Cockroaches do avoid certain odors, but the effect is small, short-lived, and easy to overstate. The best measured data comes from an Auburn University study that tested 12 essential oil compounds against German cockroaches in boxes designed to mimic a real harborage. Averaged over 14 days, the strongest compound, citronellic acid, repelled only about 45% of roaches. Peppermint's main component repelled 12%. Eucalyptus and citrus were statistically indistinguishable from an untreated control. Catnip oil is the one genuine standout: a 2002 *Journal of Economic Entomology* paper found one nepetalactone isomer significantly more active than an equivalent dose of DEET. Even so, none of this is a control method. A 2022 review in the same journal warns that "at worst, a repellent would disperse the cockroaches throughout a house or apartment," and UF/IFAS Extension notes that repellent products near bait render the bait ineffective.
The earlier version of this article said peppermint oil repels "up to 100% of German cockroaches within 24 hours of exposure," and listed eucalyptus, citrus, cucumber, and tea tree as effective. We could not find a source for the 100% figure, and the measured data points the other way: eucalyptus and citrus were among the weakest compounds tested. This version reports the numbers that exist and is explicit about what they do and do not support.
The Only Ranked Data Worth Quoting
In 2009, Alicia Kyser Phillips completed a master's thesis at Auburn University under Arthur Appel, whose lab has studied cockroach physiology for decades. She tested 12 pure essential oil components against adult male German cockroaches using Ebeling choice boxes, an apparatus that gives the insect a lit side and a dark side. Because German cockroaches strongly prefer dark harborage, any roach that stays out in the light is telling you the dark side smells bad enough to override its instincts. The thesis calls this method the better approximation of "normal cockroach habitat."
Here is the mean percentage of roaches repelled over 14 days, with the everyday scent each compound belongs to:
- Citronellic acid (citronella, lemongrass): 45.0% — the clear winner, and the only compound significantly better than everything else tested.
- Beta-pinene (pine): 22.0%
- Geraniol (rose, geranium): 19.9%
- Carvacrol (oregano): 16.0%
- Alpha-pinene (pine): 15.9%
- Linalool (lavender): 14.9%
- Eugenol (clove): 14.5%
- Thymol (thyme): 13.1%
- Menthone (peppermint, mint): 12.4%
- trans-Cinnamaldehyde (cinnamon): 12.1%
- 1,8-Cineole (eucalyptus): 9.6%
- Limonene (lemon, orange): 5.6%
- Untreated control: 3.9%
Two results deserve attention. First, limonene and 1,8-cineole were not statistically different from the untreated control. Citrus and eucalyptus, the two scents most often named in roach-repellent listicles, performed about as well as nothing. Second, even the best compound left the majority of roaches unbothered. A treatment that moves 45% of a population is not a treatment that clears a kitchen.
The effect also decayed quickly. The thesis reports that for most compounds, repellency "decreased dramatically during the first 2-3 days of the experiment." Whatever you apply on Monday is substantially weaker by Thursday.
Why Some Lists Claim Much Higher Numbers
Phillips also ran a second method, the harborage-choice test, in which roaches pick between a treated cup and an untreated one. Those numbers look far more impressive: carvacrol 76%, citronellic acid 75%, eugenol 73%, thymol 73%.
But look at the control in that experiment. Untreated cups "repelled" 52% of the roaches, simply because the insects sat where they started and never explored. The thesis says so directly: the method "lacks a stimulus (such as light) to encourage the movement of cockroaches to the treated cup; therefore, even if the treated cup was not repellent, it may have appeared repellent." Several compounds scored below the do-nothing control.
This is how a weak repellent becomes a marketing claim. Change the arena and the same oil goes from 13% to 73%. When you see a large repellency percentage quoted without a control value or a test design, it is not telling you much.
Catnip Is the Real Outlier
Catnip is the one scent with a defensible claim, and it is worth being precise about what the research says.
Chris Peterson, Leah Nemetz, Leah Jones, and Joel Coats at Iowa State tested catnip essential oil and its two nepetalactone isomers against adult male German cockroaches in a choice-test arena. Their 2002 Journal of Economic Entomology paper reports that E,Z-nepetalactone "was the most active of the compounds tested, being significantly more active to this species than equivalent doses of DEET, the essential oil, or Z,E-nepetalactone." They also found that insects with their antennae removed showed no response, confirming the effect is genuinely olfactory rather than irritant contact.
You will see this reported as "100 times more effective than DEET." That figure comes from a 1999 American Chemical Society conference presentation, not from the peer-reviewed paper, and it does not appear in the published results. The published finding is narrower: significantly more active than DEET at equivalent doses, against adult males, in a lab arena.
