Mesa Production 1958 Characterize Surface Leakage Current

# Characterize Surface Leakage at the Exposed Mesa Edge: Measuring the Problem Before Trying to Fix It

## 1. Why This Step Does Not Yet Try to Solve Anything

This step takes a sample of wafers straight from the mesa etch and measures collector-base leakage current under reverse bias as a function of time, humidity, and surface cleanliness, deliberately holding off on any attempt to fix the exposed-junction problem Step 5 identified until this line understands exactly how that leakage behaves. Fixing a problem no one has measured risks solving the wrong version of it, and surface leakage at an exposed junction is not a single number — it depends on bias voltage, ambient humidity, and time since exposure in ways that interact, so a fix validated under one set of conditions can fail completely under another. The standard way to separate surface leakage from the bulk leakage every junction has anyway is to compare the measured reverse current against what diffusion alone would predict:

$$I_{\text{total}} = I_{\text{bulk}} + I_{\text{surf}}, \qquad I_{\text{surf}} = I_{\text{total}} - I_0\left(e^{V/V_T}-1\right)$$

where $I_{\text{bulk}}$ is the leakage a clean, unexposed junction of the same area and doping would show, $I_0$ its saturation current, $V_T$ the thermal voltage, and the remainder, $I_{\text{surf}}$, is whatever this specific exposed sidewall is adding on top. Measuring $I_{\text{surf}}$ across many wafers, many humidity conditions, and many elapsed times is what turns "the mesa edge is a problem" into a quantified curve this line can actually design a defense against.

Separating What the Junction Does From What the Surface Adds measured leakage versus the bulk diode's predicted leakage REVERSE LEAKAGE VERSUS RELATIVE HUMIDITY relative humidity → Itotal Ibulk, nearly flat with humidity Isurf, grows sharply with humidity the gap between the two curves, at any humidity, is the surface's own contribution Isurf = Itotal − I₀(eV/VT−1) — a measured surplus, not a guess the fix this series tries next must be judged against this curve, not against intuition

## 2. Real Diagram: Humidity Is the Variable the Device Specification Never Mentions

A device specification states a maximum leakage current at a nominal condition. It says nothing about how that leakage changes as the ambient humidity around an unprotected junction edge rises — information this series needs precisely because the device will travel through handling, probing, and storage environments this line cannot fully control before it is finally sealed.

Three Humidity Conditions, Same Device leakage measured on the same wafer, held at three different ambient humidities DRY, 20% RH Isurf small close to the bulk value alone MODERATE, 50% RH Isurf rising a measurable multiple of bulk HUMID, 85% RH Isurf dominates bulk leakage is now the smaller term the same wafer can pass test at 20% humidity and fail it at 85%, with nothing about the device itself changed

## 3. Why This Project's Earlier Series Never Measured a Device Against Its Environment

The 1956 and 1957 series, and Step 1 through 5 of this series, each treated a device's parameters as properties of the device alone — junction depth, resistivity, critical dimension, all measurable without reference to what surrounded the wafer when the measurement was taken. Surface leakage at an exposed junction breaks that pattern: it is not a property of the device in isolation, but a property of the device *in an environment*, and the same physical device can read as reliable or unreliable purely as a function of the humidity it happens to be sitting in. That is a genuinely new kind of measurement for this project, and it is the direct, necessary consequence of Step 5's exposed sidewall — a measurement this line would never have needed to invent if the junction edge had simply stayed covered.

Step 6 does not fix anything; it produces the one curve every candidate fix this series considers next will have to be judged against, because a fix nobody has measured against this curve is a fix nobody can actually trust.

Take mesa production 1958 characterize surface leakage current further

Ask the copilot about this term, or have our engineers assess it against your process.

Ask CFSGPT about this → Request a technical assessment → Open the CFS app