Point Contact Transistor Process Flow

# Point-Contact Transistor Device Process Flow: Fabrication, Surface Inversion & Electrical Forming

The point-contact transistor — invented at Bell Laboratories in December 1947 by John Bardeen and Walter Brattain under William Shockley — was the world's first working solid-state amplifier. Unlike modern planar MOSFETs or junction transistors that rely on photolithographically patterned masks and thermally diffused dopant wells, the point-contact transistor operated without a single fabricated metallurgical junction. It was manufactured by mechanically pressing two microscopic spring-loaded metal whiskers (the emitter and collector) onto the chemically prepared surface of an n-type germanium crystal, spaced merely 50 micrometers apart, and electrically "forming" the collector via a high-current micro-discharge.

Understanding the point-contact transistor device process flow bridges the fundamental gap between laboratory solid-state physics and the birth of semiconductor manufacturing. Every modern transistor feature — gate length, source/drain contact resistance, channel transport length, and interface state passivation — originated as a mechanical or chemical processing step in the 1947–1952 Bell Labs pilot line.

---

## 1. Germanium Substrate Metallurgy & Ingot Crystallization

The first phase of point-contact transistor manufacturing centered on obtaining single-crystal germanium with tightly controlled bulk electrical properties. In 1947, metallurgy of group-IV semiconductors was in its infancy:

1. Fractional Induction Zone Refining: Raw germanium dioxide ($GeO_2$) was reduced in a hydrogen furnace at $650^\circ\text{C}$ into germanium powder, then melted in graphite boats inside high-frequency induction furnaces. To eliminate donor and acceptor contaminants, multiple passes of zone melting swept impurities to the ingot tail.
2. Controlled n-Type Doping: Pure intrinsic germanium ($E_g \approx 0.66\,\text{eV}$ at $300\,\text{K}$) was doped with pentavalent donor impurities — primarily antimony (Sb) or arsenic (As) — targeting an electron concentration of $N_D \approx 10^{14}\text{--}10^{15}\,\text{cm}^{-3}$ and a resistivity of $\rho \approx 1\text{--}10\,\Omega\cdot\text{cm}$.
3. Wafer Dicing & Lapping: Ingots were sliced using diamond-impregnated saws into rectangular wafers and diced into individual pellets measuring approximately $1.5 \times 1.5 \times 0.5\,\text{mm}$. The dicing process introduced severe mechanical crystal lattice disruption (micro-cracks, dislocations, and work damage) across the wafer surface.

---

## 2. The CP-4 Chemical Polish & Bardeen Surface Inversion

Mechanical abrasives leave a surface layer densely populated with recombination centers that quench minority carrier lifetime. The device process flow relied on a chemical etch:

The proprietary CP-4 chemical etch formulation comprised:
- $15\,\text{mL}$ Glacial Acetic Acid ($CH_3COOH$)
- $25\,\text{mL}$ Concentrated Nitric Acid ($HNO_3$)
- $15\,\text{mL}$ Hydrofluoric Acid ($HF, 48\%$)
- Several drops of liquid Bromine ($Br_2$)

The Chemistry and Physics of Surface States:
- Nitric acid oxidizes the germanium surface to $GeO_2$, while hydrofluoric acid continuously dissolves the oxide into soluble fluogermanic acid ($H_2GeF_6$), yielding a mirror-smooth specular surface.
- Crucially, John Bardeen's historic surface states theory explained why early field-effect experiments had failed: dangling bonds and chemical adatoms at the exposed germanium surface created a dense band of electronic energy levels within the bandgap ($D_{it} > 10^{13}\,\text{cm}^{-2}\text{eV}^{-1}$).
- These surface states pinned the Fermi level near the valence band edge, creating a naturally occurring p-type inversion layer right at the top face of the n-type crystal without any deliberate acceptor doping!

---

## 3. Base Ohmic Soldering & Mounting

To establish a low-resistance base connection, the unpolished bottom surface of the germanium die was soldered to a threaded brass or copper mounting stud:
- A non-rectifying ohmic contact was formed using lead-tin-antimony solder ($Sn\text{-}Pb\text{-}Sb$) at approximately $250^\circ\text{C}$.
- The antimony dopant in the solder ensured strong n-type contact to the bulk germanium, eliminating parasitic Schottky barrier formation and keeping extrinsic base resistance $r_b < 50\,\Omega$.
- The threaded stud served simultaneously as the mechanical chassis, electrical base terminal ($V_b = 0\,\text{V}$), and thermal heat sink.

