Thermal Printing

Every other printing process brings the colourant with it. Thermal printing does not: the colourant is already in the paper, sitting in two layers that are not allowed to touch. One is a leuco dye, a colourless ring-closed molecule; the other is a weak acid developer. Warm the coating past about 90 degrees Celsius, the matrix holding them apart melts, the two meet, the dye ring opens, and that spot turns black. Cool it and it stays black. So the print head has nothing to deliver. It is a row of small resistors on an alumina substrate under a hard wear glaze, and printing is just switching them. Texas Instruments shipped that as the Silent 700 in 1969: a terminal with no ribbon, no ink, no hammers and almost no noise, at a time when the alternative on a desk was a teleprinter you could hear through a wall. The interesting engineering is not the dye, it is the pulse. Heat spreads sideways through the substrate as fast as it goes forward into the paper, so if you drive an element for too long the warm patch grows wider than the dot spacing and the dots merge. Working that out sets the pulse length at a fraction of a millisecond, which in turn sets the element resistance, the drive voltage, and how many elements you can afford to fire at the same instant. It also explains the two things everyone notices about thermal prints. They fade, because anything that made the dye change once - heat, sunlight, a solvent, the plasticiser in a wallet - can go on changing it. And a long receipt goes darker towards the end, unless the firmware shortens its pulses as the head warms up.