Developing a MEMS Device with Built-in Microfluidics for ... - MDPI

micromachines Article

Developing a MEMS Device with Built-in Microfluidics for Biophysical Single Cell Characterization Yuki Takayama 1,2,3 , Grégoire Perret 1,3 , Momoko Kumemura 4,5 , Manabu Ataka 3,5 , Samuel Meignan 6,7 , Stanislav L. Karsten 8 ID , Hiroyuki Fujita 3,5 , Dominique Collard 1,3, *, Chann Lagadec 3,7 ID and Mehmet Cagatay Tarhan 1,2,3, * 1

2 3

4 5 6 7 8

*

Laboratory for Integrated Micro Mechatronic Systems (LIMMS/CNRS-IIS), Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan; [email protected] (Y.T.); [email protected] (G.P.) Univ. Lille, CNRS, Centrale Lille, ISEN, Univ. Valenciennes, UMR 8520-IEMN, 59652 Villeneuve d’Ascq, France CNRS/IIS/COL/Lille University SMMiL-E Project, CNRS Délégation Nord-Pas de Calais et Picardie, 2 rue de Canonniers, Lille, Cedex 59046, France; [email protected] (M.A.); [email protected] (H.F.); [email protected] (C.L.) Graduate School of Life Science and Systems Engineering, Kyushu Institute of Technology, 2-4 Hibikino, Wakamatsu-ku, Kitakyushu-shi, Fukuoka 808-0196, Japan; [email protected] Centre for Interdisciplinary Research on Micro-Nano Methods, Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba, Meguro-ku, Tokyo 153-8505, Japan Tumorigenesis and Resistance to Treatment Unit, Centre Oscar Lambret, Université de Lille, 3 rue Frédéric Combemale, 59000 Lille, France; [email protected] INSERM U908 Laboratory, Lille University—Science and Technologies, Building SN3, 59655 Villeneuve d’Ascq, France NeuroInDx, Inc., 20725 S Western Ave #100, Torrance, CA 90501, USA; [email protected] Correspondence: [email protected] (D.C.); [email protected] (M.C.T.); Tel.: +33-3-20-29-55-52 (D.C.)

Received: 4 May 2018; Accepted: 29 May 2018; Published: 1 June 2018







Abstract: This study combines the high-throughput capabilities of microfluidics with the sensitive measurements of microelectromechanical systems (MEMS) technology to perform biophysical characterization of circulating cells for diagnostic purposes. The proposed device includes a built-in microchannel that is probed by two opposing tips performing compression and sensing separately. Mechanical displacement of the compressing tip (up to a maximum of 14 µm) and the sensing tip (with a quality factor of 8.9) are provided by two separate comb-drive actuators, and sensing is performed with a capacitive displacement sensor. The device is designed and developed for simultaneous electrical and mechanical measurements. As the device is capable of exchanging the liquid inside the channel, different solutions were tested consecutively. The performance of the device was evaluated by introducing varying concentrations of glucose (from 0.55 mM (0.1%) to 55.5 mM (10%)) and NaCl (from 0.1 mM to 10 mM) solutions in the microchannel and by monitoring changes in the mechanical and electrical properties. Moreover, we demonstrated biological sample handling by capturing single cancer cells. These results show three important capabilities of the proposed device: mechanical measurements, electrical measurements, and biological sample handling. Combined in one device, these features allow for high-throughput multi-parameter characterization of single cells. Keywords: single cell analysis; biophysical cell characterization; bioMEMS; microfluidics; MEMS design

Micromachines 2018, 9, 275; doi:10.3390/mi9060275

www.mdpi.com/journal/micromachines

Micromachines 2018, 9, 275

2 of 14

1. Introduction Cells in a sample solution can be largely heterogeneous, consisting of various different sub-populations, even if they originate from the same tumor or cell line. Immunochemistry has been an effective means to investigate sub-populations within a cell solution especially based on specific membrane protein expressions, designated as “cluster of differentiation” (CD). Highly specific immunoaffinity-based assays represent a powerful analysis method, especially with the development of high-throughput cell cytometry. However, the use of biomarkers is still an important limitation due to practical reasons, such as availability of surface markers and corresponding antibodies and heterogeneity between specimens regarding combined marker expressions. Therefore, routine clinical tests for early diagnosis would benefit from cost and time-efficient alternatives. Cell shape and structural integrity significantly influence many biological processes. Therefore, the physical properties of cells may potentially be used to reflect the state of their health [1]. This connection between biophysics and disease has been attracting scientific research attention, especially for oncological studies [2–7] where diseased cells proliferate “uncontrollably” and disrupt the organization of tissue. When the physical properties of a cell change, the behavior of the cell, e.g., the way they spread, changes as well [8]. In human cancer cell lines, invasive cells exhibit biophysical properties that are distinct from their noninvasive counterparts with a reduction in stiffness and an increase in metastatic efficiency [2–4,9–11], especially in the case of circulating tumor cells (CTCs). Therefore, the biophysical properties of cells can be considered as biomarkers to distinguish cells and thereby used as practical label-free indicators for routine clinical examinations targeting early disease diagnosis. Numerous methods have been used to examine the biophysical properties of single cells: biomechanical assays to probe cell components [3,12,13] or single cell deformation [10,14–16], cytoadherence assays [8], and microfluidics-based assays for mechanical properties [5,17] or for electrical properties [18]. As the majority of these methods can be applied solely for the study of adherent cells, suitable options to analyze non-adherent cells, such as CTCs, are limited. Moreover, there is a trade-off between throughput and information content [7]. Certain techniques such as atomic force microscopy (AFM) allow very sensitive measurements but suffer from low-throughput, while microfluidics-based methods like deformability cytometry provide rapid measurements but with limited information content. However, using biophysical properties for practical routine tests to distinguish cells in a sample solution requires both high-throughput and multi-parameter analysis. Three main features necessary for the targeted practical system are (i) the high-throughput handling of non-adherent cells without compromising (ii) the detection sensitivity for (iii) a multi-parameter analysis. Microfluidics technology is well-suited to handle non-adherent cells and such capabilities have already been demonstrated with high-throughput techniques [5,17]. For a practical yet still highly sensitive multi-parameter detection method, on the other hand, microelectromechanical systems (MEMS) technology provides low-noise electrical detection while providing mechanical and electrical stimulation of cells. Integrating MEMS with microfluidics enables the handling and/or analysis of biological samples such as microtubules [19], DNA [20], and cells [21], even for clinical purposes [22]. MEMS devices with built-in microfluidics simplify biophysical characterization and improve throughput, even though existing examples use external actuation and optical detection [23,24]. To achieve the optimal use of MEMS, sensitive harmonic analysis can be combined with practical electrostatic actuation. However, to reach the necessary sensitivity, the MEMS elements have to be operated in air while the cell characterization is performed in liquid. This study develops a MEMS device with an embedded microfluidic channel to perform multi-parameter biophysical (mechanical and electrical) characterization. Electrostatic actuation allows both single cell stimulation and harmonic analysis for sensitive real-time measurements due to a specific device design, separating the in-liquid biological sample handling from in-air MEMS actions (actuation and detection). Here, we introduce the device features, perform real-time measurements (mechanical and electrical), and demonstrate a proof of concept for single cell handling. This device will

