[27]microgripperSU8

JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 14, NO. 4, AUGUST 2005 857

Electrothermally Activated SU-8 Microgripper forSingle Cell Manipulation in Solution

Nikolas Chronis and Luke P. Lee

Abstract—The development of a SU-8-based microgripper thatcan operate in physiological ionic solutions is presented. Theelectrothermally activated polymer gripper consists of two “hot-and-cold-arm” actuators that are fabricated in a two-mask surfacemicromachining process. The high thermal expansion coefficientof SU-8 (52 ppm C) compared to silicon and metals, allows theactuation of the microgripper with small average temperatureelevations (10 – 32 C) at low voltages (1–2 V). The polymermicrogripper can be used for the manipulation of single cellsand other biological species in solution with minimal undesiredinteractions.

[1330]

Index Terms—Polymer MEMS, polymer microgripper, singlecell manipulation, SU-8 actuators.

I. INTRODUCTION

THE development of miniaturized systems for manipulatingbiological samples in solution has become a great tech-

nological challenge for the future of genomics and proteomics[1]. Current biomanipulation tools such as optical tweezers [2]and micropipetes [3], although powerful, rely on bulky and ex-pensive setups. Alternatively, the use of microgrippers as themechanical end-effectors that securely grasp and transport themicro object to the desired location seems to be a promising ap-proach, offering robustness and manipulation flexibility withoutinterfering optically or electrically with the sample. Microgrip-pers capable of being activated in an ionic environment can beused for single cell manipulation and positioning, cell isola-tion, as well as for minimally invasive endoscopic operations.At the same time, biocompatibility issues of the mechanism thatconverts energy into mechanical motion need to be taken intoaccount.

Most of the previously developed microgrippers cannot beoperated in physiological solutions because the actuation mech-anism is not compatible with liquid operation. Electrostaticgrippers [4], [5] cannot be stimulated in electrolytic media.Thermally driven grippers [6]–[9] operate at extremely hightemperatures and high voltages (bubble formation by electrol-ysis typically occurs at in water [10]). Piezoelectricgrippers [11] are also activated at high voltages and producesmall displacements often requiring multilayer actuators or am-plification mechanisms. Shape memory alloys-based grippers[12], [13] could potentially be used in liquid environment, butthey can cycle only a few times before complete immobilization.

Manuscript received May 5, 2004; revised October 28, 2004. This work wassupported by DARPA under the BioFlips program. Subject Editor N. de Rooij.

The authors are with the Berkeley Sensor and Actuator Center, Departmentof Bioengineering, University of California, Berkeley, CA 94720 USA (e-mail:[email protected]).

Digital Object Identifier 10.1109/JMEMS.2005.845445

In pneumatically driven microgrippers [14], the actuationmechanism is not integrated with the gripper, posing an ad-ditional level of complexity in the design and fabrication ofsuch structures. Pneumatically driven microcages [15], [16],thermally activated polymer-based actuators [17], [18] andelectroactive polymer microarms based on ionic absorptionand swelling [19], [20], can operate in aqueous solutions buthave limited actuation control and are restricted to out of planemotion. This makes it difficult to manipulate and activelyposition small (5 – 40 in diameter) cells.

In contrast to previous works, our approach is based on the de-velopment of a SU-8 microgripper with integrated, electrother-mally activated, in plane, SU-8 actuators. Taking advantage ofthe structural rigidity, the chemical resistance, as well as theability to define high aspect ratio structures on SU-8 films, wefabricated a SU-8 microgripper that can operate in both liquidand air environments. A critical element in our design is thelarge coefficient of thermal expansion of SU-8 that allows theelectrothermal activation of the microgripper in ionic physio-logical solutions at low voltages and temperature increases. Theproposed actuation mechanism enables the manipulation andpositioning of single cells or other biological species with min-imal undesired interactions. This paper demonstrates for the firsttime the successful manipulation of a 10- in diameter singlecell in solution using a MEMS microgripper.

