Subgrade stabilization with lime, cement, and geogrid

Subgrade stabilization, and why the bottom decides the job

Subgrade stabilization is the work of treating a weak subgrade so it can support the pavement, instead of pumping and rutting under it. The subgrade is the natural soil at the bottom of the section, below the aggregate base and below the asphalt or concrete. Every wheel load on the surface ends up there, spread out but still real, and if that soil moves, everything stacked on top of it moves with it.

There are a few ways to do it, and they split into two families. Chemical stabilization mixes a binder into the soil, lime for clay and cement for granular and silty material, so the soil itself gets drier, stiffer, and stronger. Mechanical stabilization puts a geogrid and clean aggregate over the soft soil to bridge and confine it, with no chemistry and no cure time. When neither fits, you undercut the bad soil and replace it with stone.

The reason any of this matters is simple and unforgiving. You cannot build a lasting pavement on a subgrade that moves. A mat that pumps at the bottom fails no matter how good the asphalt is on top. The failure starts at the bottom. This guide covers finding the weak subgrade and fixing it; building and compacting the aggregate base above it, and proof rolling for density, are covered in the base and subgrade compaction guide cross-linked below.

Why does a weak subgrade fail the pavement?

A weak subgrade fails the pavement from the bottom up, and by the time it shows on the surface the damage is already structural. The soil under the section deflects under each wheel pass. The base and the mat flex with it, and asphalt only flexes so many times before it cracks.

It shows up three ways, and they are all the same problem wearing different faces. Rutting is the wheel paths sinking, because the soil under them is consolidating or shoving sideways under load. Pumping is water and fines getting worked up through the section under traffic, so you see mud or gray slurry bleeding at cracks and joints after rain. Alligator cracking, the interconnected map of cracks in the wheel path, is fatigue, the mat bending over a base and subgrade that will not hold still.

No asphalt fixes a bad bottom. A thicker mat over a pumping subgrade buys a little time and then cracks in the same pattern, because the problem was never the mat. This is the whole argument for spending money below the surface. You are paying once to keep the soil still, instead of paying again and again to patch what the soil keeps breaking.

How do you find a bad subgrade?

You find a bad subgrade with a proof roll, and the proof roll is the test that catches what the lab numbers miss. Run a loaded tandem-axle dump truck, commonly in the range of a full legal load, slowly across the prepared subgrade at a walking pace and watch the ground under and ahead of the tires. The geotech or the inspector usually calls it. Soft spots show themselves as rutting, as visible deflection, or as pumping, where the ground flexes like a sponge and water works to the surface ahead of the wheel.

Learn to tell elastic rebound from real pumping. Ground that dishes under the load and springs right back can be acceptable. Ground that keeps deflecting, ruts, or pumps water is failing and gets flagged. Common practice treats sustained deflection past roughly an inch, lateral shoving, or any pumping as a fail, but the project sets the acceptance criteria, so go by the spec and what the geotech accepts.

Back the proof roll with soil data. The CBR, the California Bearing Ratio, rates how much the soil resists a punching load and feeds the pavement design. Atterberg limits and a sieve classify the soil and tell you whether it is plastic clay, silt, or granular, which is what decides the treatment. The proof roll finds the soft spot in the field. The lab tells you why it is soft and what will fix it.

The soil decides the method

The treatment follows the soil, so read the soil first. Clay is the plastic, fine-grained material that holds water, swells when wet, and shrinks and cracks when it dries. It is sticky, it has a high plasticity index, and it is the classic problem subgrade because it loses strength as it takes on water. Lime is the chemical answer for clay.

Silt sits in between. It is fine like clay but with little or no plasticity, so it does not swell much, but it has almost no cohesion and turns to soup when it is wet. Granular soil is sand and gravel, which drains and carries load well when it is confined, but has no binder of its own and can ravel and rut when it is loose or contaminated with fines.

Cement works across granular and silty soils and can handle low-plasticity material. The general rule the geotech leans on is the plasticity index: lime for the high-PI plastic clays, cement for the granular and low-plasticity soils, and a closer look or a blend for the in-between cases. Lime does not work well on soils with little clay in them, and cement struggles to mix uniformly into wet, sticky, high-plasticity clay, which is exactly why matching the binder to the soil is the first decision, not an afterthought.

Lime for wet plastic clay

Lime is the treatment for wet, plastic clay, and it works in two stages. First it dries and modifies the soil. Quicklime or hydrated lime mixed into wet clay takes up water and triggers cation exchange and flocculation, which clumps the fine particles together. The immediate result is a soil you can work: lower moisture, much lower plasticity, and a friable texture instead of a sticky one. That alone can turn an unworkable clay into a stable platform.