Micromachines 2018, 9, 275

3 of 14

Micromachines 2018, 9, x

3 of 14

allow high-throughput characterization of individual cells in complex biological samples. This could couldpotential open potential tool alternatives routine clinical tests for early disease open tool alternatives for routinefor clinical tests for early disease diagnosis withdiagnosis improvedwith cost improved cost and time efficiency. and time efficiency. 2. Concept and Design 2.1. Principle 2.1. Principle The device consists consists of of three three main main components components (Figure (Figure 1): 1): The proposed proposed single single cell cell characterization characterization device (1) microfluidicchannel channelintegrated integratedwith with two opposing to handle a series of comb (1) aa microfluidic two opposing tipstips to handle cells,cells, (2) a (2) series of comb drive drive actuators to perform compression, and (3) a differential capacitive sensor for displacement actuators to perform compression, and (3) a differential capacitive sensor for displacement measurements. This sensor sensor is is connected connected to to aa second measurements. This second electrostatic electrostatic comb comb drive drive actuator actuator for for harmonic harmonic measurements. A flow is created in the microfluidic channel (shown in gray) with a vacuum measurements. A flow is created in the microfluidic channel (shown in gray) with a vacuum pump pump connected to the connected to the outlet outlet after after introducing introducing aa cell cell solution solution via via the the inlet. inlet. Adjusting Adjusting the the vacuum vacuum pump pump pressure of the single cell the characterization pressure enables enablescontrol control of flow the rate, flowwhich rate,allows whichforallows for positioning single cellat positioning at the area between tips. One tip, connected to an electrostatic actuator for displacement (shown in red), characterization area between tips. One tip, connected to an electrostatic actuator for displacement is used for compressing cells, whereas the other (shown in blue) is used for sensing the mechanical (shown in red), is used for compressing cells, whereas the other (shown in blue) is used for sensing properties of cells under various strain values with a displacement based on the mechanical properties of cellscompressive under various compressive strain values withsensor a displacement asensor differential capacitor (shown in light blue). After being introduced via the inlet, a cell is positioned based on a differential capacitor (shown in light blue). After being introduced via the inlet,ata the characterization tips to perform cell mechanical/electrical characterization. cell is positioned area at for theopposing characterization area single for opposing tips to perform single cell By controlling the flow rate, the captured cell can then be washed away to receive the mechanical/electrical characterization. By controlling the flow rate, the captured cell cannext thencell be for characterization. washed away to receive the next cell for characterization.

Figure 1. (a) Schematic close-up view of of thethe tips at Schematic image imageof ofthe thedevice devicewith withembedded embeddedchannel, channel,(b) (b)AA close-up view tips the handling area. at the handling area.

2.2. Design Description The device is composed of a MEMS layer and a polydimethylsiloxane (PDMS) layer. The MEMS fabricatedonon a silicon-on-insulator The top of silicon of the wafer the layer isisfabricated a silicon-on-insulator (SOI)(SOI) wafer.wafer. The top silicon the wafer forms the forms functional functional part of the proposed device (Figure 2a). In addition to being the handling layer of the part of the proposed device (Figure 2a). In addition to being the handling layer of the device, device, the bottom silicon hasfunctions: two other(i) functions: (i)mechanical providing mechanical connection between bottom silicon has two other providing connection between electrically electrically isolated elements of the sensing part and (ii) completing the microfluidic channel together isolated elements of the sensing part and (ii) completing the microfluidic channel together with the with the PDMS layer2b). (Figure The functionality of theisdevice is achieved the combination PDMS layer (Figure The 2b). functionality of the device achieved through through the combination of three of three areas. distinct areas. distinct

Micromachines 2018, 9, 275

4 of 14

Micromachines 2018, 9, x

4 of 14

Figure 2. 2. (a) image of of thethe handling areaarea where compressing and Figure (a) Scanning Scanningelectron electronmicroscope microscope(SEM) (SEM) image handling where compressing sensing tips access the microfluidic channel, (b) Schematic view of polydimethylsiloxane (PDMS) and sensing tips access the microfluidic channel, (b) Schematic view of polydimethylsiloxane (PDMS) coveralignment. alignment. cover

2.2.1. Handling Area and Microfluidic Channel 2.2.1. Handling Area and Microfluidic Channel The handling area consists of a channel formed by the frontside silicon (30 µm) of the SOI wafer The handling area consists of a channel formed by the frontside silicon (30 µm) of the SOI wafer as sidewalls, the backside silicon of the SOI wafer as the bottom surface, and the PDMS slab as the as sidewalls, the backside silicon of the SOI wafer as the bottom surface, and the PDMS slab as the top top surface. The width of the channel is 100 µm with a narrower portion of 20 µm between the surface. The width of the channel is 100 µm with a narrower portion of 20 µm between the actuation actuation and sensing tips (Figure 2a). The handling channel is accessed by four openings: an inlet and sensing tips (Figure 2a). The handling channel is accessed by four openings: an inlet and an outlet and an outlet in the PDMS slab to handle the liquid in the channel (Figure 2b) and two side openings in the PDMS slab to handle the liquid in the channel (Figure 2b) and two side openings at the sidewalls at the sidewalls for the compressing and sensing tips. As both tips need to be moveable, the PDMS for the compressing and sensing tips. As both tips need to be moveable, the PDMS surface above the surface above the characterization area is 10 µm above the top of the sidewalls, preventing any characterization area is 10 µm above the top of the sidewalls, preventing any contact with the actuating contact with the actuating elements (Figure S1). The tips are free to move after the oxide layer between elements (Figure S1). The tips are free to move after the oxide layer between the top and bottom silicon the top and bottom silicon of the SOI wafer is removed (see fabrication). The width of the elevated of the SOI wafer is removed (see fabrication). The width of the elevated PDMS area is designed as PDMS area is designed as 500 µm to allow for simple alignment of the two layers. 500 µm to allow for simple alignment of the two layers. The inlet and outlet openings in the PDMS are formed using biopsy punchers (1 mm for the inlet The inlet and outlet openings in the PDMS are formed using biopsy punchers (1 mm for the and 0.5 mm for the outlet) and thus, do not require any specific designs. Injection to the inlet is inlet and 0.5 mm for the outlet) and thus, do not require any specific designs. Injection to the inlet performed using a micropipette, while the outlet is connected to the pump with a tube. The side is performed using a micropipette, while the outlet is connected to the pump with a tube. The side openings for the actuating and sensing tips, on the other hand, are fabricated on the SOI wafer (see openings for the actuating and sensing tips, on the other hand, are fabricated on the SOI wafer (see fabrication) and require careful design consideration. Excessively wide openings may compromise fabrication) and require careful design consideration. Excessively wide openings may compromise detection sensitivity, while a very narrow opening will cause fabrication difficulties. The objective is detection sensitivity, while a very narrow opening will cause fabrication difficulties. The objective is to prevent liquid within the channel from leaking out of the backside silicon layer border forming the to prevent liquid within the channel from leaking out of the backside silicon layer border forming channel. Ultimately a gap of 5 µm between the sidewalls and the tips was chosen. This value is similar the channel. Ultimately a gap of 5 µm between the sidewalls and the tips was chosen. This value is to some previous studies using actuation at an air-liquid interface [25]. Another trade-off for a design similar to some previous studies using actuation at an air-liquid interface [25]. Another trade-off for parameter is related to the width of the tips themselves. The width was chosen as 20 µm to be a design parameter is related to the width of the tips themselves. The width was chosen as 20 µm comparable with the targeted cell dimensions. Larger tip width could improve cell-capturing to be comparable with the targeted cell dimensions. Larger tip width could improve cell-capturing efficiency, however, etching the SiO2 layer under the tips would be more difficult. efficiency, however, etching the SiO2 layer under the tips would be more difficult. 2.2.2. Compressing Compressing Side Side 2.2.2. Thecompressing compressingside sideconsists consistsof of electrostatic electrostatic comb-drive comb-drive actuators, actuators,which which provide provide the the necessary necessary The motion of the compressing tip. Depending on the dimension of the targeted cells, the displacement motion of the compressing tip. Depending on the dimension of the targeted cells, the displacement requirementfor forthe theactuator actuatorvaries. varies. Therefore, Therefore,we weuse use various various designs designs with with different different spring spring geometry, geometry, requirement spring constant, constant, and and actuator actuator teeth teeth dimensions. dimensions. Two Two different different spring spring designs designs are are used: used: crab-leg crab-leg spring −1 springs [26] and folded springs [27]. The folded spring design with a spring constant of 40 springs [26] and folded springs [27]. The folded spring design with a spring constant of 40 NN·m · m−1 provideshighly highlystable stablecharacteristics characteristicsin interms termsof of y-axis y-axis displacement. displacement. Conversely, Conversely,the thecrab-leg crab-legspring spring provides −1 design(30 (30N N·m allows longer longer displacement displacement at at lower lower design ·m−1))isissofter softer than than the the folded folded spring spring design, design, which which allows actuation voltages (Figure 3). A wider displacement range is more suitable for use with cells with highly variable diameters.

