Correction: The orange display is running at 700Hz, 130V in the video.
I realize that I may have conflated the issues of one-resistor-per-pixel and the display's ability to maintain an image throughout row scanning. They are separate problems that are both addressed by designing the panel to work on AC. Each pixel can maintain its state (on or off) by being supplied constantly with a lower "sustaining" voltage, and can be set or cleared by giving it a momentary higher or lower amplitude. The sustaining voltage allows the pixel to be emitting light or not, and its state remains because of its own impedance until updated on the next scan. In color plasma displays, separate electrodes are used for sustaining and addressing pixels, and the discharge may be sustained between coplanar electrodes instead of plane-to-plane, as in this display.
It's also a possibility that the dielectric and MgO layer only exists on one electrode (the metal), and the ITO is bare. I don't know. On this display, if all rows are electrically connected together, and all columns are connected together, and AC is applied to rows and columns, this effect does not work -- no light is emitted at all! At least some of the electrodes (ie every other column) must be left floating to emit any light, and to show this memory effect. So, driving AC plasma panels requires more waveform tricks that I do not fully understand.
Prior art patents: https://patents.google.com/patent/US7283301B2/en https://patents.google.com/patent/US20060132716A1/en
Physics coffee mug in opening shot:https://www.atomstoastronauts.com/collections/mugs
Support Applied Science on Patreon: https://www.patreon.com/AppliedScience