Leave-in or gel: what curly hair actually needs

Does healthy hair need a leave-in?

It doesn't need one, and it doesn't need to skip one. Nobody has tested it either way.

We used to say healthy hair should skip a leave-in. We were reading one study further than it goes, so here is that study and where it stops.

Hair carries a net negative surface charge in normal use. Conditioning agents are cationic, meaning positively charged, so they are electrostatically attracted to that surface and adsorb onto it. Tham and colleagues measured this directly using streaming potential analysis, tracking how four common conditioning surfactants deposit onto hair and how that changes the fiber's zeta potential. Healthy hair measured a baseline zeta potential of about −29.2 mV. Bleached hair measured about −0.9 mV.

Rinsing does not take all of it back off. The hair was rinsed before it was measured, and its surface charge was still changed, so a portion of that layer is still there. That is what "substantive" means in formulation terms. How much stays is not the same for every conditioner: the long-chain ones held on best, and one of the four tested barely held on at all. So this is a property of the ingredient, not a promise printed on the category.

It is tempting to read that forward: the conditioner is still there, so a leave-in is just a second dose. That is what we used to say. But the study tested rinse-out conditioner and never tested a leave-in. And a leave-in is not always just more conditioner. Some carry a film-forming polymer, the same kind of ingredient that makes a gel hold. Nobody has tested a leave-in against no leave-in on healthy hair. So there is no measured reason to add one, and no measured reason to drop one.

Three things are worth knowing about that study. It was paid for by a government science agency in Singapore, not by a brand, and the authors declare no competing interests. But two of the five work in industry, at the company that makes the instrument and at an ingredient supplier, so it is not a fully outside result either. The deposition ranking it finds is clean, though the drop in combing force is shown only as a graph, so no precise percentage should be quoted from it. And the combing was done on wet hair. The paper says plainly that dry hair and leave-in products were outside what it set out to test.

How does a gel actually work?

Not by slowing drying. That explanation is common and it is wrong.

Water breaks hydrogen bonds in the hair fiber. Broken hydrogen bonds are what make wet hair reshapeable. As the hair dries, those bonds reform, and they reform in whatever shape the hair is being held in at that moment. That is the entire mechanism behind wet setting, and it is why what you do while hair is drying matters more than what you do after.

A gel's film-forming polymers do two jobs against that process. First, they hold the curl clump physically in shape while the hair dries, so the hydrogen bonds reset defined rather than separated and frizzy. That is the cast. Second, once dry, the film sits over the set as a barrier that resists ambient humidity pulling the shape back apart. That is the hold. Scrunching the dried cast breaks the brittle film into a soft, defined result without undoing the set underneath.

Most leave-ins don't do that. They condition: they add slip and cut friction, but they don't form a film that holds a clump. Some do, though. If a leave-in lists a film-forming polymer, like VP/VA copolymer, an acrylates copolymer or polyquaternium-69, it can do part of a gel's job. So read the ingredients, not the name on the bottle. If your curls come out soft and undefined with a leave-in alone, the missing piece is usually a film-former, not more moisture.

What do the numbers say about gel performance?

Less than the bottles suggest. The measuring stops much earlier than you would guess.

Start with what nobody has done. No one has measured how long a gel holds past one day. The standard lab test hangs a curled swatch of hair in a humid cabinet and reads it for 24 hours, then ends. Every run longer than that we could find sits inside a patent, owned by the company that filed it. So when you see a number at 24 hours, that is where the test stopped, not where the hold stopped.

The second thing worth knowing is what the industry privately calls good. In its own patent filings, keeping 70% of a curl for 45 minutes is rated "good", and holding above 70% for three hours is rated "very good to excellent". Three hours. Meanwhile bottles on the shelf promise three and four days, and none of those promises name a study.

Against that, here are the two numbers we found, with their limits attached. Sankar and colleagues ran a dose-response on an acrylates crosspolymer gel: stiffness rose with polymer concentration, 6.75 N at 1.5% to 12.42 N at 4.5%, with curl retention of 81 to 86% at 90% relative humidity. The paper does not say for how many hours. Rigoletto and colleagues reported a polyquaternium-69 gel above 90% curl retention at the 24 hour mark in 90% relative humidity, where a commercial PVP and VP/VA gel began dropping after about an hour.