Three limits matter before you buy catnip oil. The study tested purified isomers, not a bottle off a shelf. It used males only, and the researchers noted that males appear more sensitive than females. And there is no registered catnip-based cockroach product with efficacy data behind it, so there is no tested concentration, carrier, or reapplication interval to follow.
Why Repelling Roaches Is the Wrong Goal
This is the part most scent articles skip, and it is the most important paragraph on this page.
A repellent scatters an infestation instead of shrinking it. The Auburn thesis puts it plainly: "If used incorrectly, repellent insecticides can spread an infestation to other parts of the home or building." The 2022 review by Oladipupo, Hu, and Appel reaches the same conclusion. Spray under the sink where the roaches are concentrated and some of them relocate into wall voids, appliance housings, or, in an apartment building, next door.
It ruins bait, which is the thing that actually works. UF/IFAS Extension's German cockroach guidance states that "using a product with repellent properties in an area where bait may be applied will render the bait ineffective because the treatment will contaminate the bait, making it repellent, and cockroaches will not consume it." Gel bait only works if it is the most appealing thing in the room. Scenting the area around it defeats the mechanism. If you are running a scent routine alongside bait, see our guide to how often to spray peppermint oil, which goes into why a reapplication schedule makes this worse.
Not seeing roaches is not the same as having fewer roaches. Pushed out of one cabinet, a population keeps breeding somewhere less convenient to treat. UC IPM makes a related point about aerosols and foggers, warning that they "may repel and disperse cockroaches to other areas without actually killing them."
The legitimate professional use of a repellent is narrow: as a flushing agent during an inspection, or to push roaches out of an untreatable harborage and toward a non-repellent product. That is a technician's tool used deliberately, not a bottle of oil misted along a baseboard.
The Kitchen Remedies, One at a Time
- Bay leaves. Whole or crushed bay leaves are a perennial internet answer with no supporting research on cockroaches. The active aromatic in bay is largely 1,8-cineole, which was among the weakest compounds in the Auburn ranking and no different from the control. Our full write-up on bay leaves covers what happens when people try it.
- Cucumber peels. No published research supports this at all. It appears to be folklore repeated until it sounded like a finding.
- Tea tree oil. Terpinen-4-ol was not in the Auburn panel, and we found no controlled cockroach repellency data for it. It is also toxic to cats and dogs if ingested or applied to skin.
- Citrus peels and lemon cleaners. Limonene was the weakest compound tested, barely above an untreated box.
- Coffee grounds. Frequently listed as both a repellent and a killer. We found no controlled data supporting either claim, and stale grounds are themselves an organic food source.
- Scented candles. A candle disperses a small amount of fragrance into open air, which is the least effective possible way to deliver a weak repellent to an insect living in a crack. We covered this in detail in what candle smell roaches hate.
- Mothballs. Do not. NPIC is unambiguous: mothballs may not be used for any purpose not described on the label, "besides, they have little or no effect as repellents." Using them in the open releases naphthalene or paradichlorobenzene vapor into living space, which NPIC links to headaches, nausea, eye and nose irritation, and, with longer exposure, liver and kidney damage. Off-label pesticide use is also a violation of federal law.
- Bleach and ammonia. Strong enough to bother anything, including you, and never to be mixed with each other or with acids. Neither is a cockroach control product.
The Smell That Actually Drives Roach Behavior
If you want to use cockroach olfaction against them, the interesting chemistry runs in the opposite direction.
German cockroaches aggregate around odors in their own droppings. A 2015 PNAS study by Wada-Katsumata and colleagues at NC State identified 40 volatile carboxylic acids in cockroach feces and showed that nymphs "strongly preferred the extract of control feces over the fecal extract of axenic cockroaches" — that is, feces from roaches raised without gut bacteria. A synthetic blend of six of those acids triggered aggregation on its own. The smell holding roaches in your cabinet is not your cooking. It is the accumulated droppings of the roaches already there.
That has a practical consequence. Cleaning droppings out of hinges, drawer runners, and appliance seams removes an aggregation signal, which is a more useful intervention than adding a repellent on top of it. The same logic explains how roaches locate resources in the first place, which we cover in how roaches find food.
What To Do Instead
Scent is a distraction from a sequence that reliably works.
- Find where they actually live. Sticky monitors in the corners of cabinets, under the sink, and behind the refrigerator will tell you within a few nights where the population is concentrated.
- Remove the real attractants. Grease films, crumbs in appliance seams, pet food left out overnight, and standing water in sink traps and drip pans. Our guide to what attracts roaches and how to remove each attractant walks through this room by room.
- Use a gel bait, and do not spray anything near it. UC IPM calls baits "the primary pesticides used to treat cockroach infestations," with results that "may take 7 days or longer." Many small placements beat a few large ones.
- Add an insect growth regulator if the population is established. It stops nymphs from maturing and prevents the rebound that follows a bait-only knockdown.