II. MICROGRIPPER DESIGN

SU-8 is a highly crosslinked epoxy-type photo-pat-ternable polymer with a coefficient of thermal expansion(CTE) of 52 [21], relatively large elastic modulus( ) [22], [23], and glass transition tempera-ture above 200 [24]. These properties make SU-8 a greatmaterial for building rigid mechanical structures for variousapplications. Recent studies on the growth of human cellson SU-8 substrates [25] also revealed good biocompatibility,making SU-8 useful in a variety of in-vitro BioMEMS appli-cations [26], [27]. However, the long-term stability of SU-8 incell culture media [28] and its hemocompatibility for in-vivoapplications [29] is still questionable. On the other hand, SU-8is a nonconductive polymer. Developing SU-8-based actuatorsbecomes an extremely challenging task, due to the difficulty ofimplementing and integrating a reliable actuation mechanism.

Based on these features and especially on the high CTE ofSU-8 and the high aspect ratio characteristics of SU-8 films,we developed an innovative approach for constructing an elec-trothermal SU-8 microgripper integrated with in plane SU-8actuators (Fig. 1). The key point is the addition of a thin metalresistor patterned at the bottom of the SU-8 structure.

1057-7157/$20.00 © 2005 IEEE

858 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 14, NO. 4, AUGUST 2005

Fig. 1. Design and close-ups of the SU-8 microgripper. The SU-8 and Cr/Au layers of the hot-and-cold-arm actuator are shown in the insets. The initial separationdistance of the two gripper arms is 7 �m. The flexure, cold arm, and hot arm of each SU-8 actuator are 6, 22, and 7 �m wide, respectively.

Our microgripper design was inspired by the well-es-tablished hot-and-cold-arm actuator design [30], [31]. Thehot-and-cold-arm actuator is composed of two arms of differentwidths that are joined at their free ends to form a U shape. Thehigher ohmic resistance in the narrower “hot” arm results ingreater heating and expansion than the wider “cold” arm. Thus,the tip of the actuator can move laterally in an arching motiontoward the “cold” arm side. The lateral deflection of the tip ofthe actuator can be estimated by [32]

where is the cross-sectional area of the hot arm (assumedequal to the cross-sectional area of the flexure); is the momentof inertia of the hot arm (assumed equal to the cross-sectionalarea of the flexure); is the length of the actuator; is the ratioof the flexure length to the hot arm length; is the coefficientof thermal expansion (CTE); is the center spacing betweenthe hot arm and the flexure; and is the net temperaturedifference.

The deflection of the tip is linear proportional to the coeffi-cient of the thermal expansion. A material like SU-8, will result18 times lower net temperature differences when compared topolysilicon ( ) for the same amount of dis-placement .

Polysilicon-based hot-and-cold-arm actuators are amongthe most popular thermal actuators, providing large range ofdisplacements and high forces. They operate though at rela-tively high voltages (3–10 V) and extremely high temperatures(400–700 ) [32], [33], conditions incompatible for use withany biological sample.

The SU-8 microgripper consists of two hot-and-cold-armactuators that mirror each other and are serially connected.Thermal expansion of the SU-8 is achieved through resis-tive heating of a thin Cr/Au layer attached selectively at thebottom of the suspended polymer structure [Fig. 1(a)–(b)].

The SU-8/Au/Cr structure is extremely stiff in the verticaldirection and thus moves in plane rather than out of planewhen is activated (the SU-8 layer is 67 times thicker than theCr/Au layer). The two SU-8 gripper arms extend 450 awayfrom the actuators forming a circular cell holder. The initialseparation distance between the two gripper arms is 7 . Dueto the deformability and fragileness of biological samples, thegripper is designed to hold a cell 8–10 in diameter withoutexerting any forces in the closed position. If larger samplesneed to be manipulated, then the cell holder dimensions haveto be modified accordingly.

III. FABRICATION PROCESS

The microgripper is fabricated using standard surface micro-machining techniques in a two-mask process (Fig. 2). A 0.3--thick layer of silicon dioxide is thermally grown on a bare sil-icon wafer followed by LPCVD deposition of a 0.7- -thickundoped polysilicon layer. The silicon dioxide layer providesthermal and electrical isolation while the polysilicon layer isused as the sacrificial layer. The metal heating elements and padsof the microgripper are subsequently defined by lift-off. Themetal elements consist of a 10 nm/300 nm-thick electron-beamevaporated Cr/Au film. A 20- -thick SU-8-10 layer is thenspun on the wafer and photolithographically patterned to formthe structural layer. A blind cut (using a dicing machine) is madeon the backside of the wafer to facilitate the accurate removalof the bulk of the wafer in a subsequent step (step f). The waferis then diced into chips. The individual chips are hard baked ona hot plate at 120 for 15 min and finally dry etch-releasedin xenon difluoride. Xenon difluoride has excellent selectivityto SU-8 and it produces gas phase products eliminating stictionproblems. The released microgripper is hard baked for a secondtime at 120 for 15 min to complete the crosslinking of SU-8.To avoid the creation of additional out of plane residual stressdue to the SU-8 polymerization, the released structures are heldflat with glass slides placed on top of them during the second

