Properties of common cheap UV-C LED chip

[Lou Jost]

WARNING: UV-C light is dangerous and requires protective gear!!!

 

I bought some very cheap UV-C sterilizer devices as possible sources of light for deep UV microscopy. 

 

 

Upon examination, they all seem to use more or less the same chip.

 

I wanted to test its properties and got the following series:

 

 

 

These photos were taken with a full spectrum Sony A7R and a Laowa 2.5-5x lens at 5x (a special lens with very thick glass elements, a poor transmitter of UV-A light, which is just perfect for this application) , except for the last frame. They show that there are two LEDs; the lower one emits light with center wavelength around 365nm, and the middle one emits UV-C.

 

The first picture shows the unpowered chip, and the second photo shows the chip powered up. No filters were used for these. They were manually white-balanced.

 

The third picture in the panel (filtered with an Astronomik sharp-cut filter which blocks UV and IR, plus a Zeiss T* filter to doubly block the UV light) above shows that these LEDs both emit some visible light, but the middle one doesn't emit much. 

 

The fourth picture is taken through a Hoya R72 filter and shows that there is some IR emission, mostly from the lower LED.

 

The fifth picture shows the lower LED is the one that produces the UV-A light. WIth a much longer exposure, a small amount of 365nm emission can also be detected from the middle LED.

 

The last picture in the panel was taken with a different camera, a monochrome MaxMax 8Mp camera with no coverslip, through a quartz lens. It shows that the middle LED is the one that produces the UV-C light. I also found that some light from this LED  passed a 254nm Edmund narrowband filter (40nm FWHM).

 

 

 

[Lou Jost]

I'd like to disable or cover up the 365nm diode. Any thoughts on how to do that? It probably gets really hot so it might not be such a good idea to cover it up, but maybe it could be covered by something metallic that would conduct the heat...

[Andrea B.]

Could you use electrician's tape to cover the LED with the caution that you don't have the light turned on for more than n minutes? Not sure what n should be.

 

Could you paint over the unwanted LED with some kind of black paint which, we would hope, cannot catch fire if it gets too hot?

 

Dr. Klaus once covered an unwanted aperture window on some lens with black nail polish. But I don't know if black nail polish is flammable when dried.

 

 

I just googled:  "Is dry nail polish flammable?" and got this.

Dry nail polish is not flammable though it can still catch fire.

If dry nail polish can still catch fire, doesn't that mean it is flammable? 

The proper question is probably "Is dry nail polish combustible?"

 

 

[Lou Jost]

Thanks for the suggestions Andrea. Even if something is not flammable, I suppose it could boil or decompose under high heat...Now I am thinking that it would be even better to make some kind of metal sliding cover that could switch between cobvering UV-A and UV-C, as needed. Some metals are black under UV light, I think, so they absorb the light and conduct the heat away.

[Foxfire]

Strange device. If it was cheap and the 365nm diode is not wanted, then maybe just to remove it? Seems like the UV-C one should still work if the 365nm one is removed, althoug can't tell for sure.

 

Edit: probably it's small circuitry, but maybe some swith can be integrated...   

 

[Lou Jost]

18 minutes ago, Foxfire said:

Strange device. If it was cheap and the 365nm diode is not wanted, then maybe just to remove it? Seems like the UV-C one should still work if the 365nm one is removed, althoug can't tell for sure.

As you say, we can't be sure. Electronics are complex and removing a component can change the resistances and voltages of the system.

[Doug A]

@Lou Jost nice comparison. I'm surprised there is a 365nm led. Thought the germ killing mostly happened in the UV C range. You could use black muffler paint. Motorcycles probably emit far more heat than a LED. 

Looking forward to more images.

Thanks for sharing,

Doug A

[Lou Jost]

4 minutes ago, Doug A said:

@Lou Jost nice comparison. I'm surprised there is a 365nm led. Thought the germ killing mostly happened in the UV C range. You could use black muffler paint. Motorcycles probably emit far more heat than a LED. 

Looking forward to more images.

Thanks for sharing,

Doug A

That's a great suggestion; I would never have thought of it.

 

I suspect there is some regulatory reason for the 365nm LED with some violet leakage. Without that, it would be hard to tell if the light was on, and that would be really dangerous with UV-C. It might also make the light appear more powerful, which would be a selling point for these cheap devices. I imagine most consumers are completely ignorant about the varieties of UV light and would not like a light that is completely invisible and doesn't even cause stuff to glow very much.

[dabateman]

Could you try aluminum foil? 

That should block the light and act as a heat sink to wick away some of the heat.

 

[Lou Jost]

10 minutes ago, dabateman said:

Could you try aluminum foil? 

That should block the light and act as a heat sink to wick away some of the heat.

 

Yes, I was thinking of that, and it's worth a try.But it would reflect the 365nm light back into the well containing the LEDs and I think much of it would leak out again through the uncovered part of the chip.

[colinbm]

Lou, this 275nm is in the very dangerous region of UVC, much more dangerous then 254nm.
Between 260 & 280nm is the most dangerous / damaging for living tissue, eyes & skin !
This device is made in China & is most likely fake & useless & very low powered.
My suggestion is to give up on it.
Having suggested that, perhaps black rubber will cover the 365nm LED & be resistant to the heat ?
 

