Americium-241 as Gamma Ray Source

The ionization type smoke detector has an interesting nuclear radiation source which I have removed to make a simple spintariscope earlier. Because I have mentioned how I got my Am-241 radioisotope, here I will just describe what I was doing with it using my newly bought Geiger-Muller detector. 

The GM assembly was a low-cost unit bought from GQ electronics, USA. The detector was 99% completed (even included a rechargeable battery) but without the tube itself, so I bought a US Navy surplus CBON 6107/BS-212 Geiger Muller Tube manufactured by Anton Electronic Laboratories during the 1950's as the radiation detecting element. Naturally, I suspect there might be some degradation of gaseous mix inside the tube over the inactive years especially when the tube features an alpha radiation permeable mica window but I currently have no resources to test this hypothesis.

Well, the availability of the window was actually the reason why I bought the tube in the first place since Am-241 sources emits mostly alpha radiation. My previous attempts on taking long-exposure photographs on a phosphor screen scintillated by alpha particles at different distance from the radioactive source have failed to give any substantial data (or any photographs actually showing variation of scintillation intensity for that matter). What I really want to do was simple - to observe the inverse square law in ionizing radiation. 

In fact, the set up was so simple, there is really nothing more simple it can go. Here's a photograph of my actual set up:

As you can see, two wooden blocks of same dimension was placed linearly away from each other with the block on the left sticky-tacked to the table, securing its position. The block on the right was free-moving and it is where the radioactive source was tacked onto. The block on the left is where the detector GM-tube was placed, held firm by masking tape and connected to the Geiger Counter circuit with crocodile clips. The circuit features an audio data cable which transfers the audible "clicks" to my laptop's sound card which a software was provided by the vendor was installed to count the "clicks" at a given time. 

I understand this kind of set-up was crude but I have not a linear translational stage for precise distance measurement. So this is really the most I can do now. The measurement was to record the number of "clicks" the counter register over a predetermined exposure time, then average the results for a quick radiation measurement unit called [CPM] or counts per minute, which describes the average number of nuclear events registered by the GM tube. 

First up, I measured the background radiation (i.e. without the radioactive source nearby the detector) which over 15 minutes gives me 11 CPM. Next, I measured the CPM when the source was directly 1 mm away from the GM tube, then at 2 mm, 3 mm, so on until 30 mm. The results were then subtracted 11 CPM to compensate background radiation, here was what I got:

Radioactivity at increasing source distance using Am-241 unattenuated 0.8 microcurie source. 

Notice instead of an exponential decay, the rate of decrease seems to fit Boltzmann sigmoidal function. There are apparently two "interesting" regions observed from the plot. For one, at source distance closer than 5 mm, the counts seems to "flat out" slightly less than 1200 CPM. This seems to suggest the existence of a saturation limit for count discrimination in the software provided by GQ electronics. Since the software interpret the threshold of audio signal "clicks" coming from the GM circuitry, if two successive clicks was loud and close enough, the software might not be able to resolve the coincidence thus limits the CPM resolution at higher activity levels. This I have yet again to verify, perhaps in the future with x-ray photons. 

The next anomaly is when the source distance range between 10 to 14 mm. The rate of decrease of activity seems to "slow down" which suggest a drastic drop in alpha radiation being absorbed (attenuated) by air but retains a fraction of gamma radiation which is still relatively "bright" at that point. But then again, it could just be a statistical anomaly because of the short exposure time (3 minute average).

I guess more importantly, the data possibly showed the process of nuclear decay is complicated and "dirty". A pure sample of Am-241 was known to be both alpha and gamma ray emitter, although with the latter far less significant than former. Nevertheless, I believe it could be separately measured and detected. Before I went on discussing how I do so, let us study the decay scheme of Am-241:

The figure above graphically illustrate the first decay product of Americium-241 to Neptunium-237. It shows the probability of decay channels which for alpha particles we can see it is with kinetic energy about 5.5 million electron volts. On gamma rays, Am-241 nucleus relaxes mostly on 13.9, 17.6 and 59.5 thousand electron volts, these photon energies are comparable to those of medical x-ray machines which can be shielded with appropriate measure. 

As previously mentioned, the GM tube has a fragile mica window (not seen in photos), which was protected with a black metallic cap (see photo below) having a small hole about 1.2 mm diameter measured with vernier caliper to prevent accidental puncture, thus destroying the tube. The tube itself has an outer and inner diameter of 8.8 mm and 6.4 mm respectively. To register signal for alpha particles, it will have to enter the tube through the small hole on the metal cap. Gamma rays however, depending on its energy, have the capability to penetrate the metal cap, thus the detection area will be a wider, 6.4 mm.

The GM tube was protected by a black "top-hat" metal cap, with a small hole letting alpha particles through.

