01Distinguish obscuration from an image-based measurement

NIST’s technical paper explains light obscuration through the reduction in transmitted light as a particle passes the sensing region. Flow imaging instead records images of particles moving through a viewed region and analyses those images.NIST — Optical methods for particle characterisation (opens in a new tab)

The reported diameter may be an equivalent size defined by the method rather than a direct measurement of every physical dimension. A long thin fibre and a compact particle can therefore require more context than one diameter value.

Image information can help describe morphology, but shape classification is not automatically chemical identification. An instrument label such as “protein-like” needs the classifier’s evidence and limitations.

02Consider how the liquid and particle affect optical detection

Zölls and colleagues investigated refractive-index effects in protein-particle measurements. Increasing the formulation’s refractive index reduced the measured concentration by light obscuration and flow imaging in their experiments; close optical matching could make particles undetectable to those systems.Zölls and colleagues — Refractive index and apparently invisible protein particles (opens in a new tab)

This is a limitation of contrast-dependent measurement, not evidence that the particles physically disappeared. It also does not mean that every laboratory solution produces the same effect.

Ripple and Hu studied corrections for relative bias between light obscuration and flow-imaging instruments. Their results showed improved agreement after method-specific corrections, while retaining residual differences.Ripple and Hu — Correcting relative bias in particle counters (opens in a new tab)

When methods disagree, ask whether their detection and sizing definitions are comparable before concluding that one instrument counted incorrectly.

03Read cumulative and interval counts correctly

For an original example, suppose a report gives 120 particles/mL at or above 10 µm and 30 particles/mL at or above 25 µm. These are cumulative bins: the second group is already included in the first.

The count from 10 µm up to, but not including, 25 µm is therefore 120 − 30 = 90 particles/mL, assuming matching measurements and the stated boundary convention. Adding the two cumulative counts would double-count the larger particles.

A different report might use separate intervals instead. Preserve its actual bin definitions, including whether boundary sizes are included. Similar-looking table headings can otherwise create a false comparison.

Counts per millilitre and counts per container also have different denominators. Converting between them requires a represented volume and does not establish that particles are distributed uniformly through every container.

An instrument’s lower sizing boundary is not evidence that smaller particles are absent. It identifies the region not described by that reported count.

04Preserve the preparation and sampling history

A particle measurement applies to the prepared portion that reaches the sensing region. Settling, bubbles, adsorption or changes during dilution can complicate whether that portion represents the original sample.

A preparation chosen to improve optical contrast can also change the particles themselves. Its suitability must therefore address both detection and preservation of the feature being counted.

A low subvisible count does not establish dissolved chemical purity or absence of endotoxin. The most informative report combines the measured size distribution with its method and limitations, leaving other quality attributes to their own evidence.