Jump to content


* * * * *


Discuss this article in our forums


Rudy E. Kokich

Cosmic Rays are extremely energetic subatomic particles which fly through the universe at speeds approaching the speed of light. They were discovered in the early 1900's by Wolf, Pacini, Hess, and Millikan, who carried out extensive research on ionizing radiation using shielded electrometers, which measure electric charge on a metal plate or a ball. They showed that ionizing radiation was present deep under ground and under water, that its levels were much higher at higher altitudes, and that the levels varied at different latitudes on the Earth's surface. The results suggested that this radiation originated in outer space, and was composed of charged particles which are influenced by the Earth's magnetic field. The extreme energy of this radiation was demonstrated by the fact that it penetrated without significant attenuation to electrometers enclosed in thick metal containers.

With the development in subsequent decades of high altitude air travel and space travel, research on cosmic rays, and their effects on the human body and delicate electronic circuits became ever more important. By comparing ground-based and space-based research, it was confirmed that cosmic rays are not electromagnetic radiation, but subatomic particles which can be divided into two types: primary and secondary.

Primary cosmic rays are charged subatomic particles arriving from outer space at nearly light speed velocities. 88% are protons, or nuclei of Hydrogen. 10% are nuclei of Helium composed of two protons and two neutrons, also known as alpha particles. 1% are nuclei of heavier, more complex elements in the periodic table, present in approximately the same relative abundance as in the solar system. And 1% are isolated electrons. It is not known why electrons are markedly less efficiently accelerated than atomic nuclei.

In particle physics, energy is measured in electronvolts (eV). One eV, equal to approximately 1.6 x 10-19 joules (J), is the energy change on a single electron as it moves across the electrical potential difference of 1 Volt. On the average, a primary cosmic ray particle carries the energy of 4.8 x 10-11 J, which corresponds to 3 x 108 eV, or to the velocity of 2/3 the speed of light. However, few particles move at nearly the speed of light, and carry the energy of 3 x 1020 eV, which is - incredibly - sufficient to light a 100W light bulb for 1 second. This is also equivalent to the kinetic energy of a baseball flying at 100 miles per hour. By comparison, the highest energy particles produced by the upgraded Large Hadron Collider are in the range of 13 x 1012 eV, some 23 million times weaker. Such Very High Energy Cosmic Rays are fortunately very rare, with the estimated incidence of only a single event per square kilometer of Earth's surface per century.

Main sources of the primary cosmic ray particles are presumed to be shock waves from supernova events and magnetic fields around neutron stars, pulsars, and accretion discs around black holes in active galactic centers. Since the particles are charged, and deviated by strong galactic, interplanetary, and planetary magnetic fields, the direction of the original source can not be directly determined. The Sun also serves as an intermittent source of relatively low energy cosmic ray nuclei and electrons, which are accelerated by the shock waves in the solar corona and by solar flares. During periods of high solar activity, local density of charged particles may increase between 100 and 1 million times, and last up to several days.

When primary cosmic rays enter the Earth's atmosphere, they collide with atoms and molecules in the air resulting in Secondary Cosmic Rays: showers of high energy subatomic fragments which in turn decay into other particles and gamma rays. At the Earth's surface, the most common secondary particles are Muons, negatively charged Leptons, similar to electrons, but with 207 times greater rest mass. Only 3% of the cosmic rays on the surface are primary cosmic rays.

Muons are thought to be generated with the mean energy of 6 GeV (6 x 109 eV) at the altitude of 15,000 meters. They interact very little physically with ordinary matter but, since they carry a negative charge, they do lose energy by ionizing atoms near which they pass. The longer the path through a material, the greater the energy loss. At sea level, where they constitute about half the natural background radiation, the mean energy of muons is around 4 GeV. This allows them to penetrate deep under water, and over 700 meters under ground. To fully shield from most secondary cosmic rays would require a wall of iron 2 meters, or 6 feet thick.