Micromachines 2018, 9, 275

5 of 14

actuation voltages (Figure 3). A wider displacement range is more suitable for use with cells with Micromachines 2018,diameters. 9, x 5 of 14 highly variable Micromachines 2018, 9, x

5 of 14

Figure 3. 3. Actuation Actuation characteristics characteristics of of various various spring spring designs. designs. Figure Figure 3. Actuation characteristics of various spring designs.

Four groups of comb drives (with 250 teeth per group) provide actuation. Each tooth is designed Four groups of drives (with 250 group) provide actuation. Each tooth is Four groupsof of4comb comb drives 250 teeth teeth per per each group) provide actuation. tooth is designed designed to have a width µm. A pitch(with of 8 µm between tooth results in a gapEach of 2 µm between two to have aa width of 44 µm. A pitch of 88 µm between each tooth results in aa gap of 22 µm between two to have width of µm. A pitch of µm between each tooth results in gap of µm between complimentary opposing teeth (Figure 4). The crab-leg spring design has 20 µm long teeth with 6 two µm complimentary opposing teeth 4). The spring design has 20 µm with 66 µm complimentary opposing teeth(Figure (Figure The crab-leg crab-leg spring design µm long long teeth teeth withspring µm of overlap between opposing teeth. To4).obtain a longer stroke withhas the20 more-stable folded of overlap between opposing teeth. obtain a longer stroke with the more-stable folded spring of overlap between opposing teeth. To longer stroke with the more-stable folded spring geometry, we use two different designs: 20 µm long teeth (with 4 µm overlap) for ordinary use and geometry, we use two designs: 20 teeth 44 µm overlap) ordinary use geometry, use(with two 4different different designs: 20 µm µm long longrequiring teeth (with (with µm overlap) for for(Table ordinary use and and 30 µm longwe teeth µm overlap) for applications longer displacement 1, Figure 3). 30 µm long teeth (with 4 µm overlap) for applications requiring longer displacement (Table 1, Figure 3). 30 µm long µm applications requiring longer displacement (Table 1, Figure 3).

Figure 4. SEM images of the fabricated device: (a) Displacement sensor based on differential Figure 4. SEM images oftips the accessing fabricatedthe device: (a) Displacement sensor based on device, differential capacitors, (b)images opposing handling area, sensor (c) anbased overview of the (d) Figure 4. SEM of the fabricated device: (a) Displacement on differential capacitors, capacitors, (b) opposing tips accessing the handling area, (c) an overview of the device, (d) electrostatic comb drive actuators with (e) a close-up view. (b) opposing tips accessing the handling area, (c) an overview of the device, (d) electrostatic comb electrostatic comb drive actuators with (e) a close-up view. drive actuators with (e) a close-up view. Table 1. Main parameters of device designs. Table 1. Main parameters of device designs. Table 1. Main parameters of device designs. Compressing Side Sensing Side Compressing Side SideConstant Comb Tooth SpringSensing Spring Spring Shape Spring Compressing Constant (N/m) Side Sensing Side Comb Tooth Spring Spring Constant Length/Overlap (µm) Shape (N/m) Spring Constant (N/m) Device Spring Shape Length/Overlap (µm) Shape (N/m) Spring Constant Comb Tooth Spring Constant Design 1 Crab-leg 30 20/6 Crab-leg 25 Spring Shape Spring Shape (N/m) Length/Overlap (N/m) Design Crab-leg 30 20/6 Crab-leg 25 Design 12 Folded 40 20/4(µm) Crab-leg 5 Design 23 1 Folded 40 Crab-leg 5 Design Crab-leg 30 20/620/4 Crab-leg 25 Design Folded 40 30/4 Crab-leg 25 Design Folded 40 20/430/4 Crab-leg 525 Design 32 Folded 40 Crab-leg Device Device

Design 3

Folded

40

30/4

Crab-leg

25

A typical displacement in a crab-leg spring actuator is 6 µm for an actuation voltage of 60 V. A typical displacement a crab-leg actuator is 6 µm for anspring actuation voltage 60 V. Above 60 V, the comb drive in teeth stick tospring each other. Having a higher constant, theoffolded A typical displacement in a crab-leg spring actuator is 6 µm for anspring actuation voltage offolded 60 V. Above 60 V, the comb drive teeth stick to each other. Having a higher constant, the spring design requires ~75 V to achieve 6 µm of displacement. Devices with a tooth length of 20 µm Above 60 V, the comb drive stick to each Having a higherwith springtooth constant, the spring ~75 V teeth to achieve 6 µm of other. displacement. length offolded 20 µm providedesign ~8 µmrequires of displacement before suffering from stiction atDevices ~85 V, whileathe longer teeth length spring design requires ~75 V to achieve 6 µm of displacement. Devices with a tooth length of 20 µm provide ~8 µm of displacement displacement beyond before suffering stiction at ~85 V, while the longer teeth length (30 µm) provides 14 µm (atfrom 110 V). provide ~8 µm of displacement before suffering from stiction at ~85 V, while the longer teeth length (30 µm) provides displacement beyond 14 µm (at 110 V). (30 µm) provides displacement beyond 14 µm (at 110 V). 2.2.3. Sensing Side 2.2.3. Sensing Side The sensing side consists of three elements [26,28]: (i) comb-drive actuators providing the The sensing sidemotion consistsofofthe three elements (i) comb-drive actuators providing the necessary vibrating sensing tip, [26,28]: (ii) differential parallel plate capacitors for necessary vibrating the sensing (ii)the differential parallel plate1acapacitors for displacement sensing,motion and (iii)ofa sensing tip to tip, access handling area (Figures and 4). These displacement sensing, and (iii) a sensing tip to access the handling area (Figures 1a and 4). These elements are mechanically connected while kept isolated electrically to provide optimal simultaneous

