0
0
By K. Current, K. Yuk, C. McConaghy, P. Gascoyne, J. Schwartz, J. Vykoukal, C. Andrews
Originally appears: 2005 Symposium on VLSI Circuits Digest of Technical Papers
A high-voltage (HV) SOI CMOS VLSI chip has been demonstrated to transport droplets on programmable paths across its coated surface. This HV exciter for a fluidic lab-on-a-chip system creates dielectrophoretic forces that move and help inject droplets. Electrode excitation voltage and frequency are variable: at 100V, fELECmax = 200Hz. Data communication rate is variable up to 250kHz. This 10,377μm-by-8210μm demonstration chip has a 32×32 array of nominally 100V electrode drivers dissipating 1.87W max.
Introduction
The programmable fluidic processor (PFP), a fluidic analog of a microprocessor, has a program memory that defines the movement of droplets or biological cells around the channel-less array of identical electrodes on the chip surface. Samples of various droplet sizes are accommodated and may be injected onto the chip surface above the electrode array from MEMS sample chambers to react with reagents and other chemicals. The electrode voltage excitation creates forces by employing dielectrophoresis (DEP) to move the sample and reagent droplets along arbitrary programmable paths across the electrode array [1]. These multiple droplets are moved simultaneously to the same array location and combined, forming a chemical reaction indicative of characteristics of the original sample. The resultant droplet is moved by DEP to a location for characterization. Motion and reaction results are now sensed optically off-chip. On-chip electronic location and characteristic sensing are in development. While PFP uses high-voltage excitations for droplet manipulation, several low-voltage lab-on-a-chip devices employing DEP on planar microelectrode arrays have been reported in recent years [2,3,4].
The PFP is programmable in several senses. Since the volumes and masses of various size biological cells and chemical molecules require various DEP forces to be moved, the PFP can accommodate them by having variable electrode excitation voltage amplitude, phase, and frequency. The path taken by each droplet on the array, the rate of movement across the chip surface, and the loading speed of the array memory are all variable, allowing complete flexibility for multiple reactions in chemical and biological fluid analyses.
Some advantages that this micro-scale PFP analyzer provides are the wider availability of valuable diagnostic tools; cheap, disposable units; reduced costs of chemicals and samples due to volume scaling; and versatility by testing multiple agents with a generic PFP analyzer type. Determination of chemical and cell characteristics using μL volumes, such as the sorting of cancer cells by electronic means has been demonstrated [5].
The VLSI electrode exciter chip is the engine for the programmable fluidic processor system summarized below. The chip architecture is an expandable NxN electrode array. The 10,377μm x 8210μm demonstration chip with 32×32 cell array has an electrode excitation voltage and frequency that are variable: at 100V, fELECmax = 200Hz. Data communication rate is variable up to 250kHz. The array of nominally 100V electrode drivers dissipates a maximum of 1.87W.
Theory
DEP is the movement of neutral but polarizable particles under the influence of an inhomogeneous electric field. In positive DEP, a polarizable particle within a non- or less polarizable medium is polarized in the direction of an applied inhomogeneous electric field, creating a greater net force and movement towards the higher electric field gradient. In negative DEP, the particle is non- or less polarizable than its surrounding medium, causing the particle to be pushed away from the higher electric field gradient.
DC electric fields can be used to generate DEP forces, but AC electric fields are preferred to exploit the dielectric properties of the particle. The polarization direction of the particle’s charges exhibits an AC behavior when subject to an AC electric field. However, the resulting force can vary depending on the frequency and dielectric properties of the particle. The time averaged DEP force experienced by a particle can be expressed by
where ε m is the permittivity of the medium, a is the radius of the particle, Erms is the electric field and Ke is the Claussius Mosotti factor, representing the complex permittivities of the particle [1,3,4]. The contribution of the complex permittivities of the particle, ε *p, and of the medium, ε *m, are contained in Ke, which can be expressed by
where
and
Therefore, the DEP force experienced by the particle and thus its movement in response to that force is a function of the frequency dependent dielectric properties of the particle and the medium and can be used for particle characterization and identification as well as manipulation [1,5].
