SparkFun Forums

[silic0re]

Hi there,

I'm looking for a small solid-state radiation sensor. From what I understand, radiation detectors used to use Gieger tubes, but now a good number of solid-state alternatives exist. These newer detectors typically use some kind of scintillation material where the material emits light when some kind of ionization event occurs in the material (such as being hit by a gamma ray, beta radiation, or some other types depending on the material) ( http://en.wikipedia.org/wiki/Scintillat ... physics%29 ). I think that typically the kinds of scintillation materials they use include ZnS, something called Nal(TI), and I've even been able to find some BaF2 crystals that have beautiful responses to a variety of particles (including high energy neutrons!) on eBay.

The issue is, I haven't really been able to find any end-product detector (I tend to imagine something that looks like a thermopile when I envision what I'm looking for) that I could purchase. There are a number of robotics and hobby-electronics websites that sell gieger kits, but these typically use gieger tubes which require high voltage sources. A solid state device, from the little I know about them, seems to just be a scintillator infront of a photodetector/photomultiplier. And I've been able to find products that use the devices -- 'radiation detectors' in keychains or watches -- but not the actual sensor.

Any thoughts?

thanks,
silic0re

[bigglez]

Hi there,

I'm looking for a small solid-state radiation sensor
.....
The issue is, I haven't really been able to find any end-product detector (I tend to imagine something that looks like a thermopile when I envision what I'm looking for) that I could purchase.
.....
A solid state device, from the little I know about them, seems to just be a scintillator infront of a photodetector/photomultiplier. And I've been able to find products that use the devices -- 'radiation detectors' in keychains or watches -- but not the actual sensor.

Any thoughts?

Greetings SilicOre,

You are right about the ttechnology - I've worked on X-Ray systems with scintallor crystals and PMTs (Photomultiplier Tubes), which also require high voltage (500V) of very low noise and high stability.

I've also worked on systems with solid state photo-diode scintallation detectors, which may require modest high voltage (50V) and high gain amplifiers (requiring careful design and shielding). The crystal must be in total darkness, shielded from background radiation, and compensated for temperature drift in the amplifier and photodiode(s).

I was traveling to the US from the UK on business and was asked to carry a suitcase of small scintallation crystals and film that had been radiated in an experiment. The film was to be read by a US government agency as part of an X-Ray product qualification process. I was told to keep the case with me, and not let it go through any X-Ray inspection. This was years before the current level of passenger screening. I never did learn if my samples were useable when I landed and handed them over.

A lot depends upon what kind of radiation and what levels you expect to detect. A commercial Radon gas detector (like a smoke alarm) might suffice - you can rip one apart to see how they did it. Are you looking for a broad function detector, or a specific source type? Are you looking for a threshold detector (alarm) or a radiation metrics (levels)? Is this a battery powered portable device or can it be larger and AC powered?

It's possible that common op-amps and some transistors, even high density memory or CCD imagers could be used as detectors without the scintallation crystals. There is a class of semiconductors that are said to radiation proofed (Rad-Hard) for use in spaceflight and weapons.

The gain of the electronics would have to be very high and much thought put into shielding the unwanted electrical noise and radio interference.

We have a cat that had a thyroid condition. We were offered radiotherapy as an option and sent our cat to a specialist that dosed her with 50ml of an iodine-131 isotope. She was released back to us when the radiation fell to human safe levels (about five days).

I borrowed my friend's Civil Defense radiation meter, and had no trouble pegging the pointer on the higher sensistivity scale. The cat's neck was hottest, as was the cat litter box... After a few days I was unable to move the pointer much above background noise. My friend thinks the meters were of little value and would only confirm a kill (due to poor sensitivity). Not bad for 1960s technology!

The cat is doing very well, has gained weight, and enjoyed the last couple of years post-treatment.

Comments Welcome!

