Edwin T. Carlen

Electrothermally actuated inline microfluidic valve

A normally open electrothermally actuated inline microvalve that has been developed, fabricated and tested with liquids is presented. These actuators use high volumetric expansion of a sealed patch of Paraffin heated above its melting point, providing large displacements and forces while using low power. The inline valve is surface micromachined on top of preformed flexible microfluidic channels using a low temperature fabrication process; therefore it is suitable for integration with microfluidic networks requiring actuation of a large number of independent valves under electrical control. Complete closure of sealed microchannels has been observed with power as low as 40 mW. Response times of 15 ms have been measured. Breakdown of the inline valve occurs at an upstream pressure of 23 psig.

Electrothermally activated paraffin microactuators

A new family of electrothermally activated microactuators that can provide both large displacements and forces, are simple to fabricate, and are easily integrated with a large variety of microelectronic and microfluidic components are presented. The actuators use the high volumetric expansion of a sealed, surface micromachined patch of paraffin heated near its melting point to deform a sealing diaphragm. Two types of actuators have been fabricated using a simple three mask fabrication process. The first device structure consists of a 9 /spl mu/m thick circularly patterned paraffin layer ranging in diameter from 400 to 800 /spl mu/m all covered with a 4-/spl mu/m-thick metallized p-xylylene sealing diaphragm. All fabricated devices produced a 2.7-/spl mu/m-peak center deflection, consistent with a simple first order theory. The second actuator structure uses a constrained volume reservoir that magnifies the diaphragm deflection producing consistently 3.2 /spl mu/m center diaphragm deflection with a 3-/spl mu/m-thick paraffin actuation layer. Microactuators were constructed on both glass and silicon substrates. The actuators fabricated on glass substrates used between 50-200 mW of electrical power with response times ranging between 30-50 ms. The response time for silicon devices was much faster (3-5 ms) at the expense of a larger electrical power (500-2000 mW).

Surface micromachined paraffin-actuated microvalve

Normally-open microvalves have been fabricated and tested which use a paraffin microactuator as the active element. The entire structure with nominal dimension of /spl phi/600 /spl mu/m /spl times/ 30 /spl mu/m is batch-fabricated by surface micromachining the actuator and channel materials on top of a single substrate. Gas flow rates in the 0.01-0.1 sccm range have been measured for several devices with actuation powers ranging from 50 to 150 mW on glass substrates. Leak rates as low as 500 /spl mu/sccm have been measured. The normally-open blocking microvalve structure has been used to fabricate a precision flow control system of microvalves consisting of four blocking valve structures. The control valve is designed to operate over a 0.01-5.0 sccm flow range at a differential pressure of 800 torr. Flow rates ranging from 0.02 to 4.996 sccm have been measured. Leak rates as low as 3.2 msccm for the four valve system have been measured.

Paraffin actuated surface micromachined valves

A new, active, normally-open blocking microvalve that uses the thermal expansion of a sealed, thin paraffin patch for actuation has been fabricated and tested. The entire structure is batch-fabricated by surface micromachining the actuator and channel materials on top of a single substrate. The paraffin actuated microvalves are suitable for applications requiring many devices on a single die, low processing temperatures, and simple, nonbonded process technology. Gas flow rates in the 0.1-2.0 sccm range have been measured for several devices with actuation powers less than 50 mW.

Recent progress in microfluidic devices for nucleic acid and antibody assays

Microfluidic devices are increasingly important in biomedical assays. In this paper, we review recent progress on these devices for chemical amplification, hybridization, separation, and detection of nucleic acids. Recent developments in microfluidic technologies for detection of protein binding events in immunoassay applications are also reviewed.