CHRONIS AND LEE: ELECTROTHERMALLY ACTIVATED SU-8 MICROGRIPPER 859

Fig. 2. The two-mask fabrication process of the microgripper. The blindcut (step d) facilitates the removal of the bulk of the substrate in order foroverhanging structures to be obtained.

Fig. 3. Scanning electron micrographs and close-ups of the overhangingfabricated SU-8 microgripper. The Cr/Au layer is covered by the SU-8 layerand therefore is not visible.

hard baking step. We found that the second hard baking step onthe released structures is critical to complete the polymerizationprocess. If that step is omitted, plastic deformation is observedduring operation due to the formation of extra polymer bondsin the SU-8 film. The chips are broken into two pieces using adiamond scriber leaving the overhanging structure attached toone of them. The released microgripper is shown in Fig. 3.

IV. THERMOMECHANICAL FINITE ELEMENT MODELING

Biocompatible operation of the microgripper requires acti-vation at low temperatures. We used finite element modelingsoftware (ANSYS) to determine the steady state temperaturedistribution within the microgripper and particularly the max-imum temperature change – observed in the hot armside of the actuator – and at the tip of the cell holder. Thetemperature at the tip has to be as low as possible to avoidthermal loading of the cell and the maximum temperature has

Fig. 4. Finite element modeling of the temperature field within themicrogripper (h = 200 W=m K, d = 12 �m). The temperature change atthe tip is almost negligible.

to be lower than the boiling point of the solution ( foraqueous solutions) to avoid uncontrollable bubble formation.

We constructed a two-material (the thin Cr layer was omitted)electrothermal model of the microgripper to simulate the ex-pansion of the SU-8 structure due to the resistive heating of themetal film. To simplify the simulations we assumed that all thesurfaces of the microgripper have an “overall” convection coef-ficient . Conductive heat transfer from the bottom surface ofthe device was not taken into account due to the overhangingconfiguration of the microgripper [34]. The substrate was as-sumed to be thermally grounded and thus the temperature ofthe device anchors was fixed and equal to the ambient temper-ature. Thermal radiation was omitted due to small temperatureschanges. The coefficient of the thermal expansion and elasticmodulus of SU-8 were assumed to be 52 and 4.4 GPa,respectively. The thermal conductivity of SU-8 has not been re-ported in the literature. A value of was chosenwhich corresponds to a typical value of thermal conductivitiesencountered in thermoplastics and epoxy resins [35], [36].

Fig. 4 depicts the temperature distribution within the mi-crogripper (only half of the microgripper is shown) whenthe gripper opens from its initial position. Theconvection coefficient was taken as , repre-senting an average value for the free convection coefficient inaqueous liquids. Those assumptions and results ( ,

) reproduce the scenario of grasping a 19diameter cell in solution. Due to the relatively low thermal con-ductivity of SU-8 and the high convection coefficients in liquidenvironments, the temperature drops considerably along thegripper arm. The temperature difference at the tip is negligible( ), indicating minimum interference of the temper-ature field with the biological sample. A temperature gradientis present in the vertical direction within the hot-and-cold-armactuator due to the selective heat generation at the bottom ofthe structure. That temperature gradient combined with thebimorph structure of the actuator results a maximum upwardout of plane displacement of 1.4 at the tip of the gripper.The maximum temperature change of 57.79 is observed

860 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 14, NO. 4, AUGUST 2005

Fig. 5. Dependence of the gripper opening with maximum temperaturechange. The maximum temperature change is observed near the middle of thehot arm side (Fig. 4). The gripper opening is measured from the rest position(shown in the inset).

near the middle of the hot arm and is lower than the temperaturechange that is needed to reach the boiling point.

The dependence of the gripper opening with the maximumtemperature change for three different convection coefficientsis shown in Fig. 5. A value of corresponds to atypical value for free convection coefficient in air. Higher con-vection coefficients produce larger displacements for the samemaximum temperature changes.

