Lilienfeld 1926 Measure Baseline Device Characteristics
# Measure Baseline Device Characteristics: The First Test Must Trace Failures Back to Their Step
Steps 1 through 6 built and sealed the device. Step 7 is the first time current actually flows through it in a controlled measurement, and the goal is not simply to confirm the device "works" — it is to measure the terminal-to-terminal current-voltage behavior at zero control bias and, from the shape of any deviation from ideal, trace a possible defect back to the specific construction step that would explain it. A baseline measurement that only reports pass/fail discards the diagnostic information every prior step built in.
An offset at zero voltage points to Step 5, not Step 4 or Step 3. If the I-V curve does not pass through the true origin — meaning some nonzero current flows even at zero applied voltage, or some nonzero voltage is needed to achieve zero current — the most likely cause is a built-in asymmetry between the two terminal contacts themselves (different contact resistance or, worse, a slight galvanic potential from dissimilar contact materials), which is a Step 5 defect, not a film or electrode problem.
Asymmetric slope between positive and negative sweep directions points back to Steps 2 or 3. If the terminal-to-terminal resistance differs meaningfully depending on sweep polarity, the likely cause is an asymmetric field distribution from the buried electrode — which happens when the electrode-glass bond (Step 3) is uneven, or when the fracture geometry (Step 2) introduced an asymmetric capacitive coupling across the film's length, biasing the unmodulated baseline conduction path itself.
Curvature near the origin, rather than a clean straight line, points to Step 4. A non-uniform film — thicker in some regions, thinner in others, as flagged by Step 4's own uniformity check — produces a baseline I-V curve that is not purely ohmic, because different regions of the film reach their own local saturation or depletion behavior at different applied voltages even with no control bias involved. A perfectly uniform film, by contrast, produces a clean straight line through the origin at low bias.
Comparing against the Step 6.7 pre-seal baseline closes the loop on construction history. Step 6.7 already recorded terminal-to-terminal and terminal-to-electrode resistance immediately before sealing. Step 7's first action should be comparing the fully sealed, settled device against that recorded value — if they diverge beyond measurement noise, something changed during or after sealing (curing stress, solvent attack, mechanical disturbance), and that is diagnosed before any new electrical hypothesis about Steps 1-5 is even considered.
| Step | Process operation | Input | Output | Specification | Constraint |
|---|---|---|---|---|---|
| 7.1 | Connect four-wire probes to the two exposed terminal leads | Sealed device from Step 6 | Probed device | Separate current-force and voltage-sense leads per terminal, per the four-wire method established for the 1925 series | Two-wire measurement conflates probe/lead resistance with the device's own resistance |
| 7.2 | Compare zero-bias resistance to Step 6.7 recorded value | Probed device from 7.1 | Seal-integrity-confirmed device | Within 5% of Step 6.7 baseline | A larger deviation means sealing altered the device; diagnose Step 6 before proceeding further |
| 7.3 | Sweep terminal-to-terminal voltage from negative to positive, zero control bias | Seal-confirmed device from 7.2 | Raw I-V sweep data | Sweep range matched to expected device operating range; step size fine enough to resolve curvature | Too coarse a step size masks the very curvature that carries diagnostic information |
| 7.4 | Repeat sweep in reverse direction (positive to negative) | Device from 7.3 | Second I-V sweep | Same range and step size as 7.3 | A single-direction sweep cannot detect hysteresis or directional asymmetry |
| 7.5 | Overlay both sweep directions; check for offset at V=0 | Two sweeps from 7.3, 7.4 | Offset-checked data | Zero-crossing offset within instrument noise floor | A detected offset implicates Step 5 terminal asymmetry, per the fault tree |
| 7.6 | Check for slope difference between positive and negative sweep halves | Two sweeps from 7.3, 7.4 | Slope-symmetry-checked data | Slope difference within 5% between halves | A detected asymmetry implicates Step 2 or Step 3, per the fault tree |
| 7.7 | Check for curvature deviation from linear fit near V=0 | Two sweeps from 7.3, 7.4 | Linearity-checked data | Residual from best-fit line within instrument noise near the origin | A detected curvature implicates Step 4 non-uniform film, per the fault tree |
| 7.8 | Record final baseline conductance g(0) and all deviation findings | Checked data from 7.5, 7.6, 7.7 | Complete baseline characterization record | g(0) reported with uncertainty; each deviation type explicitly flagged pass/fail | This record becomes the reference point for every later test, exactly as Step 19 established for the 1925 series |
Step 7 is where the 1926 construction protocol and the 1925 testing protocol meet. Everything from here forward — leakage measurement, bias-response testing, gain quantification — follows the same measurement discipline already established and proven across the fourteen 1925 steps. What Step 7 adds that a plain baseline test would not is the explicit link back to construction: every possible deviation this device could show has already been assigned, in advance, to the step most likely responsible for it. That assignment is what turns a simple pass/fail baseline check into a diagnostic one.