- Consider dusts for voids only. Boric acid and silica aerogel work in dry, undisturbed spaces behind and under appliances, not on open surfaces where people and pets contact them.
- Seal and exclude. Caulk gaps around pipe penetrations, fit door sweeps, and close the gap behind the sink cabinet.
The full sequence is laid out in our step-by-step plan for getting rid of roaches.
DIY or Call a Pro?
Handle it yourself if you are seeing a few roaches, the monitors show activity in one or two spots, and you own the space. Bait plus sanitation plus exclusion resolves most early German cockroach problems in three to six weeks.
Call a professional when you see roaches in daylight, find them in more than one room, live in a multi-unit building, or have run bait properly for a month with no decline. In an apartment, a unit-by-unit treatment is usually the only thing that works, because roaches move through shared walls and plumbing chases regardless of what you do inside your own door.
Frequently Asked Questions
Does peppermint oil actually repel roaches?
Weakly and briefly. Menthone, peppermint's main component, repelled 12.4% of German cockroaches on average over 14 days in the Auburn choice-box tests, with the effect concentrated in the first day or two. Whole mint oil is more toxic at high concentrations applied directly to an insect, but that has little bearing on a diluted spray on a baseboard.
What is the single most effective scent?
By measured repellency, citronellic acid, the citronella and lemongrass compound, at about 45% over 14 days. Catnip's E,Z-nepetalactone outperformed DEET in a separate choice-test study. Neither is sold as a tested cockroach control product, and neither will clear an infestation.
Do essential oils kill roaches or only repel them?
Concentrated oil applied directly to an insect can kill it. Research from the same Auburn lab found median lethal doses for corn mint oil of roughly 2.6% for American cockroaches and 3.8% for German cockroaches when applied to the insect's body. That is a laboratory dose on a roach you have already caught, which is a very different thing from a mist on a cabinet wall.
Will a strong smell keep roaches from moving in?
Not dependably. Prevention comes from removing food, water, and harborage and from sealing entry points. A scent that fades within days and that half the population ignores is not a barrier.
Are essential oil sprays safe around pets?
Not automatically. Concentrated oils can irritate skin and eyes, and several, including tea tree, are toxic to cats and dogs. NPIC notes that "no matter how natural, all pesticides should be used with care." EPA's minimum-risk exemption under FIFRA Section 25(b) means EPA did not review the product for safety or effectiveness, not that it was cleared.
Why did my roach problem get worse after I started spraying oils?
Most likely because you scattered them, contaminated your bait, or both. Repellents push roaches out of the harborage you know about and into ones you do not, and a scented surface near gel bait makes the bait unattractive. If you were baiting, stop spraying and give the bait a clean two weeks.
Can I smell a roach infestation?
Yes, at scale. A heavy infestation produces a musty, oily odor from droppings and secretions. That smell is a symptom worth investigating, not something to mask with fragrance.
Sources
- Phillips, A. K. (2009). *Toxicity and Repellency of Essential Oils to the German Cockroach*, Auburn University master's thesis — the 12-compound repellency ranking in Ebeling choice boxes and the harborage-choice method, the 3.89% untreated control, the 2-3 day decay, the corn mint oil lethal doses, and the warning that repellents can spread an infestation.
- Peterson, C. J., Nemetz, L. T., Jones, L. M., and Coats, J. R. (2002). Behavioral Activity of Catnip Essential Oil Components to the German Cockroach, *Journal of Economic Entomology* 95(2) — E,Z-nepetalactone significantly more active than equivalent doses of DEET, and the antennectomy result.
- Oladipupo, S. O., Hu, X. P., and Appel, A. G. (2022). Essential Oils in Urban Insect Management, *Journal of Economic Entomology* 115(5) — repellents dispersing cockroaches through a home, and the standardization gap created by registration exemption.
- UF/IFAS Extension: Assessment-Based Pest Management of German Cockroaches — why repellent products contaminate bait and make it ineffective.
- UC IPM Pest Notes: Cockroaches — baits as the primary pesticide approach, the 7-day-plus timeline, the caution that foggers and aerosols disperse rather than kill, and guidance on boric acid and desiccant dusts.
- Wada-Katsumata, A. et al. (2015). Gut bacteria mediate aggregation in the German cockroach, *PNAS* 112(51) — the 40 fecal volatile carboxylic acids, the preference for bacteria-bearing feces, and the synthetic six-acid aggregation blend.
- NPIC: Can I use mothballs to repel insects or animals? — label-only use, the statement that mothballs have little or no repellent effect, and the exposure symptoms.
- NPIC: Minimum Risk Pesticides and EPA: Conditions for Minimum Risk Pesticides — what the FIFRA 25(b) exemption does and does not mean.
