We keep a cat at our office. For a long time the worst part of my week was Monday morning.
With the air-conditioning running, the place was fine. Nobody commented. Then the aircon went off at the end of the day, the unit sat there overnight, and I would unlock the door the next morning and get hit in the face by it before I had taken two steps inside. Not a hint of something. A wall.
Since we started using an automatic litter box, that has stopped. I no longer smell the box at all in normal use. The one time I do smell it is when I open the waste bin to change the litter, and then it is worse than it ever was in the tray.
That is my honest experience, and I wanted to know whether it meant anything. So we went and read what has actually been measured. Some of it explained what I had noticed. One part of it says nobody has ever measured the thing I am claiming.
First surprise: the air-conditioning was never removing the smell
I had assumed the aircon was cleaning the air. It was not.
The foundational work here is Fang, Clausen and Fanger at the Technical University of Denmark, published in Indoor Air in 1998 and replicated in ASHRAE Transactions in 2000. They varied temperature and humidity across a normal comfort range and asked people to rate the air.
Odour intensity did not change significantly. The air smelled exactly as strongly at 26 °C as at 20 °C.
What collapsed was acceptability:
| Condition | People dissatisfied with the air |
|---|---|
| 20 °C, 40% RH | 19% |
| 23 °C, 50% RH | 40% |
| 26 °C, 60% RH | 85% |
Same pollutant load. A 4.5-fold swing in complaints, driven entirely by temperature and humidity. Their models tracked enthalpy — the combined heat and moisture content of the air — with correlations of 0.955 to 0.975. The mechanism is respiratory cooling: cool dry air on the lining of your airway reads as "fresh" more or less regardless of what is in it.
So the aircon was not fixing the office. It was making the same office bearable. And a split-system air conditioner recirculates room air rather than exchanging it with outside, so it was not diluting anything either.
Second surprise: the night is worse for three separate reasons
The tolerance effect above. Overnight the room drifts to Singapore ambient — around 28 °C and high humidity — which is past the 26 °C / 60% condition where 85% of people object.
The chemistry speeds up. Ammonia is the dominant litter box odorant, and how much of it leaves the litter and enters the air depends on temperature twice over. It becomes less soluble in water as things warm (Henry's law), and the ammonium/ammonia equilibrium shifts toward the volatile free base as the pKₐ falls. Working through the published constants for both effects, going from 20 °C to 30 °C gives roughly three times as much airborne ammonia from identical waste at identical pH — before any increase in the rate at which bacteria are producing it. That arithmetic is ours, derived from published values; we have not found anyone who has measured it in a litter box.
And nothing is carrying it away. In a sealed room, odorant concentration climbs toward a plateau set by the source rate divided by the ventilation rate. Halve the ventilation and you double the plateau. The useful correction to the intuition is that it does not climb all night — at typical air change rates a room is most of the way to its plateau within about six hours. What keeps rising overnight is the source itself, as bacteria multiply and the litter's pH drifts up.
Where the smell actually comes from — and why timing matters
Three findings, in order, explain why removing waste quickly should help.
Cat urine is sterile when it leaves the cat. A 2020 study in Frontiers in Veterinary Science cultured urine from ten healthy cats and found no bacterial community at all — every sample was culture-negative.
Ammonia cannot form without bacteria. Urea on its own is extraordinarily stable, with a decomposition half-life in water measured in years. The enzyme urease accelerates that by a factor of around ten thousand. Essentially all the ammonia in a litter box is bacterial.
So the bacteria arrive from somewhere — and mostly they arrive in the faeces. That is the connection almost nobody makes. The faeces are not simply a second smell sitting next to the first. They are the starter culture for the urine smell.
The sulfur side of it works differently and is worth knowing because it is so often described backwards. Cats excrete a species-specific amino acid called felinine, made in the urine by an enzyme called cauxin — cauxin produces felinine, it does not break it down. Felinine itself is odourless. It is felinine's later breakdown, into thiols such as 3-mercapto-3-methylbutan-1-ol, that produces the sharp catty note. Exactly how much of that breakdown is bacterial and how much happens spontaneously is not settled in the literature. One useful practical consequence is settled, though: felinine output is testosterone-dependent, and intact males produce far more of it than females or neutered males.
Ageing matters too. A 2019 study found cats could reliably distinguish fresh urine from the same cat's urine at 24 hours old — the volatile profile had already changed. Over two weeks, cat urine gains odorants, going from around 21 volatile compounds to 34–36. Faeces do the opposite and lose volatiles as they dry.
Now the part where I have to be careful
Everything above is a mechanism. It says that removing faeces and urine clumps within minutes rather than hours should mean less ammonia, because you have taken away both the bacterial inoculum and the substrate before much conversion has happened.
Nobody has measured it.
We searched for any study measuring odour or ammonia from a litter box as a function of how fast waste is removed. There is none. We searched for any peer-reviewed study of automatic litter boxes and odour, of any kind. There is none. Scooping frequency and odour: none. Litter change frequency and odour: none. What comes back instead is retail copy and patents, which is the reliable signature of a field with a lot of commercial activity and no science in it.
So what I have is a sound mechanism and one man's nose in one office. That is not evidence, and I am not going to dress it up as evidence. If you are deciding whether to spend a few hundred dollars on a machine partly because of what I have said here, weigh it accordingly.
The one thing that genuinely got worse
Opening the bin.