Read those as rankings, not as durations. Both come from industry, neither carries a DOI, and the test itself is gentler than daily life: the hair is rolled so it shelters itself, and nothing in the cabinet ever moves, sweats, or gets slept on. None of this means gels stop working after a day. People get multi-day hold constantly. It means the number on the bottle is a claim, and nobody has published the measurement behind it.

The practical read: PVP on its own is hygroscopic, meaning it absorbs atmospheric moisture and can make frizz worse in humidity, which is the opposite of what a hold ingredient should do. VP/VA copolymer, acrylates copolymers, and polyquaterniums hold up.

What is the difference between cast and hold?

They are independent variables and almost every product review confuses them.

Cast is the rigidity of the dried film on the hair surface. It is what locks curl clumps in place while the hair dries, so it is what produces definition. Hold is how well the style survives afterwards, particularly against humid air. Hold is measured in hours and sold in days.

They move separately. A gel can have strong humidity hold with almost no cast: polyquaternium-69 forms a soft, flexible film, which means no crunch to scrunch out, but it is not a cast former. A gel can have strong cast with only moderate humidity durability: VP/VA forms the rigid shell that locks the clumps during drying but has moderate humidity resistance on its own.

So the useful question is which one you actually need. If your curls or waves lack definition, you need cast. If your style looks right in the morning and collapses by afternoon in humid air, you need hold. If you need both, look for a formula listing VP/VA alongside a polyquaternium, which is the most cost effective upgrade in this category. "Use a gel with hold" is incomplete advice.

Can a leave-in and a gel clash?

In a beaker, yes. On hair, nobody has tested it, and a clash can even help.

Most leave-ins carry a positive charge, because that is how conditioning agents stick to hair. Many gels get their thickness from a polymer with a negative charge, like a carbomer or xanthan. Opposite charges can lock together, and in one lab test a positively charged guar and xanthan did exactly that in solution and dropped out as a dense complex. That is where the advice to never layer the two comes from.

It is not a hair result. That test was a drug carrier in a beaker with no hair in it. The closest thing to a hair test runs the other way: a company deliberately made a complex of two opposite polymers and used it to mend split ends on hair tresses. Nobody has put a named leave-in and an unchanged gel on wet hair, dried it and counted the white bits. So a clash is real chemistry and an unproven hair problem.

Our test for it is the one we give clients. Rub a little of each product together in your palm with a few drops of water, and if you see white bits, try another pair. If it looks smooth, put both on one small wet section, let it dry all the way, then scrunch it and look. A smooth palm does not prove the dry hair will be fine, because a gel can leave white bits on its own when its dried film breaks. That is practice, not a measured rule.

How do you pick a leave-in?

By what is in it and how your hair responds, not by whether your hair is healthy or damaged.

Start with the job you want done. For slip and easier detangling, look for conditioning ingredients: a cationic like behentrimonium or cetrimonium chloride, or a polyquaternium used for slip. For help holding the curl, look for a film-forming polymer. If weight is the problem, skip the ones with butters or heavy oils near the top of the list. Damaged or processed hair often gets more out of the slip, because rough hair tangles more. That is a reason to try one, not a rule.

If you use one with a gel, the leave-in goes on first and the gel on top. The gel sets the shape. Most leave-ins don't.

And if a leave-in worked once and then seemed to stop, don't blame it straight away. Check three things first. Have you been clarifying, or is product building up? Did you change how you dry, like air drying instead of diffusing? Did you get a haircut? Any one of those can change the result on its own.

You will also hear that the humectants in a leave-in pull water out of humid air and make frizz worse. We are not going to tell you that, because nobody has actually tested it. Humectants do hold water, and hair does swell as it takes on water. But no study has checked whether a humectant makes real hair frizz more than bare hair sitting in the same air. The papers people cite for it measured glycerin in a dish, with no hair in it. There is a second problem with the advice: a gel needs a humectant to keep its film flexible, or the film dries brittle and flakes. So you cannot dodge humectants by swapping a leave-in for a gel anyway. Watch how your own hair responds, and treat this one as an open question rather than a rule.