2

At the moment I am trying to build a DIY fluorescence detector using the Amplex Ultra Red probe with excitation and emmission maxima at 568 and 581nm respectively.

I just have problems choosing the right photodiode. The photodiode will have to measure emmission through a .3 cm thick glass wall in a biological. It has to be very small and Through hole. Somebody recommended me this photodiode but I can't find any details about it's specs to fit in with my local supplier.

Any help is very appreciated

Thanks in advance

Felix
  • 21
  • 3
  • Ultra Red or yellow green? Use any PD with a 580 colour filter and TIA amp then calibrate – Tony Stewart EE75 Jul 20 '18 at 12:59
  • Probe spec says “excitation at 530nm , detection at 580 nm (assumed peak not dominant CIE corrected to eyes). – Tony Stewart EE75 Jul 20 '18 at 13:09
  • Ultra Red yes. So this would to it ? https://www.digikey.ca/product-detail/en/excelitas-technologies/VTB8441BH/VTB8441BH-ND/5885860 just add a filter and TIA that's it? I do not understand the second part of your comment. I assume it is a correction to the infos I posted before. Thanks – Felix Jul 20 '18 at 13:25
  • wavelengths are usually defined as $ \lambda _D ~ or ~ \lambda _P $ D for dominant eye correction. P for true wavelength.. any PD with good specs. What power detection and BW are you using? uW? DC? or pulse? – Tony Stewart EE75 Jul 20 '18 at 13:47
  • "spec" If excitation at 530 nm results in signal saturation when the emission is read at 590 nm, you may lower the excitation wavelength to 490–525 nm." You have much to learn to save a few bucks on a DIY probe on calibration errors and wavelength shift vs saturation – Tony Stewart EE75 Jul 20 '18 at 13:59
  • You must block the emitter wavelength and pass the shifted wavelength so a high quality Edmund Scientific filter is needed. This ratio controls your Signal/Noise error – Tony Stewart EE75 Jul 20 '18 at 14:04
  • nmP is used for lab tests, nmD is used for visible LED specs, dont confuse these – Tony Stewart EE75 Jul 20 '18 at 14:11
  • I've been doing phosphorescence detection for decades. I use a dichroic beam splitter in one of the applications. But you have a rather narrow separation of wavelengths! And fluorescence may be quite fast compared to what I do. It's a lot easier to suggest useful thoughts when more is known about the application and constraints involved. For example, knowing 3mm glass says a few things, but mostly it means lots more questions to ask. What's the tau of fluorescence? Mixed taus? Or just one? Could you write a lot more? Or is that all you can discuss? – jonk Jul 20 '18 at 16:35
  • The 3mm glass is actually a chamber containing a physiological medium wiht cells and the Fluorescent probe. A colleague of mine already assembled something which works great using a green LED for excitation. I just wasn't sure which photodiode to buy, because I change supplier here in Canada and he wasn't able to give me more info about the photodiode he used. I will be connecting the DIY fluorescence detector using a Coax cable to an existing slot on the machine I am using for incubation. The commercial version available is too expensive. – Felix Jul 20 '18 at 17:00
  • @Felix the digikey VTB8441BH has an infrared-blocking filter (and so has less sensitivity below 600nM.) Better would be a plain silicon photodiode. After all, you're going to be adding a blue-blocking, red/IR-pass filter, opposite of the filter built into that PD. – wbeaty Jul 20 '18 at 22:27

3 Answers3

1

See the spectra for the probe. The absorption peak is near the emission peak.

You want to excite near the peak of the absorption, and detect near the peak of the emission.