Programmable Fluidic Processor System
The PFP electrode driver chip and MEMS structure cross-section is in Fig. 1 [1]. The MEMS structure holds the fluid reservoirs and the fluid micro-channels, and seals the edges of its reaction chamber (the open area of the MEMS structure directly above the chip electrode array) to the passivation oxide of the CMOS chip surface. The passivation oxide is coated with a thin hydrophobic layer to prevent wetting and contamination of the droplet movement surface [1]. The reaction chamber is filled with an inert, insulating oil. The fluid reservoirs have injection channels that open in the oil-filled reaction chamber. The sample fluids and test chemicals are loaded into the reservoirs and then injected into the reaction area by DEP forces [1,6]. The electronics for each electrode site lie beneath the respective electrode.
The versatile capabilities of the PFP analyzer system are designed into the PFP chip. The electrode size was selected to be 100μm squared on the top metal layer each separated by 100μm, in a regular array, based upon DEP and fluid properties [1]. The chip architecture allows arbitrary expansion of the electrode array in both directions of the array, allowing easy connection in groups of chips to provide larger droplet manipulation surfaces. For demonstration, this chip uses a 32×32 array of identical high-voltage (HV) electrodes and the appropriate standard CMOS logic.
The PFP electrode array generates the forces for DEP droplet manipulation. Each electrode driver in the array can be programmed to produce either a 0-degree (in-phase) or 180-degree (out-of-phase) phase delayed square-wave. An AC square-wave is necessary to exploit the frequency-dependant properties of the sample, while high-voltage may be necessary for correct actuation in environments exposed to shock and varying orientations. The array’s phase configuration produces controlled inhomogeneous electric fields confining a droplet to a particular electrode site or transporting it between sites [1]. For example, a fluid droplet will stay atop an in-phase electrode if the surrounding electrodes are out-of-phase. However, if an adjacent electrode is changed to in-phase and the residing electrode is changed to out-of-phase, the droplet will move away from the out-of-phase electrode towards the in-phase one. PFP array programming configures the electrodes’ signal phases, resulting in precise droplet manipulation along a defined path, sectioning of the array for parallel manipulations, and grouping of multiple electrodes for handling larger droplet volumes and combining or separating droplets [1].
32 x 32 Demonstration Architecture
Fig. 2 shows the 32×32 demonstration PFP chip architecture. There is a HV electrode driver array, and other functions are standard CMOS logic. The communications circuitry consists of an 11-bit communications shift register (SR) with gated outputs, a 5-bit row decoder, and a 5-bit column decoder, for programming the array.
A. High Voltage Electrode Digital Driver Cell
Each electrode has an underlying driver cell as shown in Fig. 3 consisting of two D latches (DL) in master-slave flipflop operation that store the phase data, addressing logic controlled by a row (ROW) line and a column (COL) line, and a HV electrode driver. The HV electrode driver consists of a 2:1 mux controlled by the flip-flop outputs C and CB, a HV NMOS transistor connected to the high voltage supply VHH through a 5MΩ load resistor and an electrode connected to VOUT. The electrode output VOUT will be out-of-phase with the reference square wave VIN if C is logic low and in-phase if C is logic high. VIN sets the frequency of VOUT. An analog high-voltage electrode driver circuit has also been demonstrated for creating high voltage sinusoidal electrode excitations.
The electrode array’s next phase configuration is set by sequentially programming each site with the new phase data and then updating all phase outputs of the array simultaneously. A target driver is programmed by feeding an 11-bit data sequence consisting of the 5-bit column address, 5-bit row address and 1-bit phase data into the DATA line of the SR controlled by CLK. After the data is set, addresses are passed to the decoders, and the phase data D is passed to the array cells by activating the gated output of the SR with the LOAD signal. The address enables only the target driver cell to accept the new phase data. When EN1B activates, the first DL of the target driver cell stores the new phase data from the D line and the memory is programmed. This entire programming sequence is repeated for each desired array cell. When all new data has been loaded, EN2B is activated, the second DL of each driver cell loads the new phase data and the output of the entire array is simultaneously updated with the new phase configuration.
B. DEP-Assisted Droplet Injection Electronics
Fluid injectors are to be mounted on the right side of the chip, overlapping the reference electrode with tips pointing close towards the last column of electrodes in the array (COL31). In the injection scheme, a DEP force creating the supplemental pressure needed to bring fluid from the injector tip onto the chip surface is generated by a potential difference between a grounded electrode in the last column of the array serving as the collecting electrode and the reference electrode exerting an arbitrary signal.
