Lab on a Chip Self-assembled magnetic bead chains for sensitivity enhancement of microfluidic electrochemical biosensor platforms †

In this paper, we present a novel approach to enhance the sensitivity of microfluidic biosensor platforms with self-assembled magnetic bead chains. An adjustable, more than 5-fold sensitivity enhancement is achieved by introducing a magnetic field gradient along a microfluidic channel by means of a soft-magnetic lattice with a 350 μ m spacing. The alternating magnetic field induces the self-assembly of the magnetic beads in chains or clusters and thus improves the perfusion and active contact between the analyte and the beads. The soft-magnetic lattices can be applied independent of the channel geometry or chip material to any microfluidic biosensing platform. At the same time, the bead-based approach achieves chip reusability and shortened measurement times. The bead chain properties and the maximum flow velocity for bead retention were validated by optical microscopy in a glass capillary. The magnetic actuation system was successfully validated with a biotin – streptavidin model assay on a low-cost electrochemical microfluidic chip, fabricated by dry-film photoresist technology (DFR). Labelling with glucose oxidase (GOx) per-mits rapid electrochemical detection of enzymatically produced H 2 O 2 . a four-channel bipotentiostat detects enzy- matically produced H O m and 500 μ m lattice spacings show a 1.7-fold increase in sensitivity in comparison to the monopolar setup. Due to the small thickness of the EDM-cut lattice, the low-cost laminated lattices show bet- ter performance.


Introduction
Since the 1990s, microfluidic biosensor systems utilizing solid micro-particles (beads) became popular.
1,2 Microfluidic systems benefit from small dimensions and reaction volumes, thus enhancing the time-to-result and lowering the analyte volume at high sensitivities. 3,4Besides polyimide, silica or agarbased beads, super-paramagnetic particles composed of magnetic nanoparticles, embedded in a polymeric matrix have become a common analytical tool in biological labs. 5icrofluidic biosensor systems largely benefit from these spherical magnetic microparticles. 6,7The possibility of magnetic manipulation, combined with their high surface-tovolume ratio enables fabrication of re-usable, highly sensitive on-chip biosensors for various applications. 8,9Additionally, magnetic bead actuation is utilized for biosensing as well as cell separation in platforms that use droplet microfluidic approaches. 10n the last 15 years, significant progress in the manipulation of magnetic beads on-chip has been achieved. 11In order to realize highly sensitive bead-based microfluidic biosensors, the active contact between beads and sample has to be maximized.3][14][15][16] Bead-based immunoassay platforms with varying channel cross-sections in combination with permanent magnets have been presented. 17Integrated electromagnets and highly complex dynamic systems linking external electromagnets and permanent magnets have also been proposed. 18,19The active manipulation of magnetic beads inside microfluidic channels offers new chances for the performance of a large variety of reliable, sensitive and low cost biological tests.Nevertheless, the complexity of the manipulation system and/or the requirements for the chip design of the aforementioned manipulation techniques still prevent from their widespread use.Enzyme-linked immunosorbent assays (ELISA) are still the most sensitive, specific and selective measurement method for the qualitative and quantitative detection of biomarkers. 202][23][24] In this work, we mimic these biological binding assays with a biotin-streptavidin model assay, labelled with glucose oxidase.Detection of biological assays is commonly done via optic absorbance or chemiluminescence measurements.

25,26
In contrast to these techniques, electrochemical detection of enzymatically produced H 2 O 2 takes advantage from fast read-out due to the intrinsic enzymatic amplification, independence of the optical chip properties and easy miniaturization of chip and measurement setup.Therefore, electrochemical detection is well suited for point of care (POC) diagnostic systems.This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins Microfluidic electrochemical biosensors, fabricated in DFR technology, benefit from low production costs and possible mass fabrication on roll-to-roll processes.They have already been proposed by Horak et al. as disposable sensors for the detection of different analytes. 28,29In this work, the previous DFR based sensor platform is extended to a bead-based sensor platform.It combines DFR technology with a new costefficient, small and universally applicable magnetic actuation principle.This is based on a soft-magnetic lattice, fabricated by lamination of soft-magnetic and non-magnetic foils into a stack.Combined with permanent magnets, these structures create a magnetic field gradient along the microfluidic channel in the underlying microfluidic chip.This leads to self-assembly of magnetic beads in chain like structures with improved contact between the beads and the solution.In comparison to other techniques this allows low-cost mass fabrication of the manipulation device without any need for complex machinery.][19] We optically observed the influence of the soft-magnetic lattice application on the bead distribution in the microchannel as well as the bead retention under flow conditions.Electrochemical validation was performed on the redesigned DFR based biosensor chip, utilizing a custom-made measurement adapter.A more detailed description of the DFR technology and the characterization of the biosensor chip is given in the Electronic Supplementary information (ESI).