[awright]

Once you have the scintillator material, your problem becomes one of detecting the light pulse and deciding how to interpret the information. I think a thermopile would be totally useless for the purpose because, as a thermal sensor, it has inherently slow response and would brobably not even see a scintillation pulse. Might be able to integrate a very large scintillation rate into a usable response, 'though, if you're looking for an alarm configuration but I don't see any advantage to a thermopile in this application. A thermopile is also generally most responsive to infrared thermal radiation, which I would guess your sciltillator does not produce.

There are large area photodiodes optimized for very high sensitivity and low noise that would probably be suitable. Devices called "photon detectors" offer the ultimate in sensitivity. I do not know whether they are generally avalanche photodiodes or just simple guarded photodiodes optimized for low noise. You can probably get app. notes with recommended circuitry for very low noise light sensing from the photodiode manufacturer. I'm sure you can get such recommended circuits in app notes from the low-noise op-amp manufacturers. Try Analog Devices, National, Freescale, and others.

Avalanche photodiodes are the solid-state equivalent of the photomultiplier tube. It is biased up to a relatively high voltage just below the threshold of spontaneous avalanche breakdown, typically in the few hundred volt range. The light pulse pushes the photodiode into avalanche and you get a current pulse out, somewhat analogous to the breakdown of a photomultiplier tube. Both devices must be in circuits that allow them to recover blocking of the applied voltage after the microsecond or so output pulse is delivered to the load.

Google on "avalanche photodiode" and read the Perkin-Elmer writeup.

You will get much less sensitivity out of a non-avalanche photodiode but you won't need the high voltage bias. There ain't no free lunch.

Peter, a decade ago I had a study of my internals that required ingestion of a radio tracer milk shake. Came home and held an antique tube-type geiger counter to my abdomen and it went crazy.

I don't think the response of non-radiation hardened ICs would be very useful as radiation detectors because the affect of radiation on ICs tends to be gradual or sudden deterioration of functionality and increase in noise level, not generation of useable pulses for each incoming particle. Moreover, the affect of radiation on the IC would be strongly dependent upon where the particles struck the IC. It would all be statistical, of course, but an IC will present a very small target compared to a photodiode of a few millimeters diameter. A scintillator, on the other hand, presents a fairly large target volume from which a relatively small photodiode can sense light pulses. I'm not sure, but I believe that unshielded photodiodes respond to impact of radioactive particles, possibly leading to some deterioration of the photodiode.
Have fun.

awright

[saipan59]

As it happens, my current project is to build a 'radiation measurement' platform based on a military surplus scintillation probe (it will also support a Geiger tube).

A commercial Radon gas detector (like a smoke alarm) might suffice

Actually no. A smoke alarm is unrelated to detecting Radon. An 'ionization type' smoke alarm *contains* a small amount of Americium-241, which is a good emitter of Alpha particles. Detecting Radon requires (for example) an appropriate type of scintillation counter, or a sensitive Geiger counter (if you are able to concentrate the air sample).

There *are* solid-state radiation detectors, but the only 'useful' ones I have heard of are VERY VERY expensive, based on highly purified Germanium.

Without spending many thousands of $, your realistic choices are these:
1) An ionization-chamber type of "survey meter", such as the old Civil Defense CDV-715. Simple and cheap, but it *only* detects VERY HIGH levels of radiation. If the meter needle moves, the radiation is high enough that you are in danger.
2) A Geiger counter, which by definition includes a Geiger-Mueller tube. I've built a couple myself, and I recently bought a mil-surplus one on eBay for $35.
3) A scintillation probe/counter. The only affordable version I've heard of uses a PMT (photomultiplier tube). All scint probes contain a 'scintillation crystal' of some sort - the most popular type for general use is sodium iodide (NaI). The 'probe' is essentially just the crystal, plus a PMT, in a VERY light-tight container. Some probes (like mine) have a preamp circuit built-in.

PMTs and G-M tubes both require high voltages at low currents. Everything that you would be detecting is relatively low levels of radiation, so will need to use one of these approaches.

If you want to proceed with making something, I can point you to specific details.