Then comes the slow part. As long as the soil stays at a high pH, the lime and the clay keep reacting in the pozzolanic reaction, forming cementitious bonds that gain strength over weeks and months. That long-term strength is what makes lime stabilization, not just lime modification. The difference is dose and intent, and the lab decides which one you are buying.

Lime is the wrong tool on a soil with little clay in it, because there is nothing for the pozzolanic reaction to feed on. It is the right tool, and often the only practical one, when the subgrade is a fat, wet clay that no amount of rolling will compact. Quicklime and hydrated lime behave differently in handling and dust, and the rate, the lime type, the mellow time, and the cure all come from the mix design and the spec, not from a rule of thumb.

Cement for granular and silty soil

Cement is the treatment for granular and silty soils, and where lime modifies clay, cement mostly adds strength. Portland cement mixed into a sandy or silty soil hydrates and cements the particles into a hard, slab-like layer. Done at a base-course rate and thickness it becomes cement-treated base, a stiff structural layer in its own right. Done lighter it stiffens a marginal subgrade enough to build on.

Cement gains strength faster than lime. The hydration starts in hours, not days, so the working window is short and the cure is quicker, which can matter on a tight schedule. It also handles a wider range of soils than lime, including the low-plasticity silts and sands that lime cannot help.

The catch with cement is cracking. A cement-treated layer is rigid, and a rigid layer shrinks as it cures and cracks under traffic if it is over-cemented or too thick. Those shrinkage cracks can reflect up through a thin asphalt mat. The fix is not to guess high on the rate. You hit the strength the design calls for and no more, because more cement is more strength and also more cracking. Cement also fights wet, plastic clay, where it will ball up instead of mixing, which is the case where the geotech sends you to lime instead.

Fly ash, kiln dust, and proprietary additives

Lime and cement are not the only binders, and the others show up where they are cheap, local, or written into the spec. Fly ash, a byproduct of coal combustion, has pozzolanic activity and is used on its own or blended with lime or cement to stabilize granular and silty soils, sometimes at higher percentages than cement alone. Cement kiln dust, a byproduct of cement manufacturing, carries free lime and works as a lower-cost stabilizer on a range of soils.

Both fly ash and kiln dust vary in chemistry from source to source, so a mix design on the actual material is not optional. A rate that worked with one batch of fly ash can underperform with the next. The lab confirms the strength gain with the specific stockpile you will use.

Proprietary liquid additives and enzyme products get marketed for subgrade work, and some have a place. Treat them with more caution than the established binders. The performance claims are not always backed by the kind of long-term field data lime and cement have, so do not specify one over a geotech recommendation, and require a mix design and acceptance testing the same as for any other treatment. If the spec names a product, follow it; if a salesman names one, send it to the geotech.

How much additive do you add?

You add the percentage the lab says, not the percentage that sounds about right. The mix design is a laboratory determination of the additive rate for the specific soil on the project, and skipping it is the most expensive shortcut on a stabilization job. Guess low and the soil never reaches strength. Guess high and you waste binder and, with cement, invite cracking.

For lime, the lab runs the soil through a procedure that finds the rate holding the soil at the high pH the reaction needs, commonly the ASTM D6276 pH method, then confirms the result with plasticity and strength testing. The target is a measured drop in plasticity index and a strength gain, not a round number. For cement and the other binders, the lab molds samples at several rates, cures them, and breaks them to find the rate that meets the design strength.

The deliverables you build to are an additive percentage, a treatment depth, a moisture target, and a strength or plasticity acceptance value. Those numbers belong to the geotech report and the project specification, and they change with the soil, so a rate from another job is a starting guess at best. Get the mix design on the soil that is actually on the site, and build to it.

The stabilization process, step by step

In-place chemical stabilization runs in a set order, and each step has a way it goes wrong. First the additive is spread across the prepared subgrade, by a spreader truck for dry lime or cement or as a slurry to cut dust, at the rate per square yard the mix design sets. Uneven spread is uneven treatment, so the spread rate is checked, often with pans or by load tracking.

Then a reclaimer or a rotary mixer cuts the binder down into the soil to the full design depth and pulverizes it, while water trucks bring the moisture to the target the mix design calls for. The reaction needs water, and the compaction needs the soil near its optimum moisture, so the water is metered, not eyeballed. The mixing has to reach the design depth uniformly; a shallow or streaky pass leaves untreated soil under a treated crust.