Experimental Fabrication of soft-magnetic lattices
Fig. 1a illustrates the fabrication of soft-magnetic lattices in lamination technique.Mu-metal foil (Vacuumschmelze GmbH & Co. KG, Germany) and double-sided sticky tape (Tesa, Germany) in thicknesses of 100, 350 and 500 µm are laminated to a multilayer stack of 13 mm.Subsequent embedment in a polymeric matrix ensures mechanical stability and a final grinding step releases the soft-magnetic lattice structure. .The investigated lattice spacings are a tradeoff between high magnetic field gradients for bead capture and small distances between bead chains improving the diffusion of analyte towards the beads.Detailed information thereof is given in the ESI.
In order to compare the lamination approach to existing structuring methods of soft-magnetic materials, an additional lattice in 350 µm spacing was fabricated by electrochemical discharge machining (EDM) out of 100 µm thick Mu-metal foil.
The fabricated devices are shown in Fig. 1b.

Actuation of magnetic beads
The magnetic force that acts on an individual magnetic bead is defined by the bead properties and the applied external magnetic field.In the saturation region, the force is given by 30 in which ݉ ሬሬሬԦ ௦௧ is the magnetization of the bead in saturation and ‫ܤ‬ ሬ ሬԦ the induced magnetic flux density.As seen in eqn.(1), for a given particle magnetization, the magnetic force is directly dependent on the gradient of the magnetic field.The viscous drag is counteracting this force and, assuming fluid flow only in the x-direction, the maximal flow velocity for bead retention is consequently given by 30 where ݉ ௦௧ is the beads saturation magnetization in xdirection, ߟ the dynamic viscosity of the medium, ‫ݎ‬ ௗ the particles radius and ‫ݔ݀/ܤ݀‬ the magnetic field gradient in xdirection.

Setup for optical investigations
For real-time monitoring of magnetic beads under flow conditions and optical visualization of bead chains, a setup out of a transparent glass capillary (Hilgenberg GmbH, Germany; I.D.: 80 µm; O.D.: 120 µm), a syringe pump (PHD2000 Harvard Aparatus, USA; syringes from Hamilton Company, USA) and the magnetic actuation as illustrated in Fig. 2a was used.The

Please do not adjust margins
Please do not adjust margins magnetic actuation system was located perpendicular to the glass capillary enabling microscopy from above.To allow easy disassembly of the capillary from the magnetic actuation system, the two 20 x 10 x 4 mm³ sized NdFeB permanent magnets as well as the soft-magnetic structure were fixed on a transparent 4 mm thick PMMA base.The glass capillary that resembles the microfluidic channel was glued onto a second, independent plate (PMMA top).Before the magnetic parts were attached, 5 µl of solution with the desired concentration of monodisperse 1.43 µm streptavidin coated magnetic beads were introduced into the capillary.The PMMA top was subsequently mounted onto its counterpart for magnetic actuation.Deionized water was withdrawn through the micro-capillary at flow rates of up to 10 mm sec -1 with the previously mentioned syringe pump.
Pictures of the bead chain formation were taken at different bead concentrations and flow velocities for the 100 µm, 350 µm and 500 µm laminated lattices.

Electrochemical measurements on-chip
As illustrated in Fig. 3c, the the measurement cell consists of a Ag/AgCl reference electrode and two Pt electrodes used as working and counter electrode in amperometric measurements.The working electrode detects enzymatically produced H 2 O 2 at 450 mV vs. Ag/AgCl.Glucose oxidase, producing H 2 O 2 from a glucose solution, was used as an enzymatic label.All solutions were prepared in ultra-pure, HPLC grade water (Alpha Aeser, USA).After introduction of the beads, 100 mM phosphate buffered saline (PBS), 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO4, 2 mM KH 2 PO 4 , at pH = 7.4 was withdrawn through the microfluidic channel at a mean flow velocity of 1.5 mm sec -1 (flow rate of 2.88 µl min -1 ).When the amperometric signal This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins reached a stable steady-state current (baseline), the 40 mM glucose substrate, diluted in 100 mM PBS (pH = 7.4), was introduced and stop-flow measurements were performed.In the stop phase, H 2 O 2 accumulates in the channel and is subsequently flushed over the working electrode when the flow is restarted.The generated current peak is visualized with the software BioMON (Jobst technologies GmbH, Germany) on a computer.

