The Detergent in Your Ice Cream

August 25, 2026 · 7 min read

Flip over the ice cream. Somewhere after the cream and sugar: cellulose gum, polysorbate 80, mono- and diglycerides. They are why it scoops smoothly, why the bread stays soft for two weeks, why the dressing never separates. They are emulsifiers. Chemically speaking, they are detergents. And for sixty years, the safety question nobody asked was what a detergent does to the one part of your body that is coated in something a detergent dissolves.

What an emulsifier is

An emulsifier is a molecule that makes oil and water stay mixed. Soap is one. The two most studied in food are carboxymethylcellulose, listed on labels as cellulose gum or CMC (CAS 9004-32-4, E466), and polysorbate 80 (CAS 9005-65-6, E433). They appear in ice cream, packaged bread, salad dressings, sauces, plant milks, and a long tail of other processed foods. They earned their place honestly: they make products smoother, more stable, and longer-lasting on the shelf.

Your intestines wear a mucus coat that keeps trillions of gut bacteria at a safe distance from your own cells. Emulsifiers are detergents. In the first controlled human trial, 11 days of one common emulsifier was enough to change the gut microbiome.

"It's just a stabilizer. It passes right through you." That is exactly the problem. It passes through the gut, which is where the microbiome lives. Emulsifiers were approved on toxicity and cancer testing done decades before anyone knew the microbiome mattered. Nobody was asked whether they were safe for your bacteria, because nobody was asking about bacteria at all.

The mucus layer, and what happened in mice

The lining of your gut is protected by layered mucus. Most of your gut bacteria never touch your intestinal wall; the mucus holds them at a distance, which is a large part of how you host trillions of them without constant inflammation.

In 2015, Benoit Chassaing and colleagues published in Nature a direct test of what emulsifiers do to that arrangement. Mice were given CMC or polysorbate 80 at low concentrations, about 1 percent, in their water. The result: bacteria ended up dramatically closer to the intestinal wall, in mucus that should have been nearly sterile. The mice developed low-grade inflammation and the mouse version of metabolic syndrome, with more body fat and worse blood sugar. Mice genetically prone to colitis developed robust colitis.

The elegant part of the study is how it pinned down the mechanism. When the researchers transferred the gut bacteria of emulsifier-fed mice into germ-free mice that had never touched an emulsifier, the recipients developed the same problems. The emulsifier was not poisoning the mouse. It was changing the bacteria, and the changed bacteria did the rest.

11 days In a double-blind controlled-feeding trial, healthy adults eating carboxymethylcellulose for 11 days showed reduced gut microbiome diversity and reduced short-chain fatty acids (Chassaing et al., 2022, Gastroenterology).

Then they fed it to people

Mice are mice. So in 2022, the same group published a human test in Gastroenterology: a double-blind controlled-feeding study in which healthy adults lived on an emulsifier-free diet, and some of them had 15 grams per day of CMC added to it for 11 days. Everything else about the diets was identical.

The CMC group's gut microbiome lost diversity. Their stool showed reductions in short-chain fatty acids, the compounds gut bacteria produce when they ferment fiber, which are broadly considered beneficial to the gut lining. And in 2 of the 7 CMC subjects, bacteria had moved into the normally sterile inner mucus layer, the same encroachment seen in the mice, and a feature researchers describe as central to gut inflammation.

Honesty about scale: this was 16 people for 11 days, the effects in most subjects were modest, and 15 grams per day sits at the high end of real-world intake. Nobody developed a disease in a week and a half, and this study does not prove emulsifiers cause disease in humans. What it establishes is narrower and still important: the mouse finding was not a mouse quirk. The mechanism operates in people, at a dose a processed-food-heavy diet can approach, on a timescale of days.

How to read a label for them

The names to know: carboxymethylcellulose, cellulose gum, polysorbate 80, and their frequent companions, mono- and diglycerides and carrageenan. Two fairness notes belong here. First, not all emulsifiers are the same molecule, and they have not all been tested; lecithin, from eggs and soy, is chemically a different thing than a synthetic cellulose derivative, and the research above tested CMC and polysorbate 80 specifically. Second, a food containing an emulsifier is not poison, and eating one is not an emergency. The research question is about daily, cumulative exposure across most of what you eat.

You cannot rinse an emulsifier off, but you can count how many meals a day contain one. The practical lever is not vigilance about any single product. It is the ratio of your week that comes out of a pot instead of a package. A dinner cooked from an onion, a chicken thigh, and a can of tomatoes contains zero synthetic emulsifiers, not because anyone read a label, but because nothing in it needed to survive a warehouse.

The system part: fewer labels to read

Here is the trap with "read every label": it adds a research project to a grocery trip that was already the most cognitively loaded errand of the week. Vigilance does not survive a Tuesday.

A planned week of home cooking changes the denominator instead. When most meals are built from single ingredients, the emulsifier count falls as a side effect, without anyone standing in aisle five squinting at E-numbers. Hestia plans that kind of week: whole meals from real ingredients, the list built from what the pantry is missing. And for the packaged goods that do earn a place in the cart, Ember, Hestia's 0 to 100 nutrition-and-processing score, shows the factors behind its number, so heavily processed products are visible as what they are. A score is a signal, not a diagnosis. But it turns the label-reading project into a glance.

See a week with a short ingredient list

Whole meals from real ingredients, pantry subtracted, cost visible before you shop.

See a sample plan

There are households where the ingredient lists are short. Not because someone memorized the E-numbers. Because most of what gets eaten there never had a label in the first place.

The bread has four ingredients because it was chosen once, deliberately, and stays on the list. The dressing is olive oil and vinegar because that takes eleven seconds.

Nobody in that kitchen is fighting the food industry. They just cook from a plan.

Week 5 changes everything. It always does.