The design process goes like this:

  1. This this an epi-fluorescence arrangement or something else? (epi fluor is where you excite from the same side as you detect) Having all the equipment on one side of the sample has advantages.
  2. Shop for a source. A laser is the usual as it's bright and has a high etendue. Relatively monochromatic. However they are not available in every wavelength (close to the absorption peak). More complex to drive (and perhaps cool). An LED on the other hand has a broad emission,easier to drive, and available in more wavelengths.
  3. Shop for an clean up filter. The cutoff wants to be near where the dye excitation / emission responses cross.
  4. If a epi system, you'll need a dichroic mirror.
  5. Shop for an emission filter to detect light in the emission range only.
  6. The photodiode is the least of your worries.
  7. If the fluorescence signal is small (it will be cause you want to look at small samples, few molecules of the dye) you might need a lock-in amplifier.
  8. A way of confirming that it's fluorescence of the dye and not stray light being detected.

The 3 mm gap to the sample is a problem. You might need some optics to excite a small volume and similar optics to capture as much of the fluorescent signal.

The filter people are Chroma, Omega and Semrock (Edmund Optics and Thorlabs do a limited range) and you need to be aware that there is a limited acceptance angle for interference filters. It's easy to spend a few $k on filters.

The design problem is mostly shopping for sources and filters.

This is the same problem as fluorescence microscopy so google "dichroic filter fluorescence microscopy"

D Duck
  • 2,217
  • 1
  • 9
  • 18
0

Go to digikey.com and search on photodiode. Narrow your search to through-hole pins and start comparing. Filter by packages. The smallest you'll find are 3 mm dia. TO-46 packages are a hair under 5 mm.

WhatRoughBeast
  • 60,521
  • 2
  • 37
  • 97
  • Thanks. Do you think that the viewing angle affects my measurements? I still have to filter by wavelength too ? 580nm or will the color filter to this job? – Felix Jul 20 '18 at 13:43
  • it depends on stray light if you need a filter – Tony Stewart EE75 Jul 20 '18 at 13:51
  • @Felix - Brace yourself for a big shock. You will need a very good bandpass fliter to look at the 580 nm emission while blocking the 568 excitation. You can try Edmund Optics, Thorlabs, or Newport. Look specifically for "fluorescence bandpass" filters. And by "very good" I mean "very expensive". Like around 200 bucks at the low end. And no, you can't use a cheap filter. Don't even think about it. – WhatRoughBeast Jul 20 '18 at 14:37
  • Look for Diocratic filters . You may need two depending on saturation results – Tony Stewart EE75 Jul 20 '18 at 15:17
  • @WhatRoughBeat knowing that the commercial alternative would require a couple of 10K 200 wouldn't be too bad. – Felix Jul 20 '18 at 17:01
0

The signal is 580 nm detected in a 10nm tolerance while the interference is passthru of the emitter at 530 nm peak.

This requires high SNR to get accurate rejection emitter while only reading the fluorescent level.

Best bet : $50 http://www.omegafilters.com/products/filters/bandpass/580df10.html

The emitter and detector beam width depends on including only the sample and Path length , but the angles and change in intensity must also satisfy your intensity range. Diffused emitter / detectors will average while clear give more resolution.

For small gap transmission I would consider two tests filtered with/without fluorescent to compare levels. The better your optical matching. , the lower the baseline and higher SNR results.

The PD sensitivity for 5mm types is around 0.5 uA/uW or mA/mW of detected power over cone area. This is attenuated by shorter than test spec nm by 50% or so.

You don’t need Blue enhance or a huge detector unless dictated by your experiment.

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 185
  • this filters seems to work for what I want to do. The machine will have to record fluorescence production of biological samples and compare between individuals and different environmental conditions. Absolute values are less important as long as I may detect variation between treatmetns and or individuals. – Felix Jul 20 '18 at 17:07
  • Ok anything else to add? – Tony Stewart EE75 Jul 20 '18 at 19:19
  • Extremely common photodiodes (PIN) used everywhere are BPW34, and S1223, see Newark and Digikey. – wbeaty Jul 20 '18 at 22:29
  • For the moment I have the information I need to do a first build. I'll let you all know how it went. – Felix Jul 23 '18 at 17:41