Like other elementary particles, the muon can exist as ordinary matter: called negative muon, or antimatter: called antimuon or positive muon. Positive muons decay 100% of the time into positrons, antineutrinos, and neutrinos. However, negative muons of ordinary matter may follow one of two paths. Some decay into electrons, neutrinos, and antineutrinos, while others may rejoin matter by capture into very tight orbits around atomic nuclei where they combine with a proton to produce a neutron and a neutrino.

Muons are unstable elementary particles with a half-life of 2.2 µs (microseconds). This means that every 2.2 millionths of a second half of all the muons spontaneously decay into more stable particles. During that time light covers only 660 meters. Virtually no muons would survive long enough to traverse the 15,000 meter trip through the Earth's atmosphere were it not for the relativistic effects of time dilation. Due to their velocity, which is near the speed of light, time on a muon passes about 40 times more slowly than in our frame of reference, and its half-life is proportionately longer. This allows it a sufficient life span to travel through the atmosphere.

Muons can be detected with cloud chambers, Geiger counters, and scintillator detectors, but can also be recorded with common digital camera CCD and CMOS chips which are sensitive to charged particles. Muon flux at the surface of the Earth averages approximately 1 particle per square centimeter per minute. The surface area of the APS-C camera sensor (22.3 x 14.9 mm) is 3.3 cm2, which means that we can expect on average 3 muon strikes on the sensor during a 1 minute exposure.

The following methods were used to obtain images listed below:

- Camera: Canon T3i, lens removed and replaced with a light-tight body cap.

- Camera set to BULB for long exposures, with remote control shutter switch attached.

- Sensitivity set to ISO 1600. Higher sensitivities result in too much thermal noise. Less sensitivity results in less distinctive muon trails.

- Image quality in camera's Menu set to maximum resolution 18M (5184x3456), JPG format.

- In camera's Menu, go to Custom Functions, C.FnII:Image, turn on Long exp. noise reduction. This feature automaticaly takes an appropriate dark exposure to eliminate defective pixels in the original photo. In spite of this, some isolated defective or hot pixels may still appear on long exposure photos at high sensitivity settings. However, it is statistically extremely unlikely for several defective pixels to appear in a cluster or a row, adjacent to one another.

- On the Canon T3i camera, turning off Live View before taking a long exposure eliminates amplifier glow.

Examples of Amplifier Glow, and Defective or Hot Pixels and Thermal Noise on magnified images

- Since most cosmic rays arrive at high angles to the horizon, the camera is placed on its back so that the CMOS sensor is horizontal, and maximally exposed to the cosmic ray flux.

- The length of exposure should be 1 to 3 minutes. Longer exposures result in more thermal noise. Shorter exposures are less likely to capture cosmic rays. Keep in mind that the camera will then automatically take a "dark" exposure of equal duration.

- Images, which appear entirely black at low magnification, are magnified about 5x (approximately 20% crop) with image processing software, slowly scanned, and examined for "dots", "clusters", and "streaks". These are further magnified to reveal muon trails similar to those shown below.

-Image processing software used here for magnification and cropping was XnView, but a number of other excellent freeware programs are suitable.

-Pixel "clusters" are caused by muons striking the sensor at high angles, while "streaks" are created by muons striking at lower angles. There is no statistical confidence that single, isolated, illuminated pixels, or "dots", are caused by muon collisions rather than residual defective or "hot" CMOS pixels. Isolated "dots" should be excluded from the muon count.

Several interesting cosmic ray studies are available to amateurs and students.

The camera may be placed upright, with the sensor in the vertical position, which should, in theory, result in more "streaks", and fewer muon trails overall.

The camera may be covered with a thick metal pot, which should not result in the decrease in the muon flux since muons readily pass through matter.

Daytime and night-time muon flux may be compared during periods of known high and low solar activity.

It is reasonable to assume that the total number of illuminated pixels in a "cluster" or a "streak" is proportional to the energy level of the incident muon. With data from a large number of muon trails it is possible to draw a histogram with the number of illuminated pixels on the X-axis, and the number of muon trails on the Y-axis. Such a histogram would reflect the distribution of relative muon energy under the given experimental conditions.

