Rate Grown Transistor 1954 Dip Seed Tip Meniscus

# Dip the Seed: Locking In the Interface Geometry Before Pulling Begins

The seed crystal, now with its tip cleanly etched by the controlled melt-back of step eight, is lowered deeper into the stable melt until the newly-exposed, pristine lattice face touches the liquid-solid interface and the first atoms begin to crystallize from the melt. In the 1951 germanium series, the seed dip is primarily a mechanical positioning step — the seed moves into contact with the melt, a thin solid-liquid interface forms, and the seed and melt reach thermal equilibrium at the dip point before pulling starts. In the rate-grown series, the dip step does all of that, but it also *locks in* the interface geometry that the melt-back in step eight created. The boundary-layer thickness $\delta$ at the newly-formed growth interface directly reflects the seed tip's geometry from step eight: if that tip is symmetric, $\delta$ is uniform across the interface; if the melt-back was uneven, $\delta$ varies azimuthally, and that variation is now frozen in as the initial condition that the pull-rate schedule will try to modulate away in the steps to follow.

## 1. Dip Depth and the Meniscus: Forming the First Solid-Liquid Interface

$$d_{\text{dip}} = d_{\text{seed position}} \quad \text{and} \quad R_{\text{interface}} = R_{\text{seed}} - \delta$$

The seed is lowered into the melt at a controlled speed until it reaches the dip depth — the depth at which the pristine, melt-back-exposed tip of the seed just touches the melt surface. At that moment, atoms from the liquid melt begin to crystallize onto the solid seed face, forming a meniscus — the curved solid-liquid interface that will advance upward as the seed is pulled. The radius of curvature of this meniscus depends on the interfacial energy and the thermal gradient, but its position at any moment reflects the seed's geometry. A seed with an asymmetric tip geometry (from uneven melt-back) will produce an asymmetric meniscus, with different curvatures at different azimuthal positions. The Burton-Prim-Slichter boundary-layer model shows that a non-uniform meniscus curvature implies a non-uniform boundary-layer thickness $\delta(\theta)$ as a function of azimuthal angle $\theta$ around the seed's circumference.

## 2. Real Diagram: Meniscus Formation and Boundary-Layer Lock-In

Seed Dip: From Immersion to Growth Interface the meniscus locks in the boundary-layer geometry that step 8 created symmetric seed tip (uniform δ) melt surface symmetric meniscus, uniform δ pulling can begin; k_eff modulation is spatially uniform asymmetric seed tip (non-uniform δ) melt surface asymmetric meniscus, non-uniform δ(θ) pulling begins with spatial dopant heterogeneity frozen in the seed's geometry from step 8 persists as a boundary-layer signature that pulling cannot erase

## 3. The Meniscus-Boundary Layer Coupling: A Geometric Lock

$$k_{\text{eff}}(v; \delta(\theta)) \text{ is now path-dependent on seed geometry}$$

Once the seed touches the melt and the meniscus forms, the boundary-layer thickness $\delta(\theta)$ at each azimuthal position on the interface is determined by the seed's tip geometry from step eight. The Burton-Prim-Slichter model links boundary-layer thickness directly to the meniscus curvature: a sharper (more curved) meniscus is associated with a thinner boundary layer, while a flatter meniscus has a thicker boundary layer. An asymmetric seed tip, created by uneven melt-back, produces an asymmetric meniscus with a radius of curvature that varies with angle $\theta$. This means the effective segregation coefficient $k_{\text{eff}}$ — which depends exponentially on the local boundary-layer thickness — is now spatially heterogeneous across the interface *before pulling even begins*. The pull-rate schedule, designed to modulate $v$ in time to control dopant incorporation, has no way to know about or correct for this spatial structure that is already locked in at the meniscus. When pulling starts in step ten, the schedule's velocity modulation will be superimposed on top of this pre-existing spatial heterogeneity, meaning the dopant profile in the finished crystal will carry a permanent signature of the seed's step-eight geometry, regardless of how carefully the pull-rate sequence is executed.

## 4. Thermal Equilibration at the Dip Point: Confirming the Interface is Stable

$$T_{\text{interface}} = T_{\text{liquidus}}(N_D, N_A) \quad \text{and} \quad \frac{\partial T}{\partial z}\big|_{\text{interface}} = \text{const}$$

Once the seed tip has touched the melt, the system must equilibrate thermally. The interface temperature settles at the liquidus temperature — the temperature at which the melt and solid are in equilibrium — which depends on both the donor and acceptor concentrations in the melt. The thermal gradient along the seed's axis must be stable and well-defined. In the rate-grown series, this equilibration phase is also an opportunity to confirm that the meniscus is stable and symmetric (or at least symmetric enough that it won't oscillate as pulling begins). If thermal oscillations remain, the meniscus will wobble, the boundary-layer thickness will fluctuate, and $k_{\text{eff}}$ will oscillate in time, frozen as striations into the crystal. The dip hold — the time during which the seed remains in contact with the melt without pulling — continues until the furnace temperature and the seed position have both settled into stable, non-oscillating states. Only then does the pull-rate schedule begin.

Thermal Equilibration During Dip Hold T_interface must settle before pulling begins; oscillations → meniscus wobble → k_eff striations time during dip hold → interface temperature oscillation amplitude dip hold begins: T still oscillating from initial contact pull-rate schedule can only begin here T stable; meniscus locked in; ready to pull any remaining thermal noise will modulate k_eff during the pull, creating permanent striations

## Dip the Seed's Place in the Process Lineage

Dipping the seed is step nine of this project's rate-grown series — after the seed tip has been cleaned by melt-back and the melt has been thoroughly stabilized, and before the actual pulling begins. It is the step that forms the first solid-liquid interface and locks in the boundary-layer geometry that step eight created. An asymmetric seed tip from uneven melt-back produces an asymmetric meniscus, which translates into spatially-variable boundary-layer thickness and spatially-variable $k_{\text{eff}}$ that the pull-rate schedule cannot later correct. Step ten, initiating the pull, is where the seed begins to move upward at the beginning velocity of the pull-rate schedule, setting the growth rate that will drive segregation.

Take rate grown transistor 1954 dip seed tip meniscus 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