Summary

A particle report can support electronic-grade chemical batch release only when the reported counts belong to a traceable and comparable measurement chain. A numerical “pass” is insufficient when the sampling location, bottle blank, particle-size channels, reporting unit, analyzed volume, or relationship between online and offline measurements is unclear. The first review should determine whether the sample represents the final filtered and packaged batch, whether blank contributions are low enough to distinguish product particles, and whether the reported channels match the purchasing specification. Online data may capture process transients, while offline data may include contamination from bottles, transfers, storage, and laboratory handling. These results should be correlated rather than expected to match automatically. No single particle limit, channel set, or sampling route is universally suitable across electronic chemicals and downstream processes.

What the Particle Report Proves—and What It Does Not

A liquid particle counter detects optical events and assigns them to particle-size channels. The result can describe the number and apparent size distribution of detected particles under the stated test conditions.

It does not normally identify:

• Particle chemistry;
• Particle source;
• Shape or hardness;
• Whether the particle will dissolve during use;
• Whether it will be retained by a point-of-use filter;
• Whether it will create a defect in the customer’s process.

For a light-scattering liquid-borne particle counter, the measured size is an optical-equivalent size influenced by the refractive indices of the particle and liquid matrix. ISO 21501-2:2019 also addresses size-setting error, counting efficiency, false counts, maximum particle concentration, sampling flow, sampling time, and sampling-volume error.

The result should therefore be interpreted as particle counts measured in a defined sample, through a defined sampling route, using a defined instrument and reporting basis.

It should not be interpreted as proof that the chemical is universally clean or suitable for every electronic application.

For batch release, the result must be connected to:

• Product code and grade;
• Production and filling batch;
• Sampling point and time;
• Final filtration stage;
• Commercial packaging;
• Analytical method;
• Particle channels and units;
• Applicable purchasing specification.

If these connections are incomplete, the data may still support process monitoring, but not necessarily commercial release.

Start With the Blank Pathway

Offline testing adds a bottle, cap, transfer route, holding period, and laboratory handling sequence between the production system and the particle counter. Each step may contribute particles.

“Bottle blank passed” is not sufficient information. The reviewer needs to know which background source was evaluated.

A useful bottle blank should represent the actual bottle material, cap system, cleaning route, bottle lot, conditioning procedure, holding time, and particle channels used for the sample.

A blank generated with another bottle type or under another handling sequence provides weaker evidence.

Blank Subtraction Is Not a Default Correction

Subtracting the blank count from the sample count can make a borderline result appear acceptable without resolving the measurement problem.

Blank subtraction is defensible only when the approved method defines:

• How the blank is prepared;
• How many blank measurements are required;
• Whether the mean, maximum, or another value is used;
• How blank variability is treated;
• How negative calculated values are handled;
• When the blank invalidates the test.

When blank counts approach the sample result or release limit, the method may not be able to distinguish product contamination from sampling contamination.

The stronger response is usually to investigate the bottle and transfer pathway, qualify a lower-background sampling system, and repeat the test under controlled conditions. Mathematical subtraction alone should not convert an incapable measurement pathway into release evidence.

Align the Particle Channels Before Comparing Results

Particle counts are comparable only when channel definitions and reporting bases are aligned.

A label such as “0.2 µm particles” may refer to:

• All particles equal to or larger than 0.2 µm;
• Particles within a defined size interval;
• A nominal instrument channel;
• A customer-defined threshold;
• A channel calibrated using a particular reference material.

These descriptions do not necessarily represent the same result.

Cumulative and Differential Counts Are Different

A cumulative result at ≥0.2 µm includes particles assigned to that channel and all larger channels.

A differential result may include only particles between 0.2 µm and the next channel.

A differential value cannot be compared directly with a cumulative limit. A supplier also cannot infer compliance at an unmeasured smaller channel from a passing result at a larger channel.

Subtracting one cumulative channel from another may produce a numerical interval, but that calculation should not automatically be treated as a validated differential measurement. Instrument response, channel resolution, and method rules still need to be considered.

Units and Volume Basis Must Match

Before two reports are compared, confirm whether the values are expressed as:

• Counts per milliliter;
• Counts per liter;
• Total counts in the analyzed volume;
• An average of several runs;
• A maximum result;
• A normalized or calculated value.

A total count from one analyzed volume cannot be compared directly with a concentration from another volume without a defined conversion.

The report should identify the actual analyzed volume, number of runs, replicate rule, and final reporting calculation. Small volumes may produce less stable estimates when particle concentrations are low or unevenly distributed.