The simulated temperature profile along the hot-and-cold-arm actuator is graphically depicted in Fig. 6 for twodifferent convection coefficients ( and

) at the same electrical power input. The de-picted profile corresponds to the temperature distribution withinthe metal layer. Due to the vertical temperature gradient, thetop surface of the SU-8 structure is expected to have a slightlydifferent profile. Although the power input is the same, thegripper opens 14.4 in air (top curve, ) and12 in aqueous solution (bottom curve, ).The areas under those curves represent the average temperaturechange of the actuator

(1)

where , , and is thelength of the flexure, hot arm, and cold arm, respectively. Therelation between and is linear with a small de-pendence on the convection coefficient (Fig. 7). For all practicalpurposes, the is approximately 1.62 times the value of

at (and at). As will be discussed in Section III, the can

be experimentally estimated by measuring the change of the re-sistance of the Cr/Au layer with temperature. Fig. 7 can then beused to back-calculate the for a given .

V. EXPERIMENTAL RESULTS

To determine the mechanical performance, we activated ourmicrogripper using dc driving voltages and measured under

Fig. 6. Predicted temperature profile along the hot-and-cold-arm actuatorfor two different convection coefficients (h = 10 W=m K and h =

200 W=m K). The area under the curve represents the average temperaturechange �T for each case.

Fig. 7. Predicted maximum temperature change versus average temperaturechange. An increase in the value of the convection coefficient from 10W=m K

to 500 W=m K results in an increase in the value of �T by 19% for thesame �T .

a light microscope the corresponding displacement response.Identifying the dependence between applied voltage and dis-placement when operation takes place in physiological mediawas of special interest, since our microgripper is intended towork in such environments. High voltages and strong electricfields can cause electrolysis, electroosmotic flow and irre-versible cell damaging. Therefore, the need for low drivingvoltages becomes a critical factor for biocompatible operation.All the measurements presented here were performed on thereleased microgripper without removing the bulk of the wafer.

Fig. 8 shows the displacement response when a dc voltageis applied in air and in Dulbecco’s phosphate-buffered saline(D-PBS) environment1. When the gripper is immersed inD-PBS, it immediately swells approximately 0.5% (measuredfrom SU-8 strain gauges that were fabricated on-chip [37]). Noadditional swelling was observed after 1 h. The displacementcorresponds to the total change in the distance between the twogripper arms when both actuators are activated, starting froman initial separation of . Operation in D-PBS solutionrequires higher voltages than operation in air due to the higher

1Dulbecco’s phosphate-buffered solution is typically used for cell washing oras a diluent for media or assays, or as an inorganic base in standard media andit is composed of physiological, inorganic salts.

CHRONIS AND LEE: ELECTROTHERMALLY ACTIVATED SU-8 MICROGRIPPER 861

Fig. 8. Gripper opening (measured from the initial separation of 7 �m shownin the inset) versus applied dc voltage in air and D-PBS environment.

convection coefficients encountered in liquids. Grasping a10- diameter cell in D-PBS requires only 1.2 V (3 ofdisplacement from rest). For the full range of displacements,no bubble formation was observed, suggesting that minimumelectrolysis took place. Alternatively, high frequency ac volt-ages can also be applied if electrolysis needs to be completelyavoided. The error bars correspond to a set of measurementsfrom four different devices. They are within 5% of the averagevalue, thus indicating great repeatability of the microgripperperformance. The gripper consumes 4.7 mW at maximumdisplacement ( ) in air operation, while the powerconsumption in D-PBS environment is 90.5 mW. The highpower consumption of the actuator in liquid (15–20 timeshigher than air) is justified by the higher convection coefficientsin liquids (typically 10–100 times higher than air). Due tothe relatively small electrical resistance of the microgripper( ), ohmic losses from the probe tips to the metal padsand also from the pads to the actuator become significant. Wecalculated that 7.5% of the total input power is lost from theabove two sources.

