How long does weed killer stay in the soil?

How long after weed killer is it safe to plant?

That is the question underneath the search, so here it is straight off the labels rather than off somebody's blog. Roundup Ultra, a 360 g/L glyphosate concentrate, prints a short replanting table on the back. A light cover of annual weeds, and replanting or sowing can commence the day after treatment. A dense cover including perennial weeds, and you leave at least 7 days for the weedkiller to move to underground roots and shoots before digging or cultivating the soil. Read the reason in that second line, because it has nothing to do with safety. The wait is there so you do not chop up a plant that is still busy carrying the dose down to its own roots. Turn the bed over on day two and you have interrupted the kill, not poisoned the ground.

The ready-to-use bottles say the same thing in plainer words. Roundup Pump 'n Go Fast Action (MAPP 14481, 7.2 g/L glyphosate acid) claims on the front that it degrades in soil to allow replanting, and explains inside that it is inactivated on contact with the soil and is then broken down by soil micro-organisms. Same instruction as the concentrate: leave the weeds 7 days before digging or cultivating, so the treatment can reach the roots. Then the residual sits an arm's length away on the same shelf and says six months. Nothing on the front of either bottle tells you which of the two you are holding.

The half-life is not the same number as the replanting interval

Here are the actual persistence figures, from the University of Hertfordshire's Pesticide Properties DataBase, which is the reference set European regulators work from. It records soil persistence as DT50, the time for half of what you applied to break down. Diflufenican, the residual half of a path and drive product, is 92 days in the lab and 113.7 days in the field, classed moderately persistent and persistent. Glyphosate is 17.3 days typical and 6.45 days in the field, classed non-persistent. Pelargonic acid, the fatty acid in the natural-looking bottles, is 1.22 days. Maleic hydrazide is 0.8 days. That ordering is why one bottle says plant tomorrow and another says wait half a year.

But now put those half-lives against the label intervals and notice that they do not match, in either direction. Glyphosate's half-life is a fortnight and its label lets you plant the next day. Diflufenican's half-life is three months and its label wants six. Neither interval is calculated from the half-life, because the two numbers answer different questions. A half-life tells you when the chemical will be gone. A label interval tells you when the next thing you plant will stop being damaged. For a systemic that binds on contact those two land miles apart in your favour, and for a residual they land miles apart against you, because a residual only has to be present in the top few centimetres in tiny amounts to stop a seed establishing. That is the entire job it was sold to do.

Why glyphosate stops working the second it touches the ground

Because it sticks, and the database puts a number on how hard. Glyphosate has a Koc of 1,424 and a Freundlich figure of 4,348, which in plain terms means it clings tightly to clay and organic matter rather than drifting about in the water between the particles. Bound up like that it cannot be drawn in by a root, and soil bacteria then break it down. That is what the Roundup wording means by inactivated on contact with the soil. It is not that the chemical vanishes on impact, it is that it is stuck fast and no longer available to a plant. It is also why glyphosate has to go on the leaf and can never be watered into the ground: put it on soil deliberately and you have wasted the whole lot.

That mechanism settles a good half of the worries people arrive with. A weedkiller that binds on contact is not going to travel sideways through the soil to the shrub two feet away, and it is not lying in wait for the bedding plants going in next month. If a border died after you sprayed the drive, the cause was spray drift onto the leaves on the day, or run-off across the surface, and not something seeping through the ground afterwards.

The part the half-life does not tell you

A half-life is only the first half. Glyphosate's DT90, the time for 90 per cent of it to go, is 322 days in the lab, and the same database classes that figure as persistent. Half of it is gone in a fortnight and the last tenth can take the best part of a year. Neither fact contradicts the other, and quoting only the friendly one is how the marketing gets away with it.

There is a second omission worth knowing about, because glyphosate does not disappear so much as turn into something else. Its major breakdown product in soil is AMPA, aminomethylphosphonic acid, and AMPA is by far the more durable of the pair: a typical soil half-life of 114 days, 419 days in the field, classed persistent and very persistent respectively. A product sold on how quickly it breaks down leaves behind a residue that outlasts it several times over.

I am not going to turn that into a scare story, because the same database records AMPA as a metabolite with a mode of action of not applicable. It is not a herbicide. It is not what stops anything growing, and it is not the reason a replanting interval exists. Plant into it and your plants will be fine. But if what you genuinely asked was how long weedkiller stays in the soil, rather than when you can plant, then the honest answer is a good deal longer than the half-life on its own suggests, and nobody selling you the bottle is going to volunteer that.

The residual is the one that costs you a border

Everything above is about a plain glyphosate. Put a residual in the tank and the arithmetic changes completely. The Weedol Pathclear label (MAPP 16524, 125 g/L glyphosate plus 20 g/L diflufenican) reads, word for word: if an area needs to be cultivated allow at least 6 months between treatment and replanting, and dig the soil well to reduce risk of damage to subsequent planting. Then, in the very next sentence: where this or other products containing diflufenican are applied in successive years, diflufenican will build up in the soil, and there is a risk of damage to subsequent plantings even if the soil is well dug. Read that twice. Digging is offered as the fix and then withdrawn as a guarantee one line later.

The number I would actually plan around is not the six months, though. It is buried further down the same label in the paragraph about getting rid of leftovers, which tells you to pour unwanted dilute product onto a waste area away from drains and gutters which is not going to be used to grow desirable plants within 2 years. Two years. That is the manufacturer's own figure for how long a patch of ground that has taken a dose of this should be treated as off limits for planting, and it is four times the replanting interval printed higher up on the same piece of card. Set that against the promise on the front of the box, which is that it prevents new weeds for up to 3 months by creating an invisible barrier, and the trade you are actually being offered comes into focus: three months of protection bought with a six-month planting ban and a two-year one on any ground you tip the dregs on.

For all that, diflufenican is a rarity rather than a fixture of the aisle. HSE's register of amateur herbicides approved for Great Britain runs to roughly 331 distinct products, and diflufenican appears in four of them. So the residual problem is not lurking in every bottle in the aisle. It is one small family of path and drive products, and the awkward part is that they are the ones marketed at exactly the surface most people are out there spraying.

Can you dig in or compost the weeds you sprayed?

With a glyphosate, yes, and the label says so outright rather than leaving you to guess: dead or dying weeds can be dug in or composted without harming future plants. Give perennials their seven days first, for the reason above, then treat the dead growth as garden waste like anything else.

The real soil-residue horror story in British gardening belongs to a different family of weedkiller altogether, and it is worth separating the two so you can stop worrying about the wrong one. Clopyralid and aminopyralid are hormone-type products that survive being eaten by an animal, composted, and rotted down, and the RHS warns that they come through green-waste composting and turn up in compost and manure sold to gardeners, where they distort and kill tomatoes, beans and potatoes. Nothing on a glyphosate bottle behaves like that. If you have read somewhere that weedkiller lingers in soil for years and wrecks the next crop, that is the story you read, and it is about a chemistry that was never on your drive.

So the check before you plant is a question, not a waiting period. Was there a residual in the bottle? Turn it over and look down the ingredients for diflufenican. If it is there, you are on the label's six months and I would treat six months as the floor rather than the target. If it is not there and you sprayed straight glyphosate, the ground is ready when the weeds are dead, and you can plant into it with a clear conscience.