How to Search Alloy Compositions by Element and Percentage Range
An alloy composition search must do more than find records containing the right element names. A defensible result preserves percentage units, applies range and logical constraints consistently, distinguishes missing values from zero, and shows why each record matched, failed, or remained inconclusive.
The Sample below shows the filtering stage on five supplied fictional Fe–Al–Mg records. The filtering result is real and reproducible, but the records were not retrieved from an external alloy, patent, or literature database.
Fe–Al–Mg composition filter
Two supplied records satisfy every composition rule; one remains incomplete because Mg was not reported.
| Record | Fe | Al | Mg | Status | Decision reason |
|---|---|---|---|---|---|
| A-101 | 82 wt% | 12 wt% | 4 wt% | Exact match | All required ranges satisfied. |
| A-102 | 76 wt% | 18 wt% | 3 wt% | Excluded | Al exceeds 15 wt%. |
| A-103 | 88 wt% | 9 wt% | – | Incomplete | Missing Mg cannot be treated as zero. |
| A-104 | 80 wt% | 10 wt% | 6 wt% | Excluded | Mg exceeds 5 wt%. |
| A-105 | 83 wt% | 8 wt% | 5 wt% | Exact match | Inclusive boundaries satisfied. |
Define elements, units, ranges, and missing-value rules before running the Skill.
The result separates two exact matches, one partial record with missing Mg data, and two exclusions. That prevents incomplete compositions from quietly passing a numerical filter.
What an alloy composition search needs to answer
Define which elements are required or excluded, which unit applies, what ranges are inclusive or exclusive, and whether conditions use AND, OR, ONLY, or MAJORITY logic. Application, patent, document, and classification filters are additional constraints, not substitutes for composition criteria.
A useful deliverable keeps one row per source record, retains exact values or reported ranges, identifies the source when evidence comes from a patent or paper, and distinguishes exact, partial, and excluded results. Common failures include mixing wt% with at%, treating an unreported element as zero, merging unrelated records, changing boundaries, or inventing classification bands.
How to search, filter, and classify alloy records
First define composition criteria and units. Next obtain records from supplied data or a suitable search source. Normalize names and units only when supported. Apply logical and numerical filters record by record, retain missing fields, add only defined classifications, preserve provenance, and review partial matches instead of forcing a yes/no result.
In this run, Fe, Al, and Mg were required in wt%; Fe had to be the majority element; Al had to fall from 8 to 15 wt% inclusive; and Mg could not exceed 5 wt%. A-105 demonstrates inclusive boundaries. A-103 demonstrates that an unreported Mg value cannot be treated as zero.
Where ordinary AI gets composition filtering wrong
Without explicit rules, general AI may normalize incompatible units, infer absent elements, merge values from different sources, or create thresholds that the user never requested.
The Alloy Composition Search Skill packages interpretation and output rules as a repeatable workflow. It works directly with supplied records. When external evidence is needed and mace-mcp is configured, it can selectively structure a query, match alloy substances, retrieve documents, and extract document-backed composition data. It stops when sufficient data is available instead of calling every tool.
Why PatsnapMaterials decisions backed by searchable evidence
The Skill keeps composition rules and output format consistent, while optional connected research paths can bring in detailed patent and scientific records when external evidence is required. This Sample deliberately isolates filtering behavior and does not claim database-backed alloy findings.Database foundationPatsnap connects patents and papers with searchable substance, composition, property, process, and application information. This materials data foundation helps teams trace a composition or requirement back to technical evidence instead of treating extracted values as context-free facts.
Patsnap OpenPatsnap Open provides integration paths for selected data and tools through APIs and MCP Servers. Database evidence supports screening and comparison; source review, test conditions, and experimental validation still determine material suitability.
Prepare, install, and run the Skill
Prepare required and excluded elements, units, ranges, logical rules, and any application, document, or classification filters. You may provide your own table. Preserve blanks or - for unknown values instead of pre-filling zero.
Use the install command beside the Sample, then give your agent a precise request:
Use $alloy-composition-search on the supplied table. Require Fe, Al, and Mg; use wt%; require Fe to be the majority element; keep Al from 8 to 15 wt% inclusive and Mg at or below 5 wt%. Do not treat missing elements as zero.For supplied data, no MCP retrieval is required. Use the optional mace-mcp route only for composition extraction, substance matching, or supporting patents and papers.
How to interpret and verify the result
Exact match means every reported requirement is satisfied. Excluded means a reported value violates a rule. Partial means available data cannot support a defensible yes/no decision. Low, Medium, and High in this Sample are user-defined labels, not universal metallurgical categories.
For patent- or paper-backed results, retain a readable reference and note the source section when available. A composition match alone does not prove identical processing, microstructure, properties, application suitability, patent scope, or freedom to operate.
Bring current evidence into your AI workflow
Connect your agent to patent, scientific, and engineering research tools for the next evidence-heavy question.
Frequently asked questions
What input does the Skill need?
Provide the required and excluded elements, units, ranges, logical rules, and any application or classification constraints.
Does missing composition data count as zero?
No. A missing value remains unknown unless the source explicitly reports zero.
Does the Skill require an MCP connection?
No connection is required to filter supplied records. External retrieval is optional when the task also requires patent, literature, or composition evidence.
Disclosure: This article describes a Patsnap product and links to Patsnap Open. Skill output supports research and does not replace qualified materials, engineering, scientific, regulatory, or legal review.