Micromachines 2018, 9, 275 Micromachines 2018, 9, x

6 of 14 6 of 14

mechanical and electrical sensing. The comb-drive actuators are driven by a lock-in-amplifier while 2.2.3. Sensing Side the phase is locked based on the output of the displacement sensor for resonance mode. The actuating sensing side consists of sensing three elements (i) with comb-drive actuators the elements of the area are [26,28]: suspended 6 crab-leg springs.providing Two different −1 −1 necessary vibrating motion of the sensing tip, (ii) differential parallel plate capacitors for displacement sensing area designs were fabricated to provide different stiffness: 5 N·m and 25 N·m . The stiffer sensing, and (iii) aenough sensingsensitivity tip to access handling areawith (Figures 1a andoscillations 4). These elements are design provides forthe measurements harmonic while static mechanically whilefrom kept the isolated electrically optimal simultaneous mechanical measurementsconnected would benefit higher sensitivitytoofprovide the softer design. and electrical The comb-drive actuators are driven by a lock-in-amplifier is Similar tosensing. the compressing actuator, four groups of comb drives (with 280while teeth the perphase group) locked on the of the displacement sensorisfor resonance mode. providebased actuation foroutput harmonic detection. Each tooth designed to have the same dimensions as the The actuating elements of the sensing area are suspended with 6 crab-leg springs. different crab-leg spring design of the compressing side: a tooth width of 4 µm, a tooth length ofTwo 20 µm, a gap −1 and 25 N·m−1 . The stiffer sensing area designs were fabricated to provide different stiffness: 5 N · m of 2 µm and an overlap of 6 µm between two consequent opposing teeth. Although the sensing tip is design enoughofsensitivity forduring measurements harmonic while static actuatedprovides with amplitude only 0.2 µm normal usewith (3 Vp-p actuationoscillations voltage), standard 5 µm measurements would benefit from the higher sensitivity of the softer design. displacement (at ~50 V) can be provided if required. Similar to the compressing fourof groups of comb drives (with 280 teeth group)movable provide The displacement sensor,actuator, comprised two stationary electrodes with a per common actuation harmonic detection.capacitors, Each toothworks is designed to have the same dimensions the(10 crab-leg electrode for forming two identical in differential mode [26,29]. Eighty as teeth µm × spring design of the compressing side: a tooth width of 4 µm, a tooth length of 20 µm, a gap of 575 µm; width × length) form parallel-plate capacitors with an initial gap of 40 µm. The movable 2electrode µm and is andesigned overlap of µm between opposing teeth. Althoughwith the an sensing tip of is to 6have a gap of two 5 µmconsequent between both stationary electrodes overlap actuated with amplitude of only 0.2 µm during normal use (3 V actuation voltage), standard 5 µm p-p 525 µm (Figure 4). Working in differential mode, the gap between one of the electrode pair increases, displacement ~50 V)the canother be provided if pair required. while the gap (at between electrode decreases during measurements. The displacement sensor, comprised of two stationary electrodes with a common movable electrode two identical capacitors, works in differential mode [26,29]. Eighty teeth 2.3. MEMSforming Fabrication Process (10 µm × 575 µm; width × length) form parallel-plate capacitors with an initial gap of 40 µm. The device fabrication starts with a standard photolithography process (Figure 5) on the The movable electrode is designed to have a gap of 5 µm between both stationary electrodes with frontside of an SOI wafer (30/2/350 µm-thick). Photoresist is spun on the wafer as a mask for deep an overlap of 525 µm (Figure 4). Working in differential mode, the gap between one of the electrode reactive ion etching (DRIE), and 30-µm-deep structures (tips, comb drive actuators, capacitive pair increases, while the gap between the other electrode pair decreases during measurements. sensors) are defined. After protecting the etched structures on the front surface with a photoresist layer, a 100-nm-thick (Al) mask is patterned on the backside surface as an etch mask for 2.3. MEMS Fabrication aluminum Process another DRIE process stopping at the buried oxide layer (BOX) and forming 350-µm-deep structures. The device fabrication starts with a standard photolithography process (Figure 5) ontothe frontside The final process is the removal of the SiO2 BOX layer with hydrofluoric acid (HF) release the of an SOI wafer (30/2/350 µm-thick). Photoresist is spun on the wafer as a mask for deep reactive ion structures. As a result, the opposing tips are suspended to move freely for single cell characterization etching and 30-µm-deep structures (tips, comb drive actuators, capacitive sensors) are defined. (Figure (DRIE), 5). AfterThe protecting the cover etchedisstructures the afront surface with a photoresist layer, a 100-nm-thick PDMS top fabricatedon using standard soft-lithography process. SU8-2015 (with a aluminum (Al) mask is patterned on the backside surface as an etch mask for another DRIE process height of 10 µm, MicroChem, Westborough, MA) is patterned on a silicon wafer and used as a mold stopping the buried oxide layer (BOX) andare forming 350-µm-deep structures. final is for PDMS.atThen, the channel inlet and outlet opened using biopsy punchers The (1 mm forprocess the inlet the layer with hydrofluoric (HF) and to release the structures. As a device result, 2 BOX andremoval 0.5 mm of forthe theSiO outlet). Finally, the PDMS cover isacid aligned assembled on the MEMS the opposing tips are suspended to move freely for single cell characterization (Figure 5). to form the channel (Figure 2b).

Figure 5. Fabrication process of the device. (a) Frontside photolithography, (b) Frontside silicon Figure 5. Fabrication process of the device. (a) Frontside photolithography, (b) Frontside silicon etching, etching, (c) Protecting frontside structures, (d) Backside Al deposition, (e) Backside photolithography, (c) Protecting frontside structures, (d) Backside Al deposition, (e) Backside photolithography, (f) Al (f) Al etching, (g) Backside etching, SiO2 removal. etching, (g) Backside siliconsilicon etching, (h) SiO(h) 2 removal.

3. Setup and Operation 3.1. Experimental Setup

Micromachines 2018, 9, 275

7 of 14

The PDMS top cover is fabricated using a standard soft-lithography process. SU8-2015 (with a height of 10 µm, MicroChem, Westborough, MA) is patterned on a silicon wafer and used as a mold for PDMS. Then, the channel inlet and outlet are opened using biopsy punchers (1 mm for the inlet and 0.5 mm for the outlet). Finally, the PDMS cover is aligned and assembled on the MEMS device to form the channel (Figure 2b). 3. Setup and Operation Micromachines 2018, 9,Setup x 3.1. Experimental

7 of 14

Experiments are are performed performed on on an an upright upright microscope microscope stage stage (VH-S30B, (VH-S30B, Keyence Keyence Corporation, Corporation, Experiments Osaka, Japan). A long-working distance objective (VH-Z50L, Keyence Corporation) and camera Osaka, Japan). A long-working distance objective (VH-Z50L, Keyence Corporation) and aa camera (Infinity 3, 3, Lumenera LumeneraCorporation, Corporation,Ottawa, Ottawa, ON, Canada) the experiments. experiments. (Infinity Ontario, Canada)are areused usedto to monitor monitor the The fabricated devices are connected to peripheral electronics and fluidic equipment, which are controlled and (version 16, 16, National Instruments Corporation, Austin, TX, USA) anddriven drivenby bya aLabVIEW LabVIEW (version National Instruments Corporation, Austin, TX, program. Prior to positioning on the on microscope stage, the assembled device is mounted on a printed USA) program. Prior to positioning the microscope stage, the assembled device is mounted on a circuit board connected with aluminum wires (Figure printed circuit(PCB) boardand (PCB) and connected with aluminum wires 6a). (Figure 6a).

Figure 6. (a) An assembled device on printed circuit board (PCB) mounted on the setup, (b) Schematic Figure 6. (a) An assembled device on printed circuit board (PCB) mounted on the setup, (b) Schematic view illustrating electrical connections of a device. view illustrating electrical connections of a device.

Actuating the compressing tip requires applying a potential difference between the combs of the Actuating the compressing tip requires applying a potential between the combs of electrostatic actuator. This potential difference is provided withdifference a function generator (33500B, the electrostatic actuator. This potential difference is provided with a function generator (33500B, Keysight Technologies Inc., Santa Rosa, CA, USA) and a high voltage amplifier (WMA-100, Falco KeysightBV, Technologies Rosa, CA, USA) and a high voltage amplifier (WMA-100, Falco Systems Katwijk aanInc., Zee,Santa The Netherlands) to maintain higher voltages. As one of the actuation Systems BV, Katwijk aan Zee, The Netherlands) to maintain higher voltages. As one of the actuation electrodes is also used for electrical sensing, this electrode is grounded virtually by the electrodes is also amplifier used for electrical sensing, this electrode is grounded by signal the transimpedance transimpedance needed for electrical measurements and thevirtually actuation is applied on amplifier needed for electrical measurements and the actuation signal is applied on the other electrode the other electrode (Figure 6b). (Figure The6b). sensing side has a more complex structure to provide mechanical oscillations for harmonic The sensingsimultaneously side has a moreperforming complex structure to provide mechanical oscillations for (PLL) harmonic analysis while electrical measurements. Phase-lock loops are analysis while simultaneously performing electrical measurements. Phase-lock loops (PLL) are essential to perform harmonic oscillations for mechanical measurements. A lock-in-amplifier (Model essential to perform harmonic oscillations for mechanical measurements. A lock-in-amplifier (Model 7230, AMETEK, Inc., Berwyn, PA, USA) drives the comb-drive actuations at a constant phase 7230, AMETEK, Berwyn, PA, USA) drives the readings. comb-drive actuations at a constant phase according to theInc., differential capacitive sensor Two stationary electrodes (C1 according and C2), to the differential capacitive sensor readings. Two stationary electrodes (C and C ), forming 1 2 inputs oftwo forming two identical capacitors with the movable electrode (C0), are connected to the the identical capacitors with the movable electrode (C ), are connected to the inputs of the lock-in-amplifier 0 lock-in-amplifier after passing through low-noise current-to-voltage (A/V) preamplifiers (Signal after passing through low-noise current-to-voltage (A/V) preamplifiers (Signal (provided Recovery, model Recovery, model 5182). C0, on the other hand, is connected to a power source by the 5182). lockC , on the other hand, is connected to a power source (provided by the lock-in-amplifier) to be polarized 0 in-amplifier) to be polarized with a constant voltage. The lock-in-amplifier uses sensor measurements to drive the actuators at the resonance frequency. The tip of the sensing area is connected to another lock-in amplifier (HF2LI, Zurich Instruments Ltd., Zurich, Switzerland) for electrical measurements. An electrical signal is applied on the sensing tip, while the compressing tip is connected to a transimpedance preamplifier (Zurich Instruments HF2TA) that feeds the input of the lock-in amplifier.

