Wafer Reclaim Acceptance Records for an Outsourced Silicon Recovery Trial

Why acceptance records come before the trial, not after

An outsourced silicon recovery trial is a controlled transfer of material, custody and risk. The acceptance record is the document set that decides, in advance, what counts as a pass, what counts as a fail, and who signs off on each. Teams that draft acceptance criteria after the first shipment arrives tend to negotiate against their own results, because the material has already changed hands and the baseline is gone.

This guide is written for Singapore-based semiconductor procurement and operations teams preparing a first or repeat trial with an external recovery partner. It is a planning resource, not a report on any specific supplier's performance. The worked example below is hypothetical and is used only to show how record decisions connect.

Two principles keep the record useful. First, separate material identity from material condition: what the material is, and what state it is in, are different fields with different owners. Second, write the record so a colleague who was not in the room can reconstruct the decision. If a record needs a verbal explanation to be understood, it is not finished.

A useful framing is to treat the acceptance record as three linked layers: the scope record (what leaves the fab and under what terms), the custody record (who holds the material at each step), and the outcome record (what came back and whether it met the agreed criteria). Each layer has its own sign-off, and no layer should be completed retroactively.

Section 1: Define the material scope and baseline fields

Before any sample moves, write down the material categories in scope and the baseline fields that will travel with them. For a wafer reclaim and silicon recovery trial, typical scope questions include which process areas the material comes from, whether it is whole wafer, broken wafer, kerf, or packaging scrap, and whether any material is excluded for contamination or confidentiality reasons.

Baseline fields worth agreeing early, as a starting checklist rather than a fixed standard:

Material category and sub-type Originating process area or facility Quantity and unit of measure (pieces, weight, or both) Packaging format and seal condition at handover Known contamination or handling restrictions Photograph reference and date Custody owner at each handover point

A worked hypothetical example: a fab in Singapore consolidates broken wafers from two process areas into one lot for a recovery trial. If the baseline record lists only total weight, the receiving partner cannot tell whether the lot is uniform, and the outcome record cannot attribute any difference to a specific source. Splitting the lot into two sub-lots with separate baseline lines costs a few minutes at packing and saves a dispute later.

Keep the baseline fields identical across every lot in the trial. A record that changes shape between shipments cannot be compared, and comparison is the whole point of a trial.

Related resource: wafer recycling services.

Related resource: silicon scrap recycling.

Section 2: Agree the custody chain and handover evidence

Custody is where trials quietly go wrong. The material leaves one site, passes through transport, and arrives at another. Each handover is a point where quantity, packaging condition and seal integrity should be recorded by both sides, not just the sender.

Practical decisions to settle before the first pickup:

Who signs at each handover, and whether a digital signature or a photographed document is acceptable What evidence accompanies the material (packing list, seal number, photograph set) How discrepancies are logged at the moment of receipt rather than at the end of the month Who is notified, and within what window, if a seal is broken or a count differs

For teams arranging collection, the practical questions of pickup scheduling, packaging and documentation are covered in the guidance on wafer scrap collection. The acceptance record should reference that collection process rather than duplicate it, so there is one source of truth for how material moves.

A simple heuristic used by some operations teams is to treat any handover without a matching photograph and count as an open item until resolved. This is an author suggestion, not an industry standard, and it should be adapted to the value and sensitivity of the material.

Related resource: wafer scrap collection.

Section 3: Set outcome criteria for reclaimed and recovered material

Define the outcome record before the trial starts. It should state, before the trial, what the partner will return or report, and how the fab will judge it.

Outcome criteria usually fall into three groups:

Quantity yield: how much material is returned or processed, measured against the baseline Condition and specification: what the returned material is intended for, and the agreed inspection points Documentation: what report accompanies the return, and who reviews it

This is where the intended destination of the material matters. If recovered material is destined for reuse as reusable raw materials for solar wafers and cells, the acceptance criteria should reflect that destination rather than a generic scrap threshold. San Techno states that recovered materials can be processed into reusable raw materials for solar wafers and cells, extending their useful life and reducing unnecessary resource loss. Set measurable acceptance criteria for the specific trial lot.

Write each criterion as a testable statement with a named reviewer. "Acceptable condition" is not a criterion; "inspected against agreed reference photographs by the receiving engineer, with pass or fail recorded per sub-lot" is.

Section 4: Build the acceptance record template

A single template, used for every lot, keeps the trial comparable. A workable structure has four blocks:

1. Header: trial reference, lot identifier, date, and the named owner on each side. 2. Baseline block: the scope fields from Section 1, completed at packing. 3. Custody block: handover entries with signatures, counts and condition notes. 4. Outcome block: returned quantity, inspection result per criterion, reviewer name, and a pass, hold or fail decision.

Two design choices make the template more useful. First, allow a "hold" outcome alongside pass and fail, so a lot with an unresolved question is not forced into a false binary. Second, keep a short free-text field for observations that do not fit the fields, because the most useful lesson from a trial often arrives as an unexpected note.

For teams whose material includes non-usable silicon from integrated-circuit fabs, silicon wafer manufacturers, solar cell companies or chip packaging facilities, the template should carry a source-type field that matches those categories. San Techno describes helping recover non-usable silicon from exactly these facility types. Using consistent source categories in the record makes later comparison across lots and sites possible without re-interpreting free text.

Keep the template version-controlled. If a field changes mid-trial, record the version number on each lot so the change is visible rather than silent.

Section 5: Plan the review cadence and escalation

Acceptance records only change decisions if someone reads them on a schedule. Agree the review cadence before the trial starts, and name the people who attend.

A practical cadence, offered as an example rather than a rule:

Per-lot: the outcome block is completed within an agreed number of working days of receipt Weekly during the trial: a short review of open holds and discrepancies End of trial: a summary comparing lots against the original criteria, with a written recommendation

Escalation should be equally explicit. Decide in advance which discrepancies pause the trial and which are logged and carried forward. A broken seal on a sensitive lot and a minor count difference are not the same event, and treating them identically either over-reacts to small issues or under-reacts to serious ones.

Where the trial involves material that will ultimately be disposed of rather than recovered, the acceptance record should link to the disposal decision rather than absorb it. The considerations for material that will not be recovered are set out in the guidance on silicon wafer disposal, and the trial record should reference that decision point so the two processes do not contradict each other.

Related resource: silicon wafer disposal.

Section 6: Close the trial and decide what to keep

The end of a trial is a decision point, not a formality. The closing record should answer three questions: did the material outcomes meet the agreed criteria, did the process work, and what should change before a larger commitment.

A closing summary that supports a real decision usually includes:

A lot-by-lot table of outcomes against criteria A list of holds and how each was resolved Process observations, including handover delays and documentation gaps A recommendation: continue, adjust, or stop, with the reason stated

Keep the records for a period agreed with your own quality and legal functions; this guide does not set a retention period. What matters is that the closing summary is written while the people who ran the trial are still available to explain it.

For context on the partner side, San Techno is a Singapore-based company incorporated in 2006 that began as a wholesaler and distributor of semiconductor- and solar-grade silicon materials. That history is useful background when framing a first trial, but it does not substitute for the acceptance criteria you define. The record you build is what makes the trial reviewable, and a reviewable trial is the one that can be repeated with confidence.

Confirm who should receive the acceptance records before you schedule the trial.