Photomultiplier Tube Suppliers

[AntiProtonBoy]

Could someone point me to reliable suppliers of photomultiplier tubes? Been looking at eBay stock, but they seem to be a hit and miss in terms of reliability, information/datasheets, etc.

I'm looking at some of the Hamamatsu range, would be good to know suppliers that stock them.

[mikeselectricstuff]

These are specialised items so unlikely many are held in stock. Contact hamamatsu for your local rep.

[ejeffrey]

Also, be prepared to wait.  Buying anything from Hamamatsu takes forever.

[mikeselectricstuff]

Also, be prepared to wait.  Buying anything from Hamamatsu takes forever.

This sort of stuff is typically made to order

[plesa]

Could someone point me to reliable suppliers of photomultiplier tubes? Been looking at eBay stock, but they seem to be a hit and miss in terms of reliability, information/datasheets, etc.

I'm looking at some of the Hamamatsu range, would be good to know suppliers that stock them.

Hi,
I know only about two companies which are producing PMT tubes nowadays.
Hamamatsu Photonics
Electron Tube www.electrontubes.com

On ebay you can find some old tubes from Burle or Photonis ( they cease production PTM few years ago) datasheets are still available.
But be carefull, the old tube can have high dark current and the brekdown voltage can be be smaller due to helium glass permeability.
It probably does not matter you you are just make some experiments.
The delivery time for Hamamatsu is about 2 months.

[AntiProtonBoy]

Thanks for your replies. Yeah, I'm kinda worried about the degraded condition of some decommissioned tubes.

This sort of stuff is typically made to order

Really? I'd imagine they will cost a fair bit, too.


Side note, what other options do I have as alternatives for PMTs? Amplification and rise time is a big factor here.

[ejeffrey]

This sort of stuff is typically made to order

Could be, but even for stuff you would imagine to be fairly standard (InGaAs photodiodes) they typically take 2-3 months minimum.  Maybe they don't maintain stock of anything, and when you order they just wait until they are making a batch for a big customer and make them in a batch.

Side note, what other options do I have as alternatives for PMTs? Amplification and rise time is a big factor here.

Avalanche photodiodes (APDs) are the other option.  They are a bit nosier than PMTs, more sensitive overall, and more sensitive in the NIR especially.  They have much smaller active areas -- typically 100 micron x 100 microns.  They also require a high bias voltage, but not as high as PMTs.  The biggest manufacturer of APDs is also Hamamatsu   However, they are considerably cheaper and more likely to be in stock with distributors, at least if you aren't to picky and are willing to take whatever is available.

Do you want photon counting or linear amplification?  What wavelength are you interested in, and what is your source?  What kind of risetime do you need?

[plesa]

The price of PMT depens on type, but lot of standard type cost is about 200-500 EUR.
If you are starting with PMT I recomend to read following Hamamatsu book.
sales.hamamatsu.com/assets/applications/ETD/pmt_handbook_complete.pdf

[AntiProtonBoy]

This may sound crazy, but I'm dreaming about constructing a series UV experiments, such as a sensitive spectrometer, or a UV reflectance imaging device. Therefore, I will require a sensor with linear amplification, with a spectral response at least covering the UV-A and B region (280 nm); the higher the better. Typical sources could include diffuse reflection illuminated by UV lamps and outdoor sunlight. I'm also looking into using UV band pass filters and UV diffraction grating, which I'd expect to contribute losses in light intensity, hence the need for good light amplification. Data sampling rate will be at 60 MSPS, so that means rise times of ~8 ns or better is needed. As far as I see, most tubes are good in that respect.

@plesa, yeah I have been reading that Hamamatsu book, good source of info.

[ElektroQuark]

Hi:

I have a used PMT. It's OK. It from Hamamatsu, from an atomic absortion unit. If you are interested PM me.

[Time]

You won't find a PMT with a flat response over such a large range as 280-400 nm.  You will have to correct for your sensors response and the transmittance of your optics if you are interested in actual intensity accuracy.  If you are really only interested in relative accuracy of the spectral content you might be able to get away without correction.

http://www.thorlabs.com/NewGroupPage9.cfm?ObjectGroup_ID=1295

http://www.thorlabs.com/NewGroupPage9.cfm?ObjectGroup_ID=2909

Spectroscopy is not a cheap science, sorry.  Having a fast response and flat response over such a wide wavelength seems a little like having your cake and eating it too.

[Time]

To get your flat response I would use a diffraction grating projected onto a CCD.  You can get fairly cheap CCDs these days.  Or maybe some kind of photodiode array.  Just something that will let you map the diffraction pattern.

[AntiProtonBoy]

@EdoNork, Thanks for the offer. I'll consider it after doing a bit more research.

You won't find a PMT with a flat response over such a large range as 280-400 nm.  You will have to correct for your sensors response and the transmittance of your optics if you are interested in actual intensity accuracy.

Yeah I realise that. I could always make corrections in the FPGA during acquisition, assuming I have an accurate response curve data for the PMT in question.

