Efficient Blue and White Perovskite LEDs via Manganese Doping
Loosely inspired by XKCD’s Thing Explainer, in the following hopefully coherent narrative we’re going to try to break down our recent paper in a non-jargony way. Wish us luck! Paper link: https://doi.org/10.1016/j.joule.2018.08.005
First, let’s talk about color. This picture represents the colors you can see. For your phone/TV/FancyWatch/whatever to show them, you need a red, green, and blue LED (circles in the picture). By powering them at different levels you can make any color between the three circles. In that way, with three LEDs, you can make almost every color. So every screen has red, green, and blue emitters.
We decided to study perovskites as our light emitter. These materials are awesome – they’re bright, have sharp colors, and are inexpensive to make. Mixing a few materials together in the right ratios and you get the structure above at left – a lattice of lead atoms surrounded by halides, with cesium in the voids. For this work, we’re making nanocrystals – tiny crystals on the nanoscale surrounded by insulating ligands. At right, you can see what the nanocrystals actually look like at the nanoscale using an electron microscope. They’re roughly 20 nanometers wide – 500,000 of them would fit across your fingernail. We were hopeful that these tiny crystals would be powerful enough to act as the light emission engine of our LEDs.
A lot of smart people have done very cool research on these materials and have shown that red and green LEDs can be very efficient. But as we mentioned above, you need red, green, AND blue LEDs to access all the colors. But blue LEDs are way worse – over 100x worse! This is a huge problem, and we set out to discover why they were so bad and how we could fix them.
We immediately found the big issue: these blue materials aren’t very good at emitting light, which turns out to be a bit of a problem in a light emitting diode. We measure how good materials are at emitting light using a measurement called photoluminescence quantum yield or PLQY. 0% means the material emits no light whatsoever, and 100% means it emits light perfectly. We find that red and green perovskites have PLQYs of 70-90%. Our blue materials? 9%. Think about it like pouring a glass of beer. Red and green materials barely spill a drop, while blue materials are getting beer everywhere except the glass, making a giant mess and wasting most of the beer. We wanted to figure out why we were making a mess and how to fix it.
Then it got weird. Unrelated to LEDs, several groups had started adding manganese as a dopant to these nanocrystals to do some cool magnetic stuff. The Mn takes energy from the perovskite and emits it with an orange color. But all the groups noted that the amount of light coming out of the perovskite increased rather than decreased! Going back to the beer analogy, this would be like a friend asking you to split the little beer you had left in your glass, but somehow after sharing you had more beer in your glass than when you were drinking alone! Like I said, weird. We wondered if we could take advantage of this bonus brightness.
So we made some nanocrystals. The right two vials above have LOTS of manganese, and you can clearly see its orange emission. But if we look at the three left vials (which have 0, 0.1%, and 0.2% Mn) we see that as we added the dopant, we got much brighter materials with the nice pure blue we wanted. Success!
Of course, we can formalize things a little bit. We measured the PLQY and saw that it went up to 28%. So when our manganese friend asks to split the beer, they’re really making sure a lot less spills, and there’s plenty available for both glasses. (What’s actually happening? We suspect that the Mn helps repair defects in the crystals, but research in that area is ongoing, by us and others).
To measure our LEDs, we use a metric called external quantum efficiency or EQE. That’s the number of photons that we get from the device, divided by the number of electrons we put into it. Perfect materials would be 100% - one photon for every electron. In reality, in these structures a lot of the light is trapped, and so the best you can do is about 25%. That’s the efficiency of the OLEDs in cell phones, for example, and red and green pervoskites are starting to approach this value. At this blue wavelength, however, perovskites are much lower. In fact, our devices without Mn (black line in the picture) reached a maximum efficiency of 0.50%, which was the highest for perovskite nanocrystals at the time. When we add Mn, we see the value shoots up to 2.12% due to the improved performance from the perovskite nanocrystals. These devices are bright and efficient, showing that blue has the potential to be competitive with red and green!
Finally, we decided to make a white LED that could eventually be used for room lighting. To do that, we add downconverters in front of our LED. These perovskites absorb the blue light and re-emit it as red or green. When properly mixed, the three colors together look white. The middle pictures are our downconverters under room light and UV light, showing how good they are at downconversion. When the whole system is built, we see a blue, red, and green emissive peak, which is properly mixed to give white, as shown in the picture.
We hope this shows that blue perovskites are not the dumpster fire they are often assumed to be. They have the potential to be just as good as red or green, and when we pair all three together, have a lot of promise to make the next generation of emissive TVs and phones.
Again the paper can be found here: https://doi.org/10.1016/j.joule.2018.08.005. Thanks for tagging along on this journey! Any questions can be addressed to congreve@stanford.edu.