Final Project

Breathalyzer for Lung Cancer Detection

For my Senior thesis project I'm working with the Aizenberg lab at Harvard to sense Volatile Organic Compounds indicative of lung cancer in human breath with chemiresistive sensors. Research has shown that you can distinguish between someone with lung cancer and a healthy control based on the signature of VOCs in one's breath. This has been done largely via Gas Chromatography Mass Spectrometry which is expensive and inaccessible equipment.

Thesis Project Scope

The goal of the project is to detect VOCs indicative of lung cancer with chemiresistive sensors to see if lung cancer can be diagnosed easily and inexpensively. The project is chemistry-focused... we're doing testing on arrays of sensors with various VOCs to see if we can achieve a binary classification with 80% accuracy. We've set up a testing rig in the lab and are testing with single gasses/refining our sensor selection. We're soon to move on to mixtures of gasses. This is a difficult problem for a number of reasons. The chemiresistive sensors are not specific--they're sensitive to multiple VOCs. We will have to employ some complex data processing to tackle this. Also, the concentration of the VOCs in the breath is less than the sensors are able to register. This will require preprocessing/selective concentration of the target VOCs which is very difficult. We're hoping to address this with adsorption and thermal desorption or pressure-based concentration.

HTMAA Scope

For my HTMAA project I want to work on device-ifying this technology. I want to start by very simply making an alcohol breathalyzer with an MQ3 sensor. I'm also considering making a breathalyzer with a Hydrogen Sulfide sensor that could detect if you have garlic breath. Ultimately I want to use what I learn in this class and apply it to my thesis, allowing me to create a hand-held breathalyzer-like device that could theoretically be used for lung cancer diagnosis. This is substantially different than the focus of my design project from the lab. This is also certainly preemptive given that we're not sure if we can even detect lung cancer with the chemiresistive sensors, but I think it would be a great project. I would become familiar with system integration, miniaturization, and Embedded System Development. It would also allow me to test with pre-processing human breath to remove noise/moisture.

Midterm Review

Diagram:

Diagram of final extruded model Diagram of final extruded model

Early Prototyping:

First prototype image Second prototype image Third prototype image Fourth prototype image

Tasks to be completed:

Schedule:

Final Project Build

For my final project, I decided to build a breathalyzer. This idea is adjacent to my thesis project. I'm working on my thesis with the Aizenberg lab at Harvard, trying to create a device capable of diagnosing diseases in human breath samples. The device would contain an array of chemiresistive sensors. Mixtures of gasses (or actual breath samples) would be passed over the sensors, and time series data from the sensors would be analyzed by a neural network to determine which gasses are present and whether the breath sample is indicative of a certain disease.

Thesis project concept block diagram

Diagram showing the basic concept of the thesis project.

I started by purchasing a breathalyzer and dissecting it to understand how it works and how its components are laid out.

Dissected breathalyzer

Dissected breathalyzer components.

With the extruded view from earlier on this page in mind, I created a second iteration of the device housing. This version included mounting holes, access holes, and slots for specific components.

Device housing design

Device housing with mounting holes and slots.

Front plate design

Front plate design for the device.

Back plate design

Back plate design for the device.

I then began conceptualizing how the circuit would work based on the input and output devices.

Circuit design diagram

Conceptualized circuit for the device.

I started by breadboarding each component individually and then integrated all components onto a single breadboard to test the planned circuit and begin working on the code.

Breadboard with integrated components

Breadboard setup with all components integrated.

From there, I moved on to designing the circuit in KiCad.

KiCad design

Circuit design in KiCad.

I then milled the PCB from a copper plate using the Roland SRM-20 milling machine.

Milled PCB

Milled PCB.

After that, I reflow soldered the components onto the PCB and passed it through the reflow oven.

Reflow soldered PCB

Reflow soldered PCB.

Once the PCB was ready, I mounted all the components and plugged them into their respective headers. Below is an image of the fully integrated device.

Fully integrated device with components visible

Fully integrated device with components visible.

Completed device with front plate added

Completed device with the front plate added.

During this process, I had to do some debugging to refine the code. Below is the complete code I wrote in the Arduino IDE to control the device:

You can download the complete code here.

The code worked well, and the image below explains its functionality.

Code functionality explanation

Diagram explaining the functionality of the device code.

The device worked well overall, though I encountered some challenges with sensor drift. To calibrate the device, I sat down with six beers, periodically blowing into it to create a function that converts MQ3 readings to BAC values.

Here is the final video of the project:

Reflection

This project was a long and difficult process, but it was incredibly rewarding. It gave me valuable hands-on experience with system integration, circuit design, and debugging. I'm confident that the skills I developed here will help me progress on my thesis and future endeavors.

Alternative Ideas

Laser cut and soldered Stained Glass Lamp

I've been interested in making glass art for a long time. I was inspired by the laser Pidgeon-cage to want to make a stained glass lamp. I think I would be able to cut the glass using the laser cutter or the water jet, and then I could use copper foil and special solder to put it together.

Inductive Charger

The thought of inductive charging has been pretty interesting to me, particularly in its application to the automotive industry. I thought that this might be a fun project to embark on.

Wearable Technology

Nicotine Injector (Anthony's Idea!)

Anthony has the idea of making a dermal nicotine injector. I think this is an interesting idea for drug delivery, and it would be fun to work together on. I wonder if this could be used to wean people off nicotine.

Earring

I asked myself the question "At what point are you a cyborg?" I figured you are a cyborg if you have technology inside of your body. Thus if you have a pacemaker I think you are a cyborg. I think that an earring also meets this criteria.

I'm not sure what my wearable tech earring would do exactly, but I think perhaps it could have a microphone or camera in it, and it could listen to everything you say and every conversation you have. This is reminiscent of 'AI pins' I've seen recently. I'm not exactly sure what they do. Anyways theoretically this earring could record every conversation you have/everything you hear. It could turn important conversations into simple bullet points for later reference, it could remember specific dates and times and add them to your calendar, it could tell you how to improve in social situations, and maybe it could record your voice and the type of stuff you say and train an LLM based on you to impersonate you. These are all just ideas I'm throwing around... not sure if I could really implement them.