Introduction

Research has been a pretty integral part of my work and design experience. I’ve been able to train on industry equipment at smaller scales, design a system from scratch and build it up for evaluation, and tackle my specific role in collaboration with others. Also interestingly, it’s really taught me the importance of storytelling - every step of the research process requires it. Proposals to secure funding, stump speeches to get collaborators, attending conferences, and lastly publishing. Then the process starts all over again.

I got involved in research pretty early on - a stint assisting with animal surgeries in an immunology lab, and a taste of independent additive manufacturing research. After declaring for ECE, I began my current research: making and testing biosensors, TEP devices, and OLEDs.

Note: This section will only include brief project context and parts that I was directly involved with in NEDL. there’s a lot more going on not shown that I’d be more than happy to answer any questions about here. All figures were made by me.

Nano-Biosensors

Designing, fabricating, and characterizing biosensors is my primary project. For better or for worse, the comparison I’ve found that helps people understand the best what I do is with Theranos. The conceptual goal is to create a device that uses trace amounts of blood for rapid disease diagnosis - the major difference is that mine has evidence of working (and that I unfortunately did not receive millions in funding). Here is a holistic physical description:

The device consists of 3 parts: a flexible PET substrate, gold electrodes, and a polyaniline sensing channel. Polyaniline, a p-type organic semiconductor, is functionalized with a series of crosslinkers to attach anti-BNP receptors. The target molecule is BNP (The target molecule is b-type natriuretic peptide), which is a protein released by the heart when its ventricles undergo strain - as such it is a well established biomarker for early indication of cardiovascular disease.

biosensor model

Theory

At its core, the device is an ion selective field effect transistor, or ISFET. Using the standard transistor analogy, the ISFET acts like a valve controlling the flow of current, except its operated with target molecule concentration instead of direct gate voltage. They’re usually used as pH sensors given the ease of measuring H+ concentration, but this idea can be extended to any charged molecule including different biomarker proteins with the inclusion of an organic channel path to accommodate functionalization of antibodies.

ISFET mechanism

Above is a simplified visualization of what is happening. A blood sample is placed over the device, and an amount of target molecules proportional to their concentration conform to the antibodies and induce a polarization in the semiconducting layer. This increase in mobility can be observed by measuring the current flow in response to an applied electrode voltage. Using this result, current response can be mapped directly to biomarker concentration.

Realization & Testing

The biosensors are fabricated with a bottom-up process in an ISO4 (class 10) cleanroom at the NFCF. To begin, PET is cut in the shape of standard 4” wafer. Lithography involves standard spin-coating and maskless exposure. The first layer is a mixture of gold-chromium electrodes optimized for flexibility deposited via electron beam deposition. The second layer is the polyaniline bridge which is chemically synthesized in an ice bath to precisely control thickness. Antibodies are deposited directly on and left to adhere for 6 hours before a solution of bovine serum albumin (BSA) is deposited to increase specificity 30 minutes before testing. All solutions were diluted in phosphate buffer solution to optimize for Debye-screening length.

PET wafer	Au-Cr electrodes	PANI film	Functionalization

Pictured above is a wafer containing 150 sensors. For testing, a 2uL droplet of UPMC patient plasma is deposited directly onto the polyaniline channel. A potentiostat generates a linear voltage sweep across the electrodes. The gate is biased with -0.4V by an external power source. The recorded current and reference charge response is then analyzed. Below is an example output from samples across 80 patients:

Current Response	Reference Charge Response

Results & Analysis

Initial testing was conducted with artificially spiked samples, and after success subsequent samples were collected from UPMC patients through the UPMC Cardiovascular Institute. All results here are in association with the latter phase.

Baseline Testing	Blind Sample Testing

First a batch of healthy subject samples were tested as a control to establish a baseline. Then, a mix of healthy and at-risk patient samples were blindly tested. the results on the right show linear separation between samples depending directly on sample BNP concentration, with relatively little spread outside data classes.

Flexibility Query	Hemolysis Query

Additional device testing was also conducted to evaluate the potential of the sensors as a wearable. As a thin film device capable of structural flexibility, the effect of different bending radii were shown to have insignificant effect on signal fidelity, with strong restoration ability. The size of the sensors ensure that no smaller radii were needed. On the right is a brief characterization to observe the effects of hemoglobin concentration of sample conductivity. The figure shows a difference heat-map, where the darkest red displays a difference of 0.4pA from the green, indicating that iron in blood is strongly solvated and inconsequential to measurements

Principle Component Analysis	Quadratic Discriminant Analysis

An additional 84 patient sample responses were analysed with PCA and QDA. Results demonstrated high success with binary classification of data. Accuracy should be well above 95% when multiple redundant samples are taken (given 2uL per sample). Additionally, this result could easily be used to develop a visual classifier based on output signal strength.

Micro-TEP Devices

This is a secondary project that I assist a colleague with. TEP stands for thermo-electric power, or simply converting heat gradients directly into electrical energy. Naturally, the aim is to create devices to place in areas with steep temperature changes to generate power. The project is currently in the optimization and formula tweaking phase.