Biotinylated GOx model assay
Streptavidin-coated magnetic beads (Microparticles GmbH, Germany) with a diameter of 1.43 µm were used for all experiments.The biotinylated GOx (bGOx) model assay takes advantage of the strong affinity of biotin to streptavidin. 29hen getting in close proximity, bGOx binds to the streptavidin-coated magnetic beads.Dilution of the magnetic beads was done by diluting the stock solution (10 mg ml -1 bead-content) 10-fold and 100-fold in 0.1 M PBS.To experimentally validate the functionality of the fabricated soft-magnetic lattice structures, the appropriate bead dilution was washed with 10 mM phys.PBS (pH = 7.4) and incubated for 30 min with up to 1 µg ml -1 bGOx.After subsequent washing with wash buffer (phys.PBS with 1% Tween 20) on a magnetic holder (Dynal technologies, Norway), 3.5 µl of the solution were introduced to the biosensors by capillary filling and the chips were mounted onto the measurement adapter.To prevent from unspecific adsorption during on-chip immobilization-tests, the chips were blocked for 1 h with 1% bovine serum albumin (BSA) in phys.PBS and washed with 200 µl of wash buffer.Then, the 0.1 mg ml -1 bead dilution was introduced to the micro-capillary with the syringe pump, incubated with 200 ng ml -1 bGOx on-chip and washed by withdrawal of 20 µl wash buffer previously to measurements.

Simulation
Simulation of the magnetic flux inside the microfluidic channel was done with the open source program FEMM (D. C. Meeker, Finite Element Method Magnetics, Version 4.0.1) in a distance of 100 µm from the soft-magnetic lattice.This value represents the distance between the DFR cover layer and the center of the microfluidic channel in the biosensor chip.Employing 10 x 20 x 4 mm³ permanent neodymium-iron-boron (NeFeB) magnets (Webcraft GmbH, Germany) and a 350 µm spaced laminated soft-magnetic lattice, the 2-D FEM simulations show formation of a magnetic field gradient in the microfluidic channel (Fig. 4).The utilized setup consisted of a permanent magnet underneath the biosensor chip and the lattice structure as well as a second permanent magnet applied to the channel cover.According to eqn. 2, capture of 1.43 µm sized magnetic beads (Microparticles GmbH, Germany) takes place at the local maxima of the magnetic flux gradient.Hence, the shown combination of lattice and magnets retents the 1.43 µm sized magnetic beads up to a flow velocity of 3.44 mm sec -1 .

Lamination of soft-magnetic lattices
The fabrication of soft-magnetic lattices by lamination proved to be fast and facile.Lattice spacing down to 100 µm were fabricated without any noticeable issues.The advantages of this technology are low costs, high structure thicknesses, fast processing and huge flexibility in design compared to other suitable techniques like EDM, etching, puls-lasering, etc.Additionally, no external equipment is necessary during fabrication and no heat is introduced.

Structures and properties of the bead formation
Fig. 5a shows, that application of the soft-magnetic lattices results in the formation of magnetic bead clusters.These consist of multiple bead chains with the distance between individual clusters being defined by the lattice spacing.The bead structures tend to accumulate at the top of the channel due to the enlarged distance to the permanent magnet on this side of the micro-capillary in the optical setup.However, this effect is averted on the electrochemical measurement adapter since the magnetic parts are directly applied to the channel resulting in a highly symmetric setup.Figure 5b, c and d represent the self-assembled beadstructures at bead concentrations of 1 mg ml -1 , 0.1 mg ml -1 and 0.01 mg ml -1 .The higher the bead concentration, the larger the bead clusters and the better is the channel perfusion.The resistance of these bead structures against fluid flow was investigated using a 350 µm laminated lattice.Deionized water as substrate was supplied and the mean flow velocity was increased stepwise.On-line visualization enabled determination of the maximum flow velocity for bead retention.Fig. 5e and f show that bead chains withstand flow velocities up to 3 mm sec -1 and cover a significantly higher fraction of the channel cross-section than in the monopolar setup (Fig. 5g).Furthermore, these optical results match with the analytical estimations based on the This journal is © The Royal Society of Chemistry 20xx J. Name., 2013, 00, 1-3 | 5 Please do not adjust margins Please do not adjust margins magnetic simulations, yielding a theoretical maximum flow velocity of 3.44 mm sec -1 .