To get your flat response I would use a diffraction grating projected onto a CCD.

Interesting suggestion, but I see some potential problems with that.

Are you saying that CCDs actually have a relatively flat spectral response in the aforementioned UV range? Would be cool if they did.

My other concern is that the array resolution of cameras may not be sufficient enough to reveal fine spectral lines. Hence the need of a fast PMT (or perhaps an avalanche diode). That way I can make fine-grained data acquisition around 60 mega-samples in real-time. Having said that, the spectral resolution might be also limited by the diffraction grating itself.

Project construction could be something like this:

1. UV light enters through a pinhole;
2. Next stage is a UV band pass filter, like Schott UG11;
3. Filtered light hits a UV diffraction grating for spectral breakdown;
4. A 100 ~ 3000 RPM polygon scanning mirror reflects parts of the spectrum via a slit;
5. Light from slit exposes the PMT with high gain;
6. Signal from PMT gets converted/amplified from a current signal and to a voltage signal;
7. Signal passes through a 70 MHz low pass filter;
8. A high speed ADC mezzanine card converts voltages at 60 MSPS, at 16-bit resolution;
9. FPGA controls scanning mirror RPM for sync, does real time correction to sampled data, buffers it in RAM, perhaps applies some entropy coding, etc;
10. Buffered data is streamed over USB for my pleasure.




Yeah, not a cheap project, as I have to buy just about everything, minus the polygon mirror. On the flip-side, these components have good reuse value for other projects.

[Time]

You dont need the response curves of anything.  You just setup your optics and detector.  Get a laboratory source of known spectral content and intensity.  Use this to calibrate the over all system.  Thats how its done in the industry.  No need to correct for the attenuation of each individual part.

The resolution is based more on your gradient and where your optical plane is with regards to your mirror, like you said.  I have used spectrometers that are relatively wideband (50-180nm) and into the deep UV that uses a CCD as the detector.  CCD is probably the most common type of detector for a spectrometer.  Like I said, you could get away with a photodiode array at those near visible wavelengths though.

[ejeffrey]

There are other options you should consider.  PMTs are kind of bitchy -- they require high voltage, they are large, fragile, and expensive, and they die instantly if exposed to a pathetically small amount of light.  Even when they are powered down, bright lights can temporarily or permanently damage them.  For single element detectors there are some exotic semiconductors like gallium phosphide that make good UV photodiodes.  Even some silicon detectors can operate down to 200 nm if they have a UV grade window.  For imaging or linear array (i.e., for a spectrometer), another option is UV sensitized sensors.  You can coat a silicon detector with a fluorescent material that converts UV to visible photons.  This works very well and is most commonly used for x-rays, but I have seen it used for cameras in the 200 nm range: presumably because the available camera sensors that met the other requirements didn't have good UV sensitivity.

If you really need the high speed and low light sensitivity, I suggest looking at APDs.  However, ordinary photodiodes are extremely versatile devices and should not be discounted.  At low frequency they can measure pico- or even femtowatts, while they can be quite fast if you have a moderate amount of power.  Plus silicon CCDs and linear arrays are widely available at low cost.  In a spectrometer application a linear array is usually collecting 1000x as much light as a single detector with a monochromator would!  That can make up for a lot of deficiencies.

[plesa]

There are some silicon photodiodes for DUV/EUV detection 115-215nm, these detectors are used eg in litography machines. But as I know they are not commercially available.
Still there is lot of information missing. I don't thik that we are talking about designing all from scratch.
The HT supply including voltage dividers are commercialy available from Hamamatsu including magnetic shield, this makes the implementation much more easier.
The TIA for PMT will be easy to construct ( depends on the bandwidth).

The diode needs quite good TIA amplifier including biasing circuit ( e.g. femto.de). The diodes without housing are more sensitive than PMT tubes.
I'm working with these detectors every day but for photon/electron detection systems.

[mikeselectricstuff]

BTW Edmund Optics carry some Hamamatsu packaged PMT products, as well as avalanche photodiodes and some other funky photodetectors.

[tecman]

I've worked and also experimented with many, many PMTs in my career.  Despite what may have been said, they are not bitchy and will no be ruined by bright light.  Some will experience a loss of sensitivity if powered up and overdriven (bright light) but this is temporary, unless too much anode current is drawn.  This can be mitigated with good circuit design. 

PMTs are the best device for sensitivity, linearity and low noise.  Yea you need high voltage, but today that not a big deal.

I have several dozen PMTs at home.  Many were purchased on eBay.  You can find good deals there if you know the models.  Burle, Hams and RCA are good brands.  Stay away from the russian stuff.

paul
 

[AntiProtonBoy]

Thanks again for your suggestions (and the link, Mike). Lots think about now.

The use of UV sensitive scintillator is also a good idea, I should have thought of that. Only down side is their decay time, but I might give em shot with a APD combo.

Regarding PMT power supplies, yep I've seen those compact 1 kV DC / DC converters made specifically for that purpose, or PMT modules with the power supply integrated. They looked pretty neat.