[Laena] and her colleagues at the La Trobe Institute for Molecular Science in Melbourne, Australia used a Raspberry Pi to make a low-cost electrochemiluminescence (ECL) detector to measure inflammation markers, which could be used to detect cardiovascular disease or sepsis early enough to give doctors a better chance at saving a patient’s life.
ECL reactions emit light as a result of an electrically-activated chemical reaction, making them very useful for detecting biochemical markers in blood, saliva, or other biological samples. ECL setups are fundamentally fairly straightforward. The device includes a voltage reference generator to initiate the chemical reaction and a photomultiplier tube (PMT) to measure the emitted light. The PMT outputs a current which is then converted to a voltage using a transimpedance amplifier (TIA). That signal is then sampled by the DAQCplate expansion board and the live output can be viewed in ppLOGGER in real-time.
Using the RPi allowed the team to do some necessary, but pretty simple signal processing, like converting the TIA voltage back to a photocurrent and integrating the current to obtain the ECL intensities. They mention the added signal processing potential of the RPi was a huge advantage of their setup over similar devices, however, simple integration can be done pretty easily on most any microcontroller. Naturally, they compared their device to a standard ECL setup and found that the results were fairly comparable between the two instruments. Their custom device showed a slightly lower limit of detection than the standard setup.
Their device costs roughly $1756 USD in non-bulk quantities with the PMT being the majority of the cost ($1500). Even at almost $2000, their device provides more than $8000 in savings compared to ECL instruments on the market. Though cost is much more than just the bill of materials, we like seeing the community making efforts to democratize science, and [Laena] and her colleagues did just that. I wonder if they can help us figure out the venus fly trap while they’re at it?
If that includes a Rpi then that’s pretty cheap 🤣
I wonder if this is a special PMT. The last time I bought one from Hamamatsu, some years ago it was between 600 oder 700Euro. But of course, this company makes you feel poor. :-)
Olaf
TFA says it IS a Hamamatsu, but also includes the high voltage supply, which usually is at least as expensive as the PMT: It’s their Model H7826-01 (which is a really cute little thing!).
TFA also says they align the sensor under the PMT with magnets… I wonder what that does to the electron trajectories in the PMT, notoriously sensitive to magnetic fields. They don’t mention this.
The great thing about a PMT (over, a solid-state equivalent) is the enormous collecting area of the photocathode: it’s not hard to collect over a steradian of solid angle: fantastic for detecting emitted photons like this. Very, very expensive to get similar performance with solid state detectors.
$350 for PMT+ HV power supply+amplifier. Reasonable enough price?
https://www.ebay.com/itm/155259806019
$350 for a used pile of parts, pulled from old equipment, of unknown provenance, and shipped from Poland?
No, not reasonable.
Understood.
Time to make a YT video how to connect it all to Rpi, add a zero to the price ;)
A silicon photomultiplier would make more sense here I think. We use the Hamamatsu ones (20-90 dollars) and they’re very competitive with Hamamatsu PMTs, but dirt cheap:
https://github.com/OpenSiPM/sipm-bias-control
If the system is like a laser or microscope application, with light collimated well enough (more precisely, its étendue is small enough) to effectively use the (usually) small SiPM detector aperture (3% of the area of PMT used here), AND the photon flux is high enough to be well above the SiPM’s characteristcally high dark current then, yes, a SiPM is certainly a good choice for cost & space-constrained applications.
This application has an isotropic source (so, difficult to collimate to a small detector), and its lower detection limit is set by the dark current of the detector. So, here, the PMT is the better choice, for both reasons.
Unfortunately we tried some photodiodes and their performance was just not suitable, the photon flux of the chemical background and the low concentrations of interest is not high at all so you need a really sensitive detector (as Paul said really well above). But there’s been some fantastic developments recently towards cheap and still effective alternatives to traditional PMTs, so I hope my work gets superseded soon :)
Hi folks, Laena here! So thrilled this picked up more interest and I hope the below helps.
I took the prices off the quotes we received from an overseas supplier, in AUD. Then converted to USD. The exchange rate has fluctuated a lot over the last 10 years (I think this module we bought 5 years ago). Did not include the ‘Australia tax’ i.e. large shipping costs…! The module we used includes HV as Paul has said, which is why the cost is a bit higher (unfortunately).
As for the magnets, they were really tiny and weak magnets something like an inch or two from the actual PMT (bare minimum to keep the case closed if it got knocked, essentially). We didn’t notice any issues related to this.
We considered photodiodes and cheaper alternatives but when you’re looking at really low detection limits where the background ‘signal’ is dominated by a chemical response (as in this case), rather than the photodetector’s response, then you need a really sensitive detector – we did try a good photodiode and it was pretty hopeless in comparison. There’s exciting work coming out on cheap and still pretty good alternatives, which will be a game-changer for ECL.
More information about the DAQCplate as well as other Pi-Plates can be found at: https://Pi-Plates.com.
The free ppLOGGER program has been deprecated and replaced by PIPLATElogger.