Characterization of the electrochemical biosensor platform
Wafer-level fabrication of the DFR biosensor chips was followed by their characterization.A detailed description is given in the ESI.It was found that the biosensor chips have a linear and reproducible H 2 O 2 -sensitivity of 6 nA mm -2 µM -1 at a flow velocity of 1.5 mm sec -1 .Washing with wash buffer verified the ability to remove almost all beads from the microfluidic channel.Therefore, re-usability of the chips is also possible.
Stop flow measurements produce a detectable signal peak resulting from the accumulation of enzymatically produced H 2 O 2 in the microfluidic channel.It was proven, that the peak charge and peak current in stop-flow measurements are proportional.Thus, the peak current can be taken as the measurement signal.

Electrochemical measurements with pre-immobilized beads
After characterizing the sensor platform, evaluation of the magnetic actuation with soft-magnetic lattices had to be done.Electrochemical measurements were performed with beads that were pre-immobilized with different bGOx concentrations in standard lab tubes.Fig. 6 shows the results of the measured continuous flow signal (baseline), the 2 min.and the 5 min.stop-flow peak current densities after introduction of 3.5 µl of bead solution into the microfluidic capillary and subsequent magnetic actuation.The different magnetic actuation setups were experimentally tested with a bead-concentration of 0.1 mg ml -1 .A significant increase in the signal was seen, when soft-magnetic lattices were applied.Lamination was found to be the superior technology for lattice fabrication compared to EDM, even though EDM preserves higher surface qualities.The reason is the enormous lattice thickness of the laminated devices (several mm) that cannot be achieved by EDM or other high precision fabrication techniques.The 100 µm lattice spacing showed significantly less signal compared to the 350 µm and 500 µm spacings.This is related to the distance between beads and soft-magnetic lattice, which is in the same size as the lattice spacing.This leads to a decrease in the amplitude of the magnetic field gradient in the microfluidic channel.Hence, bead capture and bead-chain formation cannot be guaranteed.Different concentrations of bGOx were subsequently applied to investigate the effect of the lattice-application over a concentration range from 0.32 ng ml -1 to 1 µg ml -1 of bGOx during incubation with magnetic beads in lab tubes.Fig. 7a shows the 5 min.stop-flow peak current densities of the aforementioned measurements.It was seen, that the 350 and 500 µm laminated soft-magnetic lattices were able to enhance the electrochemical signal of pre-immobilized beads by a factor of 1.7 at a three times lower standard deviation.This is explained by the formation of magnetic bead chains and clusters leading to a more homogeneous bead distribution in the channel and improved contact between beads and solution.This effect was seen over the whole concentration range and for two different bead concentrations of 1 and 0.1 mg ml -1 .Besides, it was found that the sensitivity of the platform could be tuned by increasing the introduced bead amount.Nevertheless, an increased bead concentration This journal is © The Royal Society of Chemistry 20xx Please do not adjust margins Please do not adjust margins without allocation of a lattice structure also leads to higher measurement errors.This can be explained by the accumulation of beads at the end of the permanent magnet in the monopolar setup.The limited space at this location results in a loss of beads that are forced in larger distances to the magnet and can thus no longer withstand the flow.Microtiter-plates are a commonly used and well established technique for the detection of biological assays.Therefore, pre-immobilized beads were measured on microtiter-plate format and on the magnetic bead-based electrochemical biosensor platform.In contrast to the chip platform, the microtiter-plate format is limited to bead concentrations of up to 0.1 mg ml -1 due to the optical properties of the beads caused by their high iron oxide content.For comparability, all tests shown in Fig. 7b were done with a bead concentration of 0.1 mg ml -1 .It was seen, that our platform shows the same signal characteristics as the microtiter-plate format.On both formats, the limit of detection (LOD) was 8 ng ml -1 .
Additionally, the measurement duration on-chip was reduced to 2 min on the electrochemical biosensor system.Furthermore, only 3.5 µl of bead solution and analyte volume are necessary, compared to 100 µl used on the microtiterplate format.