---

## 4. Dual Cat's Whisker Alignment & Spacing Geometry

The defining mechanical challenge of point-contact manufacturing was placing two microscopic metal electrodes within tens of micrometers of each other:

1. Cat's Whisker Material: Phosphor-bronze (an alloy of copper with $3.5\text{--}10\%$ tin and up to $0.4\%$ phosphorus) or beryllium-copper wire with a diameter of $5\,\text{mil}$ ($127\,\mu\text{m}$) was chosen for its high elastic spring constant and metallurgical doping properties.
2. Electrolytic Pointing: The wire ends were electrolytically etched or ground into sharp chisel points with tip radii under $5\,\mu\text{m}$.
3. The Polystyrene Wedge Slit: In Brattain's original assembly, gold foil was cemented over the apex of a triangular polystyrene wedge, and a microtome razor blade was used under a binocular microscope to cut a slit across the apex, separating the foil into two electrically isolated contacts spaced by:

4. Spring Preloading: Each whisker incorporated an S-shaped or C-shaped bend. When lowered onto the germanium die by micromanipulators, the spring flexure applied a controlled mechanical preload of $0.1\text{--}0.2\,\text{N}$, sufficient to penetrate residual native oxide while avoiding plastic deformation of the sharp tip.

---

## 5. Electrical Pulse Forming: The Secret of $\alpha > 1$

As mechanically placed, two cat's whiskers on germanium act merely as two back-to-back point-contact diodes. They provide negligible power gain and current amplification $\alpha < 0.5$. The critical manufacturing step that converted the assembly into an active transistor was electrical forming:

The Forming Discharge Recipe:
- The emitter was left open or forward biased, while a high-voltage capacitor ($C = 0.1\text{--}1.0\,\mu\text{F}$) charged to $50\text{--}100\,\text{V}$ was discharged through a current-limiting resistor ($100\text{--}500\,\Omega$) directly through the collector whisker in the reverse direction.
- The instantaneous discharge drove a peak current $I_{peak} > 1\,\text{A}$ through a microscopic contact area ($A < 10^{-7}\,\text{cm}^2$), creating a transient current density:

Micro-Alloy Physics and Carrier Multiplication:
1. Local Joule Melting: The thermal pulse momentarily heated a microscopic hemispherical zone ($r_{melt} \approx 2\text{--}5\,\mu\text{m}$) directly beneath the collector tip beyond germanium's melting point ($938.3^\circ\text{C}$).
2. Phosphorus Doping / Thermal Acceptor Conversion:
- Rapid quench cooling created a high density of crystal defects and thermal acceptors, producing a localized p-type region.
- When phosphor-bronze wire was used, donor phosphorus atoms from the bronze simultaneously alloyed into the melted cap, creating a localized $n^+$ layer.
3. The Shockley Hook Mechanism: The resulting $p\text{-}n\text{-}p\text{-}n$ micro-structure acted as a "hook collector": holes injected by the emitter accumulated at the collector space-charge barrier, lowering the potential barrier for electron emission from the metal. This produced an extraordinary current multiplication factor:

Current gain $\alpha > 1$ allowed point-contact transistors to operate as single-device bistable flip-flops, negative-resistance oscillators, and high-frequency amplifiers that early junction transistors could not match.

---

## 6. Silicone Potting & Type A Hermetic Cartridge Encapsulation

The exposed germanium surface and point contacts were exceptionally sensitive to ambient moisture, mechanical shock, and atmospheric contamination:

• Surface Passivation: The active die and whisker assembly was impregnated with silicone wax or high-viscosity polyethylene grease. This moisture-excluding matrix prevented adsorbed water dipoles from altering the surface-state potential and drifting the operating point.
• Cartridge Packaging: The assembly was hermetically sealed inside a cylindrical brass cartridge (the Western Electric Type A transistor, $0.2\,\text{in}$ diameter by $0.5\,\text{in}$ length). The cartridge featured two insulated eyelet pin leads for the emitter and collector, with the brass body serving as the common base terminal.

---

## Point-Contact vs. Grown Junction vs. Modern Planar Process Comparison

---

## Why the Point-Contact Process Flow Was Superseded

The point-contact transistor device process flow was an extraordinary triumph of applied physics, but its reliance on mechanical cat's whiskers and electrical pulsing proved its undoing:
1. Mechanical Shock Vulnerability: A single physical drop or thermal cycle could dislodge the whisker tips by a few micrometers, permanently destroying the $s < L_p$ geometry.
2. Forming Non-Uniformity: Pulse forming varied wildly depending on surface oxide thickness, contact force, and instantaneous spark discharge characteristics, leading to severe lot-to-lot yield loss.
3. High 1/f Flicker Noise: The unpassivated surface states and irregular micro-alloy interfaces generated enormous electronic noise.

By 1951, William Shockley's grown-junction transistor and subsequent alloy-junction and planar photolithographic processes replaced mechanical whisker placement with chemical diffusion and optical projection. Yet every fundamental concept — minority-carrier injection, diffusion length constraints, surface-state passivation, and ohmic base contact engineering — was pioneered in the six-step point-contact process flow.