There are several venues where tourists may visit deep under ground. One is the nickel mine in Sudbury, Ontario. Another is Ruby Falls in Chattanooga, Tennessee. If an opportunity presents to take measurements, muon flux and mean energy should be significantly lower deep under ground than on the surface.

Conversely, taking measurements during high altitude airplane flights should show more numerous and more energetic muon trails and, possibly, an occasional primary cosmic ray trail.

  • noisejammer, Snickersnee, LB16europe and 6 others like this


I think the main issue with CCD detection of cosmic rays is that the sensor is a small 2D type. Cloud chambers offer a 3D view with a nice time persistence of the particle trajectory visibility .

I may have read there are or were some cosmic ray detectors that looked like a stack of CCDs , with a computer reading all the "slices" in the stack and creating a 3D rendering by connecting the places where each individual 2D CCD array showed cosmic ray hits. Somewhat similar to CT scan imaging .

Even then, the number of slices is discrete rather than a continuum and the integration intervals are also discrete , while a cloud chamber can show multiple hits each with its own intrinsic timing . The CCD or whatever imaging sensor  has thermal noise that builds up to higher levels for longer integrations while the cloud chamber doesn't build up noise, it just waits with the same noise level regardless of how long the wait.

    • Lrsafari likes this
Mar 21 2016 07:53 PM

Interesting article but I think long exposure noise reduction is probably the worst possible way to correct for hot pixels and thermal noise in the presence of a transient signal.

To see why, note that the shutter doesn't impede muons meaningfully so that you can expect the same number of muon strikes (statistically at least) to occur in your dark frame. This introduces noise in your dark frame that is identical to your signal. The only way to correct this is to make a lot of dark frames and use sigma combine to eliminate random variation from the systematic noise present in the sensor.


Of course, the camera's temperature isn't controlled so getting this to work probably means some juggling of levels.  The idea would be to make your master dark is representative of your camera when the image is captured.

Very cool, thanks for posting this; I'll have to look for these with my cooled KAF-16803 sensor.  I.I. Rabi's reaction to the discovery of the muon was memorable: "Who ordered that?!".  I remember quantum physics courses at MIT long ago.  Fortunately, there was (and ) is not much in the subject that can actually be solved during the course of an exam; you need a lot of computer time as soon as you get to helium.  One thing thrown my way was to solve for "muonium";  a hydrogen analog in which the proton gets a muon instead of an electron as its atomic partner.  With the same charge as an electron, but 207 times its mass, its energy states are a lot different.  I could solve it then; now, not so much.  


One related bit: I expect that traditional silicon solar cells would also detect cosmic rays.  You might want to back bias them, and use a very sensitive transimpedance amp.  I installed 17 kilowatts of panels at our homes over the last year, but their micro inverters prevent you from using them to detect anything.  I have meant for some time to try a small panel and see if I could see anything.  A bit of our local uranium ore should show up as well.


All the best,



Very cool ! When I worked Aerospace (building satellites) Optical Metrology, we typically had lots of KAF16803s running continuous collects 24/7/365. Our data were of course loaded with Cosmic Ray Hits, which we had software seek out and expunge.  I got fascinated by these quantum events, so wrote my own subroutine to save them to "Toms Cosmic Rays." I put select "interesting ones" in a Rogue's Gallery that the other scientists would see. My favorites were "Cat's Claws" where a spray of particles apparently hit. The most common seemed to be plain vanilla point events though. This is a great "Astronomy!" thing one can do with any camera - even the broken ones! Just set it on continuous acquisition with shutter closed and play with the 1x1 download times. The other settings shouldn't matter if you collect Raw 16 bit cooled camera and process later for best visibility. Fun project that costs nothing and don't even have to go outside!  PS You can get 3-D "Cloud Chamber" by mounting a stack of very large chips like a layered cake. But that is expensive and can't cool them (easily).  Tom Dey

    • phaxman and Rich_B like this
Mar 22 2016 10:11 AM

I have done this inadvertently but came to the same conclusions.  You can see long one when the rays are parallel with the focal plane.  taking darks gets rid of the problem but it was upon inspection of the darks that i realized what was happening.

