboxyl magnetic beads with Con A. (b) (A) Yeast cells were decorated with the Con A modified magnetic beads. (B) The yeast–magnetic bead complexes were captured at the areas around the nickel powder@PDMS pillars in a microfluidic chip. The scale bar in (A) is 10 μm and in (B) is 50 μm.

Fig. 6(b)-(B) shows that the yeast cell–magnetic bead complexes were captured in the microfluidic chip with nickel powder@PDMS pillars at magnetic field-1. It is obvious that the yeast cell–magnetic bead complexes were only captured around the nickel powder, at both sides, along the direction of the external magnetic field. This yeast cell–magnetic bead array could be useful for an environmental toxicity screening system.²⁷ In addition, this is only a model to prove that our approach could be used for cell capture and selection. We believe that this simple and robust method has great potential applications such as enrichment of circulating tumour cells (CTCs) or drug release systems in microfluidic chips.

Conclusions

In this paper, we reported a simple and robust approach to control the magnetic field distribution in microchannels by using nickel powder@PDMS pillars. A numerical simulation method and red fluorescent magnetic nanoparticle capture experiments were used to study the magnetic field distribution under two kinds of different external magnetic fields. These results were matched well and proved the validity of this method. In addition, yeast cell–magnetic bead complexes were captured around the nickel powder@PDMS pillars in this magnetic field controllable microfluidic chip. We believe that this approach has great potential in practical applications in cell sorting, immunoassays or drug release system in microfluidics.

Acknowledgements

This work was supported by the National Basic Research Program of China (973 Program, 2011CB933600), the 863 program (2013AA032204), the National Natural Science Foundation of China (21175100), and the Program for New Century Excellent Talents in University (NCET-10-0656).

Notes and references

1. L. H. Reddy, J. L. Arias, J. Nicolas and P. Couvreur, Chem. Rev., 2012, 112, 5818–5878.
2. E. Q. Song, J. Hu, C. Y. Wen, Z. Q. Tian, X. Yu, Z. L. Zhang, Y. B. Shi and D. W. Pang, ACS Nano, 2011, 5, 761–770.
3. X. Yu, H. S. Xia, Z. D. Sun, Y. Lin, K. Wang, J. Yu, H. Tang, D. W. Pang and Z. L. Zhang, Biosens. Bioelectron., 2013, 41, 129–136.
4. Y. H. Tennico, D. Hutanu, M. T. Koesdjojo, C. M. Bartel and V. T. Remcho, Anal. Chem., 2010, 82, 5591–5597.
5. N. Pamme, Lab Chip, 2006, 6, 24–38.
6. M. A. M. Gijs, F. Lacharme and U. Lehmann, Chem. Rev., 2010, 110, 1518–1563.
7. Y. J. Liu, S. S. Guo, Z. L. Zhang, W. H. Huang, D. M. Baigl, Y. Chen and D. W. Pang, J. Appl. Phys., 2007, 102, 084911.
8. R. Q. Zhang, S. L. Liu, W. Zhao, W. P. Zhang, X. Yu, Y. Li, A. J. Li, D. W. Pang and Z. L. Zhang, Anal. Chem., 2013, 85, 2645–2651.
9. A. Csordas, A. E. Gerdon, J. D. Adams, J. R. Qian, S. S. Oh, Y. Xiao and H. T. Soh, Angew. Chem., Int. Ed., 2010, 49, 355–358.
10. T. Deng, M. Prentiss and G. M. Whitesides, Appl. Phys. Lett., 2002, 80, 461–463.
11. Y. J. Liu, S. S. Guo, Z. L. Zhang, W. H. Huang, D. Baigl, M. Xie, Y. Chen and D. W. Pang, Electrophoresis, 2007, 28, 4713–4722.
12. J. D. Adams, U. Kim and H. T. Soh, Proc. Natl. Acad. Sci. U. S. A., 2008, 105, 18165–18170.
13. Y. L. Liu, J. D. Adams, K. Turner, F. V. Cochran, S. S. Gambhir and H. T. Soh, Lab Chip, 2009, 9, 3604.
14. X. Yu, X. Feng, J. Hu, Z. L. Zhang and D. W. Pang, Langmuir, 2011, 27, 5147–5156.
15. J. Hu, M. Xie, C. Y. Wen, Z. L. Zhang, H. Y. Xie, A. A. Liu, Y. Y. Chen, S. M. Zhou and D. W. Pang, Biomaterials, 2011, 32, 1177–1184.
16. H. Y. Xie, M. Xie, Z. L. Zhang, Y. M. Long, X. Liu, M. L. Tang, D. W. Pang, Z. Tan, C. Dickinson and W. Zhou, Bioconjugate Chem., 2007, 18, 1749–1755.
17. H. Y. Xie, C. Zuo, Y. Liu, Z. L. Zhang, D. W. Pang, X. L. Li, J. P. Gong, C. Dickinson and W. Z. Zhou, Small, 2005, 1, 506–509.
18. L. L. Huang, P. Zhou, H. Z. Wang, R. Zhang, J. Hao, H. Y. Xie and Z. K. He, Chem. Commun., 2012, 48, 2424–2426.
19. G. P. Wang, E. Q. Song, H. Y. Xie, Z. L. Zhang, Z. Q. Tian, C. Zuo, D. W. Pang, D. C. Wu and Y. B. Shi, Chem. Commun., 2005, 4276–4278.
20. E. Q. Song, G. P. Wang, H. Y. Xie, Z. L. Zhang, J. Hu, J. Peng, D. C. Wu, Y. B. Shi and D. W. Pang, Clin. Chem., 2007, 53, 2177–2185.
21. D. C. Duffy, J. C. McDonald, O. J. A. Schueller and G. M. Whitesides, Anal. Chem., 1998, 70, 4974–4984.
22. A. L. Gassner, M. Abonnenc, H. X. Chen, J. Morandini, J. Josserand, J. S. Rossier, J. M. Busnel and H. H. Girault, Lab Chip, 2009, 9, 2356–2363.
23. J. R. Qian, X. H. Lou, Y. T. Zhang, Y. Xiao and H. T. Soh, Anal. Chem., 2009, 81, 5490–5495.
24. M. Donolato, A. Torti, N. Kostesha, M. Deryabina, E. Sogne, P. Vavassori, M. F. Hansen and R. Bertacco, Lab Chip, 2011, 11, 2976–2983.
25. A. Touhami, B. Hoffmann, A. Vasella, F. A. Denis and Y. F. Dufrene, Microbiology, 2003, 149, 2873–2878.
26. R. C. Montijn, J. van Rinsum, F. A. van Schagen and F. M. Klis, J. Biol. Chem., 1994, 269, 19338–19342.
27. J. Garcia-Alonso, R. F. Fakhrullin, V. N. Paunov, Z. Shen, J. D. Hardege, N. Pamme, S. J. Haswell and G. M. Greenway, Anal. Bioanal. Chem., 2011, 400, 1009–1013.

---

Footnote

† Electronic supplementary information (ESI) available: Culture of yeast cells. Conjugating Con A to the magnetic beads. Optical imaging of the nickel powder@pillar arrays. Simulation of the magnetic field distribution in the microchannels under magnetic field-1 and magnetic field-2. Numerical simulation of |(B·∇)B| under magnetic field-2. Control of the magnetic field distribution by nickel powder@PDMS pillars. See DOI: 10.1039/c3ra47902a

---