Friday Night Lights: Taking A Look Inside A Night Vision Image Intensifier I² Tube

    When talking about image intensification (I²) with regards to analog night vision we are talking about the heart of night vision, the Image intensifier I² tube. The I² tubes are what performs the actual magic of amplifying light so we can see in the dark without the use of flashlights.

    Looking Inside An Image Intensifier I² Tube

    My wife works for ZEISS in their EM (electron microscope) RMS (Research Microscopy Solutions) division. The SEM (Scanning Electron Microscopes) that my wife uses are good for looking at really small samples but they do not really penetrate and look at things inside an object like an X-ray system can. At the ZEISS office, they have an Xradia Context MicroCT. Last October, ZEISS had an open house for the family members of ZEISS employees.

    Our daughter looking through a ZEISS LM (Light Microscope) during the ZEISS open house.

    For the demonstration, they showed a 3D X-ray scan of a spider taken with a ZEISS Xradia Context MicroCT. Once scanned they can manipulate the 3D image like a CAD drawing and peel back the scan in all 3 dimensions. So you can look at cross sections of the spider without having to physically cut the spider. The wheels were turning on how I could use this technology for some Chenanigans.

    Fast forward to December 27th. My wife’s co-worker had some free time to indulge my curiosity about night vision. I brought a dead image intensifier I² tube for him to scan. It is an old tube that I got for free that no longer fires up. We were not sure what an X-ray would do to a working image intensifier so I did not want to use my newer and better image intensifier tubes.

    This image tube was made by Intevac. They were the company that got the first ANVIS contracts. They produced the first 10160 tubes that were used in the AN/AVS-6. For those not familiar with what I just said, ANVIS or AN/AVS-6 are dual tube night vision goggles used exclusively for flying aircraft. The 10160 tubes are a type of image intensifier I² tube that does not have manual gain.

    This particular tube came out of an old B.E.Meyers Dark Invader OWL and is a Gen 3 Omni II era tube. For some more information regarding Omni contract classifications check out this post by David aka CJ7Hawk on

    My wife’s co-worker, Robin, had to mount the tube on a stand so the Xradia could scan it. He used scotch tape to hold the tube in place. I asked him if the stands need to be made of any special material. He said no but most of his job is fabricating some sort of stand just to hold the sample in place for the Xradia to scan it. However, he does not have special materials so it becomes more “arts and crafts” at work trying to use what is available to make said stands. So I recommended that he should get some LEGO elements so he could build a custom stand. Once he is done he could break apart the elements and reuse them for a future project.

    Below is the computer interface for the ZEISS Xradia.

    Once the Xradia is setup and closed you can initiate a preview scan. This is for moving and orienting the sample so when you scan it you can get the image you want.

    See that solid mass in the middle of the image tube? That is made of glass and the Xradia cannot see through it. That is the fiber twist. Keep reading to learn what that is.

    The final scan takes a few hours but the results are amazing. Below is a cross section of the Image intensifier I² tube. I can move a slider in the software to peel back layers of the image and look inside the Image intensifier I² tube.

    Image intensifier I² tube cross section

    The giant circle in the center is made of glass fibers.

    While the cross section above is neat, it just looks like a boring medical X-Ray. Here are some exported images from the ZEISS Xradia software that show the image tube in a better position and light. You can move the 3D scan in all 3 dimensions as well as assign color palettes for certain densities.

    While these images are pretty I did not really know what I was looking at. For that I refer you to David, CJ7Hawk of and his post about what is inside a tube.

    For those who think analog night vision should be cheap like digital cameras, think again after you read the information below.

    Inside an Image Tube.


    Getting a feel for the inside of an image tube can be difficult if you’re one of the many who has never taken one apart. This is a quick article on what exists beneath the plastic boot that hides what lies within.


    Although we (mostly) all kind of know what’s inside of an image tube, have you ever considered what it all means? Having a good image in your mind helps you understand the technology, it’s limitations and the scale on which things occur.


    So for those interested, here’s a brief journey through a modern Gen3 image tube. And for those who like to have a yardstick to measure things by, the average thickness of a human hair is about 100 microns, or 100,000nm or 1 million Angstroms.


    The Photocathode.

