Micrologic 1960 Isolate Reverse Biased Junction Islands
# Isolate Devices Sharing a Substrate by a Reverse-Biased Junction, Not by Distance: Separation Without an Etch
## 1. Why Cutting a Mesa Around Every Device Was Never Going to Scale to a Circuit
This step surrounds each transistor's collector region with a diffused isolation ring of opposite polarity, held reverse-biased in every operating condition the circuit will see, so that neighboring devices sharing the same piece of silicon are kept electrically separate by a depletion region rather than by a mesa trench cut between them. Mesa isolation, the technique this project documented across the 1956 and 1958 series, isolates by physically removing silicon between devices — a strategy that worked when a wafer carried a handful of identical, widely spaced transistors, but becomes impractical the moment a single chip needs several differently sized, closely packed components wired into one circuit, because a mesa trench around every device would eat into area a layout this dense cannot spare. A reverse-biased junction isolates without removing anything, provided the isolation stays wide enough that neighboring depletion regions never meet:
where $W_{\text{dep}}$ is the depletion width at reverse bias $V_R$, $V_{bi}$ the junction's built-in potential, $N_a$ and $N_d$ the doping concentrations on each side, and $s_{\min}$ the minimum spacing this layout must leave between two isolated islands so their depletion regions never overlap under the highest reverse bias the circuit will apply. Two islands placed closer than $s_{\min}$ would have their isolation junctions merge electrically under worst-case bias, connecting two devices this layout intended to keep apart.
## 2. Real Diagram: A Cross-Section With No Trench At All
Where a mesa cross-section from 1956 or 1958 shows a visible step down through the silicon between devices, this cross-section shows a continuous, flat substrate the whole way across — the only thing separating the two transistor regions is a change in doping polarity, invisible to the eye and effective only because it stays reverse-biased.
## 3. Why Mesa Isolation Was Correct for 1956 and 1958 and Wrong for This Series
The 1956 drift-transistor series and the 1958 mesa production line both isolated devices by cutting silicon away, and that strategy was entirely appropriate for their purpose: each wafer carried many copies of the *same* device, spaced however far apart a dicing operation needed, with no requirement that any two devices ever cooperate electrically. This series' layout, by contrast, needs several *different* components placed close enough together that a short run of patterned metal can connect them into one function — the isolation strategy has to coexist with proximity, not prevent it. Junction isolation is the first technique in this project's history that achieves separation while permitting the kind of tight, cooperative placement a real circuit, rather than a field of identical devices, actually requires.
Step 3 does not protect a junction the way 1959's series did, and it does not cut a mesa the way 1956 and 1958 did; it keeps two devices electrically apart using nothing but a bias condition this circuit is already committed to maintaining for entirely different reasons.