That is not a flaw in the machine, it is conservation of mass. A sealed bin cuts the rate at which odorants enter the room, so the room's concentration falls in proportion. Meanwhile the bin's own headspace climbs toward equilibrium with its contents, with no dilution at all. When you open it, that accumulated bolus comes out at once.
The automatic box did not make the smell go away. It moved it into a container I open once a week instead of a tray I stood over twice a day. That trade suits me enormously. It might not suit someone with a smaller flat and nowhere to take the bin.
(No one has measured this either. Sealed litter pails: no peer-reviewed evidence. Nappy pails: the published diaper research is about chemicals emitted from the nappies themselves, not about containment.)
Four things the evidence does support
1. Neuter the cat. Felinine and cauxin output are testosterone-dependent, and intact males excrete substantially more of the sulfur precursor. This is the cheapest odour intervention available and it is the best-evidenced one on this page.
2. The litter itself matters more than most people think. A 2023 study measuring ammonia emission rates found traditional clay litter at 728 mg/m²·h against 21–32 mg/m²·h for barley-based litters with probiotics — a more than twenty-fold difference, with odour reported as 70–90% lower. That is a single conference-proceedings study and we have not been able to verify every detail of it, so treat the magnitude as indicative. But the direction is supported independently: a 2022 paper in the Journal of Feline Medicine and Surgery showed a treated litter cutting urease activity by over 97% and taking headspace thiols to undetectable, against 59–110 ng above commercial products.
3. Zeolite is the best-evidenced additive. Natural clinoptilolite adsorbs ammonia and amines well, and it is the material in the one litter-specific study above that produced measured results. We do not sell one.
4. Activated carbon has a problem nobody mentions. Its adsorption capacity falls as humidity rises, and at high humidity water condenses in the pores. A litter box in Singapore is exactly the warm, wet environment where carbon performs worst — industrial systems cool and dry the air before the carbon bed for precisely this reason. Carbon also saturates silently: a filter that has reached capacity looks identical to a fresh one and does nothing. If your machine has a carbon filter, the replacement schedule is not optional.
And a caution about a product we sell ourselves
We sell a baking soda litter deodoriser. Having read the chemistry properly, our own product page overstates what it can do, and we are going to fix it.
Baking soda works by being a mild base that neutralises acidic odorants — the volatile fatty acids, which are mostly a faecal smell. The one published study we could find, on food waste rather than litter, found about 70% odour reduction at roughly 50 g per kg of waste. The same study warns that going above 100 g/kg pushes pH past 9.0 and increases ammonia release.
That matters here because raising pH is exactly the wrong direction for ammonia: the volatile free-base fraction rises tenfold for every pH unit. A litter box is an ammonia problem more than an acid problem. So generously dumping baking soda into a tray is, on the published chemistry, capable of making the dominant smell worse rather than better.
Our page currently says the product neutralises "acidic urine odour" and controls "ammonia smell". On what we can find, the first is the wrong description of cat urine odour and the second may be backwards at high dose. It also gives no dosage at all, which is the part that actually matters, since the published benefit and the published harm are separated only by how much you use. We are rewriting it with a stated dose. No one has studied baking soda in a litter box, so we will say that too.
One last thing, about what your cat thinks
A 2017 study in Behavioural Processes did something clever: it separated the smell of waste from the physical presence of waste, using real waste, fake waste, and odour alone.
Cats avoided the physical obstruction — real or fake clumps and faeces. The odour alone produced no avoidance at all.
So scooping matters to your cat, and it matters for a reason that is not the one on the packaging. The smell is your problem, not theirs.
If you want the technical companions to this piece: Automatic Litter Boxes and Tofu Litter covers what the specifications decide, 6 Litres of What? covers why you cannot compare litter prices in Singapore, and How to Get Rid of Cat Smell in an HDB Flat is the practical version for a home rather than an office.
Written September 2026. Sources: Fang, Clausen & Fanger, Indoor Air 8:80–90 (1998) and ASHRAE Transactions 106(2) (2000), on temperature, humidity and perceived air quality; Balboni et al., Frontiers in Veterinary Science 7:438 (2020), on the sterility of healthy cat urine; Ray, Saetta & Boyer, Environmental Science: Water Research & Technology 4:87–98 (2018), on urea stability and urease; Miyazaki et al., Chemistry & Biology 13:1071–1079 (2006), on cauxin and felinine; Suzuki et al., Journal of Chemical Ecology 45:579–587 (2019), on urine ageing; Banik, Koziel & Li, Separations 8:15 (2021), on cat urine and faecal volatiles; Robins et al., Journal of Feline Medicine and Surgery 24:123–130 (2022), on litter treatment and urease; "Ammonia and GHG Emission Rates from Traditional and Vegetable Cats' Litter", SHWA 2023 proceedings; Qamaruz-Zaman et al., Waste Management 35:187–193 (2015), on baking soda dosage in food waste; Ellis, McGowan & Martin, Behavioural Processes 141:284–290 (2017), on litter box appeal; Persily & Polidoro, NIST Technical Note 2213 (2022), for the single-zone mass balance; Sander, Atmospheric Chemistry and Physics 15:4399–4981 (2015), for Henry's law constants. The 20 °C to 30 °C ammonia calculation is our own derivation from those published constants and has not been measured in a litter box. The account of our office is one person's experience and is not a controlled observation.