Micromachines 2018, 9, 275

8 of 14

with a constant voltage. The lock-in-amplifier uses sensor measurements to drive the actuators at the resonance frequency. The tip of the sensing area is connected to another lock-in amplifier (HF2LI, Zurich Instruments Ltd., Zurich, Switzerland) for electrical measurements. An electrical signal is applied on the sensing tip, while the compressing tip is connected to a transimpedance preamplifier (Zurich Instruments HF2TA) that feeds the input of the lock-in amplifier. The channel outlet on the microfluidic PDMS cap is connected to a flow sensor with tubing before reaching the vacuum pump (AF1, Elveflow, Paris, France, Figure 6a). This equipment is monitored and controlled by the LabVIEW program. Micromachines 2018, 9, x 3.2. Liquid Handling

8 of 14

As cell cell characterization characterization can can only only be be performed performed in in aa dedicated dedicated area, area, targeted targeted cells cells inserted inserted via via the the As channel inlet inlet have have to to be be transported transported and and positioned positioned between between the the tips. tips. Applying Applying aa negative negative pressure pressure channel at the outlet of the channel creates a flow and adjusting the pressure level changes the flow speed to to at the outlet of the channel creates a flow and adjusting the pressure level changes the flow speed control the motion of cells in the channel. control the motion of cells in the channel. After assembling assemblingthe thePDMS PDMScover coverand andthe theMEMS MEMSdevice, device,we we connect outlet channel After connect thethe outlet of of thethe channel to to the pump with tubes (Figure 7ai). At first, we fill the channel and the connection tube with water the pump with tubes (Figure 7ai). At first, we fill the channel and the connection tube with water (Figure 7aii–iv). 7aii–iv). Due Due to to the the small small dimensions dimensions (5 (5 µm µm gap gap between between tips tips and and the the sidewalls), sidewalls), high high surface surface (Figure tension at the handling area maintains the stability of the air-liquid interface. As a result, the liquid tension at the handling area maintains the stability of the air-liquid interface. As a result, the liquid inside the channel does not leak out (Figure 7av). Moreover, by injecting another solution in the inlet, inside the channel does not leak out (Figure 7av). Moreover, by injecting another solution in the inlet, we can can exchange exchange the the solution solutionin inthe thechannel channelwithin withinseconds seconds[20,30]. [20,30].This Thisproperty propertyisisimportant important we forfor a a variety of purposes, for example washing, surface treatment, and drug testing. This is demonstrated variety of purposes, for example washing, surface treatment, and drug testing. This is demonstrated by replacing replacing water water inside inside the the channel channel with by with aa blue-dye blue-dye solution solution (Figure (Figure 7b). 7b).

Figure Figure 7. 7. (a) (a) After After assembling assembling the the PDMS PDMS cover cover on on the the MEMS MEMS device device (i); (i); the the formed formed channel channel is is filled filled with liquid. Adjusting the pressure with the pump, the liquid enters the channel (ii); reaches with liquid. Adjusting the pressure with the pump, the liquid enters the channel (ii); reaches the the opposing area (iv); andand finally, completely fillsfills the the channels; (b) opposing tips tips(iii); (iii);goes goesthrough throughthe thehandling handling area (iv); finally, completely channels; The liquid exchange capability of of thethe device is is tested (b) The liquid exchange capability device testedwith withwater waterand andaablue bluedye dyesolution. solution. (i) (i) While While the the channel channel is is filled filled with with water; water; (ii,iii) (ii,iii) aa blue blue dye dye solution solution is is injected injected at at the the inlet inlet and and within within seconds seconds (iv) is completely completely replaced. replaced. (iv) the the liquid liquid in in the the channel channel is

3.3. Mechanical Detection Method By avoiding the usage of the MEMS elements in liquid, we achieve a higher system quality factor when performing sensitive harmonic analyses [26,29]. In short, the sensing part of the device oscillates harmonically with a series of comb-drive actuators (0.2 µm of actuation at 3 Vrms). These actuators are driven at the resonance frequency of the system by a lock-in amplifier (Signal Recovery, AMETEK 7270 DSP). Due to the mechanical connection (between the sensor, the actuator and the

Micromachines 2018, 9, 275

9 of 14

3.3. Mechanical Detection Method By avoiding the usage of the MEMS elements in liquid, we achieve a higher system quality factor when performing sensitive harmonic analyses [26,29]. In short, the sensing part of the device oscillates harmonically with a series of comb-drive actuators (0.2 µm of actuation at 3 Vrms ). These actuators are driven at the resonance frequency of the system by a lock-in amplifier (Signal Recovery, AMETEK 7270 DSP). Due to the mechanical connection (between the sensor, the actuator and the sensing tip), actuators oscillate the central electrode (C0 ) of the differential capacitive sensor. Detecting the capacitance between the movable common electrode (C0 ) and the other two fixed electrodes of the differential capacitive sensor (C1 and C2 ), a LabVIEW-controlled PLL keeps the sensing arm at resonance throughput the experiment. Capturing a biological sample between the tips changes the spring constant of the system, which can be detected as a change in the resonance frequency by the PLL driven by the lock-in amplifier. To improve the detection performance of the displacement sensor, C0 is polarized with a constant voltage V0 (3 V) and the fixed electrodes are kept grounded [28]. During harmonic oscillations, the change in the gap between the mobile and fixed parallel plate electrodes creates dynamic currents that are collected from C1 and C2 . Amplified with low-noise current-to-voltage preamplifiers (with a gain of 108 ), the sensor outputs are fed into the lock-in-amplifier for real-time measurements. 3.4. Electrical Detection Method Electrical insulation of the various elements of the sensing area is crucial for simultaneous mechanical and electrical detection. By connecting through the backside silicon, the SiO2 layer keeps the sensing elements mechanically attached while providing electrical access to each individual element. This allows us to apply a sinusoidal signal at the actuator (3 Vrms at the resonance frequency of the system) and to read the sensor outputs in parallel with the electrical measurements performed at the tip of the sensing arm. The electrical detection is performed between the sensing and the compressing tips. A 1 Vp-p signal is applied on the sensing tip, while the compressing tip is connected to the virtual ground of a transimpedance amplifier. The dynamic current passing through the amplifier is converted to voltage and measured in real-time. Moreover, by sweeping the frequency at the sensing tip, the frequency response of the system can be analyzed. 4. Device Performance and Results Hereafter, we show some results to investigate the performance of the proposed device. Starting with the frequency response of the device, we performed mechanical and electrical measurements using different sample solutions. Finally, we demonstrated the biomaterial-handling capabilities of the proposed device to confirm the capability of single cell analysis. 4.1. Frequency Response and Real-Time Analysis To characterize the behavior of the device during harmonic analysis, the frequency response of the system was monitored with the lock-in-amplifier (Figure 8). Compared to the initial device characteristics (resonance frequency of 1195 Hz with a quality factor (Q-factor) of 8.9), the assembled device (with PDMS cover) showed a slight increase in the resonance frequency (1200 Hz) and a decrease in the Q-factor (6.3) due to the increased damping as a result of the PDMS slab. Filling the channel with liquid also changes the frequency response. Although the effect of the liquid in terms of mass is negligible compared to the total mass of the mobile part of the sensing arm, the surface tension due to the air-liquid interface in the handling region did affect the spring constant of the system. As a result, the resonance frequency was increased to 1220 Hz with a similar Q-factor (6.8). Observing the changes in the mechanical properties of biological samples requires real-time monitoring of the system response. To achieve this, we performed repeated PLLs controlled by

