Featured

What is a Collision Cross Section (CCS), and why should environmental chemists care?

Panyawut Tonanon

05 Jun 2026 • 6 min read

Photo by National Cancer Institute / Unsplash

Every environmental chemists are familiar with sugar-cube-in-Olympic-pool analogy, which generally goes this way: if we dissolve a sugar cube in an Olympic-size pool, the concentration of sugar comes down to 1.6 ppb (ug/L). For sensitive instruments the like of typical LC-MS or more sensitive ones such as LC-MS/MS or LC-QTOF, achieving ppb-level sensitivity is straightforward. However, unwanted signal such as dirty sample or instrument noise can also be amplified, which reduce the confidence in signal identity.

When we identify a compound by mass spectroscopy, we usually pin it to two numbers: its mass-to-charge ratio (m/z) and its retention time (rt). Together, those two handles let you say "this is probably compound X" with reasonable confidence.

But "reasonable" isn't always good enough. In environmental chemistry, the same m/z can hide a dozen different molecules — isomers with identical formulas but very different behaviors. And retention times drift the moment you change columns, gradients, or sample matrices.

That's where the third dimension comes in: collision cross section (CCS). Think of it as a fingerprint of shape. It's a number, measured in square angstroms (Å²), that describes how an ion tumbles through a buffer gas under a weak electric field. For a given molecule in a given drift gas, it's basically a constant. And unlike retention time, it doesn't care about your column.

In this post I'll walk through what CCS actually is, how it's measured, and why it has quietly become one of the most useful additions to environmental analytical chemistry in the last decade.

The problem with two dimensions

The standard LC-MS workflow gives us two handles on identity: m/z (what the molecule weighs, divided by its charge), and RT (when it elutes from the column). Both are useful, yet imperfect.

m/z is brilliant at ruling molecules out. If your compound's nominal mass is 156, it isn't glucose. But m/z can't tell you whether that 156 ion is a part of sulfadimethoxine (the most important sulfa antimicrobial in aquaculture), or the instrument noises (Thurman et al., 2004). This problem can partly be circumvented by looking at the fragmentation pattern (this is where LC-MS/MS comes in), or the exact mass at higher resolution (where LC-QTOF shines).
RT, on the other hand, is more selective. Different compound with different interaction against the column should, theoretically, has different retention time. However, the retention time is pretty much depends on everything: column chemistry, mobile phase composition, gradient, temperature, and even sample matrix. A method that works in clean standard solution can shift noticeably in a wastewater extract.

For targeted analysis, where we have a pure standard and a well-behaved matrix, these two are usually enough. But the questions environmental chemists increasingly need to answer are non-targeted: what's in this sample that I wasn't looking for, and is it dangerous? For that, two dimensions of evidence are not enough.

What a collision cross section actually is?

Imagine a single ion floating in a tube filled with nitrogen gas. You switch on an electric field. The ion starts drifting towards the detector.

How fast it drifts depends on how often it bumps into gas molecules along the way. A small, compact ion slips through easily. A bulky, floppy ion tumbles and bounces, losing energy to each collision. The drift time we record is a direct readout of the ion's rotationally averaged surface area (the area it presents to the buffer gas as it tumbles).

This area, converted through the Mason-Schamp equation, is the collision cross section. The drift velocity is proportional to the electric field strength, multipled by the ion's mobility constant (K). K itself depends on the ion's charge, the buffer gas, the temperature, and the ion's collision integral with the gas. CCS is the experimentally convenient shorthand for that collision integral.

A few properties of a measured CCS value:

It's a psychicochemical constant of the ion, not a chromatographic artefact.
It's reproducible across instruments (mostly)
It's matrix-independent in a way that retention time is not.
It encodes three-dimensional shape, not just mass.

That last point is what makes CCS so useful for environmental chemistry, where we routinely deal with isomers that have very different toxicities.

Why this matters in environmental chemistry?

Three concrete use cases, ordered roughly by how much difference CCS makes in a typical workflow:

1. Cleaning up messy matrices. A wastewater extract can have thousands of features in a 30-minute gradient. Many of them co-elute. Adding an IMS dimension spreads those co-eluting ions across drift time as well, improving signal-to-noise and dropping detection limits by 5–10× in published workflows. For trace organic micropollutants at ng/L concentrations — the equivalent of dissolving a sugar cube across all the water in an Olympic swimming pool — that headroom is the difference between "detected" and "missed."

2. Separating isomers of the same exact mass. The textbook case: cocaine and its hydroxylated metabolites in wastewater-based epidemiology. They share many nominal m/z values, but their CCS values are different. When the matrix is messy and the concentrations are tiny, that extra dimension is what lets you trust a peak assignment.

3. Killing false positives. A suspect-screening workflow flags thousands of features against a database. Most of them are not actually what the database says they are. Adding a CCS match — predicted or measured — roughly halves the false-positive rate in published comparisons. For a high-throughput lab, this is the single biggest practical argument for adopting the technique.

The 6PPD-quinone story is a useful worked example. 6PPD-quinone is the salmon-killing transformation product of 6PPD, a common p-phenylenediamine antioxidant used in tire rubber. It has been linked to acute mortality in coho salmon at sub-µg/L concentrations, and is now a regulatory priority contaminant in several jurisdictions. 6PPD and 6PPD-Q have different exact masses, but related antioxidant species and transformation products cluster around similar m/z values. CCS is what lets a non-target screen tell them apart confidently — and tells you when a feature you thought was 6PPD-Q is actually some other isomer with the same mass.

The catch

There are two honest caveats.

First, the experimentally measured CCS database is tiny compared to the chemical universe. A few thousand carefully measured values is impressive work, but the number of "known unknowns" environmental chemists care about is in the tens of thousands. This is exactly where machine learning comes in. Tools like DeepCCS, support vector regression models, and trained artificial neural networks now predict CCS from a molecule's SMILES string with accuracy comparable to inter-laboratory variability. That is its own post — and it's coming.

Second, IM-MS alone does not identify unknown compounds. It is a filter and a confirmer, not a structure-elucidation tool. The trifecta works when at least one of m/z, RT, or CCS is anchored against a known standard or a database entry. Without that anchor, all you have is a shape.

Closing

CCS is not a magic bullet. It is, however, a third orthogonal identifier that — for the first time — makes small-molecule environmental analysis genuinely robust against matrix effects and isomer interferences. For labs that are already running LC-HRMS, adopting it is a software upgrade and a method-development project, not an instrument purchase.

If you want a sense of where the field is heading, watch two things in the next few years: the growth of community CCS databases (the Hinnenkamp OMP library, the CCSbase project, and the NORMAN Suspect List Exchange integrations); and the maturation of ML-predicted CCS for high-throughput suspect screening.

Both of those will get their own posts. Subscribe if you want to be there when they drop.