Pete

[saipan59]

I've been able to find products that use the devices -- 'radiation detectors' in keychains or watches -- but not the actual sensor.

Those gadgets are mostly a 'dosimeter' type of thing - they contain some material that changes color when exposed to radiation, to let you know that you've been nuked. The type of thing that a nuke worker would wear on his clothing.

OR, there are some very small Geiger counter type devices being imported from Russia these days. They contain a very small G-M tube and the associated electronics, they're cheap, they are totally uncalibrated, and are probably not very sensitive (but still useful for determining if something is 'hot' or not).

Pete

[saipan59]

...and FYI, here is my earlier post regarding my previous Geiger counter project:
viewtopic.php?t=6345

Pete

[bigglez]

A commercial Radon gas detector (like a smoke alarm) might suffice

Actually no. A smoke alarm is unrelated to detecting Radon. An 'ionization type' smoke alarm *contains* a small amount of Americium-241, which is a good emitter of Alpha particles. Detecting Radon requires (for example) an appropriate type of scintillation counter, or a sensitive Geiger counter (if you are able to concentrate the air sample).

There *are* solid-state radiation detectors, but the only 'useful' ones I have heard of are VERY VERY expensive, based on highly purified Germanium.

Greetings Pete,

I might not have been clear in my comment. I believe there are continuous radon detectors (like this one) that function "like a smoke detector" meaning they are for residential use. Not that they use the same detector scheme. Reverse engineering one of these will reveal what technology is used to continuously monitor low radiation levels.

Comments Welcome!

[awright]

Interesting comments above.

I have to correct my comment about the operation of PM tubes and avelanche photodiodes. I simply must stop trying to post things at 2 am while trying to stay awake.

Both the PM tube and the avelanche photodiode operate in a linear mode, not normally a pulse output mode. A G-M (geiger) tube puts out a pulse as I described for each incident particle that initiates ionization of the gas in the tube. That's why the instruments that use G-M tubes are called "counters." They count the pulses. In that process, the amplitude and duration of each pulse is identical and the G-M tube cannot provide information on the type or energy of particle that initiated the pulse. Discrimination of particle types is built into the tube by selection of tube wall or window material and thickness.

PM tubes and avelanche photodiodes, on the other hand, both have high internal gain due to the avelanche of electrons in which the initial electron emitted by the detector is accelerated and knocks several additional electrons free and each of those is accelerated and knocks several more electrons free. This multiplication of the initial electron into thousands or tens of thousands that are eventually collected and make up the output signal is a linear process with very high gain, but it remains linear and relatively controllable. That is, the output follows the input, albeit with high gain. The amount of multiplication of electrons, or the gain of the device, is a function of the voltage applied to the device so you need a well regulated high voltage source.

The avelanche that occurs in a G-M tube is, on the other hand, uncontrolled internally and for these devices require a quenching gas inside the tube and an appropriate external circuit that limits the current to allow the tube to recover blocking. I was in error attributing that quenching reauirement to PM tubes and avalanche photodiodes, although you do have to control current magnitude in them.

I hope this is clearer and that I have not further confused the issue. Sounds like the other posters here are much more knowledgeable on the topic tha I am.

awright

[saipan59]

I might not have been clear in my comment... (snip)

Ah, I see - my apologies!
The device you pointed out probably (I'm guessing) uses circuitry similar to this:
http://www.maxim-ic.com/appnotes.cfm/an_pk/2236
Armed with this info, I should amend something I said earlier, which is that to measure Radon requires a scint probe or G-M tube. Since a commercial Radon detector like this uses a photodiode (they say so), I think the distinction is in *time*. These detectors count relatively rare events over a period of hours, whereas with a scint probe-based counter setup a measurement can be made more or less in 'real time'.