    The photocathode is a piece of glass around 5mm thick that’s clear on one side and on the other they grow a Gallium Arsenide Crystal ( GaAs ) that acts as the photocathode. That is, it converts photons into electrons.
    This is straight photoelectric effect. Gallium Arsenide is a semiconductor and you can make electronics out of it. They make very expensive solar panels out of the stuff also. Excellent for satellites and the like where high efficiency solar cells are needed, but the cost of satellites makes using this expensive stuff worth it.
    They grow the photocathode right there on the glass faceplate like how mist forms on glass when you blow on it, except it doesn’t evaporate. It grows… Just vapor but it builds up over time.
    And they build it up like a layer-cake because it’s not just a single piece of GaAs.
    You might think that in itself is unusual, but there’s more to it than that. First of all you know how they coat optics for riflescopes and the likes? Well that glass window is an optic, so they use a coating, approx 1/4 wavelength ( eg, around 200nm ) of Silicon Nitride to act as an antirelfective surface and pass the light on towards the photocathode.
    The next layer to go down is an Aluminium Gallium Arsenide layer about 5 microns ( 5000nm ) thick. This acts as a bandpass filter ( above 600nm ).
    Next they grow a single crystal of zinc-doped GaAs that’s between 1 and 1.8 microns thick. This is the active layer. Then finally they add a monolayer ( think 1 atom thick ) of Cesium and apply oxygen to convert this to Cesium Oxide. It’s this change that causes the band-bending of the GaAs and results in the negative electron affinity ( NEA ) that I mentioned in the Gen4 discussion.
    That’s a simplification of the process, but consider how thick this photocathode is… Then compare it to a Gen2 photocathode that is about 300nm ( 0.3 micron ) thick. Big difference eh?
    The MCP.
    The Microchannel Plate is the best-known part of Gen2 and Gen3 image intensifiers. These have been around for decades but even now research is going on to improve them.
    They take an etchable glass and surround it with a lead-silicate glass, create a boule and draw it through a glass-fiber furnace. They take a bunch of these glass fibers and put them together ( usually into a hexagonal shape ) and draw them once again and once more stack these together and fuse them. Because the fibers get progressively smaller each time they draw them, they end up with a rod made from millions of glass fibers.
    They then slice this at an angle – usually about 6 degrees and into slices approximately 0.3 to 0.4mm thick. These are then etched to dissolve the etchable glass and this produces millions of small holes in the plate at the same angle at which the bundle was sliced. These are the microchannels.
    They coat the outsides with a conductive layer of nichrome and the MCP is finished, ready to use ( after a little scrubbing, but that’s not for this explanation today ).
    HOWEVER, the Gen3 MCP also has an ion barrier film. They float laquer on water and create a self-supporting laquer film by lowering the water onto a supporting frame and letting the laquer dry. Then ( in a vacuum ) they place this on the MCP, coat it with Aluminium Oxide to a thickness of about 30 Angstrom ( 3nm or 0.003 microns ) and sinter this by baking to create an aluminium oxide film over the input side of the MCP.
    Kind of like wetting pastry, coating with spices and baking away the moisture.
    At 30 angstrom the film will stop any electrons that don’t have enough voltage to push their way through. This is known as the dead-voltage and the result is that not all electrons get through the film. The film stops damaging positive ions that are developed in the MCP during operation from reaching the photocathode, but the downside is that some electrons are stopped from getting to the MCP also.
    This effectively reduces the sensitivity of the tube by around 50 percent in a standard filmed tube.
    You can see the edges of the MCP in the ZEISS Xradia scan.
    The Screen.
    There’s a fair bit about screens that aren’t widely known. The technology dates back to vacuum tubes before WW2 and it’s basically the same as the old CRT TV sets that everyone on this forum knows well.
    They call them phosphor screens because they convert electrons to light, but that doesn’t mean they use phosphorus.
    In fact, they make them underwater. Imagine mixing all the stuff you need and shaking it up, pouring it into a big glass with the screen substrate and allowing it to settle. That’s how they do it. All the phosphor particles settle on the glass or fiber plate that forms the screen.
    Add some sodium silicate or potassium silicate as a binder, some barium acetate as an electrolyte and when it’s settled, let it dry in the air. The optimum thickness for a screen using 8 micron particulates is around 8 micron’s thick. This will allow a resolution of over 100 lp/mm – sufficient for the resolution of the MCP.
    Then they deposit an aluminium coating over the screen in argon gas to provide suitable conductive medium and to keep the light from getting back to the photocathode for obvious reasons.
    There are other ways to do it, but this gives you an example and probably impresses on you that these guys would make good chefs.
    There are quite a number of screens that might be used on a gen3 tube. Two common ones are P20 and P43.
    Phosphor screens exhibit two forms of function. Fluorescence and Phosphorescence. Fluorescence is when it lights up when struck by electrons. Phosphorescence is how long it remains lit up afterwards.
    P20 is a yellow-green phosphor with a peak output at 560nm. It decays to about 10% brightness in 0.2ms.
    P43 is also a yellow-green phosphor with a peak output at 544nm and decays to about 10% brightness in 1.2ms.
    This is probably why P43 is more popular as a screen. It’s extended phosphorescence allows the image to appear brighter to the user under extremely low light conditions.


