Selected Publications
High-temperature microfluidic devices (such as gas chromatography microcolumns) have traditionally been fabricated using photolithography, etching, and wafer bonding which allow for precise microscale features but lack the ability to form complex 3D designs. Metal additive manufacturing could enable higher complexity microfluidic designs if reliable methods for fabrication are developed, but forming small negative features is challenging-especially in powder-based processes. In this paper, the formation of sealed metal microchannels was demonstrated using stainless-steel binder jetting with bronze infiltration. To create small negative features, bronze infiltrant must fill the porous part produced by binder jetting without filling the negative features. This was achieved through sacrificial powder infiltration (SPI), wherein sacrificial powder reservoirs (pore size similar to 60 mu m) are used to control infiltrant pressure. With this pressure control, the infiltrant selectively filled the small pores between particles in the printed part (pore size similar to 3 mu m) while leaving printed microchannels (700 mu m and 930 mu m) empty. To develop the SPI method, a pore filling study was performed in this stainless-steel/bronze system with 370 mu m, 650 mu m, and 930 mu m microchannel segments. This study enabled SPI process design on these length scales by determining variations in pore filling across a sample and preferential filling between different sized pores.
A miniaturized short-wavelength infrared spectrometer for use with diffuse light was created by combining a thin form factor carbon nanotube composite collimator, a linear variable filter, and an InGaAs photodiode array. The resulting spectrometer measures 3 mm × 4 mm × 14 mm and shows a significant improvement in resolution over a spectrometer without the collimator when used with diffuse light. Its small size and high throughput make it ideal for applications such as wearable optical sensing, where light from highly scattering tissue is measured. Plethysmographic measurements on the wrist were demonstrated, showing rapid data collection with diffuse light.
The ability to monitor a varying impedance has a range of applications, including the measurement of biological properties using bioimpedance analysis. For this type of impedance monitoring, the human heartbeat plays a role, motivating a desire to monitor pulsatile impedance changes. A four-point circuit for pulsatile impedance monitoring is designed, simulated, and built on a PCB. The circuit design is described. The circuit's ability to measure constant impedance across frequency and extract lumped element values is characterized. Using a photoresistor setup, the circuit's response to pulsatile impedance variation ranging from 500 Ω to 70 kΩ is measured and analyzed. The measured circuit settling time for an impedance change as large as 70 kΩ is 40 milliseconds, sufficient speed for heartbeat-rate pulsatile impedance monitoring.
Estimating the elastic modulus and strength of heterogeneous films requires local measurement techniques. For local mechanical film testing, microcantilevers were cut into suspended many-layer graphene using a focused ion beam. An optical transmittance technique was used to map thickness near the cantilevers, and multipoint force–deflection mapping with an atomic force microscope was used to record the compliance of the cantilevers. These data were used to estimate the elastic modulus of the film by fitting the compliance at multiple locations along the cantilever to a fixed-free Euler–Bernoulli beam model. This method resulted in a lower uncertainty than is possible from analyzing only a single force–deflection. The breaking strength of the film was also found by deflecting cantilevers until fracture. The average modulus and strength of the many-layer graphene films are 300 and 12 GPa, respectively. The multipoint force–deflection method is well suited to analyze films that are heterogeneous in thickness or wrinkled.
Directed placement of DNA origami could play a key role in future integrated nanoelectronic devices. Here we demonstrated the site-selective attachment of DNA origami on gold dots formed using a pattern transfer method through block copolymer self-assembly. First, a random copolymer brush layer is grafted on the Si surface and then poly (styrene-b-methylmethacrylate) block copolymer is spin-coated to give a hexagonal nanoarray after annealing. UV irradiation followed by acetic acid etching is used to remove the PMMA, creating cylindrical holes and then oxygen plasma etching removes the random copolymer layer inside those holes. Next, metal evaporation, followed by lift-off creates a gold dot array. We evaluated different ligand functionalization of Au dots, as well as DNA hybridization to attach DNA origami to the nanodots. DNA-coated Au nanorods are assembled on the DNA origami as a step towards creating nanowires and to facilitate electron microscopy characterization of the attachment of DNA origami on these Au nanodots. The DNA hybridization approach showed better DNA attachment to Au nanodots than localization by electrostatic interaction. This work contributes to the understanding of DNA-templated assembly, nanomaterials, and block copolymer nanolithography. Furthermore, the work shows potential for creating DNA-templated nanodevices and their placement in ordered arrays in future nanoelectronics.
Hollow cathode plasmas are common extreme ultraviolet (EUV) lamps used for material characterization. However, the relatively high pressure of the plasma can affect downstream instruments, as well as absorb the EUV. EUV windows are difficult to fabricate due to EUV’s strong interaction with all materials. We present a carbon nanotube (CNT) microfabricated window composed of multiple high aspect-ratio columns in parallel. The open areas allow wide bandpass transmission, while the walls restrict gas flow. We model the CNT window transmission as a weight function on the light from of a Mcpherson 629-like hollow cathode helium plasma in visible wavelengths. We model the CNT window differential pumping as a series of columns between two chambers of different pressures.