rate grown transistor 1954 load crucible compensated melt dopants

# Load the Crucible: A Container That Won't Upset a Melt Carrying Two Dopants Simultaneously

Graphite, inert enough for germanium's 937°C melt, remains the right crucible material for this process as it was in 1951. But what crucible cleanliness now has to protect is not a single dopant concentration set by step 2's charge preparation — it's a carefully balanced donor and acceptor pair, both already dissolved in the liquid, that will interact with whatever residual contamination a dirty crucible introduces. A contaminated graphite vessel won't announce itself; the compensated melt will still look like molten germanium, and the contamination will only surface later as an unexplained, unscheduled shift in the junction properties that the later pull-rate schedule was designed to control.

## 1. Crucible Contamination Now Acts on a Two-Dopant System, Not a Single-Species Target

$$N_{\text{net}}(\text{crucible contam}) = N_D^{\text{(step 2)}} - N_A^{\text{(step 2)}} + \Delta N_{\text{metals}} - \Delta N_{\text{boron}}$$

In the 1951 germanium series, loading the crucible introduced a hidden risk: residual dopant or metallic impurity from a previous melt redissolves into the new charge, adding an uncontrolled term to the intended donor concentration. This step's contamination threat was fundamentally a one-dimensional problem — more donor where fewer was intended, or fewer donor where more was needed. Here, residual contamination can skew the *difference* $N_{\text{net}}$ in two independent ways: metallic contaminants (iron, copper, aluminum) act as recombination centers that shorten lifetime, while boron residue (or other p-type impurities) from an earlier melt sits next to the acceptor species already loaded in step 2. The compensated melt's junction-encoding mechanism depends on the pull-rate-driven ratio of effective segregation coefficients for donor and acceptor; crucible contamination that shifts either concentration independently is indistinguishable, at the parametric-test stage later, from a pull-rate-modulation defect.

## 2. Real Diagram: Why a Dirty Crucible Mimics a Later Pull-Rate Error

Crucible Contamination and Pull-Rate Errors: Hard to Distinguish After the Fact both shift the balance between donor and acceptor incorporation; both hide in later parametric tests clean crucible + intended pull schedule k_eff(v) ratio controlled; junction forms as designed N_D and N_A in product both at intended values dirty crucible + intended pull schedule melt composition shifts; ratio altered before pulling at parametric test: looks like a pull-rate error the first introduces a step-4 error masquerading as a step-10+ error; the second is an actual step-10+ error both show up the same way later, so step-4 crucible cleanliness has to be non-negotiable

## 3. Graphite Crucible Preparation and the Residue Problem

$$\sigma_{\text{electrical}} = \sigma_{\text{intended}} + \sigma_{\text{residue}} + \sigma_{\text{leach}}$$

The crucible is high-purity graphite, chosen for its chemical inertness toward molten germanium, identical to the 1951 series. But preparation for *this* process carries an additional discipline: because two dopant species are present simultaneously from the moment the crucible is filled, any residual dopant or other electrically active impurity from a previous melt has to be removed completely. A crucible that last hosted a heavily doped n-type charge, or one that received an acceptor-dopant pellet drop in a prior run, carries microscopic boron or gallium residue on its inner walls. When the new compensated melt dissolves that residue, the acceptor concentration $N_A$ rises above the carefully-calculated step-2 value, tilting the melt toward p-type even before a single joule of furnace energy is applied.

Crucible cleaning follows a rigorous protocol: oxidation in air to convert residual metallic dopants to oxides (which sublime away more completely than the dopant itself), followed by vacuum outgassing and a final dimensional-thickness verification to ensure the vessel has not been eroded by prior melts into a shape that would disrupt thermal contact or cause uneven heating.

Dopant Incorporation Rate (Effective Segregation) vs. Crucible Contamination k_eff(D) and k_eff(A) both shift independently when residual dopants alter N_D or N_A crucible residue level → effective segregation k_eff for each dopant k_eff(donor), expected trajectory k_eff(acceptor), expected trajectory clean crucible: both at intended point dirty crucible: both drift off-target unpredictably unpredictable drifts in both effective segregation coefficients are read as pull-rate errors downstream

## Load the Crucible's Place in the Process Lineage

Loading the crucible is step four of this project's rate-grown series — following the preparation of the seed crystal and the establishment of compensated melt composition. It is the step where crucible material selection (high-purity graphite, unchanged from 1951) and rigorous cleaning preserve the two-dopant balance established in step 2 against residual contamination that would otherwise masquerade as a later, step-10+ pull-rate error. Step five, establishing furnace atmosphere, begins the thermal profile that will melt the charge.

Go deeper with CFSGPT

Get AI-powered deep-dives, save terms, and run advanced simulations — free account.

Create Free Account