Electrochemical measurements with on-chip immobilization
To prove the efficiency of the increased active contact between beads and analyte under the application of softmagnetic lattices, on-chip immobilization of beads was performed.Therefore, 10 µl of the streptavidin-coated magnetic beads were introduced into the microfluidic channel of the biosensor chip with the syringe pump.Then, the magnetic actuation system was applied and 10 µl of analyte were withdrawn through the channel at a mean flow velocity of 1.5 mm sec -1 .When the analyte was located at the immobilization area, flow was stopped for a 15 min incubation step.After 5 min washing with 0.1 M PBS, the 40 mM glucose substrate was supplied at a mean flow velocity of 1.5 mm sec -1 and electrochemical stop-flow measurements were conducted.Fig. 8 compares the generated signals for a bGOx concentration of 200 ng ml -1 incubated with and without application of the soft-magnetic lattice.It was seen, that the increase in platform sensitivity evoked by lattice-application is higher than the 1.7-fold increase in case of pre-immobilized beads.The reason is the improved analyte capture in the channel due to the bead chain formation, which cannot be investigated with pre-immobilized beads.A more than 5-fold amplification was achieved with the 350 µm laminated softmagnetic lattice compared to the monopolar setup.

Conclusion
In this work, we presented a magnetic bead-based biosensor platform, which combines DFR technology together with softmagnetic lattices.The soft-magnetic lattices are produced by lamination of soft-magnetic and non-magnetic foils.Lamination as fabrication technique for soft-magnetic lattices benefits from low costs and allows the realization of lattice thicknesses in the mm-to cm-range.Nevertheless, the lower limit in the lattice spacing is defined by the distance between beads and lattice structure (chip cover).Optical investigation of the self-alignment of magnetic beads in chains and clusters This journal is © The Royal Society of Chemistry 20xx J. Name., 2013, 00, 1-3 | 7 Please do not adjust margins Please do not adjust margins revealed possible bead retention up to flow velocities of 3 mm sec -1 in deionized water.Furthermore, it was verified that the distance in between and the size of bead structures is tunable by the amount of introduced beads and the spacing of the soft-magnetic lattice.
Compared to the monopolar setup, the application of laminated soft-magnetic lattices with a 350 µm lattice spacing shows a more than 5-fold increase in the current signal for onchip immobilization tests using a bGOx model assay on magnetic beads.Even for pre-immobilized beads, the formation of magnetic bead chains and clusters resulted in an almost 2-fold signal enhancement over the whole measured concentration range of bGOx.Laminated soft-magnetic lattices as a magnetic actuation method can be applied to a huge variaty of magnetic beadbased microfluidic biosensor platform and represent a cheap and facile solution to significantly increase the sensitivity of such platforms.
AP 8525R (DuPont TM , USA) in 50 µm thickness is used as the substrate for wafer-level processing of the biosensor chip.Photoresist ma-N 1420 (Micro Resist Technology GmbH, Germany) is spin-coated onto the PI-substrate and structured.A 200 nm platinum layer was then evaporated onto the substrate under cleanroom conditions.Metal patterns were released in a subsequent lift-off-process.The next steps were spin coating and lithographic structuring of a 5 µm thick SU-8 3005 (MicroChem Corp. USA) isolation layer, defining the electrode sizes and an additional well for the stop barrier.The microfluidic channels were subsequently realized with the DFR material Pyralux® PC1025 (DuPont TM , USA) in a thickness of 64 µm.Then the Ag/AgCl reference electrode is fabricated by electrodeposition of silver, followed by a chloridation step.The physical barrier is realized by dispensing a small drop of 1% Teflon (DuPont TM , USA) solution into the predefined wells in the SU-8 isolation layer of the chip.After evaporation of the solvent a thin Teflon layer is formed, that is sufficient to act as a flow barrier during capillary filling.Finally, the cover-layer with the fluidic inlet and outlet ports are realized in an additional DFR layer.To prevent from bending effects due to thermal expansion effects during baking, symmetry was achieved by laminating two layers of DFR on the backside of the wafer.The chips are thereafter cut into duplicates and backed in the oven for 3 hours at 160 °C.The resulting flexible biosensor chips (Fig.3a) have a size of 25.65 x 10.20 mm².The working electrodes size is 0.311 mm², the channel width 500 µm and the microfluidic channel up to the stop barrier, the bead capture region, holds approximately 3.2 µl of analyte.A comparable processing based on a different DFR material has already been shown by Horak et al.24Magnetic, electrochemical and microfluidic setupThe sensor is mounted on a custom-made adapter prior to measurements.Fig.3bpresents this measurement adapter for electric and fluidic connection to the periphery as well as magnetic actuation.Fast and reliable electrical connection to a four channel bi-potentiostat (Jobst technologies GmbH, Germany) is realized.The setup enables facile chip connection and parallel measurement of up to four biosensor chips.The fluidic connection is established by screwing a PMMAconnector onto the PMMA base.In contrast to pressure seals, the employed 0.125 " vacuum cups (Nordson EFD, Germany) are more flexible and allow tight sealing without the risk of channel compression or pinch off.