Any electrode in COL31 can be used for collection as each has a programmable injection driver, which pulls the output to a low DC voltage close to ground when activated. The injection driver circuitry is shown connected to its corresponding COL31 electrode in Fig. 4 and is composed of a HV NMOS, a flip-flop and addressing logic. Each injection driver shares the address of its adjacent driver cell but receives data from the INJ input. Programming an injection driver with a logic high INJ data turns on the HV NMOS, which pulls the electrode output towards ground. The required signal can then be applied onto the injection reference electrode to induce droplet injection.
Test Measurements and Results
The die photo of the PFP chip is shown in Fig. 5. A 1.0μm high-voltage, high-resistance, 1 poly, 3 metal, SOI CMOS fabrication technology is used.
Fig. 6 shows the electrode output at location (31,31) for two programming sequences. The top waveform shows the output square-wave change phases and then resetting to the initial phase during a 180-degree to 0-degree phase program and memory reset at 100V at 100Hz. The bottom waveform shows the output transitioning from a square-wave to ground state during a 180-degree to ground state program at 100V at 100Hz. Additional testing has shown operation of up to 200Hz at 100V and higher speeds are possible in the absence of a test probe.
Fig. 7 shows a droplet of 180nL phosphate buffered saline in a 1-bromododecane medium being manipulated by the PFP chip coated with ~5μm of SU-8 topped with a monolayer of FluoroPel 1604 with microvinyl granules about 4μm across [1]. From t=0s, the droplet is moved downward until t=10s, to the right until t=42s, upward until t=88s, and to the left until t=115s. Table 1 shows the specifications and measured performance of the PFP analyzer chip.
Conclusion
A VLSI Programmable Fluidic Processor Excitation chip has been demonstrated to precisely control the movement of droplets across its surface. This 10,377μm x 8210μm 32×32 demonstration chip is fabricated in a standard high-voltage 1μm SOI CMOS technology capable of 100V operation, and thus may be inexpensively mass-produced. Detection of droplet location and evaluation of droplet characteristics are presently done optically off-chip. On-chip electronic droplet location detection and electronic evaluation of droplet characteristics are under development. These developments will soon allow the deployment of inexpensive wristwatch sized field chemical assay devices.
Acknowledgment
Research funded in part by DARPA contract N66001-97-C-8608 and the Army Research Office DAAD19-00-1-0515. The authors gratefully acknowledge contributions to this research by Harriet Lam, Alec Wong, and Peter Krulevitch.
References
[1] P. Gascoyne, J. Vykoukal, J. Schwartz, T. Anderson, D. Vykoukal, K.W. Current, C. McConaghy, F. Becker and C. Andrews, “Dielectrophoresis-based programmable fluidic processors,” Lab on a Chip, pp. 299-309, 2004.
[2] N. Manaresi, A. Romani, G. Medoro, L. Altomare, A. Leonardi, M. Tartagni, and R. Guerrieri, “A CMOS Chip for Individual Cell Manipulation and Detection,” IEEE Journal of Solid-State Circuits, Vol. 38, No. 12, pp. 2297-2305, December 2003.\
[3] J.R. Keilman, G.A. Jullien and K.V.I.S Kaler, “A SoC bioanalysis platform for real-time biological cell analysis-on-achip,” Proceedings of The 3rd IEEE International Workshop on System-on-Chip for Real-Time Applications, IEEE, 2003.
[4] Y. Li and K.V.I.S. Kaler, “DEP based cell separation utilizing planar microelectrode array,” 2002 Annual Report Conference on Electrical Insulation and Dielectric Phenomena, IEEE, 2002.
[5] P Gascoyne, X Wang, Y Huang, F Becker, “Dielectrophoretic separation of cancer cells from blood ,” IEEE Trans Industry Applications, V 33, n 3, pp. 670 – 678, May-June 1997.
[6] J. Vykoukal, J. Schwartz, F. Becker, P. Gascoyne, “A programmable dielectrophoretic fluid processor for droplet-based chemistry,” Micro Total Analysis Systems 2001 , pp. 72-74, 2001.
