I Observed Brownian Motion At Home
Casual Physics Enjoyer
SubscribeSign in
I Observed Some Brownian Motion At Home<br>Making a better dataset for diffusion relations with milk<br>CasualPhysicsEnjoyer<br>Jun 01, 2026
Share
Last time I wrote about Brownian motion. If you put a particle in a noisy bath, then the particle's distance from where it started moves like the square root of time. I really wanted to test this to check. So I looked at videos online of Brownian motion and tried to use tracking algorithms to measure the distance that the particle travels.
But the problem that I got is that the videos were not high quality, so the tracking algorithm was really janky. What happened was that the TrackPy seemed to think that one particle was multiple particles. It also had poor 'memory', and split the trajectory of one particle into another. Another problem is that the videos weren't long enough or clear enough. Most of the videos I found were 3 minutes at most. And I wanted to make a longer dataset.<br>So I tried to do this myself, and record some Brownian motion. The original experiment done by Brown involved pollen particles, but these were quite heavily processed so I didn't have the equipment. The size required for thermal effects to be meaningful is around on the order of 1 to 10 micrometers.<br>So there were a few options. My first idea was to try using 1 micrometer polystyrene balls. But these were expensive! Another option was to try graphite cells, which were also expensive. But then I came across Michael de Podesta's blog using milk globules in water! So I'm trying to replicate it at home to do some similar imaging.
First I got a Swift T380 microscope off Amazon. I then also got the camera insert so I could attach it to my computer. Then I got some extras - a calibration slide, microscope slides, and coverslips.
There are two binoculars, which take swappable 10× and 25× wide-field eyepieces. These pair with the four objectives on the turret (4×, 10×, 40× and 100×) to give six magnifications in total — 40×, 100×, 250×, 400×, 1000× and 2500×. And then the port at the top allows me to slot in the camera in.<br>The first annoying part was getting the camera to work with my MacBook. Swift is a Chinese company and the driver that I had to install had some software that wasn't trusted by Apple. It also didn't work and I wasn't comfortable with using it. But Claude Code realised that my Mac still was detecting the camera somehow (through some obtuse commands in Terminal), and so I realised that I could open QuickTime player and have the camera open through there.<br>So I tried the calibration slide with the camera, looks like it worked!
Then I tried some shrimp antenna.
Ok so now I wanted to get the milk particles. So I diluted regular Tesco whole milk in 10:1. (I'll try homogenized in a different run).
I then pipetted it onto a microscope slide, dropped a coverslip on top. For small stuff like droppers, lens cleaners, slides and slide covers, cheap stuff from China or Amazon works well.
Now, after lots of trying - here were the settings that worked for me:<br>I closed the light iris for higher contrast. The milk globules are nearly transparent, so stopping the condenser's iris diaphragm right down is what makes them pop out against the background. Then I brought the stage right up to the lens - I raised it until the objective almost touched the slide. Then I really slowly moved the focus knob.<br>This was 10×. I could see some stuff but not that well.
And then 40×, it became a bit clearer, ready for tracking.
And then 100×, without immersion oil. You can really start to see the Brownian motion here, even though there is some drift. I think immersion oil would work better, but it's messy, and I need to buy a proper lens-cleaning set.
My milk wasn't homogenised, so I don't know how big these are. I also now need to try the tracking algorithms on this footage to see if there's any improvement. More in the next post?
Share
Discussion about this post<br>CommentsRestacks
TopLatestDiscussions
No posts
Ready for more?
Subscribe
© 2026 Casual Physics Enjoyer · Privacy ∙ Terms ∙ Collection notice<br>Start your SubstackGet the app<br>Substack is the home for great culture
This site requires JavaScript to run correctly. Please turn on JavaScript or unblock scripts