methodology — accession, analysis, retention

identifier scheme

Identifiers take the form AM-YYYY-NNNNN-X:

The five-digit sequence has on three occasions in the registry's history exceeded its nominal capacity within a single calendar year. In each case, the convention has been to continue indexing from 100000 and to flag the affected records with a footnote in the corresponding annual addendum. The catalogue's parsers tolerate a six-digit sequence; older consumer software may not.

passes

Most accessioned specimens are taken through pass A and pass B. Pass C is performed on roughly one specimen in five and is reserved for material whose Raman response is ambiguous or absent. Pass D is rare; it requires arrangement with an external facility and is undertaken only where compositional ambiguity remains after pass C, or where the registrar judges the specimen to merit the cost.

intake and preparation

Specimens received at an intake address are weighed, measured, photographed against a scale bar, and provisionally indexed against the day's intake list. A small fragment, where sacrifice is acceptable to the submitter, is set aside for the analytical bench; the remainder is sealed in an inert polymer envelope, given a temporary reference, and shelved pending pass A.

Where the submitter has declined sacrifice, only non-destructive techniques are applied (pass A, pass B at reduced laser power, and — at the registrar's discretion — backscatter photography). The fact that a specimen has been accessioned non-destructively is not separately marked in the public record but is inferable from the absence of density figures or polished-section imagery.

Submissions arriving in unsuitable packaging — loose powder in unsealed envelopes, hygroscopic salts in cardboard, items of unannounced weight in excess of 2 kg — are accepted but not analysed until repackaged at the registry's expense, and the corresponding intake date is recorded as the date of repackaging.

raman acquisition protocol

Pass-B spectra are acquired in the open-laboratory configuration using a 532 nm diode-pumped solid-state source at nominal 80 mW (attenuated to 8–24 mW at the specimen depending on observed photodecomposition risk), with a 100× long-working-distance objective. Acquisition is 10 × 5 s per spectral window with two windows per record (100–1300 cm⁻¹, 1100–2400 cm⁻¹), recombined and baseline-corrected per the registry's published baseline algorithm (rolling-ball, ball radius 80 cm⁻¹). 785 nm is used in cases where the 532 nm spectrum is dominated by fluorescence; the choice is recorded in the public notes.

Calibration of the spectrometer's wavelength scale against a neon-discharge reference is performed at the start of each analytical day and at any time the temperature of the spectrometer enclosure varies by more than 2 °C from the last calibration. Calibration logs are retained for the lifetime of the spectrometer and are available on written request.

peak reporting

Peak positions in the public record are reported to the nearest cm⁻¹ after the baseline correction described above. A peak is reported only if its prominence exceeds three times the local noise standard deviation in a window of ±40 cm⁻¹. Shoulders, doublets observable only on second-derivative inspection, and features within 5 cm⁻¹ of the laser notch filter are not reported. The full corrected spectrum is retained internally and may be requested for any record by citing its accession identifier.

The published peak vector is therefore a deliberate simplification of the underlying measurement and is intended for cross-referencing rather than quantitative reanalysis. The registry does not warrant the suitability of the published peaks for any specific quantitative purpose.

cross-correlation

Records sharing two or more reported Raman peaks within ±4 cm⁻¹ are flagged internally as a candidate cross-reference. The candidate is reviewed by the registrar, with reference to the specimens' compositional and geographic metadata, and either entered into the cross-reference graph or dismissed with a brief note in the internal log. The current state of the cross-reference graph is reported on the operations monitor and changes asynchronously as new records are accessioned or revised.

A cross-reference is descriptive: it asserts that two records share a noteworthy spectral feature in common, not that they derive from the same locality, the same submitter, or the same species. Higher-order inferences from the cross-reference graph are the responsibility of the reader.

anomaly classification

Records exhibiting features inconsistent with the registry's provisional identification are flagged with an anomaly class between I and V. The full enumeration is given on the classifications page; in summary:

The class assignment of a record may change over its lifetime as new analytical passes are completed or as cross-references accumulate. Promotion from one class to the next is recorded as a record revision; demotion is rare and is documented in the corresponding annual addendum series.

submitter anonymity

Submitters are identified internally by a six-character truncation of a SHA-256 hash of their submission packet metadata. The full hash is retained but not published. A value of REDACTED in the submitter field indicates that the submitter has invoked their right to post-hoc anonymisation under the registry's standing data policy. Records bearing the synthetic submitter hash ai-intake are produced by the registry's automated intake pipeline and have no human submitter; their analytical content has been reviewed by the standing registrar before publication.

retention

Once accessioned, records are retained indefinitely. Withdrawal does not delete a record; it removes it from active circulation and flags it as historical. Records are revised rather than overwritten; revision history is available on request.

Physical specimens are retained at climate-controlled humidity (40–55 % RH, 16–22 °C) in inert polymer envelopes within labelled archival drawers. Specimens of known instability (deliquescent, photoreactive, or radioactive in trace) are held under additional environmental controls and are annotated as such in the internal record; the controls themselves are not described in the public metadata.

Disposal of an accessioned specimen — whether by destructive analysis, return to submitter, transfer to a competent authority, or, in rare cases, ordinary loss — is recorded as a terminal event in the record's revision history. The record itself remains; only the physical specimen ceases to be available for further pass-B or pass-C work.

indexing integrity

Each catalogue entry is associated with a content-addressed identifier (the proof-hash), computed as a SHA-256 digest over the canonical serialisation of the record's primary metadata (accession identifier, designation, formula, locality, submission date, submitter hash, and record revision) concatenated with the proof-hash of the immediately preceding entry under the canonical submission ordering. The first record in the registry is bound to a fixed registry-wide constant, designated the genesis-hash, derived from the registry's founding date.

This construction allows independent verification that no historical record has been silently altered: any modification to a single entry — including a single character of its serialisation — invalidates the proof-hash of that entry and, by transitive dependence, every proof-hash that follows. The complete sequence is exposed alongside each public record and is republished in full at each periodic export of the catalogue. Outside readers wishing to verify the chain may recompute it from the genesis-hash given below; no privileged access is required.

The integrity construction described here was introduced with the 2014 reorganisation of the catalogue and applied retroactively to all prior records in their then-current form. Records accessioned before that date carry, in addition to their proof-hash, an archival reference to the printed accession book in which the original entry was made.

per-record file inventory

Each public record exposes a small, fixed set of files addressable from its detail page:

Internal-only attachments (raw spectra, calibration logs, intake correspondence) are referenced by their internal hash but are not exposed in the public file inventory.

citation

For academic citation the registry recommends the following form:

anonymousmaterials registry. Record AM-YYYY-NNNNN-X (designation, locality). Retrieved YYYY-MM-DD from anonymousmaterials.com/sample/AM-YYYY-NNNNN-X.

Citation of multiple records may be abbreviated by listing the accession identifiers in ascending order without further qualification. The registry does not assign DOIs and does not request that one be assigned externally.

contact

Methodological queries — calibration practices, peak reporting, integrity-chain verification, anything that would be answered by an experienced registrar — should be directed through the same correspondence channel as ordinary public inquiries (see about). Replies are not signed. Where a registrar is in any doubt about the proper handling of a query, the matter is held for the next scheduled internal review.