A physicist explains what really happens when a smell reaches you and how researchers are building an electronic nose to copy it
The short answer: yes, a tiny bit of it is. When you smell poop, microscopic molecules that have evaporated from it travel through the air and land in your nose. It’s not whole bits of poop, just the smallest molecules that come off it. But those molecules are what create the smell.
Nobody enjoys that whiff, but it raises a surprising question about how smell works, and it has led researchers to build a machine that can smell.
So what exactly is in your nose?
Everything we smell releases tiny odor molecules into the air. Poop gives off very small molecules that evaporate and drift into your nose.
Only the very smallest molecules escape. Some things release very few molecules, so we hardly smell them, while others release plenty, and we notice them right away.
Why does warm stuff smell stronger?
Temperature plays a big role. Heat makes molecules evaporate more easily, so more of them reach your nose. That’s why:
- Warm coffee smells stronger than cold coffee.
- Fresh poop smells stronger than old poop, because it’s still warm.
- Perfume smells nice for a long time. It’s a mix of scent molecules and others that evaporate very easily, which carry the scent along with them.
How do we recognize a smell?
Here is the surprising part: a single molecule usually doesn’t carry a smell on its own. You recognize a scent when a whole combination of molecules arrives together.
At the ends of the nerves in your nose are receptors, the receivers for scent molecules. When a molecule attaches to a receptor, a small electrical current sends a signal to your brain.
To picture this, imagine a board with five lights, each representing a receptor:
- Nail polish remover is mostly acetone, essentially one type of molecule. One light switches on.
- Coffee contains an enormous number of different molecules. Perhaps three lights switch on.
Real noses are far more complex. Humans have around 300 different types of receptors, and scent molecules bind to each of them to a greater or lesser degree. In the light metaphor, the lights don’t just turn on or off, they burn brighter or dimmer.
Your brain reads that pattern of lights and recognizes it as a particular scent. That’s how we can tell thousands of smells apart, from fresh rain to an orange, even blindfolded.
Learning new smells
When you meet an unfamiliar scent, your brain has no pattern to match it to. You might think, “spices, but which ones?” Once someone tells you it’s patchouli, you learn it remarkably quickly and recognize it next time.
Why build an electronic nose?
Human noses are impressive, but they have limits. Some smells are too subtle to detect, and sometimes smelling is dangerous or disruptive. An electronic nose could help with:
- Detecting diseases at an early stage
- Finding explosives, at airports or on trains
- Investigating crime scenes, where a person or dog walking in could disturb the evidence
- Locating people buried under rubble after an earthquake
Technical physicist Sissi de Beer of the University of Twente is working on exactly this.
How the electronic nose works
The first version is a simple blue box containing four different sensors. Each acts as an artificial receptor. As with the five lights, the power lies in the combination of sensors, which each react differently to the same scent molecules.
To make each sensor respond differently, the team develops a unique coating for each one. Under a microscope, these coatings look like forests of nanohairs. A human hair is about 100 micrometers wide, and these structures are roughly 1,000 times smaller. That matters because scent molecules are just as tiny and need to fit between the nanohairs.
Each coating absorbs different amounts of different molecules, and the sensors measure this. The result is a pattern, like the pattern of lights in our noses. And like a human nose, the electronic version learns which pattern belongs to which scent.
Even with only four sensors, it can already recognize lavender, pine scent, and even different whiskies.
What’s next?
The four sensors are just the beginning. The team is investigating how many sensors are needed to cover a wide range of smells. They also want to find out where a scent comes from, so they’re working on mounting the nose on self-driving cars and drones to track down the source.
Sissi’s ultimate goal is a sensor as good as a dog’s nose. And who knows, one day it might even warn you about the portable toilet you’d better not step into.
Why can’t we make a machine that releases smells on demand?
We have screens for images and speakers for sound, so why isn’t there a device that produces any smell you ask for? Because smell is far harder to recreate than sight or hearing.
Every color on a screen can be made by mixing just three lights, because our eyes have three types of color-sensitive cells. Sound is a combination of frequencies and volume. Smell has no such shortcut. With around 300 types of receptors, each scent is a complex pattern, and there is no small set of “base smells” that can recreate all the others.
There is also a physical problem: a smell is matter, not information. A screen only has to produce light, but a smell machine has to release the actual molecules, which means cartridges of chemicals that run out. And unlike an image, a smell doesn’t switch off instantly. The molecules linger in the air and mix with the next scent, and how strongly they reach your nose depends on temperature, airflow and distance.
Touch is a bit easier, but not by much. It can be imitated with physics alone, such as vibrations, small motors or ultrasound, which is why phones and game controllers already give us haptic feedback. Still, our skin senses pressure, texture, temperature and stretching all at once, so truly convincing touch remains a challenge.
That is why the electronic nose is such an important first step. By learning to recognize the pattern of a scent, it starts to turn smell into something that can be measured and stored. Recognizing a smell is much simpler than recreating it, but it may be the first step toward doing both.
This article is based on a video from the University of the Netherlands featuring Sissi de Beer of the University of Twente.