MOSFET 1962 Mobile Ion Contamination Threshold Instability

# Confront Mobile-Ion Contamination: A Threshold Voltage That Can Drift After the Device Is Finished

## 1. Why a Device That Passes Every Earlier Test Can Still Fail Later

This step applies a sustained gate bias at elevated temperature to a finished device and re-measures its C-V curve with the identical technique Step 6 already established — not to check whether the device was built correctly, which Steps 1 through 6 already confirmed, but to check whether trace alkali-metal ions, inadvertently introduced somewhere during fabrication and trapped inside the oxide bulk itself rather than at its interface, are mobile enough under the device's own operating field to drift and change the threshold voltage after the device has already left the line. Step 1's own surface-state density, $D_{\text{it}}$, is fixed the moment the oxide finishes growing — a property of the interface established once and never changing again. Mobile ionic contamination is a fundamentally different kind of defect: charged ions sitting anywhere within the oxide's own thickness, free to drift toward either interface under a sustained electric field, with their migration accelerated sharply by heat — meaning a device that measured a perfectly stable threshold voltage the day it was tested can still show a different threshold voltage weeks later, under nothing more than its own normal operating bias held at an elevated ambient temperature.

$$\Delta V_T \approx -\frac{Q_m}{C_{\text{ox}}}$$

where $Q_m$ is the areal density of mobile ionic charge that has drifted to accumulate near the oxide-silicon interface under sustained bias and heat, and $C_{\text{ox}}$ is the same oxide capacitance Step 2 fixed and Step 5 and Step 6 both already depended on — the worst case, with all the mobile charge swept fully to the interface, produces the largest possible shift, and because $Q_m$ grows over time as more ions drift and redistribute, $\Delta V_T$ is not a fixed error to be characterized once, but a moving target this step exists specifically to bound.

The Same C-V Curve, Shifted Sideways by Stress identical device, before and after sustained bias at elevated temperature gate voltage, VG → Cmeas as fabricated, VT as measured in Step 6 after bias-temperature stress, VT has shifted left ΔVT ΔVT ≈ −Qm / Cox — the shift grows as more ions drift and accumulate the curve's shape is unchanged; only its position along the voltage axis has moved

## 2. Real Diagram: Ions Scattered Through the Oxide, Then Swept to One Side

The cross-section below shows the oxide before and after bias-temperature stress — mobile ionic contamination initially scattered through the oxide's own bulk, then driven by the sustained field to accumulate as a sheet of charge near one interface.

Scattered Ions, Then a Field-Driven Sheet of Charge before and after sustained bias at elevated temperature AS FABRICATED mobile ions scattered through the oxide silicon, threshold voltage as measured in Step 6 AFTER BIAS-TEMPERATURE STRESS ions swept into a sheet at one interface silicon, threshold voltage has shifted the device was never miswired or mis-measured — the oxide itself changed state under normal operating stress

## 3. A Defect Fixed at Growth, a Defect That Moves After Manufacture

Step 1's surface-state density and this step's mobile ionic contamination are easy to mistake for the same category of problem, since both are electrical defects associated with the oxide and both disrupt the gate's intended control over the channel — but the two are fundamentally different in character. $D_{\text{it}}$ is fixed the instant the thermal oxide finishes growing; a wafer with low $D_{\text{it}}$ has low $D_{\text{it}}$ for the rest of its life, a property this project measured once in Step 1 and never needed to re-measure. Mobile ionic contamination has no such permanence. The ions responsible are not a property of the interface at all — they are foreign contamination sitting somewhere inside the oxide's own bulk, free to drift under precisely the kind of sustained electric field and elevated temperature a device experiences during ordinary continuous operation, which means the threshold voltage this series spent five steps establishing and verifying is not necessarily the threshold voltage the same device will show after months in the field. No bipolar device this project has documented has ever needed a test for a property that changes after manufacture under nothing more than normal use; a diffused junction, once formed, does not redistribute itself under bias the way a contaminant trapped in an insulator can.

Step 7 does not look for a defect introduced by a mistake in this series' own process steps; it looks for contamination this process cannot fully exclude, and measures how much of this device's own promised behavior that contamination can quietly take back after the device has already shipped.

Take mosfet 1962 mobile ion contamination threshold instability 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