After mixing comes compaction, with a padfoot roller working from the bottom up and a smooth drum finishing the surface, to the density the spec requires. Then the layer cures, kept moist or sealed so it does not dry out and lose the reaction, for the time the spec sets. Lime-treated plastic clay adds a mellowing step between mixing and final compaction, covered next, because the clay needs time to react before you seal it up.

Why lime-treated clay has to mellow

Mellowing is the rest period between the first mixing and the final mixing and compaction on lime-treated plastic clay, and it is the step that separates a clay job that works from one that does not. After the lime is mixed in, the clay needs time for the lime to penetrate the clods and break down the plasticity all the way through, not just on the surface of each lump.

Common practice gives plastic clay a mellow period in the range of 24 to 72 hours, kept moist, before a second mixing pass and final compaction. The wetter and more plastic the clay, the longer it tends to need. The mix design and the spec set the actual time, and some specs let you shorten it if testing shows the plasticity index has already dropped below the target.

Skip the mellow and you compact clods that are only treated on the outside. They look fine going down and then break apart and soften from the inside later, and the layer never reaches the strength the design assumed. Cement does not get this step; it sets too fast to wait, which is part of why cement and lime are not interchangeable even when both could chemically work.

Geogrid bridges the weak spot, no chemistry

Geogrid is mechanical stabilization, not chemical, and that is the whole point of it. A geogrid is a stiff polymer grid laid on the subgrade with aggregate compacted into and over it. The aggregate locks into the grid apertures and cannot spread sideways, so the stone layer behaves stiffer and spreads load wider than the same stone with no grid. It confines the aggregate, it helps separate the stone from the soft soil, and it bridges across weak spots the way reinforcement carries a load across a gap.

The advantage over chemical treatment is time and weather. There is no binder to mix, no moisture window, no mellowing, and no cure. You roll out the grid, place and compact the stone, and build on it the same day. On a soft, wet site where you cannot get a chemical reaction to behave, or where the schedule has no room for a cure, geogrid is often the move.

Geogrid does not turn bad soil into good soil; it lets a stone section work over soil that would otherwise pump and rut. It is commonly used to build a working platform over very soft ground, and it has bridged over soils as poor as peat. The grid type, the aperture, the aggregate, and the layer thickness come from the geotech design, because the grid and the stone are engineered together, not picked off a shelf.

Separation geotextile stops the mud from pumping up

A separation geotextile is a fabric laid between a soft, fine subgrade and the clean aggregate above it, and its job is to keep the two from mixing. On a wet, fine-grained subgrade with no fabric, the stone presses down into the mud under traffic while the mud pumps up into the stone. The clean base fouls with fines, loses its strength and its drainage, and the section fails even though you placed good stone.

The fabric stops that two-way contamination. The base stays clean, the fines stay down, and the stone keeps the strength you paid for. On very soft ground a separation geotextile is often paired with a geogrid, the fabric handling separation and filtration while the grid handles confinement and reinforcement, because they do different jobs.

Pick the fabric for the function. A separation and filtration job wants the right fabric class and opening size for the subgrade fines, which the geotech specifies. A heavy nonwoven that filters well is not the same product as a stiff geogrid that reinforces, and using one where the design called for the other is a common and costly substitution.

Undercut and replace when treatment will not work

Sometimes the answer is to dig the bad soil out and replace it with stone. Undercutting means excavating the soft or wet subgrade down to firm material and backfilling with compacted aggregate, building a stone section where the weak soil used to be. It is the direct fix, and on a small soft spot or a saturated area that will not dry, it is often faster and more certain than chemistry.

Undercut earns its place when treatment is not practical. A localized pocket of soft soil in an otherwise good subgrade is cheaper to dig out than to mobilize a reclaimer for. A subgrade so wet it will not hold a moisture target fights both lime and cement. Buried organics, debris, or an old fill that will keep settling cannot be stabilized in place and have to come out.

The trade-off is haul and material. Undercut means trucking the bad soil off and trucking stone in, which is double the trucking and a disposal cost, and the deeper and wider the undercut, the faster that bill climbs. On a large soft area, that is exactly the math that pushes the decision toward in-place stabilization. A separation fabric in the bottom of an undercut keeps the new stone from pumping back down into whatever soft soil remains below.

Lime, cement, geogrid, or undercut: which one?

The choice comes down to the soil and the cost, and the geotech report usually makes the call before you get to the field. Match the method to what the soil is and what the site will let you do. Lime is for wet plastic clay. Cement is for granular and silty soil. Geogrid is for bridging soft ground fast with no cure time. Undercut is for small, wet, or fouled areas where treatment will not take.

