What does “log reduction” mean in air cleaning performance claims?

Key points

• A log reduction describes how much a measured concentration has dropped on a base-10 scale. Each “log” is a ten-fold reduction.
• As a rough conversion: 1-log = 90%, 2-log = 99%, 3-log = 99.9%, 4-log = 99.99% reduction.
• Log reduction is meaningful only when you know what was measured, how it was sampled, and over what time period.
• In air cleaning reports, log reduction often refers to microbial inactivation or removal, which are not the same thing.
• A strong claim pairs log reduction with the test method, starting level, detection limits, and checks for by-products.

You will often see performance statements such as “3-log reduction”, “99.9% kill”, or “4-log inactivation” in air cleaning and infection-control contexts. These numbers sound precise, but they are easy to misunderstand if you do not know what a “log” means.

This article explains log reduction in simple terms, shows how it relates to percentage reduction, and highlights the details that make a log claim meaningful in air purifier test reports.

What is “log reduction”?

A log reduction is the logarithm (base 10) of the ratio between the starting level and the final level of something measured.

In plain language:

• 1-log reduction means the final level is 10× lower than the starting level.
• 2-log reduction means 100× lower.
• 3-log reduction means 1,000× lower.
• 4-log reduction means 10,000× lower.

For a formal definition from a regulatory technical guidance document (not specific to air purifiers but widely used for disinfection terminology), the U.S. EPA notes the common interpretation that one log removal corresponds to 90% removal.

How does log reduction translate to percentages?

This is the conversion people usually mean:

The key thing to remember is that the differences are not small. Moving from “99.9%” to “99.99%” is not an extra 0.09% in a casual sense—it is a ten-fold difference in what remains.

Why starting levels matter

Log reduction depends on where you start. Here is a simple example:

• Start with 1,000,000 units (of bacteria, virus, particles—whatever is being counted).
• A 3-log reduction (99.9%) leaves 1,000 units.
• A 4-log reduction (99.99%) leaves 100 units.

So, a claim can sound impressive and still leave a measurable remainder depending on starting concentration and the sensitivity of the measurement.

This is also why good reports state:

• the initial concentration, and
• the final measured level or detection limit.

If the report only shows “% reduction” without showing starting level and detection limits, the claim is harder to interpret.

In air cleaning reports, does “log reduction” mean removal or inactivation?

This is a common source of confusion.

Removal

“Removal” usually means the organism or particle is no longer detected in the air sample because it has been captured (for example, in a filter or on a surface).

Inactivation

“Inactivation” usually means the organism is no longer viable (able to replicate or form colonies), even if it is still physically present somewhere in the system.

A report might show a large log reduction in viable counts because organisms were inactivated, even if the device also captured them. Or it might show a reduction because the organisms were physically removed from air by filtration.

Helpful resource in Indoor Air Knowledge Hub: How to read air purifier laboratory test reports.

What measurements produce a “log reduction” number?

You will typically see log reductions derived from one of these:

• CFU counts (colony forming units) for bacteria and some fungi (viability-based).
• Plaque assays / TCID50 for viruses or viral surrogates (viability-based).
• Particle counts or mass concentration (not usually framed as “log reduction”, but mathematically similar).

Log language is most common for microbes because the numbers can span large ranges and because many test standards use log reduction as a pass/fail or performance threshold.

What should you look for in a “log reduction” claim?

A log number by itself is not enough. When you see “3-log reduction”, check the surrounding details:

1) Over what time period?

Is it “3-log in 5 minutes” or “3-log in 2 hours”? If a report only provides an end point, ask for the time curve.

2) Under what conditions?

Key conditions often include:

• airflow rate / fan setting
• chamber size and mixing
• humidity and temperature
• whether the test was single-pass or room-decay style

3) What organism was used?

Different organisms vary in susceptibility and test handling. “A virus surrogate” is not identical to “all viruses”, and bacterial species vary too.

4) What does “not detected” mean here?

If results end at “not detected,” the detection limit matters. “Not detected” can mean:

• it truly reached near zero, or
• it dropped below what that method can measure.

5) Was secondary pollution assessed where relevant?

If a technology uses high-energy chemistry (e.g., UV/PCO, plasma/ionisation), it is reasonable to ask whether by-products (such as ozone or VOC fragments) were measured in the same test programme. For more information, see: Secondary Air Pollution.

How does log reduction relate to CADR and ACH?

These metrics answer different questions:

• Log reduction tells you “how much the measured concentration dropped” in a specific test context (often a microbe or a defined contaminant).
• CADR (Clean Air Delivery Rate) tells you “how much particle-cleaned air the device delivers per unit time.”
• ACH / eACH helps translate clean-air delivery into room size language (room volumes per hour).

If a claim is framed as “log reduction,” you should still ask how it connects to air delivery and room use. A device can achieve a strong single-pass inactivation number in a rig, yet have limited impact in a large room if the clean-air flow rate is low for that space.

Summary

A log reduction is a ten-fold scale for describing how much a measured level has dropped. It is a legitimate and widely used way to report performance, especially for microbial tests, but it only becomes meaningful when paired with the test method: what was measured, how it was sampled, over what time, at what airflow, and with what detection limits.

When you read a “3-log” or “99.9%” claim, the best follow-up question is simple: 3-log of what, measured how, and under what operating conditions?