To clean hits from work-related images, We would take the median (vs the mean) of MANY darks to expunge cosmic ray hits from the darks. That was very effective. If you are willing to dedicate a camera to cosmic rays, it turns out that Short exposures are best. Although very few images will show hits, the ones that do will be very clean and free of thermal noise. The hits are of course temporal point events, so the signature of a hit is not compromised at all --- indeed improved and quantitative. Backouts should still be applied though, mostly to normalize out systematic biases associated with each pixel. A really complete cal would even characterize and apply the offset, gain and nonlinearity or each pixel. Certainly not needed for pretty pictures though. Ummm ... Oh yeah --- if doing short exposures, make a logic pattern-recognition filter to tell you which ones to examine. e.g. local way brighter than median works, cluster of hot pixels also works... lots of ways to do it. A really fancy filter throws out all the losers and stables the keepers to enjoy... Don't tell my boss I was doing that!  Or just say I am finicky to clean up the images.  Tom Dey

Thank you for very interesting comments.

Perhaps one of you can explain the meaning of COLOR in cosmic ray images. I presume that particle energy (speed) translates into pixel BRIGHTNESS and the NUMBER of illuminated pixels in a cluster or a streak. But I am at a loss to explain which property of the particle results in the variety of distinct colors.

Also, I would like to see some of your images, especially those of primary cosmic rays, 3% of which are supposed to reach ground level.


An interesting article, thanks.  Similar to what happens when a camera is exposed to a radioactive source emitting gamma rays, but the Muon tracks are tracks and not point-hits.  The only thing is, if noise at high ISO is an issue, one of the cameras using a Sony sensor like Nikon, Sony, Pentax or Olympus will do a better job than Canon, whose sensors are running a few years behind the pack at the moment.

Catching up here: It's important to keep in mind here that a DSLR still "thinks" it is receiving  RGB photons thru a lens. But the cosmic ray hits are energetic particles that plow right in and kick electrons around (interpreted by the camera as "photo-electrons"). The color-cal of the cam (more gain applied to R and B, less to G) etc. etc. are all percolating in the background and can't be turned off! This will result in "faux-color" images that would be difficult/impossible to interpret. Most ideal would be a good monochrome chip output as uncalibrated RAW 1x1 (unbinned). Even better would have the chip outside with minimal housing around it (which is getting kinda nuts). A lot going on there. Regardless,  I still favor just collecting them and enjoying the events.  Tom

Thanks for this informative article. I've marveled at the frequency with which these events appear on my 550D subs, very apparent every time I blink a stack in PixInsight.


I was reading up on drizzle integration and saw a raw sub from the HST that appeared to be lousy with these artifacts. Google showed me they shut down some instruments on the HST while over the South Atlantic Anomaly because of elevated cosmic ray events there. We're lucky to be relatively protected down here.







did any of you guys here about DECO (digital electronic cosmic observing)?


app for droid phones 


info is collect sent to university  and data reduction is done for you

then results sent back

Cosmic ray observing that is

How do you know they are muons as opposed to something else? I assume if they turn out to be more than a single pixel then it is not a camera issue. So I guess it would have to be some kind of particle event. I'm just curious how we know they are muons versus some other cosmic ray event. Incidentally, I think I just nabbed a couple myself. I have a Canon 5D Mark III. These are both from a single 2 minute exposure.




    • rekokich likes this

Statistically, at the Earth's surface, 97% of particles are secondary cosmic rays (muons and antimuons), while 3% are primary cosmic rays (protons, alpha particles, nuclei of heavier elements, and electrons). Since all particles have a different ratio between mass and electrical charge, I assume they can be identified in a cloud chamber by the degree of deviation within a standardized magnetic field.

I captured numerous cosmic ray images, and they all look relatively similar. I don't think the camera is suitable for particle identification.