Pete

[silic0re]

Quite the interesting discussion this has generated! To clear up from above, I wasn't thinking of using a thermopile/scintillator material as a radiation sensor, but rather when I just think of what a solid-state radiation sensor might look like, I think of something like a thermopile, with a photodetector encased in a can with a scintillation material on the outside. Though really, I suppose that the entire device would have to be enclosed in a light-tight enclosure, so I have no idea why I have that thought...

Back on track, here is the watch that I mentioned: http://www.gammawatch.com/ . It's actually rather small, and gives a measurement in Sievert's that looks to have two to three significant digits. I really do wonder what sensor they are using!

A quick do-it-yourself project I noticed a long while ago is also here:
http://www.noah.org/science/x-ray/detector/. The method is essentially to coat a regular photodiode with a scintillation material such as ZnS, then encase the diode in something light-tight. One of the issues with this simple coated photodetector method is what kind of sensitivity one would achieve, and having to calibrate it...

thanks
silic0re[/url]

[saipan59]

I'm thinking that an important limitation of a photodiode-based solution is 'sensitivity'. Reading the specs for the 'Gamma Watch', I note that at low levels, it takes 20 minutes to give a reading. That tells me that it is only able to detect events relatively rarely. And this is in agreement with photodiode-based Radon detectors, which need 24 hours or so to make a good measurement.

Perhaps another way of looking at it is: If a fairly cheap solid-state detector was effective for general radiation detection and measurement, then everybody would be doing it.

So we're still left with G-M tubes (for detecting events), and scint probes (for detecting events much smaller events, and measuring their magnitudes).

Pete

[silic0re]

That's a very good point, and upon rereading the watch's website I noticed it used the term 'averaged over x number of minutes', but also that it's unable to detect short pulses (say, from an x-ray machine). This would seem to suggest that they're either averaging the signal from the photodiode over quite a period, or that they (for some reason) only sample from it for a short time.

In either case, I'm still really curious what kind of sensor they use, and if it's commercially available. I think I'd be interested in a photodiode based sensor if it was small and inexpensive, even if the sensitivity isn't anywhere near that of a gieger tube. Trading off sensitivity for size and voltage requirements seems okay right now.

(Also, just to see, I've sent the GammaWatch people an e-mail to see if they might be willing to say what type of sensor their watch uses).

[saipan59]

Their web site says it's a photodiode... They also say what energy levels it's supposed to be sensitive to - the numbers they give are pretty good, but the 'hit rate' is probably really low.

If you're going to build that kind of sensor yourself, be sure to read the page I indicated earlier (from Maxim), because it points out the difficulties (requirement for low input noise in the op-amp, etc.).

After I get a little farther along with my scint probe project, maybe I'll try using a PIN photodiode and ZnS to see what I can see.

Pete

[silic0re]

A very quick response from Environmental Instruments Canada Inc., the creators of the Gamma Watch. It appears as though they use a straight photodiode without a scintillator!

Hi Peter,

we use a photodiode w/o scintillator. The photon (or electron) interacts directly with the depletion layer of the diode w/o scintillating. You can couple a scintillator to a photodiode, but the signal will be 10 times lower, but you see more photons. It is already hard enough to see the very small signal above the thermal noise.

Any photodiode (go to digikey and find a cheap one) will do. Bigger surface area means detecting more photons but smaller pulse per photon (same amount of charge but higher capacitance). The difficulty is the amplification circuit. I won't share ours, but you can find a few on the hamamatsu website (see circuits 9 and 10 in http://sales.hamamatsu.com/assets/appli ... amples.pdf ) . Please let me know if you get them to work

[awright]

Wow! That's a really polite and helpful response from the watch maker to someone obviously not about to buy his product. His points seem valid to me.

I suspect that the long measurement time is merely to integrate the events over a long enough time to get well above the noise. Since the sensor area is relatively tiny in most inexpensive photodiodes, they would have to perform some math to get from the few events they detect in the small sensor to some dosage for the human body or to some standardized measurement units and target area.

What is it you are actually trying to accomplish, silicOre? What radiation do you want to measure? Post-apocalypse radiation safety?

awright