    The Fiber Twist.
    ( as shown above, shorter on the MX10130 that does not have a twist in the screen FO plate. The twist can be easily seen in the MX10160 on the right. )
    Remember the description of how the MCP is made? Well the fiber twist is made the same way, except it isn’t etched.
    Fiber Plate is an invention from the 1950’s and it was what made the first starlight scopes possible. It also allowed the glass to be shaped and this reduced the distortion of the tubes as well.
    Well, they grab a piece of rod made from the stuff, heat it and twist it 180 degrees. This allows the image to be inverted which simplifies the optics of the scope in which the tube is to be used.


    If you scroll up to the photo of the ZEISS Xradia control monitor you can see the preview side profile of the dead Intevac tube. There is a dark mass that the XRAY has difficulty seeing through. That is the glass fiber twist.

    Power Supply.
    The power supply is a separate component, but it too fits in the package. It takes 3v and runs this through an oscillator and after building up the voltage to a higher level, it’s put through a cockroft/Walton generator ( ) to build up the voltages to the required level.
    There are three outputs (and one ground) from a power supply.
    1. The Screen. This is kept at a potential of around 5000v.
    2. The MCP input. This is typically kept at a potential between -200v and -800v.
    3. The Photocathode. This is typically kept at a potential around 800v lower than the MCP input – so around -1600v.
    The MCP output is kept at ground potential.
    In thin-filmed tubes, the photocathode is at about -600v beneath the MCP input and in filmless, it’s approximately -200v below. This voltage variation is most likely necessary to ensure that most electrons make it through the film.
    Power consumption of the power supply is around 40ma at 3v. Not much.


    Now you have all the parts, it’s time to put them together. Under a vacuum, they use bellows to hold the pre-assembled parts apart while they activate the GaAs photocathode ( The Cesium oxidising process ) and then they fit the parts, separating the sections with ceramic rings holding apart metal rings that keep everything in place.
    Indium ( a soft metal like solder ) seals allow for air-tightness and the process is complete. Small metal tabs stick out through the metal rings and these are soldered to the power supply. The tube is filled with an insulating compound, an elastomer made from silicon. This is also known as the potting.
    Then the tube is housed in a suitable form ( eg, Thin Anvis ) for it’s intended purpose.
    Inside, the parts are now very close. The photocathode is within a few 10ths of a millimetre of the MCP and the screen is just a few mm away.
    This is why the tubes can be damaged by shock. The MCP can flex enough to actually touch the Photocathode. With the voltages involved, this will burn a hole in the photocathode and damage the tube permanently.
    Well, I hope this has been interesting. I’ve simplified this process as much as I can. If you have questions, please let me know.
    If you would like to get deeply into what makes an image tube tick, I’d recommend a book called “Image Tubes” by Illes P Csorba. It’s not in print anymore but often shows up on as a secondhand book.




    Thanks, David, you are a wealth of knowledge. Hopefully you, the reader, have learned something new about night vision and gained a better understanding of why analog night vision is expensive. They are labor intensive and according to many sources, this technology has gone as far as it can. Digital night vision is evolving at an exponential rate but it has not surpassed this technology . . . yet.


    Nicholas C

    Steadicam Gun Operator
    Night Vision & Thermal Aficionado
    Flashlight/Laser Enthusiast
    USPSA competitor

    Any questions please email him at [email protected]