The table below is the quick sort, not the spec. The actual method, rate, depth, and acceptance live in the geotechnical report and the project documents, and on a real site the answer is often a combination: lime a clay platform, then geogrid the softest pockets, then undercut the spots that fail anyway. Cost and schedule break the ties. In-place stabilization usually wins on a large area; undercut wins on a small one.

Lime is caustic: protect the crew

Lime is caustic, and it burns. Quicklime and hydrated lime are strongly alkaline, and on damp skin or in the eyes they cause chemical burns, not just irritation. Wet lime against the skin under a glove or a boot top can burn before anyone notices, because there is no immediate sting like an acid. The crew handling lime wears eye protection, gloves, long sleeves, and respiratory protection per the product's safety data sheet, and there is water on site to flush exposure right away.

Dry lime and dry cement are also a dust problem. The dust is an inhalation and eye hazard, and a fugitive dust nuisance that drifts off site and onto traffic and neighbors. Crews control it by spreading on calm days, slurrying the lime instead of spreading it dry where the site demands it, and keeping people upwind. The spreader operator and the ground crew are the ones at risk, so the PPE is not optional.

Quicklime deserves extra respect, because it generates heat when it hits water. It can heat up against wet skin and in confined storage. Handle and store it to the manufacturer's instructions, and treat any lime burn as a burn, with flushing and medical attention, not as ordinary dust on the skin.

Testing and QC: depth, rate, uniformity, strength

Quality control on a stabilized layer checks four things, because a stabilized layer can look finished and still be wrong underneath. Depth: the binder has to be mixed to the full design depth, checked by digging or coring the treated layer, because a treated crust over untreated soil fails like untreated soil. Rate: the additive went down at the design percentage, verified from spread checks or load tracking, since a lean spot is a weak spot.

Uniformity and moisture: the mix is even across the width and at the moisture the design wants, because a streaky mix leaves soft lanes. Compaction: the layer hits the density the spec requires, by the same nuclear gauge or sand cone methods used on any compacted layer. Strength is the proof. For cement and many lime jobs the lab molds and breaks samples for unconfined compressive strength, the UCS, to confirm the layer reached the design value after curing.

The acceptance numbers, the depth tolerance, the rate, the density, and the UCS or plasticity target are all set by the geotech report and the spec, and they vary with the soil and the design. Document each one as you go. A stabilization layer that is buried under base and pavement cannot be inspected later, so the QC record made during construction is the only evidence the bottom was built right.

Proof roll it again before the base

After the subgrade is treated and cured, you proof roll it again, and this re-proof roll is the hold point that everything below the surface comes down to. The first proof roll found the problem. This one confirms the fix. Run the loaded truck across the stabilized subgrade the same way and watch for the same fail cues: rutting, deflection, and pumping. There should be none.

If the treated subgrade still pumps or ruts under the truck, it is not ready for base, full stop, no matter what the calendar says. A spot that still moves gets reworked, re-treated, or undercut before any stone goes over it. Sealing a moving subgrade under a base and a mat just buries the failure where you cannot reach it.

Once the re-proof roll passes and the density and strength tests confirm it, the subgrade is a platform you can build on, and the aggregate base goes down. The proof roll on the base itself, the Proctor density on the stone, and the compaction record before paving are the next stage, covered in the base and subgrade compaction guide cross-linked below. The handoff between the two is this passing re-proof roll.

Follow the geotech report and the spec

The geotechnical report is the authority on a stabilization job, and it specifies the treatment for a reason. The geotech bored the site, classified the soil, ran the strength tests, and chose the method, the additive, the rate, and the depth to suit what is actually in the ground. That report is the engineering behind every number you build to, and the field changes the engineer's design at the field's peril.

When the soil you uncover does not match the report, that is not a license to improvise; it is a reason to call the geotech. Conditions vary between borings, and the engineer would rather adjust the design than find out later that the field guessed. The same goes for the rate, the depth, and the cure: when the spec and a habit disagree, the spec wins, and when the spec is silent, the geotech fills the gap.

Everything in this guide hedges back to those two documents on purpose. The rates, the depths, the mellow time, and the cure are written by topic here because they belong to the project, not to a rule of thumb. Identify the weak subgrade with a proof roll, match lime to clay and cement to granular soil by a lab mix design, build to the geotech report, and re-proof roll before the base. Those four hold across jobs. The numbers do not.

Cost and schedule: stabilize in place vs haul out and in

In-place stabilization is often cheaper than digging out and replacing, and the reason is trucks. Undercut and replace means hauling the bad soil off to a disposal site and hauling clean stone back in, two truckloads of trucking for every load of material, plus the disposal fee and the cost of the imported stone. On a large soft area, that bill is large and it is mostly hauling.