However, I would really like to understand which particle property is responsible for the variation in pixel color.


I'll have to look into that color thing some more. I'm actually a physicist and have some background in particle physics, but I'm a theorist so my knowledge of how they interact with the chip itself is limited. Either way, this is a very cool experiment to do with my students. Thanks so much for bringing this to our attention!

    • rekokich and drbilliii like this

I took these pictures with my Canon 70D. Exposures were 60-90 minutes (ISO 1600) and the camera was in freezer (-18 Celcius) for three days. I believe these are traces of muons. The longest trace is 182 pixels, almost 1 millimeter long! The most bizarre trace is the one that seems to be curved!?



    • rekokich and drbilliii like this

I am asking for opinions on an idea.  If I take ASA 400 B&W film (supposedly pushable to 3200 ASA), and just keep it in black plastic as thin as keeps light out, will I get tracks (speckles or streaks) of radioactive particles (cosmic rays...I hope I'm not near any pitchblende) upon developing it?


My hope was to run trials with Nd magnet on a stack of 3 or 4, 4x5 sheets, and a similar stack without the potent magnet, but thought someone might know already or have a good idea.

@latune  You going to do it again?  I hope so.  I liked your curved trace at upper right, but it looks like two deflections to my eye...and that's bizarre enough!  To me, the nine pictures have multiple oddities, and the long straight trace in the group is just as odd as any other.


Any thoughts on why you guessed they were muons?

Beta particles (electrons) gamma rays and neutrons might all be visible impacting a CCD/CMOS chip.  There are youtube videos showing this.  If you have a granite countertop in your home, an old radium watch dial or alarm clock, or are in close proximity to a smoke detector, it seems they could impact your camera's sensor too.

    • rekokich likes this


I like your images. Please send us more.

It is a good idea to place the camera into a freezer to minimize sensor noise during long exposures.

Keep the camera in a sealed plastic bag to minimize condensation, and allow it to warm up upon removal from the freezer before opening the bag.


    • Nicole Sharp likes this
Nicole Sharp
Jul 27 2018 09:23 PM

How do you know they are muons as opposed to something else? I assume if they turn out to be more than a single pixel then it is not a camera issue. So I guess it would have to be some kind of particle event. I'm just curious how we know they are muons versus some other cosmic ray event. Incidentally, I think I just nabbed a couple myself. I have a Canon 5D Mark III. These are both from a single 2 minute exposure.





Use shielding.  High-energy muons can penetrate through thick concrete, but many other forms of ionizing cosmic radiation cannot.  Place the camera in a basement under some aluminum foil perhaps.

Nicole Sharp
Jul 27 2018 09:24 PM


I like your images. Please send us more.

It is a good idea to place the camera into a freezer to minimize sensor noise during long exposures.

Keep the camera in a sealed plastic bag to minimize condensation, and allow it to warm up upon removal from the freezer before opening the bag.



Placing the camera in a refrigerator or freezer is not a bad idea, particularly to use as a shield from other forms of background radiation.

Nicole Sharp
Jul 27 2018 09:27 PM

I wonder though if you place the camera next to a bushel of bananas, if you can detect the radiation emitted.  Bananas are often one of the most potent radiation sources in a typical household.  If you try to eat two thousand bananas, you might get sick :-/ .

I've got some 1/8th inch sheet lead.  That might work. 

Wow! And this is taken just with the help of a regular cam. I recall that some time ago I wanted to take a picture of the sky using a regular digital camera. Needless to say it failed: the stars that you can see with your eyes and enjoy the beauty of those small shining spots, appeared to be almost invisible in the photo. However, after that I started searching the Internet to find more on this and came across some interesting stuff. For instance, when a photographer is going to take a picture of the night sky it is required to set a long exposure. This is because we need to allow the light coming from the faraway stars to reach the cam. Thus the longer the exposure is, the more light will reach the cam and the more stars it would be possible to capture. Otherwise, the stars remain invisible to the cam.

Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics