How best to clean a sensor?

[Ultrapurple]

I have a nice camera but I think the previous owner may have left it lying around without a lens for a while, as the sensor window is somewhat mucky. The level of contamination is sufficient that it shows up (rather too well for my liking) on low-contrast scenes, though it's fine in high-contrast images. I have tried an air duster, with little effect, and I've also very, very, very gently brushed the edge of the sensor with a Q-Tip - to no avail. (I only tried the edge because if I'm going to cause damage I don't want it in the centre of the image!).

Does anyone have any tried-and-tested methods of cleaning sensor windows like this? I'm thinking in terms of "use a spectacles cloth moistened with 99.99% isopropyl alcohol" (possibly not a good idea) rather than just "treat it like a DSLR sensor". If it was a DSLR I'd send it to a local camera shop! I have a suspicion that the manufacturer, FLIR, would possibly charge a little more than I'm comfortable with (though I'm sure they'd do a fabulous job).

So, is there someone who knows exactly what to do and is prepared to share their experience? Please?

[Fraser]

Ask FLIR for a quote on sensor cleaning. You might be surprised. Then again FLIR might well charge a lot as it might include a camera test etc. No harm in asking.

From my perspective you need to consider the physics of the situation. The window can be scratched quite easily if a dust particle gets trapped between the window and the cloth. It just acts like sand paper. Will a mild fine scratch matter ? In my experience it will not effect imagery even at 640 x 480 as it is outside the focus point and so fine. The idea behind cleaning a sensor using the ‘drag clean’ soft cotton wool and IPA method is to very gently dissolve and capture contaminants without apply pressure to the sensor window, so avoiding scratches. The cotton is swept across once and discarded. The process is repeated until all contaminants have been removed. A totally clean sensor window can actually be gently ‘polished’ with a lens grade micro fibre cloth but you need to be certain the window is clean first. I use an inspection microscope.

Q Tips soaked in IPA are useable but only at the edges where dirt collects after drag cleaning. They are too hard and apply too much pressure if high levels of contamination are present. Remember, it is important to not press a sharp spec of dust into the windows surface as that is how the scratches form.

In a perfect world the sensor would be removed and its window washed under running cleaning solution to literally lift and wash away the contaminants. Such a process is expensive. Alternatives are special optical foam cleaners that use carefully selected chemicals to loosen the contaminants and the foaming action to lift them away from the surface for gentle removal with an absorbent material. Foam cleansing of optics has been around for many years and applies no real pressure to the surface being cleaned. This has now extended to cleaning cars paintwork ! The use of specialist foaming cleaners for cars has been around in industry for years but now the public have taken to the idea as well.

So in short, if you want the job done to a high standard by those who know best and do such cleaning all the time, send the camera to FLIR. The alternative is very careful use of soft cotton wool and IPA. Acetone is also used but I use IPA. I do not recommend any form of chamois pad or dense cleaning cloth, both can trap dust and cause scratches.

At the end of the day, it is your camera, your money and your risk  If you already send a SLR in for cleaning, why not at least enquire of FLIR. I would be more wary of cleaning an SLR though.... lots of risks involved there ! I estimate a fee of £250 from FLIR for such a task.

Another thing to consider is what the contamination actually is and is it ‘sticky’. Some contaminants can be very challenging to remove without applying pressure !

Fraser

[Bill W]

Work up in 'attack' slowly, as the likelihood of damage goes up as you proceed...
1 - IPA

2 - Acetone

3 - Sticky stuff remover
https://www.lakeland.co.uk/8976/Sticky-Stuff-Glue-Residue-and-Oil-Remover-250ml
(then IPA wash)

I have even heard of brake disk cleaner being suggested ..... by a sensor manufacturer 
https://muc-off.com/products/disc-brake-cleaner

There are swabs available rather than cotton buds -
https://www.coleparmer.co.uk/i/texwipe-tx761-cleanroom-swab-0-27-l-knit-polyester-tip-pp-handle-100-pk/3367760
One to apply/wipe, one to remove


Bill

[Ultrapurple]

Thanks both.

Bill - was it that specific brake cleaner that was recommended?

Until recently GBL was a common brake cleaner in the UK - it's quite a solvent - is largely unavailable now because it was often used to spike drinks (particularly in conjunction with alcohol, only a small quantity is required to induce unconsciousness, amnesia and, more often than not, major breathing disturbances similar to very severe sleep apnoea).

Some say it was a good way of dealing with insomnia, provided you're happy with the prospect of possibly never waking up again. To be clear: I do not recommend ingesting the stuff!

[Bill W]

Thanks both.

Bill - was it that specific brake cleaner that was recommended?

Yes, it was Muc-off 400ml aerosol.  They even had the MSDS linked, it contains nearly every solvent known to mankind.
I guess brake cleaner is intended for outdoor use by default, so the idea of it indoors seems particularly crazy

Bill

[flolic]

I repair photographic equipment for a living, and in my shop brake cleaner is the first choice for cleaning lens optical elements and digital camera sensors.
Second choice is the IPA.
I am using "Variac" brake&clutch 500ml aerosol by Henkel/Loctite. That stuff evaporates very quickly and does not leave any residue. Just put few drops on the cleanroom wiper and clean the surface.

[Fraser]

Just so users of some brake cleaning fluids/sprays are aware.......

https://envirofluid.com/articles/tetrachloroethylene-a-deadly-danger-in-brake-cleaner/

Fraser

[Fraser]

Looks like many modern common brake cleaning sprays contain just Naphtha and CO2 propellant. A common solvent with fast evaporation and little, if any, residue.  Note that some brake cleaners have up to 30% IPA included with the Naphtha. Check the MDS before buying if you want Light Naphtha only.

Fraser

[Fraser]

I have been looking through different brake cleaning products and found that many state 60-100% Light Naphtha on the MDS but do not state what the other possible 40% could be. Not something I would trust on an expensive, irreplaceable lens. Some product MDS documents clearly state a significant percentage of IPA is present. I checked and almost pure (99.8%) Light Naphtha is available for use in laboratories. I think, given the choice, I would buy 99.8% Naphtha in preference to brake cleaner. Why ? Well ask yourself what the manufacturer of brake cleaner is expecting you to do with it  . Cleaning dirty metalwork and brake pads, thats what .... hardly as demanding of a quality controlled product as an optical cleaning solution ! So whilst it CAN be used, I will stick to laboratory grade chemicals, as I always have, when working on lenses 

Fraser

[Bill W]

Here's the Muc-Off MSDS extract


ethanol
25-<50%

Acetone
10-<20%

butane (containing < 0.1% butadiene (203-450-8))
10-<25%

propane
10-<25%

1-methoxy-2-propanol
3-<10%

propan-2-ol
2.5-<3%

n-butyl acetate
1.0-<2.5%


Bill

[Fraser]

So basically an aerosol acetone but with other stuff added for good measure. Personally I subscribe to the ‘less is more’ philosophy and if a single component cleaning chemical is known to work, I use it. Note the large percentage ranges given in the MSDS ..... it is like they just throw some solvent into a big container and use it with little control over content balance. Nope I will be sticking to Pure Acetone, Pure IPA and now Pure Naphtha 

If Muc-Off works well though, no big issue, but a combination of more solvent types can open the possibility of more chance of interaction with surrounding materials. I do not spray solvents onto sensors or lenses for the same reason..... overspray can get where it can cause issues. In aerosol can solvents the propellant is often Butane and/or Propane which can also act on surfaces in undesired ways.

Keep it simple, keep it safe   

Fraser

[Ultrapurple]

Well, I now have a choice of 96° ethanol, denatured, or 99.9% isopropyl alcohol. I have obtained some lint-free swabs (which look generally like single-ended, anorexic Q-tips), and I have tried cleaning a scrap lens.

So far, so good. I started with the ethanol but found the swabs difficult - they don't appear designed to be used end-on; side-on is much better. End-on they cover less than 1mm. So I swapped to Q-tips to work on a larger area at a time. (Remember - it's a scrap lens so I can afford to be lax).

Using the Q-tips with a reasonable amount enthusiasm (several grams of force, and rubbing), I got a lot of black muck off the lens. I took care only to work on a portion of the lens so I could check 'cleaned' vs 'dirty'. And of course see if I'd obviously lifted any coatings...

The result appeared good. There was some apparent residue but repeating the process mostly cleared that up.

I plan to have another go on another piece of scrap, coated germanium lens and see what happens if I use the proper swabs. Ultimately I will only be cleaning a small area - the surface of the sensor window - so as long as I can make progress and not scratch or otherwise damage the surface, I might have a go at the outside of the sensor window. If I confine my initial efforts to parts that aren't critical to image quality I'll feel more comfortable and all being well I'll be able to leave it for a while to check that nothing nasty happens after a day or two.

I'll report back.

[eKretz]

I clean my lenses and sensors with a q-tip wrapped with a Kimwipe lens cleaning cloth and acetone. Change the q-tip and rotate the cloth for every swab/wipe on sensitive stuff.  That method has never scratched anything I've used it on and works very well.

Note that MSDS percentages aren't usually indicating that sometimes it might be one mix percentage and sometimes another; they are just being intentionally vague with their percentages so that their product can't easily be exactly copied.

[cnxunuo]

you might have took it too seriously....I have cleaned many many of my 640x512, LWIR and MWIR using brute force level pressure and tap water with whatever shirt I an wearing at the day, they all came out fine. Thermal imaging sensors has pathetically low resolutions, unlike DSLR, therefore even a bunch of micro scratches wont do anything to the image....

with one exception, boson has exposed bond wire, care must be taken not to touch them when cleaning. if it's a tau2 or ULIS, give it a good polish and blow the dust off. Dust left on sensor surface will have impact on circular artifact.

[Fraser]

Just some comments on using “tap water” .......

1. Depending upon where you live, tap water can contain significant solids is suspension. An example is water from my home town of Bath. It contains very high levels of lime in solution that is a very effective abrasive on delicate surfaces ! It will also evaporate leaving lime marks on a surface. Rubbing the lime off causes severe scratching.

2. If it is desired to use water for cleaning or even mixing with IPA to form a more effective clearner, it is wise to use distilled water as that contains no solids in suspension. Such water is cheap so why not use it ?

3. Modern thermal cameras can use a variety of materials for lenses and windows. It should be remembered that Germanium is a metal and that water definitely causes it to oxidise at the surface. If water gets under the hard carbon protective layer of a lens it will cause corrosion that lifts the AR coating off of the lens. ‘spider web’ corrosion is common in cameras that have had extended exposure to damp conditions. If a lens comes into contact with water, it should be carefully cleaned and dried with warm air to avoid any water remaining in contact with the lens coating for an extended period of time. If the AR coating is already compromised by a scratch, even greater dare with water is justified. Chalcogenide glass IR lenses contain Germanium but may not be a vulnerable to oxidisation issues cause bu water.

4. Those of us who were involved in buying new thermal imaging cameras over our careers have been trained to treat Germanium lenses with enormous respect and care. The reason is very simple, some of the lenses I purchased cost £25K and could be ruined by inappropriate cleaning. No one wants to get the reputation of being the idiot who wrecked a £25K lens ! We actually sent our cameras and lenses to AGEMA and FLIR for cleaning. The value of the cameras and lenses justified such action. Modern lenses are smaller and cheaper. They may also have more resilient coatings on them and the risk of damaging such a lens may be lower. All I know is that when repairing someone else’s property I take enormous levels of care with it, as I would my own kit. The idea of scratching a camera lens when such can easily be avoided is abhorrent to me. That is just me though. Others may think differently and know from experience that minor scratching is not a big issue at these resolutions. I prefer to keep my lenses and windows scratch free though

Fraser

[cnxunuo]

Yeah probably UK tap water isn't good for the job. In China, tap water usually go thru a reverse osmosis (RO) treatment before hitting each household tap, which essentially makes them de-ionized water. I found that the lens coating is generally much easier to damage compared to sensor coating, maybe this is because a lens has multiple air-Germanium interface,  which needs good coating, but sensor presumably came with a worse performing but more robust coating of some sort.
An extreme example with coating on lens that I messed up due to west Yorkshire's horrible tap water is a nice 50mm, with huge image circle covering 640x480 on 25um sensor, the rear element thick coating just curled up and fall off after I rubbed the dirt off.

[Cat]

Any news regarding cleaning of the sensor window?
Someone left his fingerprint on the window of this sensor, probably an ULIS Pico640.
The image is blurry or unsharp on the left, but the fingerprint should be on the opposite.
Attached are two images with a tower in the distance. Of course teardown and pictures will follow...

[Fraser]

Cat,

Many thermal cameras operate with an inverted microbolometer to match the inverted output of the lens assembly. As such a mark on the right of the microbolometer window will appear on the left of the output image and a mark at the top of the window will appear at the bottom of the output image. A designer can design the optical block to present an upright image to the microbolometer to avoid the inversion of the FPA, but there is normally no need for such and the additional optics can increase cost, weight and complexity.

Fraser

[Ultrapurple]

If it's a god quality thumbprint then perhaps the FBI could track down the culprit? 

[Cat]

Success! 
The sensor window was cleaned with 99,9% IPA and Q-tips with additional cotton wool.
The IPA was taken with a new syringe directly from the storage container to avoid contamination. Cotton wool was wrapped around the Q-tips to get them softer and to cover more area. The bulb blower was used to remove any non sticky particles. The cotton wool was soaked with IPA and rolled on the sensor window. After a few swabs the fingerprint dissapeared but the drying IPA left a slight film on the window after drying with the bulb blower. A second fluffy Q-tip was slightly damped with IPA and removed the residue, the IPA was drying in a few seconds.
The liquid on the edges of the ceramic package is probably silicone oil from the thermal pad and removed with IPA afterwards.

I tested my available solvents on a glass plate before, and IPA was the lesser evil. Everything else left even more residue:
99% Ethanol (denaturated with Bitrex), Brake cleaner (acetone free) and Teslanol SP (special cleaner normally used for switches and advertised as "no residue") left a film.

Many thermal cameras operate with an inverted microbolometer to match the inverted output of the lens assembly. As such a mark on the right of the microbolometer window will appear on the left of the output image and a mark at the top of the window will appear at the bottom of the output image.

This is what I assumed, the fingerprint was on the side with the sharp image. The focus was not set correctly and there might be a small problem with coplanarity, but an object in focus stays sharp on both sides.

[_Wim_]

Success! 
The sensor window was cleaned with 99,9% IPA and Q-tips with additional cotton wool.
The IPA was taken with a new syringe directly from the storage container to avoid contamination. Cotton wool was wrapped around the Q-tips to get them softer and to cover more area. The bulb blower was used to remove any non sticky particles. The cotton wool was soaked with IPA and rolled on the sensor window. After a few swabs the fingerprint dissapeared but the drying IPA left a slight film on the window after drying with the bulb blower. A second fluffy Q-tip was slightly damped with IPA and removed the residue, the IPA was drying in a few seconds.
The liquid on the edges of the ceramic package is probably silicone oil from the thermal pad and removed with IPA afterwards.

I tested my available solvents on a glass plate before, and IPA was the lesser evil. Everything else left even more residue:
99% Ethanol (denaturated with Bitrex), Brake cleaner (acetone free) and Teslanol SP (special cleaner normally used for switches and advertised as "no residue") left a film.

Many thermal cameras operate with an inverted microbolometer to match the inverted output of the lens assembly. As such a mark on the right of the microbolometer window will appear on the left of the output image and a mark at the top of the window will appear at the bottom of the output image.

This is what I assumed, the fingerprint was on the side with the sharp image. The focus was not set correctly and there might be a small problem with coplanarity, but an object in focus stays sharp on both sides.

Thanks for the detailed description. Good to know for other optics cleaning also!

[johnelot]

is the window on the sensor made out of germanium ?   i looked at a couple thru the microscope and they seem to have tiny gold flecks in the window material 

[Fraser]

The window is normally AR coated mono-crystal Germanium or Silicon in a LWIR Microbolomter. There are more exotic window materials but I have not seen them on a microbolometer window.

Fraser

[jeffreythe00]

This is a pretty expensive option if you aren't regularly cleaning optics but I have found that this product works very well for cleaning laser optics. I've also used it to clean my Therm-App germanium lens and sensor window without damage. https://www.photoniccleaning.com/ Just thought I'd add my 2 cents

[Uho]

The photo shows spots. Dirt on the stain matrix gave. I wiped the matrix with a napkin. The spots have disappeared.