The Lowest Channel Requires the Strongest Method Review

Smaller particle channels are often more sensitive to:

• Optical background;
• False counts;
• Bubble interference;
• Matrix effects;
• Sensor cleanliness;
• Flow instability;
• Counting efficiency;
• Concentration limits.

For counters with minimum detectable particle sizes between 30 and 100 nm, SEMI C77 provides a method for evaluating counting efficiency under defined conditions.

A smaller reported channel is not automatically stronger evidence. It becomes useful only when the instrument, matrix, concentration range, and verification status support reliable measurement at that channel.

Online and Offline Results Are Comparable Only Under Defined Conditions

Online monitoring and offline bottle testing observe different parts of the material pathway.

Online measurement may capture:

• Process variation;
• Filter behavior;
• Short contamination events;
• Filling-line changes;
• Time-dependent process recovery.

Offline measurement may include:

• Sampling-port contamination;
• Bottle and closure contribution;
• Transfer contamination;
• Storage effects;
• Settling or agglomeration;
• Bubble formation;
• Laboratory handling.

The two results do not need to be numerically identical. They need to be technically correlated.

A meaningful comparison requires:

1. The same production batch;
2. A traceable sampling time or time window;
3. Defined online and offline sampling locations;
4. The same or technically mapped particle thresholds;
5. The same cumulative or differential basis;
6. Aligned count units and analyzed-volume basis;
7. Documented sample conditioning and holding time;
8. Suitable instrument verification for both systems.

Without these conditions, a difference may reflect the measurement pathway rather than the chemical batch.

Which Qualification Stage Does the Result Cover?

A clean laboratory sample does not automatically qualify a production filling route or commercial package.

At the laboratory stage, the main question is whether the method can measure the chemical matrix with acceptable background, repeatability, and channel performance.

A production-equivalent trial should determine whether the intended filtration outlet, filling point, transfer hardware, and commercial package introduce a meaningful change in particle results.

Commercial release requires batch-specific evidence from the qualified configuration. A change in the final filter, sampling port, bottle, bulk container, filling line, instrument, channel configuration, or analytical method may require an impact assessment before previous qualification data are reused.

A detailed discussion of commercial packaging qualification is available in the guide to PFA and HDPE packaging validation for semiconductor wet chemicals.

Decision Matrix for Conflicting Particle Results

The following patterns can guide an investigation, but they do not establish a root cause by themselves.

The matrix should be applied together with the approved analytical method and quality agreement. It should not replace a documented deviation investigation.

The Lowest Particle Count Is Not the Strongest Evidence

Release reviews often focus too quickly on which laboratory or instrument reports the lower particle count.

A lower number may result from:

• A cleaner sampling point;
• A larger reporting threshold;
• A differently conditioned sample;
• A selected averaging window;
• A lower-background bottle;
• Particle settling before analysis;
• Different treatment of bubbles or false counts.

A higher number may represent actual contamination, but it may also include bottle, transfer, or instrument background.

The overlooked issue is measurement equivalence.

Two laboratories can disagree while both results are valid for their own sampling pathways. They can also agree while testing the same unrepresentative sample, with neither result describing the commercial package.

The strongest release evidence is not the lowest count. It is the result with the clearest connection to:

• The released batch;
• The relevant process point;
• The required particle channels;
• Controlled blank contributions;
• The final filling and packaging route;
• A predefined acceptance rule.

For R&D, this means defining particle thresholds that relate to the intended process rather than requesting the smallest available channel by default.

For production, it means preserving traceability between online events, filter status, filling time, and sampled containers.

For quality, it means preventing retesting from becoming repeated measurement until a passing result appears.

For procurement, it means confirming that the supplier’s evidence covers the commercial grade, package, and delivery configuration rather than only a specially prepared qualification sample.

Release Boundary and Action Order

Release may be supported when the batch and filling lot are traceable, blank contributions are controlled, channels and units match the specification, and the tested sample represents the qualified filtration, filling, and packaging route.

Release should be held when:

• The sampling point is unknown;
• The blank can explain a material share of the result;
• The required channel was not measured;
• Units or analyzed volume are missing;
• Online and offline results use non-comparable thresholds;
• Replicate variation remains unexplained;
• A method, filter, filling-line, or packaging change has not been assessed.

The review sequence should be:

1. Confirm the release specification, channels, units, and calculation basis;
2. Verify batch, sampling point, time, and package traceability;
3. Review instrument, bottle, transfer, and handling blanks;
4. Match online data to the offline sampling window;
5. Investigate unresolved differences before approval.

When submitting an RFQ to ChemicalCell, state the chemical and grade, intended process stage, required particle channels, count limits and units, sampling point, commercial packaging, destination, and whether online, final-package, or consecutive-batch particle data are required.