Micromachines 2018, 9, 275

10 of 14

a LabVIEW program. Before comparing different samples and solutions, we tested the stability of the detection method. Although the mechanical characterization of cells requires less than 30 s, we monitored the system for over five min (Figure S2). Confirmation of the system stability allowed us to perform mechanical measurements in various sample media. Micromachines 2018, 9, x

10 of 14

Micromachines 2018, 9, x

10 of 14

Figure 8. Frequency response of the device in different conditions showing (a) the amplitude and (b)

Figure 8. the device in different conditions showing (a) the amplitude and theFrequency phase shift ofresponse the sensor of readouts. (b) the phase shift of the sensor readouts. 4.2. Mechanical Measurements

4.2. Mechanical Measurements Mechanical detection performance of the systemconditions was tested by measuring the change in the Figure 8. Frequency response of the device in different showing (a) the amplitude and (b) the phase shift of the sensor readouts. resonance frequency in solutions with different viscosities and surface properties. We compared the

Mechanical detection performance of the system was tested by measuring the change in the mechanical response of the device in glucose solutions of various concentrations ranging from 0.1% resonance frequency in solutions with different viscosities surface properties. We compared the 4.2.10% Mechanical Measurements to (w/v). Using the vacuum pump, these solutions wereand consecutively injected after filling the mechanical response of the device in glucose solutions of various concentrations ranging from 0.1% channel with water. The PLL measurements real-time of the the Mechanical detection performance of theallowed system was testedobservation by measuring thechanges change in in the solution between the tips. to 10% (w/v). Using the vacuum pump, these solutions were consecutively injected after filling the resonance frequency in solutions with different viscosities and surface properties. We compared the Increasing glucose concentration resulted in an increase of the resonance frequency (Figure 9). in the channelmechanical with water. The ofPLL measurements allowedofreal-time observationranging of thefrom changes response the device in glucose solutions various concentrations 0.1% Taking the water measurements as the base value, we obtained a resonance frequency shift filling of 1.4 Hz, to 10% (w/v). Using the vacuum pump, these solutions were consecutively injected after the solution between the tips. 2.4 Hz, and 8.3 Hz for glucose concentrationsallowed of 0.1%,real-time 1% and 10%, respectively 9a)inwith channel with water. The PLL measurements of the(Figure changes the Increasing glucose concentration resulteda in an increase ofobservation the isresonance frequency (Figure 9). stable characteristics (Figure S3). Similarly, decrease in amplitude observed with increasing solution between the tips. Taking the water measurements as base value, we obtained a resonance frequency shift of 1.4 Hz, glucose concentration (Figure 9b,the Figure S4). Compared to the initial measurements, a decrease Increasing glucose concentration resulted in an increase of the water resonance frequency (Figure 9). 2.4 Hz, and 8.3 Hz for glucose concentrations of 0.1%, 1% and 10%, respectively (Figure 9a) with of 5 mV, 7 mV, and 10 mV in the amplitude was observed due to higher damping. These Taking the water measurements as the base value, we obtained a resonance frequency shift of 1.4 Hz, stable measurements that the proposed device of isin capable observing changeswith in(Figure theincreasing mechanical characteristics (Figure Similarly, a decrease amplitude is observed 2.4 Hz, and 8.3show HzS3). for glucose concentrations 0.1%, 1%ofand 10%, respectively 9a) withglucose characteristics at the handling area. stable characteristics a decrease amplitude is observed with increasing concentration (Figure 9b, (Figure FigureS3). S4).Similarly, Compared to theininitial water measurements, a decrease of concentration (Figure 9b, Figurewas S4). Compared theto initial water measurements, a decrease 5 mV, 7 glucose mV, and 10 mV in the amplitude observed to due higher damping. These measurements of 5 mV, 7 mV, and 10 mV in the amplitude was observed due to higher damping. These show that the proposed device is capable of observing changes in the mechanical characteristics at the measurements show that the proposed device is capable of observing changes in the mechanical handling area. characteristics at the handling area.

Figure 9. (a) Resonance frequency of the system increased with increasing glucose concentration while (b) the amplitude decreased.

4.3. Electrical Measurements An important propertyfrequency of the proposed deviceincreased is the ability performglucose electrical measurements Figure 9. (a) Resonance of the system with to increasing concentration

Figure 9. (a)with Resonance frequency of the system increased with glucose concentration while while (b) the amplitude decreased. together mechanical characterization. To demonstrate this increasing capability, we injected solutions with (b) different the amplitude decreased. ionic strengths in the channel and examined the electrical properties. Similar to the glucose 4.3. Electrical Measurements An important property of the proposed device is the ability to perform electrical measurements together with mechanical characterization. To demonstrate this capability, we injected solutions with different ionic strengths in the channel and examined the electrical properties. Similar to the glucose

Micromachines 2018, 9, 275

11 of 14

4.3. Electrical Measurements An important property of the proposed device is the ability to perform electrical measurements together with mechanical characterization. To demonstrate this capability, we injected solutions with Micromachines 2018, 9, x 11 of 14 different ionic strengths in the channel and examined the electrical properties. Similar to the glucose measurements, we filled the channel first with water and then injected different molar concentrations measurements, we filled the channel first with water and then injected different molar concentrations of NaCl (0.1 mM–10 mM) consecutively. Using a lock-in-amplifier, we applied 1 V to the sensing tip of NaCl (0.1 mM–10 mM) consecutively. Using a lock-in-amplifier, we applied 1p-p Vp-p to the sensing and amplified the signal obtained from the compressing tip with a transimpedance amplifier (with tip and amplified the signal obtained from the compressing tip with a transimpedance amplifier (with gain of 1055 ). Measurements resulted in increasing voltage waveform amplitudes for increasing ion aa gain of 10 ). Measurements resulted in increasing voltage waveform amplitudes for increasing ion concentration when when the the potential potential difference difference was was applied applied at at 44 kHz kHz (Figure (Figure 10a) 10a) indicating indicating an an increase increase concentration in the the current current passing passing through through the the tips. tips. In In addition addition to to these these real-time real-time measurements measurements (Figure (Figure S5), S5), we we in could also obtain the frequency response of the system by sweeping the frequency from 100 Hz to could also obtain the frequency response of the system by sweeping the frequency from 100 Hz to 100 kHz kHz (Figure (Figure 10b, 10b, Figure Figure S5). 100 S5).

Figure 10. The amplitude of the signal (a) at 4 kHz and (b) during sweeping the frequency in different Figure 10. The amplitude of the signal (a) at 4 kHz and (b) during sweeping the frequency in conditions. different conditions.

4.4. Biological Sample Handling 4.4. Biological Sample Handling Besides the mechanical and electrical measurements, a key feature of the proposed device is the Besides the mechanical and electrical measurements, a key feature of the proposed device is ability to handle biological samples, e.g., single cells. We examined the single cell handling capability the ability to handle biological samples, e.g., single cells. We examined the single cell handling of the device by injecting solutions of fixed cancer cells. The human SUM159PT breast cancer cell line capability of the device by injecting solutions of fixed cancer cells. The human SUM159PT breast cancer was purchased from Asterand (Detroit, MI, USA). The cells were cultured in F12 medium (Invitrogen cell line was purchased from Asterand (Detroit, MI, USA). The cells were cultured in F12 medium Corporation, Carlsbad, CA, USA) supplemented with 5% fetal bovine serum (Lonza Group, Basel, (Invitrogen Corporation, Carlsbad, CA, USA) supplemented with 5% fetal bovine serum (Lonza Group, Switzerland), streptomycin (100 µg/mL), penicillin (100 units/mL), insulin (5 µg/mL) and Basel, Switzerland), streptomycin (100 µg/mL), penicillin (100 units/mL), insulin (5 µg/mL) and hydrocortisone (1 µg/mL) (Invitrogen). Prior to the experiments, subconfluent cell culture were hydrocortisone (1 µg/mL) (Invitrogen). Prior to the experiments, subconfluent cell culture were trypsinized, resuspended in single cell solution, and then fixed with 4% Paraformaldehyde (10 min, trypsinized, resuspended in single cell solution, and then fixed with 4% Paraformaldehyde (10 min, RT), and rinsed with phosphate buffered saline (PBS). RT), and rinsed with phosphate buffered saline (PBS). We controlled the vacuum pump and the compressing actuator to capture single cells. After We controlled the vacuum pump and the compressing actuator to capture single cells. After injecting the cell solution in the channel, we created a flow with pump and applied 110 V at the injecting the cell solution in the channel, we created a flow with pump and applied 110 V at the compressing actuator electrodes to narrow the gap between the tips. This decreases the width of the compressing actuator electrodes to narrow the gap between the tips. This decreases the width of the channel from 20 µm to ~6 µm (Figure 11a). Thus, cells with a mean diameter of ~16 µm could not channel from 20 µm to ~6 µm (Figure 11a). Thus, cells with a mean diameter of ~16 µm could not pass through. When a cell was stopped at the handling area (Figure 11b), we immediately stopped pass through. When a cell was stopped at the handling area (Figure 11b), we immediately stopped the flow and decreased the applied voltage to widen the gap between the tips and position the cell the flow and decreased the applied voltage to widen the gap between the tips and position the cell between the tips (Figure 11c). With the cell thusly captured, we could then apply the compression between the tips (Figure 11c). With the cell thusly captured, we could then apply the compression signal at the compressing actuator (Figure 11d) to examine the characteristics of the captured cell. signal at the compressing actuator (Figure 11d) to examine the characteristics of the captured cell. After After measurements, the compressing tip was moved back to the initial (open) position and the flow measurements, the compressing tip was moved back to the initial (open) position and the flow was was restarted to remove the captured cell and prepare for another cell to characterize. restarted to remove the captured cell and prepare for another cell to characterize.

Micromachines 2018, 9, 275 Micromachines 2018, 9, x

12 of 14 12 of 14

Figure 11. Sequential Sequential photos photos demonstrate demonstrate single single cell cell capturing. capturing. (a) potential difference difference Figure 11. (a) Applying Applying aa potential between the compressing actuator electrodes narrows the gap between the tips; (b) The solution is between the compressing actuator electrodes narrows the gap between the tips; (b) The solution is kept kept flowing until a cell arrives at the handling area; (c) Then, the flow rate and the potential difference flowing until a cell arrives at the handling area; (c) Then, the flow rate and the potential difference between the compressing compressingactuator actuatorelectrodes electrodes decreased until is positioned between the between the areare decreased until thethe cellcell is positioned between the tips; tips; (d) Finally, the flow is completely stopped and cell compression is performed. (d) Finally, the flow is completely stopped and cell compression is performed.

An important point to note was the stability of the air-liquid interface at the sensing tip. An important point to note was the stability of the air-liquid interface at the sensing tip. Although Although the compressing arm moved towards the sensing tip while compressing the cell, the airthe compressing arm moved towards the sensing tip while compressing the cell, the air-liquid interface liquid interface at the sensing tip was not affected. This stability is a crucial point for reliable and at the sensing tip was not affected. This stability is a crucial point for reliable and sensitive detection sensitive detection with the proposed device. with the proposed device. 5. 5. Conclusions Conclusions Separating handle and and Separating handling handling and and sensing sensing elements elements of of aa microfabricated microfabricated device device is is the the key key to to handle analyze biological samples without compromising precise MEMS performance. We demonstrated a analyze biological samples without compromising precise MEMS performance. We demonstrated MEMS device with a built-in microfluidic channel to perform single cell biophysical characterization. a MEMS device with a built-in microfluidic channel to perform single cell biophysical characterization. The built-in built-in channel, channel, requiring The requiring no no assembly assembly actions actions between between the the MEMS MEMS and and microfluidics microfluidics elements, elements, not only provides higher-throughput for analysis but also improves sensitivity byintegration allowing not only provides higher-throughput for analysis but also improves sensitivity by allowing integration of microfluidic and MEMS elements at a much finer assembly resolution. Moreover, of microfluidic and MEMS elements at a much finer assembly resolution. Moreover, simultaneous simultaneous electrical and mechanical measurements allow various to beastargeted, such electrical and mechanical measurements allow various parameters to beparameters targeted, such size, stiffness, as size, stiffness, viscous losses, membrane fluidity, membrane capacitance, and cytoplasm viscous losses, membrane fluidity, membrane capacitance, and cytoplasm resistivity. The device resistivity. device sensitive enough to distinguish differences between liquids accordingand to is sensitiveThe enough to is distinguish differences between liquids according to their mechanical their mechanical and electrical properties. As demonstrated with cell compression, the device is electrical properties. As demonstrated with cell compression, the device is capable of capturing capable of capturing a single cell and stimulating it mechanically. Sensitive real-time measurements a single cell and stimulating it mechanically. Sensitive real-time measurements and highly controllable and highly stimulation controllableallow mechanical stimulation allow at single cell compression characterization at Together different mechanical single cell characterization different levels. compression levels. Together with the electrical measurements, the device allows for practical multiparameter analyses of single non-adherent cells to perform routine clinical tests for early disease diagnosis with improved cost and time efficiency.

Micromachines 2018, 9, 275

13 of 14

with the electrical measurements, the device allows for practical multi-parameter analyses of single non-adherent cells to perform routine clinical tests for early disease diagnosis with improved cost and time efficiency. Supplementary Materials: The following are available online at http://www.mdpi.com/2072-666X/9/6/275/s1, Figure S1: Schematic view of the assembly, Figure S2: Real-time stability measurements, Figure S3: Real-time frequency monitoring in glucose solutions with different concentrations, Figure S4: Real-time amplitude monitoring in glucose solutions with different concentrations, Figure S5: Real-time electrical measurements. Author Contributions: Y.T., G.P., D.C., C.L. and M.C.T. performed the experiments. Y.T., M.K. and M.A. fabricated the devices. G.P., D.C. and M.C.T. developed the setup and control software. Y.T., G.P., M.K., S.L.K., H.F., D.C. and M.C.T. developed the design and concept of the engineering part. S.M. and C.L. provided cells and worked on the biological relevance. Y.T., G.P., D.C., C.L. and M.C.T. analyzed the data. Y.T., G.P., M.K., M.A., S.M., S.L.K., H.F., D.C., C.L. and M.C.T. contributed to the manuscript. Acknowledgments: Authors would like to thank the VLSI Design and Education Center (VDEC, The University of Tokyo, Japan) for the mask production, IRCL (Institut pour la Recherche sur le Cancer de Lille, France) for hosting SMMiL-E facilities and V. Menon for critical reading of the paper. Y.T. thanks the University of Lille and the SIRIC ONCO-Lille, and M.C.T. thanks Région Hauts-de-France for financial support. Conflicts of Interest: The authors declare no conflict of interest.

References 1. 2. 3. 4. 5.

6. 7. 8. 9. 10.

11.

12.

13. 14. 15. 16.

