The MicroLED Monitor thread

Exordium01

Ars Praefectus
3,977
Subscriptor
Then that would be an extra intermediate step. Because the pitch on a wafer is dramatically smaller than it will be on a display.

So you would then need to transfer to yet another medium with a different pitch that matches your display, or are you claiming you then use a wafer sized printer, to precisely place a mass of pixels sparsely selected individually out of the dense pack, and keep reposition, and sparsely selecting them from the remainder? That's insane.

The reality is that there isn't yet any one way to do this, and there are multiple options including even "fluidic" assembly being tried.

Apple spent 10 years and Billions on this, with watches as the the first "easy" target, and apparently recently cancelled all that massive resource spend because they determined it couldn't be done profitably.

So again, there are no products that do true MicroLED with any of these advanced transfer techniques. Those are still research stage.

All we have is essentially macroScale pick and place LED Walls.
So you admit that what you are describing isn’t microLED. Those LED walls aren’t pitched or sold as microLED and they aren’t even using individual microLED pixels because they are significantly higher power devices with higher luminosities.

You get to the pixel pitch you want by dilating the transfer film in x and y after substrate removal.
 
So you admit that what you are describing isn’t microLED. Those LED walls aren’t pitched or sold as microLED and they aren’t even using individual microLED pixels because they are significantly bowed power devices with higher luminosities.

You get to the pixel pitch you want by dilating the transfer film in x and y after substrate removal.

No you aren't reading. As I said there is no real MicroLED tech. What they call MicroLED is full normal pick and place.

For real MicroLED, you still need something analogous to pick and place. Just at massive scale and precision.

Apple aquired LuxView. Their tech was using electrostatic charges to pick up the microLEDs for placement.

This technique discusses encapsulating each microLED in a polymer ball to sort and place them.


All the transfer techniques are pick and place analogues, and none are successful enough for commercialization.
 
This is the issue right here. Pick and place is a transfer technique, but all transfer techniques aren’t analogous to pick and place. You have the relationship backwards.

Really you are going to say a technique, that picks up microLEDs with static electricity, and places them where where they need to be, isn't analogous with standard Pick and Place machines?

:rolleyes:
 

cogwheel

Ars Tribunus Angusticlavius
6,691
Subscriptor
This is the issue right here. Pick and place is a transfer technique, but all transfer techniques aren’t analogous to pick and place. You have the relationship backwards.
All transfer techniques related to MicroLEDs are analogous to pick and place. MicroLEDs do not use patterned bulk chemical deposition to create emitters directly on the substrate. They all use prefabricated components (the LEDs) that are placed on a substrate in a specific location and orientation by some type of gripping device.

This is the very definition of MicroLED, so I'm not sure what you're trying to say.
 
  • Like
Reactions: ScifiGeek

Exordium01

Ars Praefectus
3,977
Subscriptor
All transfer techniques related to MicroLEDs are analogous to pick and place. MicroLEDs do not use patterned bulk chemical deposition to create emitters directly on the substrate. They all use prefabricated components (the LEDs) that are placed on a substrate in a specific location and orientation by some type of gripping device.

This is the very definition of MicroLED, so I'm not sure what you're trying to say.
The processes are more analogous to film transfer processes than individualized pick and place. Pick and place is a specific process and has a very specific meaning and these processes aren’t it.

Are either of you in the semiconductor industry?
 
  • Like
Reactions: continuum

cogwheel

Ars Tribunus Angusticlavius
6,691
Subscriptor
The processes are more analogous to film transfer processes than individualized pick and place.
So what positions and orients the LED components on the film prior to transfer to the final substrate?

Pick and place is a specific process and has a very specific meaning and these processes aren’t it.
As far as I can tell, the difference is one of scale and type of gripper. MicroLED obviously won't/doesn't use actual pick and place machines since they're too coarse, but I haven't read anything that suggests that a similar machine that operates on a smaller scale isn't what is used.

If you have anything to link that backs up the idea that there is no precision orientation and placement step somewhere in the process, please do so.
 

Exordium01

Ars Praefectus
3,977
Subscriptor
So what positions and orients the LED components on the film prior to transfer to the final substrate?


As far as I can tell, the difference is one of scale and type of gripper. MicroLED obviously won't/doesn't use actual pick and place machines since they're too coarse, but I haven't read anything that suggests that a similar machine that operates on a smaller scale isn't what is used.

If you have anything to link that backs up the idea that there is no precision orientation and placement step somewhere in the process, please do so.
I did some microCPV stuff in a previous life. I did not do transfer printing but have talked to people who have. One common method that I have described here is to take the wafer of devices you want transfer printed and apply a transfer tape. After that, you either singulate devices or remove the substrate. Then you stretch the film until you get the desired device pitch and the. You bond the whole thing in one go. It looks, feels, and tastes like a film transfer process and nothing like pick and place. Pick and place is not commercially viable (though to be fair, neither was CPV).

There is absolutely tight alignment tolerance and orientation precision, but that’s not a new concept in the semiconductor industry.
 
  • Like
Reactions: continuum
Then you stretch the film until you get the desired device pitch and the.

That seems ridiculously prone to issues, since your required stretch might by 10x original spacing or more, and you need essentially perfect placement and alignment, on a nano scale.

And as I indicated, film transfer is NOT the only method being researched. There is static electricity pick and place, coating them in polymer balls and sieve printing, fluidic assembly techniques, etc...
 
Last edited:

Exordium01

Ars Praefectus
3,977
Subscriptor
That seems ridiculously prone to issues, since your required stretch might by 10x original spacing or more, and you need essentially perfect placement and alignment, on a nano scale.