The average increase in temperature of the microgripper actu-ator can be estimated from the change in resistance of the Cr/Auresistors

(2)

where is the temperature coefficient of resistance (TCR)of the Cr/Au film. The TCR was experimentally found to beequal to 3.36 . The resistance of the microgripper hada value of at room temperature (25 ).The maximum resistance change for the full range of motionwas approximately 11.5% for operation in D-PBS, resulting inan average temperature change of 34 (Fig. 9). The max-imum temperature change back-calculated from Fig. 7( ) is, therefore, (20 lowerthan the boiling point of water assuming initial room tempera-ture of 25 ). For the same amount of displacement higher av-erage temperatures were measured for operation in D-PBS thanin air. Those findings contradict the simulation results (see alsoFigs. 5 and 7), where higher convection coefficients result inlower for the same amount of displacement. We suspect

Fig. 9. Gripper opening versus average temperature change�T . Immersedin D-PBS, the gripper opens 10 �m at�T = 28 C. The simulation resultsmatch the experimental data in D-PBS operation for h = 200W=m K.

Fig. 10. Time-resolved electrical resistance measurements in D-PBSenvironment. The steady-state upon cooling from the maximum openingposition is reached after 9 ms. The heating process is 20 times slower (200 ms).

that the main reason for the discrepancy is the small separation(11 ) between the hot and cold arms which causes conductivethermal heating of the cold arm through the surrounding mediaAn additional error in the simulation results may be attributedto the unknown value of the thermal conductivity of SU-8.

We also obtained the thermal time response of the micro-gripper operating in D-PBS by time-resolved electrical resis-tance measurements (Fig. 10). A single step input (0.2 to 2 Vand 2 to 0.2 V) was applied to the microgripper and the cor-responding change in the resistance was recorded. The inputvoltage (2 V) corresponds to the maximum gripper opening(12 at 10.5% resistivity change). Upon heating, the gripperreaches a steady-state after approximately 200 ms (up to 10 mstime response is shown in Fig. 10) while the cooling processis completed after 9 ms. The maximum driving frequency for acomplete actuation cycle and full range of motion is therefore4.8 Hz.

The short-term operation of the gripper was remarkably reli-able. When the gripper was activated in liquid environment at50% of its maximum displacement at 30 Hz sinusoidal input for1 h ( ), no degradation in performance was no-ticed. In addition, no adhesion problems of the metallic layerand the SU-8 were observed.

862 JOURNAL OF MICROELECTROMECHANICAL SYSTEMS, VOL. 14, NO. 4, AUGUST 2005

Fig. 11. Schematic drawing of the experimental setup. The overhangingmicrogripper is fully immersed in the aqueous solution containing Hela cells.It is connected to a dc power supply and to an external micromanipulator thatis used to position the cell in front of the gripper.

Fig. 12. Manipulating a single Hela cell in solution. (a) Initial position. (b) Thegripper opens and approaches the cell. (c) The gripper grasps the cell. (d) Thecell is aligned on third column of the chessboard. (e) The gripper opens andleaves the cell. (f) The gripper returns to its original position.

VI. MANIPULATING SINGLE CELLS IN SOLUTION

The ability of our microgripper to manipulate single cellsin vitro was experimentally verified using the setup shown inFig. 11. We patterned a chessboard using photolithography ona gold film that was previously deposited on a silicon chip. Thegold chessboard had 8 square patterns equally spaced. Thesilicon chip was mounted on a XYZ micromanipulator that al-lowed us to move the chessboard toward the gripper. During theexperiment the microgripper was held fixed on the chuck of aprobe station and powered by a dc power supply. The cells usedin the experiment were Hela cells, cultured in Dulbelco’s Mod-ified Edge Medium and diluted in D-PBS.

Fig. 12 shows the successful gripping and positioning of asingle cell on the third column of the chessboard. The cell is ini-tially a few tens of microns away from the chessboard and thegripper is in its initial closed position [Fig. 12(a)]. The gripperis subsequently activated and the chessboard containing the cellapproaches the open gripper [Fig. 12(b)]. The gripper closes,gripping the cell [Fig. 12(c)]. The silicon chip then starts movingso the cell is aligned to the third row of the gold chessboard[Fig. 12(d)]. The gripper finally opens, releasing the cell andthe chessboard returns to its initial position [Fig. 12(e),(f)]. Al-though the manipulation of micro objects in aqueous mediaseems like a simple task, it possesses certain challenges. At suchsmall scales, viscous forces become dominant when comparedto inertial forces. If no additional forces are present (e.g., adhe-sion forces of the cell to the substrate, friction) then the micro

object is dragged by the fluid motion. We observed that phenom-enon when trying to approach the cell with our microgripper.The motion of the microgripper creates fluid motion that dragsthe cell away. A partial solution to the problem is to move thegripper slowly toward the cell, thus reducing the velocity fieldof the fluid. Another approach is to attach the cell to the sub-strate using appropriate adhesion molecules but that requires anextra step.