Lim, C.T.; Zhou, E.H.; Quek, S.T. Mechanical models for living cells—A review. J. Biomech. 2006, 39, 195–216. [CrossRef] [PubMed] Suresh, S. Biomechanics and biophysics of cancer cells. Acta Mater. 2007, 55, 3989–4014. [CrossRef] Cross, S.E.; Jin, Y.-S.; Rao, J.; Gimzewski, J.K. Nanomechanical analysis of cells from cancer patients. Nat. Nanotechnol. 2007, 2, 780–783. [CrossRef] [PubMed] Ketene, A.N.; Schmelz, E.M.; Roberts, P.C.; Agah, M. The effects of cancer progression on the viscoelasticity of ovarian cell cytoskeleton structures. Nanomed. Nanotechnol. Biol. Med. 2012, 8, 93–102. [CrossRef] [PubMed] Tse, H.T.K.; Gossett, D.R.; Moon, Y.S.; Masaeli, M.; Sohsman, M.; Ying, Y.; Mislick, K.; Adams, R.P.; Rao, J.; Di Carlo, D. Quantitative Diagnosis of Malignant Pleural Effusions by Single-Cell Mechanophenotyping. Sci. Transl. Med. 2013, 5, 212ra163. [CrossRef] [PubMed] Lee, G.Y.H.; Lim, C.T. Biomechanics approaches to studying human diseases. Trends Biotechnol. 2007, 25, 111–118. [CrossRef] [PubMed] Guck, J.; Chilvers, E.R. Mechanics meets medicine. Sci. Transl. Med. 2013, 5, 212fs41. [CrossRef] [PubMed] Discher, D.E. Tissue Cells Feel and Respond to the Stiffness of Their Substrate. Science 2005, 310, 1139–1143. [CrossRef] [PubMed] Jonas, O.; Mierke, C.T.; Käs, J.A. Invasive cancer cell lines exhibit biomechanical properties that are distinct from their noninvasive counterparts. Soft Matter 2011, 7, 11488–11495. [CrossRef] Suresh, S.; Spatz, J.; Mills, J.P.; Micoulet, A.; Dao, M.; Lim, C.T.; Beil, M.; Seufferlein, T. Connections between single-cell biomechanics and human disease states: Gastrointestinal cancer and malaria. Acta Biomater. 2005, 1, 15–30. [CrossRef] [PubMed] Guck, J.; Schinkinger, S.; Lincoln, B.; Wottawah, F.; Ebert, S.; Romeyke, M.; Lenz, D.; Erickson, H.M.; Ananthakrishnan, R.; Mitchell, D. Optical deformability as an inherent cell marker for testing malignant transformation and metastatic competence. Biophys. J. 2005, 88, 3689–3698. [CrossRef] [PubMed] Mathur, A.B.; Collinsworth, A.M.; Reichert, W.M.; Kraus, W.E.; Truskey, G.A. Endothelial, cardiac muscle and skeletal muscle exhibit different viscous and elastic properties as determined by atomic force microscopy. J. Biomech. 2001, 34, 1545–1553. [CrossRef] Wang, N.; Butler, J.P.; Ingber, D.E. Mechanotransduction across the cell surface and through the cytoskeleton. Science 1993, 260, 1124–1127. [CrossRef] [PubMed] Chu, Y.-S.; Dufour, S.; Thiery, J.P.; Perez, E.; Pincet, F. Johnson-Kendall-Roberts theory applied to living cells. Phys. Rev. Lett. 2005, 94, 028102. [CrossRef] [PubMed] Shojaei-Baghini, E.; Zheng, Y.; Jewett, M.A.S.; Geddie, W.B.; Sun, Y. Mechanical characterization of benign and malignant urothelial cells from voided urine. Appl. Phys. Lett. 2013, 102, 123704. [CrossRef] Fritsch, A.; Höckel, M.; Kiessling, T.; Nnetu, K.D.; Wetzel, F.; Zink, M.; Käs, J.A. Are biomechanical changes necessary for tumour progression? Nat. Phys. 2010, 6, 730–732. [CrossRef]

Micromachines 2018, 9, 275

17. 18. 19. 20.

21.

22.

23. 24.

25. 26.

27. 28.

29. 30.

14 of 14

Zheng, Y.; Wen, J.; Nguyen, J.; Cachia, M.A.; Wang, C.; Sun, Y. Decreased deformability of lymphocytes in chronic lymphocytic leukemia. Sci. Rep. 2015, 5, 7613–7615. [CrossRef] [PubMed] Cho, Y.; Kim, H.S.; Frazier, A.B.; Chen, Z.G.; Shin, D.M.; Han, A. Whole-Cell Impedance Analysis for Highly and Poorly Metastatic Cancer Cells. J. Microelectromech. Syst. 2009, 18, 808–817. Tarhan, M.C.; Yokokawa, R.; Jalabert, L.; Collard, D.; Fujita, H. Pick-and-Place Assembly of Single Microtubules. Small 2017, 16, 1701136. [CrossRef] [PubMed] Tarhan, M.C.; Lafitte, N.; Tauran, Y.; Jalabert, L.; Kumemura, M.; Perret, G.; Kim, B.; Coleman, A.W.; Fujita, H.; Collard, D. A rapid and practical technique for real-time monitoring of biomolecular interactions using mechanical responses of macromolecules. Sci. Rep. 2016, 6, 28001. [CrossRef] [PubMed] Baëtens, T.; Perret, G.; Takayama, Y.; Kumemura, M.; Jalabert, L.; Meignan, S.; Lagadec, C.; Fujita, H.; Collard, D.; Tarhan, M.C. A practical single cell analysis method for mechanical characterization of cancer cells. In Proceedings of the IEEE 30th International Conference on Micro Electro Mechanical Systems, Las Vegas, NV, USA, 22–26 January 2017; pp. 608–611. Perret, G.; Lacornerie, T.; Manca, F.; Giordano, S.; Kumemura, M.; Lafitte, N.; Jalabert, L.; Tarhan, M.C.; Lartigau, E.F.; Cleri, F.; et al. Real-time mechanical characterization of DNA degradation under therapeutic X-rays and its theoretical modeling. Microsyst. Nanoeng. 2016, 2, 16062. [CrossRef] Sugiura, H.; Sakuma, S.; Kaneko, M.; Arai, F. On-Chip Method to Measure Mechanical Characteristics of a Single Cell by Using Moiré Fringe. Micromachines 2015, 6, 660–673. [CrossRef] Chang, D.; Sakuma, S.; Kera, K.; Uozumi, N.; Arai, F. Measurement of the mechanical properties of single Synechocystis sp. strain PCC6803 cells in different osmotic concentrations using a robot-integrated microfluidic chip. Lab Chip 2018, 18, 1241–1249. [CrossRef] [PubMed] Park, J.; Nishida, S.; Lambert, P.; Kawakatsu, H.; Fujita, H. High-resolution cantilever biosensor resonating at air–liquid in a microchannel. Lab Chip 2011, 11, 4187–4193. [CrossRef] [PubMed] Yamahata, C.; Collard, D.; Legrand, B.; Takekawa, T.; Kumemura, M.; Hashiguchi, G.; Fujita, H. Silicon Nanotweezers With Subnanometer Resolution for the Micromanipulation of Biomolecules. J. Microelectromech. Syst. 2008, 17, 623–631. [CrossRef] Legtenberg, R.; Groeneveld, A.; Elwenspoek, M. Comb-drive actuators for large displacements. J. Micromech. Microeng. 1996, 6, 320–329. [CrossRef] Lafitte, N.; Haddab, Y.; Le Gorrec, Y.; Guillou, H.; Kumemura, M.; Jalabert, L.; Collard, D.; Fujita, H. Improvement of silicon nanotweezers sensitivity for mechanical characterization of biomolecules using closed-loop control. IEEE/ASME Trans. Mechatron. 2014, 20, 1418–1427. [CrossRef] Yu, S.; Fry, S.N.; Potasek, D.P.; Bell, D.J.; Nelson, B.J. Characterizing fruit fly flight behavior using a microforce sensor with a new comb-drive configuration. J. Microelectromech. Syst. 2005, 14, 4–11. Montasser, I.; Coleman, A.W.; Tauran, Y.; Perret, G.; Jalabert, L.; Collard, D.; Kim, B.J.; Tarhan, M.C. Direct measurement of the mechanism by which magnesium specifically modifies the mechanical properties of DNA. Biomicrofluidics 2017, 11, 051102. [CrossRef] [PubMed] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).