[Lou Jost]

1 hour ago, colinbm said:

Lou, this 275nm is in the very dangerous region of UVC, much more dangerous then 254nm.
Between 260 & 280nm is the most dangerous / damaging for living tissue, eyes & skin !
This device is made in China & is most likely fake & useless & very low powered.
My suggestion is to give up on it.
Having suggested that, perhaps black rubber will cover the 365nm LED & be resistant to the heat ?
 

Thanks for the cautions. I didn't know this range is more dangerous than the 254nm range; I figured the more energetic light would have been worse.

 

I don't have other options here at the moment. I had also bought a professional 290nm LED, heatsink, and driver from Boston Electronics, but I can't get it to work.

 

As my photos show, these cheap Chinese LEDs are definitely not fake, in the sense that they demonstrably emit useful light in the wavelength range that they claim. Though they are weak, they may be strong enough for some applications, hopefully good enough for transmitted light microscopy. That's my hope anyway. I have several of those wonderful hundred-year-old Zeiss Monochromat quartz objectives, corrected for 275nm and 254nm, and I am thinking of using one as an objective and another as a condenser. I'm experimenting with tiny fused silica hemispheres to put on top of the LED to send light into the condenser. It is a low-budget version of Jonathan's path towards UV microscopy.

[colinbm]

Thanks Lou
222nm Excimer tube lamps are said to be safe for humans.
What UVC illumination is Jonathan using ?
Good luck with your UV Microscopy, I am looking forwards to your results.

[Lou Jost]

1 minute ago, colinbm said:

Thanks Lou
222nm Excimer tube lamps are said to be safe for humans.
What UVC illumination is Jonathan using ?
Good luck with your UV Microscopy, I am looking forwards to your results.

Jonathan has done a lot of 313nm work, but wants to go lower.

[colinbm]

29 minutes ago, Lou Jost said:

Jonathan has done a lot of 313nm work, but wants to go lower.

Me too, but I can only do UVIVF.

[lukaszgryglicki]

222nm - wow, my previous "target" - unfortunately China lamp was broken and not working, and I lost around $275 :(

 

[Foxfire]

2 hours ago, colinbm said:

222nm Excimer tube lamps are said to be safe for humans.

that is interesting information - 222nm safe for humans, 275nm very dangerous. 
Upon checking, while radiation around 275nm may be the most dangerous, especially for the eyes, still, what is suggested (SCHEER 2017) is that 222 nm UV-C radiation at low doses (below the bacteriostatic effect threshold) is capable of inducing both erythema and cyclobutane pyrimidine dimers formation in human skin. So, it would be better to be cautious when some product advertisement says 222nm is safe for humans. 

 

Adding note @Lou Jost about using aluminium foil that this gold-yellow surface of the chip may be conductive

[JMC]

3 hours ago, Lou Jost said:

Jonathan has done a lot of 313nm work, but wants to go lower.

I've done some work at 254nm as well. There's a summary here which I need to update - https://www.ultravioletphotography.com/content/index.php?/topic/5456-uvc-safety-warning-microscopy-at-254nm-build-and-images/

 

I'll go and update that now.

[ulf]

From the second image it is clear that the LED "chip" has two chips one for UV-A and one in the center for UV-C.

Such chips has different forward voltages and must be driven separately. 

 

One can also see that the chip cavity is covered with a quartz window, so there is no risk of short circuiting of that area only are covered, but there will be leakage around the coverage.

 

The driver IC is also mounted by the LED-component on the flower shaped heat spreader metal-PCB.

 

If is difficult to tell if it would be possible to disable only the UV-A LED by cutting a track on the metal-PCB 

[Lou Jost]

2 hours ago, ulf said:

From the second image it is clear that the LED "chip" has two chips one for UV-A and one in the center for UV-C.

Such chips has different forward voltages and must be driven separately. 

 

One can also see that the chip cavity is covered with a quartz window, so there is no risk of short circuiting of that area only are covered, but there will be leakage around the coverage.

 

The driver IC is also mounted by the LED-component on the flower shaped heat spreader metal-PCB.

 

If is difficult to tell if it would be possible to disable only the UV-A LED by cutting a track on the metal-PCB 

Yes, that's all correct. There is a quartz window over it.

[Lou Jost]

6 hours ago, JMC said:

I've done some work at 254nm as well. There's a summary here which I need to update - https://www.ultravioletphotography.com/content/index.php?/topic/5456-uvc-safety-warning-microscopy-at-254nm-build-and-images/

 

I'll go and update that now.

Thanks for the update! I had read the original version some time ago but didn't remember it in detail...Your posts are all extremely useful path-breaking work!

[Lou Jost]

I'm re-reading that post now, Jonathan. Based on your post a few months ago I bought the E17-socket bulbs you mentioned, with the drivers mentioned by Colin. But they look so crudely made that I suspect they would have a lifetime measured in minutes, judging from the amount of filament sputter on the lamp after one use. How long do these really last?

[colinbm]

Mine have a few uses, but I have a few too.
The sputter is Mercury.
PS, I put reflectors on mine from cheap garden LED lights 

[Lou Jost]

42 minutes ago, colinbm said:

Mine have a few uses, but I have a few too.
The sputter is Mercury.
PS, I put reflectors on mine from cheap garden LED lights 

[JMC]

Lou, I've had about 12 hours out of mine now with no signs of it failing yet. I'm just using a bench powersupply to power it.