And as I indicated, film transfer is NOT the only method being researched. There is static electricity pick and place, coating them in polymer balls and sieve printing, fluidic assembly techniques, etc...
Only one of those you mentioned is pick and place and it was developed by a nothing company Apple bought for pocket change.
 
Only one of those you mentioned is pick and place and it was developed by a nothing company Apple bought for pocket change.

Here a summary page, there is also magnetic transfer, and while Apple didn't pay much for LuxVue. They reportedly spent 3 Billion dollars and 10 years on research after.

Bottom line for all methods, is you need some way to take fabbed pixels from them their individual color wafers and place them with extreme precision on display substrate that will almost always have drastically different pitch, so they can be directly transferred, thus a multitude of options being researched:

MicroLED.png
 

cogwheel

Ars Tribunus Angusticlavius
6,691
Subscriptor
One common method that I have described here is to take the wafer of devices you want transfer printed and apply a transfer tape. After that, you either singulate devices or remove the substrate. Then you stretch the film until you get the desired device pitch and the. You bond the whole thing in one go. It looks, feels, and tastes like a film transfer process and nothing like pick and place.
OK, now I understand what you meant by "dilating the film in x and y after substrate removal".

I'd probably call it a "film transfer process" without using the word "print", since "print" is usually used (at least when we're not talking specific subindustries where the term has its own private definition) when you want to transfer bulk materials instead of discrete oriented components.

That seems ridiculously prone to issues, since your required stretch might by 10x original spacing or more, and you need essentially perfect placement and alignment, on a nano scale.
No, not really. MicroLED is very much at the micro scale, not the nano scale. Phone screens have pixel pitches in the 50-60µm range, so the LEDs subpixels for an equivalent microLED screen would still be over 10µm apart. A 1µm positional variance wouldn't be visible to the naked eye, so you could have oversized bonding pads to handle less than perfect precision.
 
  • Like
Reactions: continuum

Exordium01

Ars Praefectus
3,977
Subscriptor
OK, now I understand what you meant by "dilating the film in x and y after substrate removal".

I'd probably call it a "film transfer process" without using the word "print", since "print" is usually used (at least when we're not talking specific subindustries where the term has its own private definition) when you want to transfer bulk materials instead of discrete oriented components.

The industry calls those processes transfer printing. There almost certainly will never be a microLED product with discrete oriented components. It’s just not commercially viable for the prices they will be able to charge.
 
No, not really. MicroLED is very much at the micro scale, not the nano scale. Phone screens have pixel pitches in the 50-60µm range, so the LEDs subpixels for an equivalent microLED screen would still be over 10µm apart. A 1µm positional variance wouldn't be visible to the naked eye, so you could have oversized bonding pads to handle less than perfect precision.

You still need ridiculous precision in a process based on stretching film in two dimensions. You need to film transfer at the original pitch the pixel element from three different color wafers, then stretch each of those films perfectly and evenly and perfectly match each other so that each film you lay down can perfectly matched subpixels right near each other with big even space between each group.

I'm very skeptical of this perfect stretch a material in two dimensions like that. It seems like the pixels themselves would be prone to detachment while stretching the film.
 

cogwheel

Ars Tribunus Angusticlavius
6,691
Subscriptor
You still need ridiculous precision in a process based on stretching film in two dimensions. You need to film transfer at the original pitch the pixel element from three different color wafers, then stretch each of those films perfectly and evenly and perfectly match each other
Sure, but we're talking modern materials science for the film, probably aided by applied softening/stretching agents during the stretching process, done on precision stretching machines, not drawing on a rubber band with a sharpie and then stretching it and pressing it against your finger. It's plausible to be economical for this use, unlike pick and place. It'll still cost more than OLED, bit it isn't obvious it'll cost so much more that it isn't viable for any application.

so that each film you lay down can perfectly matched subpixels right near each other with big even space between each group.
Uh, what? Subpixels have never been grouped like that, from actual ink on paper processes, to CRTs, to LCDs and plasma and OLEDs. Why would microLED have strongly grouped, very uneven spacing instead of the even to only slightly uneven spacing we've had for ages?
 
Sure, but we're talking modern materials science for the film, probably aided by applied softening/stretching agents during the stretching process, done on precision stretching machines, not drawing on a rubber band with a sharpie and then stretching it and pressing it against your finger. It's plausible to be economical for this use, unlike pick and place. It'll still cost more than OLED, bit it isn't obvious it'll cost so much more that it isn't viable for any application.

So far I'm not buying this stretch the substrate. All the MicroLED print transfer, are done with either selective pickup, or selective application, not stretching the substrate.

Uh, what? Subpixels have never been grouped like that, from actual ink on paper processes, to CRTs, to LCDs and plasma and OLEDs. Why would microLED have strongly grouped, very uneven spacing instead of the even to only slightly uneven spacing we've had for ages?

Fair point. I guess it's because I've seen things like this, that show smaller emitters on the same pitch. The problem is, we don't actually have any practical MicroLED displays:

microled-v-standard-diagram.jpg
 

Exordium01

Ars Praefectus
3,977
Subscriptor
All the MicroLED print transfer, are done with either selective pickup, or selective application.
This statement is false. Tons of work is going into mass transfer because the technology requires it in order to be commercially viable.

I’ve been telling people here that microLED tech is not “just around the corner” for the past five years. It comes up all the time in the Mac Ach. One of the reasons why it isn’t close is because of the bulk/mass transfer problem. I’m not saying it’s easy, I’m just saying it is necessary. Otherwise we’re going to be sticking with OLED which just keeps getting better.
 
Last edited:
This statement is false. Tons of work is going into mass transfer because the technology requires it in order to be commercially viable.

It's still a form of mass transfer. Many are transferred in parallel. Just that they are selected in a sparse pattern because the spacing/pitch needs to expand.