High-aspect ratio vertical comb-drive actuator with small self-aligned finger gaps

A vertical comb-drive actuator with thin, high-aspect ratio comb fingers and small self-aligned gaps is presented. Key to the actuator design is the self-aligned, offset comb-drive fingers, which are fabricated with small gaps (/spl les/2 /spl mu/m) using a single lithography step. The offset comb fingers are fabricated using two thick conducting layers separated by a thin dielectric layer. The lower fingers are formed from the device layer (20 /spl mu/m) of an silicon-on-insulator (SOI) substrate, while an in situ -doped polysilicon layer (20 /spl mu/m), deposited in an epitaxial reactor (Epipoly), is used for the upper comb fingers. The Epipoly films have been optimized and characterized for application as structural and electrical components. The offset comb fingers are formed using a combination of deep-reactive ion etching (DRIE), thin oxide barrier layer growth, and an isotropic dry silicon etch (XeF/sub 2/) of selected areas of the Epipoly layer. The actuator has been implemented in a high fill-factor (>90%) micromirror array for optical telecommunications applications. Large continuous scan angles (/spl plusmn/10/spl deg/) with actuation voltages <60 V have been measured with no pull-in phenomena observed.

Automatic generation of thin film process flows. I. Basic algorithms

This paper is the first in a series of two papers describing the algorithms used in the development of MISTIC (Michigan synthesis tools for integrated circuits). MISTIC is a planar device process compiler that generates process flows for thin film devices from schematics of their structure. This software uses a laboratory specific database of process recipes to produce process flows for a specific set of laboratory resources (furnaces, etchers, lithography equipment, etc.) and generates process statistics that help to choose the most suitable process flow in a comparative manner. The process compiler is augmented by several auxiliary modules: a device builder, process viewer, and database editor thus forming a self-contained process design environment. This paper concentrates on the algorithms used to construct process flows from schematic device representations. The compiler algorithms first extract a directed graph representation of the device organization stored in the form of a restricted square boolean matrix. This matrix is used to generate linear ordered lists of device layers which serve as footprints for the construction of process flows. Process flows are then constructed from these lists through a series of conversions, expansions, and insertions of process steps.

Automatic generation of thin film process flows. II. Recipe generation, flow evaluation, and system framework

This paper is the second in a series of two papers describing the methodology and algorithms used in the development of MISTIC (Michigan synthesis tools for integrated circuits). Part I [see ibid., vol. 12, no. 1, Feb. 1999] discussed the basic topological algorithms used to produce generic sequences of processing steps required for the fabrication of a given device structure. Part II discusses the expansion of these sequences into complete process flows. This procedure involves the selection of specific recipes from a set of available processing resources and the calculation of recipe parameters. These processing resources are stored in a database central to the MISTIC system framework. Since many process flows are generated for a given device, the paper also discusses the calculation of suitable figures of merit. The capabilities of the MISTIC system are demonstrated with a BiCMOS example. The MISTIC system framework which contains the basic compiler and several supporting modules: a device builder, process viewer, and database editor is also presented.

VLSI implementation of a feature mapping neural network

Modification of Kohonens self-organizing feature map algorithm and its dedicated parallel hardware implementation are the focus of this paper. This work is motivated by the need to implement a 5/spl times/5 neural network using digital standard cells and high level VLSI system design tools. The neural net considered is a two layered, feed forward architecture that learns relationships among unknown input data patterns. The prototype system consists of 25 processing units (neurons). Each processing unit operates at 10 MHz. Communication among processing units is accomplished using a broadcast bus. Performance of the system is estimated to be 110,000 iterations per second.<<ETX>>

Statistical model for spatial correlation in thin film deposition and reactive growth

This paper presents a statistical method for the estimation of thickness variations present across a wafer lot in low pressure chemical vapor deposition (LPCVD) and reactively grown films. The method uses experimental thickness data to construct a unified Karhunen-Loeve expansion based model that captures both deterministic and random thickness variations. The model uses a set of quadratic interpolation functions fitted to mean spatial data to approximate the deterministic nonuniformity and a few normalized random variables to represent run-to-run fluctuations. This model therefore retains the spatial correlations present between different deposition and growth steps in a process necessary for the estimation of parametric yield and permits the calculation of distribution functions over different lot populations (wafer, die, point, etc.). Models for spatial correlations in LPCVD oxide, nitride, polycrystalline silicon, and thermal oxide growth were constructed from a data set of 35000 thickness measurements recorded from a total of 40, 25-wafer runs. In each case, the model gives good predictions (90-95% confidence) with just one or two random variables.