Specialty Chemical Impurity Testing

Silicon Purity Testing

Sterling Analytical provides silicon purity testing across the full range of commercial silicon grades — metallurgical, solar, and electronic — quantifying boron, phosphorus, and metallic impurities at the sensitivity each grade actually demands by ICP-OES and ICP-MS. Few materials span as wide a purity range under a single name as silicon does: metallurgical-grade silicon at roughly 98–99% purity and electronic-grade polysilicon at up to 99.999999999% (11N) purity are both called “silicon,” but they represent analytical testing problems that are barely related to each other.

This range is genuinely enormous, and understanding where your material sits on it — and which specific impurities actually matter at that tier — is the starting point for scoping silicon purity testing correctly. A panel and sensitivity level appropriate for metallurgical-grade material would be completely inadequate for solar-grade polysilicon, and testing intended for solar-grade material would, in turn, fall well short of what electronic-grade semiconductor feedstock requires.
Silicon Purity Testing

Understanding the Silicon Purity Hierarchy

Silicon purity is conventionally expressed in “N” notation — the number of nines in the purity percentage — and the jump between grades represents orders-of-magnitude differences in both required purity and acceptable impurity profile, not incremental refinement.
A useful way to make 9N purity concrete: it represents roughly one foreign atom for every billion silicon atoms — a level of purity that puts electronic-grade silicon among the purest industrial materials produced anywhere.

Why Boron and Phosphorus Get Special Treatment — And Why They're Genuinely Different From Other Impurities

This is worth understanding specifically, because boron and phosphorus occupy a conceptually different role than most other silicon impurities, and that role actually flips depending on which stage of the silicon supply chain you’re looking at.
Boron and phosphorus are both dopants — elements deliberately introduced into silicon during crystal growth at carefully controlled concentrations to create p-type or n-type semiconductor material, the foundation of how transistors and photovoltaic cells actually function. At the finished wafer stage, boron or phosphorus presence isn’t contamination at all; it’s the entire point, added precisely and intentionally. But at the polysilicon feedstock stage — before controlled doping ever happens — any boron or phosphorus already present in the raw material is an uncontrolled, unwanted contaminant that will distort the carefully engineered doping profile a manufacturer is trying to create later in the process.
This is why boron and phosphorus get such intensive, specific attention in feedstock-stage silicon purity testing across both solar and electronic grades: it’s not that these two elements are more toxic or more difficult to remove than other contaminants in some general sense, it’s that their electrical activity in silicon is so significant that even trace amounts measured in parts per billion can meaningfully alter the material’s electrical behavior in ways that compete with and corrupt the intentional doping process downstream. Testing for boron and phosphorus specifically, often at tighter sensitivity than the broader metallic impurity panel, reflects this genuinely distinct electrical significance rather than a generic “test for everything” approach.

Matrix & Digestion

Silicon’s digestion behavior varies meaningfully by grade and physical form, and the preparation approach has to match both.

Our approach:

What We Test For

Element panels are scoped to the specific grade and application — a solar cell manufacturer’s concerns are genuinely different from a semiconductor wafer producer’s, even though both are nominally testing “high-purity silicon.”

Method Selection: Matching Sensitivity to Grade

This is where being direct about method capability matters most, since using an insufficiently sensitive method for a given grade produces a result that looks complete but doesn’t actually demonstrate what the specification requires.
We’ll scope the method to your actual grade and specification rather than defaulting to a single approach across genuinely different sensitivity requirements.

Why Purity Requirements Keep Tightening

It’s worth understanding that “electronic grade” isn’t a fixed target — it’s a moving one, and what qualified as adequate purity for semiconductor manufacturing some years ago doesn’t necessarily meet current leading-edge requirements today. As device geometries continue shrinking and transistor densities increase, the tolerance for any given level of metallic or dopant contamination correspondingly tightens, since smaller device features are more sensitive to a given absolute level of contamination than larger ones were.
This matters practically for ongoing supplier qualification and specification management: a silicon source that comfortably met specification for a given application several years ago may no longer be adequate for current-generation requirements, even without any change in the material itself — the bar has simply moved. This is part of why periodic re-verification, not just one-time qualification, makes sense for silicon feedstock used in applications where purity requirements continue to evolve.

Common Contamination Issues We Identify

Who Uses This Service

Sample Quantity & Packaging

Required sample size: 5–20 grams of representative material, depending on grade and required sensitivity.

Packaging guidelines:

Turnaround Time & Pricing

Standard turnaround: 3–5 business days for metallurgical and solar-grade ICP-OES testing Extended turnaround may apply for electronic-grade ultra-trace panels; rush service available on request
Pricing starts from $150 per sample for metallurgical-grade testing, with solar and electronic-grade ultra-trace panels priced according to required sensitivity and element scope.

What You Receive

Clients receive a detailed analytical report suitable for grade verification, specification compliance, and commercial documentation.

Your COA includes:

All results are supported by CRM-traceable calibration, with duplicates and matrix spikes performed on each analytical batch — particularly important for solar and electronic-grade material where results directly support high-value commercial and technical decisions.

Methods & Standards

Sterling Analytical applies methods scaled to silicon grade and required sensitivity:

Explore related services:

Request a Quote

Submit your sample details and intended silicon grade to receive a tailored quote and recommended analytical approach.

Frequently Asked Questions

It's used to verify boron, phosphorus, and metallic impurity content across metallurgical, solar, and electronic-grade silicon, supporting grade verification, specification compliance, and supplier qualification.
Metallurgical grade is roughly 98-99% pure. Solar grade ranges from 6N to 9N (99.9999% to 99.9999999%). Electronic grade reaches 9N to 11N (up to 99.999999999%). Each represents orders-of-magnitude tighter purity and a genuinely different testing requirement, not incremental refinement.
Boron and phosphorus are dopants intentionally added during controlled crystal growth to create semiconductor behavior. At the raw feedstock stage, before controlled doping happens, any boron or phosphorus already present is an uncontrolled contaminant that can distort the intended doping profile, which is why these two elements receive such specific, sensitive attention.
ICP-OES is generally adequate for metallurgical and standard solar-grade testing. ICP-MS becomes necessary for premium solar-grade and electronic-grade material, where required sensitivity drops to sub-ppb levels ICP-OES cannot reliably reach.
It tightens. As semiconductor device geometries shrink and transistor density increases, tolerance for any given contamination level correspondingly tightens, meaning material that met specification several years ago may no longer be adequate for current requirements.
5–20 grams of representative material, depending on grade and required sensitivity.
Standard turnaround is 3–5 business days for metallurgical and solar-grade testing, with extended turnaround possible for the most demanding electronic-grade ultra-trace panels.