VII. CONCLUSION

Microgrippers capable of manipulating biological samplesin physiological liquids is a missing piece from the biologicaltoolbox. The reported design, based on electrothermally acti-vated SU-8 hot-and-cold-arm actuators aims to fill that gap. Thepolymer microgripper is fabricated in a standard two-mask sur-face micromachining process. The high coefficient of thermalexpansion of SU-8 (52 ) makes the operation of themicrogripper at average temperatures changes less than 32possible while the thermal loading of the sample is minimized(the temperature change at the cell holder is negligible). Drivingvoltages below the electrolysis point are sufficient to open thegripper up to 11 m. Such features make the proposed actuationmechanism compatible with liquid operation. Our microgrippercan potentially be used in any type of biological assay wheresample manipulation is required. Cell positioning on a patchclamp array [38] and single cell isolation/manipulation [39] forgenetic engineering experiments are two examples of great prac-tical interest. The duration of these experiments (typically lastseveral hours) is also compatible with the excellent short-termstability of the SU-8 actuators. The SU-8 microgripper oper-ating in an air environment can alternatively be used for variousmicromanipulation applications such as electronic componentshandling and optical microparts assembly.

ACKNOWLEDGMENT

The authors acknowledge J. Diamond and C. Ionescu-Zanettifor their help with the Hela cells and the time-resolved resistivitymeasurements.

REFERENCES

[1] E. W. H. Jager, O. Inganas, and I. Lundstrom, “Microrobots for microm-eter-size objects in aqueous media: Potential tools for single-cell manip-ulation,” Science, vol. 288, pp. 2335–2338, 2000.

[2] A. Ashkin, “Optical trapping and manipulation of neutral particles usinglasers,” in Proc. Natl. Acad. Sci., vol. 94, May 1997, pp. 4853–4860.

[3] R. Kumar, A. A. Kapoor, and R. H. Taylor, “Preliminary experimentsin robot/human cooperative microinjection,” in Int. Conf. IntelligentRobots and Systems (IROS 2003), vol. 3, 2003, pp. 3186–3191.

[4] C. J. Kim, “Silicon electromechanical microgrippers: design, fabricationand testing,” Ph.D. dissertation, Univ. California, Berkeley, 1991.

[5] C. J. Kim, A. P. Pisano, and R. S. Muller, “Silicon-processed over-hanging microgripper,” J. Microelectromech. Syst., vol. 1, pp. 31–6,1992.

[6] C. G. Keller, “Microfabricated silicon high aspect ratio flexures forin-plane motion,” Ph.D. dissertation, Univ. California, Berkeley, 1998.

[7] C. G. Keller and R. T. Howe, “Hexsil tweezers for teleoperated micro-assembly,” in Proc. 10th Int. Workshop on Micro Electro MechanicalSyst., Nagoya, Japan, 1997, pp. 72–7.

[8] P. C. Shiang and H. Wensyang, “An electro-thermally and laterallydriven polysilicon microactuator,” J. Micromech. Microeng., vol. 7, pp.7–13, 1997.

CHRONIS AND LEE: ELECTROTHERMALLY ACTIVATED SU-8 MICROGRIPPER 863

[9] W. H. Chu and M. Mehregany, “Microfabrication of tweezers with largegripping forces and a large range of motion,” in Proc. Solid-State Sensorand Actuator Workshop, 1994, pp. 107–110.

[10] C. Neagu, H. Jansen, H. Gardeniers, and M. Elwenspoek, “The electrol-ysis of water: An actuation principle for MEMS with a big opportunity,”Mechatronics, vol. 10, pp. 571–81, 2000.

[11] M. C. Carrozza, A. Menciassi, G. Tiezzi, and P. Dario, “The develop-ment of a LIGA-microfabricated gripper for micromanipulation tasks,”J. Micromech. Microeng., vol. 8, no. 2, pp. 141–3, Jun. 1998.

[12] M. Kohl, E. Just, W. Pfleging, and S. Miyazaki, “SMA microgripper withintegrated antagonism,” Sens. Actuators A-Phys., vol. A83, no. 1–3, pp.208–13, May 2000.