With the previous result, I decided to check if I could measure significant counts after "filtering" out the alpha particles emitted from Am-241 by placing plastic sheets (as attenuator) in front of the tube. The sheet was a simple plastic projector slide cut out into 25 square centimetre pieces, I weighed them with a sensitive scale to obtain its density. Then I measured the background-compensated counts by successively adding attenuators at a fixed source-to-detector distance (10 mm), this was what I obtained:

When a "filter" was introduced, the counts drastically reduced but stay relatively the same for increasing thickness.

Without any attenuation, the GM counter reads slightly less than 800 CPM. When a piece of plastic was introduced between the source and detector, the count reduced dramatically to about 120 CPM, which is still relatively (about 10 times) higher than background radiation. I followed up by stacking up to 4 sheets of plastic, which persisted in similar counts. This suggest the radiation has relatively high penetration power, possibly gamma ray photons. 

With this in mind, I proposed finding the inverse square relation using attenuated Am-241 source simply by placing a plastic "filter" in front of the radioactive source. By measuring the counts starting 7 mm, in 3mm increment steps, it took a total of 2.5 hours to get a full set of data and here was the plot:

Inverse square law for gamma radiation emitted from Am-241, 0.8 uCi source. 

It fits nicely to an exponential decay profile. According to theory, for every linear multiple of distance away from the source, the intensity of radiation, in our case, CPM, should drop by the factor of the squared distance. Which is quite a dramatic drop (a very steep decay curve) and my data is far from adhering this idealized case. 

In addition to the different penetrating strength of low-energy gamma rays emitted from Am-241 sources, an intelligent guess would be the intensity of gamma radiation was tempered by the absorption and scattering mechanism on the plastic sheet. Besides, as mentioned, the energy range of gamma rays emitted by Am-241 falls within 100 kiloelectron volt, not to mention some are in the soft x-ray range which readily interacts with low density materials such as polymers. 

For the purpose of this study, every gamma ray photon matters (pun unintended), therefore choosing the right attenuator I think, plays a critical role in demonstrating the inverse square law relationship. I reckon fully shielding the GM tube with a 3 mm thick aluminium plate to block away other low-energy gamma rays allowing the majority 59.5 keV photons to pass through, probably will get a more theory-consistent curve with that configuration. 

Experimental Proof of Planck's Constant using LED

I received a lovely book as a birthday gift from a friend whom I told once I was so keen to get myself a copy but couldn't afford at that time. Authored by a father-daughter duo, they are more than simple enthusiast for reproducing some of the greatest work in 20th century physics. Exploring Quantum Physics through Hands-on Projects really provides all the little ideas (and some not so little) for the home-experimenter to verify some of the important works in physics which has earned its pioneers their Nobel prize.

In this article, I'll give an account of an experiment I've done according to that was described in chapter 4. The basic principle discussed was quantization which the authors called "the core for quantum physics". In it, they described a simple but effective "visualization" of Planck's constant at work using LED of different colours. 

A simple circuit can visualize Planck's constant at work - proving quantization as part of nature's behavior.  

The theory suggest that a single coloured LED (not white) emit almost (but not quite) monochromatic light. Assuming they do, the energy required to generate particles of light (photons) streaming out of the LED, have energy described by simple equation:

E = hf = (hc)/wavelength , c is speed of light in vacuum

Which states the energy, E, of a single particle of light is directly proportional to its frequency under a physical constant called the Planck's constant, h. 

Now because the colour of light is associated with its wavelength (and hence frequency), not surprising it will require increasing amount of energy to power LED that generates light with short wavelength (or high frequency). The Prutchi duo indeed suggest the constant h can be approximated by measuring the wavelength (or frequency) of light and voltage required to "start" LEDs. This is only true by assuming the energy of the photon is a direct consequence of the electrons transit from a higher level to a lower level corresponding to the materials inside the LED:

eV = hf = (hc)/wavelength , assuming E = eV

where e is the charge of an electron and V is the voltage required to "switch on" the LED. 

I have five LED of different colour which I have their wavelength measured using an Ocean Optics USB spectrometer. The results were normalized for all the LED emissions and graphically represented below:

The spectral profile of LEDs. They emit light with one peak wavelength corresponding to their colour.  

Specifically, the spectral emission peaks are 631 nm, 595 nm, 561 nm, 527 nm and 456 nm giving uncertainty of 2 nm for spectrometer calibration. The wavelengths correspond to their appearance: red, orange, yellow-green, green and blue respectively. 

By connecting the LED in series with a potentiometer powered by a 9 V battery, the "switch on voltage" can be measured by placing a multimeter probe between the LED electrodes while adjusting the potentiometer such that the LED start to emit light. 

When I did the experiment, I found a major flaw in its argument. The authors of the book themselves explained that this method can only approximate the constant h due to the very nature of LED themselves. To say light emitted from electron transition due to externally applied electric potential inside the "active region" of the LED is an oversimplification of the real case. 

1. The emission of light, as I found, happens even in as low as 0.07 V. As long as the LED is properly connected to their terminals (sign not reversed), a slight potential difference across it will cause a tiny current flowing through the diode which cause it to emits light. 