Mixing a binder into the soil that is already there skips both haul legs. You bring in the lime or cement, which is a fraction of the tonnage of replacing the soil with stone, mix it in place, and the weak soil becomes the structure instead of going to a landfill. On a big footprint that is a real saving, which is why chemical stabilization tends to win on area and undercut wins on small, isolated spots.

Schedule cuts the other way and has to be in the math. Lime-treated clay needs mellow and cure time that undercut and geogrid do not, so a treatment that is cheaper in dollars can be slower on the calendar. On a tight schedule, geogrid or undercut can beat lime even at a higher material cost, because they let you build the same day. Run both numbers, dollars and days, and let the geotech-approved options compete on the real constraint.

Common mistakes

• Building on a pumping subgrade and trusting the asphalt to carry it.
• Using the wrong additive for the soil, such as cement on wet plastic clay where it balls up instead of lime.
• Skipping the lab mix design and guessing the additive rate from another job.
• Giving lime-treated clay no mellow time, then compacting clods treated only on the outside.
• Mixing the binder short of the design depth, leaving untreated soil under a treated crust.
• Not re-proof-rolling the treated subgrade before placing the aggregate base.
• Ignoring the geotech report and improvising the method, rate, or depth in the field.
• Placing clean stone on a soft fine subgrade with no separation fabric, so the mud pumps up and fouls the base.

Field checklist

What to document

A stabilized subgrade gets buried under base and pavement, so the record made during construction is the only proof the bottom was built right. Capture what the geotech specified and what you actually did, side by side, so a reviewer can see the treatment matched the design.

Record the soil classification and the geotech method, the additive type and rate, the treatment depth, the moisture target and what you achieved, the mellow time for lime, the cure time, the compaction results, and the strength or plasticity acceptance test. Record the first proof roll, what it found, and the passing re-proof roll after treatment. Note who accepted each hold point. If the field deviated from the report, record the geotech approval that allowed it.

Standards and references

The governing document on any specific job is the project geotechnical report and the specification, because they are written for the soil that is actually on site. They set the method, the additive, the rate, the depth, the mellow and cure times, and the acceptance criteria, and they override any general figure in this guide.

Behind them sit the test methods and the agency specs. ASTM and AASHTO cover the soil classification, the Atterberg limits that flag plastic clay, the CBR for bearing strength, the moisture-density relationships for compaction, and the unconfined compressive strength used to accept a treated layer. For lime, ASTM D6276 is the pH method commonly used to estimate the lime proportion for stabilization. The state DOT specifications and the FHWA geotechnical guidance carry the construction requirements for in-place chemical stabilization and for geosynthetic subgrade stabilization.

Industry references from the lime and cement associations give practical construction manuals for lime-treated and cement-treated soils, and the geogrid and geotextile manufacturers publish design guidance for their own products. Cite them as guidance, confirm the current edition, and let the geotech report and the project spec control. The exact rates, depths, and cure times belong to the project, so verify them there before you build.

Units, terms, and conversions

Subgrade stabilization carries a vocabulary that shifts between the geotech report, the spec, and the supplier, so the same idea reads a few different ways across a project.

Additive rate appears as a percentage by dry soil weight and also as pounds per square yard for a given depth, which is how a spreader is set. Treatment depth is in inches. Strength shows up as unconfined compressive strength in psi or kPa, and bearing as CBR in percent. Plasticity is the plasticity index, a unitless number from the Atterberg limits. Compaction is a percentage of a Proctor maximum dry density at a target moisture content.

Subgrade

The natural soil at the bottom of the section that carries the base and pavement above it

Proof roll

Driving a loaded truck over the subgrade to find soft spots by rutting, deflection, or pumping

Pumping

Water and fines worked up through the section under traffic, a sign the subgrade is moving

Plasticity index (PI)

The range of moisture over which a soil is plastic; high PI means a plastic clay

Mellowing

The rest period after first mixing that lets lime react through plastic clay before final compaction

Pozzolanic reaction

The slow lime-clay reaction at high pH that forms cementitious bonds and long-term strength

UCS

Unconfined compressive strength, the lab value used to accept a treated layer

CBR

California Bearing Ratio, a measure of subgrade bearing strength used in pavement design

Geogrid

A stiff polymer grid that confines aggregate and bridges soft soil, a mechanical not chemical fix

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is proof rolling?
• What is crusher run or dense-graded aggregate?
• What type of retaining wall is best?
• When does a retaining wall need an engineer?
• How thick should asphalt be?