[13] I. Roch, P. Bidaud, D. Collard, and L. Buchaillot, “Fabrication and char-acterization of an SU-8 gripper actuated by a shape memory alloy thinfilm,” J. Micromech. Microeng., vol. 13, pp. 330–6, 2003.

[14] S. Butefisch, V. Seidemann, and S. Buttgenbach, “A new micro pneu-matic actuator for micromechanical systems,” in Proc. 11th Int. Conf.Solid-State Sens. Actuators, vol. 1, Berlin, Germany, 2001, pp. 722–5.

[15] J. Ok, C. Milton, and K. Chang-Jin, “Pneumatically driven microcagefor micro-objects in biological liquid,” in Proc. 12th Int. Conf. MicroElectro Mechanical Syst., 1999, pp. 459–463.

[16] Y. Lu and C. J. Kim, “Micro-finger articulation by pneumatic paryleneballons,” in Proc. 12th Int. Conf. Solid-State Sensors, Actuators and Mi-crosystems, 2003, pp. 276–279.

[17] C. Ho-Yin and W. J. Li, “A thermally actuated polymer micro roboticgripper for manipulation of biological cells,” in Proc. Int. Conf. Robot.Automation, 2003, pp. 288–293.

[18] , “A polymer-based micro thermal actuator for micromanipulationsin aqueous environment,” Int. J. Nonlinear Sciences Numer. Simulation,vol. 3, pp. 775–778, 2002.

[19] W. Jennifer, L. Zhou, and J. L. Wen, “MEMS-fabricated ICPF grippersfor aqueous applications,” in Proc. 12th Int. Conf. Solid-State Sensors,Actuators, and Microsyst., p. 2003.

[20] E. W. H. Jager, E. Smela, and O. Inganas, “Microfabricating conjugatedpolymer actuators,” Science, vol. 290, pp. 1540–5, 2000.

[21] H. Lorenz, M. Laudon, and P. Renaud, “Mechanical characterization ofa new high-aspect-ratio near UV-photoresist,” Microelectron. Eng., vol.41–42, pp. 371–374, Mar. 1998.

[22] H. Lorenz, M. Despont, N. Fahrnl, N. LaBianca, P. Renaud, and P. Vet-tiger, “SU-8: A low-cost negative resist for MEMS,” J. Micromech. Mi-croeng., vol. 7, pp. 121–124, Sep. 1997.

[23] L. Dellmann, S. Roth, C. Beuret, G. A. Racine, H. Lorenz, M. Despont,P. Renaud, P. Vettiger, and N. F. de Rooij, “Fabrication process of highaspect ratio elastic structures for piezoelectric motor applications,” inProc. Int. Conf. Solid-State Sens. Actuators, 1997, pp. 641–644.

[24] N. C. LaBianca and J. D. Gelorme, “High aspect ratio resist for thickfilm applications,” in Proc. SPIE – The Int. Soc. Opt. Eng., 1995, pp.846–852.

[25] G. Voskerician, M. S. Shive, R. S. Shawgo, H. von Recum, J. M. An-derson, M. J. Cima, and R. Langer, “Biocompatibility and biofouling ofMEMS drug delivery devices,” Biomaterials, vol. 24, pp. 1959–1967,2003.

[26] Y. Choi, R. Powers, V. Vernekar, B. Frazier, and M. LaPlaca, “High as-pect ratio SU-8 structures for 3-D culturing of neurons,” in Proc. ASMEInt. Mechan. Eng. Congress and Exposition, (IMECE 2003), 2003.

[27] E. W. H. Jager, C. Immerstrand, K. H. Peterson, K. E. Magnusson, I.Lundstrom, and O. Inganas, “The cell clinic: Closable microvials forsingle cell studies,” Biomed. Microdev., vol. 4, no. 3, pp. 177–187, 2002.

[28] M. Li, J. D. Glawe, H. Green, D. K. Mills, M. J. McShane, and B.K. Gale, “Effect of high-aspect-ratio microstructures on cell growthand attachment,” in 1st Ann. Int. IEEE-EMBS Special Topic Conf.Microtechnol. Med. Biol., Lyon, France, 2000, pp. 531–536.