2. When the voltage is increased to about 1 V, the LED gets progressive brighter and at 1.75 V for LED designated "red", the brightness suddenly increased indicating "switching on". This sudden "switch on" voltage varies with LED of different colour and it is the parameter which we measure. 

The two points above actually describes the character of a semiconductor diode which I will not explain its operating mechanism in detail. Briefly, the wavelength of light emitted is actually dependent on the band-gap energetics of the materials between a p-n junction. The voltage I measured was actually the "forward bias voltage drop, Vd" corresponding to the point when the voltage-current curve starts to exhibit exponential behavior. 

Photograph taken by myself showing the "active region" in a green LED. Notice the tiny bright cubic crystal sitting on the reflective cavity supported by the cathode anvil "flag". Two thin wires connects the crystal are the cathode and anode providing energy (electric potential) to create light.

Nevertheless, I tabulated the data and plot a graph that include both a "theoretical" computation assuming [eV = (hc)/wavelength] and "experiment" points:

Notice how far the experiment results deviates from theory which assumes energy needed to generate light, E = eV even after including measurement uncertainties. But because 5 experiment points cannot statistically justify the trend, I calculated the Planck's constant, h, based on the result obtained from the 5 LEDs.

The average turns out actually not bad when compared to the actual value of Planck's constant: 6.626E−34 Js. Which differs only about 0.2% considering an experiment based on so many assumptions. It need to be noted that this apparently "good" result is also derived from statistics (since we averaged it), so given if we have more LEDs which emits light in wavelength other than the five I currently have, the "new" averaged result might not be so beautiful. 

This experiment tells us one important thing: that quantization of energy in the form of light, is detectable even in a slightly complicated solid-state system such as the LED. Different colours of light indicated by its frequency regardless of where it comes from, whether from our star, the lamp in our house to the screen we touch can be regarded as little "packets" of energy emitted from them. Each little "packet" is an energy signature that is in the multiple of Planck's constant. They typically enters our eye in huge numbers having different energy individually for different colours. These little particles or "corpuscles" as Sir Isaac Newton calls it, brought us a beautiful universe through a sensory perception we call sight. 

Fluorescence in Highlighter Pigments

It took me a while to publish these new photographs thanks to work and all those things associated with bad publications. Luckily, my homemade set-ups never stopped during these times and I now managed some time to whip up a quick but interesting post. 

An acrylic carrom striker puck imbued with fluorescent material glows green when exposed to near ultraviolet light. The light source used for this photo was from a commecially inexpensive near-violet LED.

About a month ago, I received a complimentary 5-colour-in-1 highlighter as a door gift from an exhibition. Since I only use yellow and blue for all my highlighting jobs, I decided to use the pigments contained within these highlighters for the so-called UV photography. There are a few types of UV photography techniques but this article in particular focuses on a phenomena called "fluorescence". 

Fluorescence is a form of luminescence. An object is fluorescent if it emits light after absorption of light in another wavelength. Typically, light of shorter wavelength will be absorbed by a fluorescent material and re-emitted as light in longer wavelength. Fluorescent is familiar to us. Most people know shining ultraviolet (shorter wavelength) from commonly available "black-light" on bank notes will reveal strips of glowing colours typically red or orange (longer wavelength). These strips are essentially fluorescent paints that was printed on bank-notes for the purpose counterfeit identification.

Now as far as I am familiar with highlighters, I know they contain fluorescent pigments to exhibit bright colours. Pyranine is a known fluorescent dye that was added into yellow highlighters but I was curious to know what pigment is responsible for other colours or if they could fluoresce at all. So I disassembled the new 5-in-1 highlighter, removed the felt tip and washed all the spongy core with 20 ml of deionized water. It appears water is an excellent solvent for all the colours and I kept them separated with a glass vial for a while. 

Lately I came across a very cheap source of ultraviolet light. Apparently near UV light sources was already introduced commercially in the form of inexpensive light emitting diodes. These LED typically made from mixing different portions of Gallium Nitride (GaN) and Indium Nitride (InN) are able to release continuous near-violet UV light peaked at 390 nm. I bought my light source with MYR 5 from a night market and was able to demonstrate the following:

One colour did not glow under UV. Does it require light of a shorter wavelength to exhibit fluorescence? 

The picture above shows 5 different highlighter colours mentioned earlier. They were diluted to approximately the same concentration to about 3 ml each colour and kept in glass sample bottles (top). When shined with near violet light, only 4 colours appears to fluoresce (bottom). Perhaps blue highlighter needed another light source with wavelength shorter than 390 nm to cause it fluoresce? Needless to say, the photograph describes the relative intensity of fluorescence emission from the other 4 dyes pretty well. Based on visual observation, the green dye could actually share same fluorescent material with yellow. But we will never truly know unless we get a spectroscopic data from these samples.