[29] B. A. Weisenberg and D. L. Mooradian, “Hemocompatibility of ma-terials used in microelectromechanical systems: Platelet adhesion andmorphology in vitro,” J. Biomed. Mater. Res., vol. 60, no. 2, pp. 282–91,2002.

[30] H. Guckel, J. Klein, T. Christenson, K. Skrobis, M. Laudon, and E. G.Lovell, “Thermo-magnetic metal flexure actuators,” in Proc. Solid-StateSens. Actuator Workshop, 1992, pp. 73–75.

[31] J. H. Comtois and V. M. Bright, “Surface micromachined polysiliconthermal actuator arrays and applications,” in Proc. Solid-State Sens. Ac-tuator Workshop, 1996, pp. 174–177.

[32] R. Hickey, M. Kujath, and T. Hubbard, “Heat transfer analysis and op-timization of two-beam microelectromechanical thermal actuators,” J.Vacuum Sci. Technol. A-Vacuum Surfaces & Films, vol. 20, pp. 971–974,May 2002.

[33] H. Qing-An and N. K. S. Lee, “Analysis and design of polysiliconthermal flexure actuator,” J. Micromech. Microeng., vol. 9, pp. 64–70,1999.

[34] N. D. Mankame and G. K. Ananthasuresh, “Comprehensive thermalmodeling and characterization of an electro-thermal-compliant microac-tuator,” J. Micromechan. Microeng., vol. 11, pp. 452–462, 2001.

[35] L. Guerin, M. Bossel, M. Demierre, S. Calmes, and P. Renaud, “Simpleand low cost fabrication of embedded microchannels by using a newthick-film photoplastic,” in Proc. Transducers, 1997, pp. 1419–1422.

[36] F. Lin, G. S. Bathia, and J. D. Ford, “Thermal conductivities of powder-filled epoxy resins,” J. Appl. Polymer Sci., vol. 49, pp. 1901–1908, 1993.

[37] N. Chronis, “Lab-on-a-chip mechanical and optical components for themanipulation and detection of biological species,” Ph.D., Univ. Cali-fornia, Berkeley, 2004.

[38] T. F. Kosar, N. L. Stucky, C. Chen, K. J. Kim, and A. Folch, “Nanoholearrays for parallel patch-clamping and focal delivery of biochemical fac-tors to cells,” in Proc. 7th Int. Conf. Micro Total Analysis Syst., 2003, pp.25–28.

[39] S. Yu and B. J. Nelson, “Autonomous injection of biological cells usingvisual servoing,” Lect Notes Contr. Inf., vol. 271, pp. 169–178, 2001.

Nikolas Chronis received the B.S. and Ph.D. degreesfrom Aristotle University, Greece, and the Universityof California, Berkeley, in 1998 and 2004, respec-tively, both in mechanical engineering.

In 2000, he joined the Berkeley Sensor and Ac-tuator Center, University of California, Berkeley, asa graduate student researcher under the supervisionof Luke Lee. His research interests include polymerMEMS actuators and micromachining, microfluidicsand optical microsystems for lab-on-chip applica-tions.

Luke P. Lee received both the B.A. degree inbiophysics and the Ph.D. degree in applied scienceand technology: Applied Physics (major)/ Bioengi-neering (minor) from the University of California atBerkeley (UC Berkeley).

He has more than ten years of industrial experiencein integrated optoelectronics, holographic lithog-raphy, high-power surface emitting laser diodes,superconducting quantum interference devices(SQUIDs), and nanomagnetic bioassays. His currentresearch interests are quantum plasmonic integrated

nanocrescents (QPINs) for biomolecular imaging, biologic application specificintegration circuits (BASICs) for quantitative systems biology and moleculardiagnostics. At UC Berkeley, he has developed polymer micromachining forBioMEMS, biophotonic MEMS, bioPOEMS (biopolymer optoelectrome-chanical systems) for lab-on-a-chip and microscale-confocal-imaging-arrays(micro-CIAs), nano-POEMS for single molecule detection and manipulation,neural interfaces, and nanogap DNA junctions-based bioelectronics. He hasauthored and coauthored over 120 publications on nanoplasmonic SERSprobes, cell-based biochips, bioMEMS, biophotonics, integrated optics, mi-cromachined surface emitting laser diodes, SQUIDs, biomagnetic sensors, andnanogap DNA junctions for label-free DNA detection.