Asphalt compaction window: temperature, rolling, and density

The compaction window, and why it closes

The compaction window is the time and temperature range in which you can roll a hot-mix asphalt mat to target density. It opens when the mat comes off the screed hot and workable. It closes when the mix cools to the point that the aggregate locks up and the roller stops moving it, no matter how many passes you make.

Density is not added cold. It is built while the binder is fluid enough to let the stones slide past each other into a tighter arrangement. Once the binder stiffens, the skeleton is set. You can run the roller until the cows come home and you will polish the surface without closing another air void.

So the whole job is a race against the mat cooling. Lay it hot, get the rollers tight behind the paver, and put the passes on while the heat is still in the mat. Everything else in this guide is about how much time that race actually gives you, what steals it, and how to know you won before the cores come back.

Why does asphalt have to be compacted while it's hot?

Asphalt has to be compacted hot because density comes from the aggregate rearranging, and the aggregate only moves while the binder is soft. The binder is a thermoplastic. Hot, it flows and acts as a lubricant that lets the stones shift under the roller into a denser packing. As it cools it stiffens, grips the aggregate, and freezes the structure in place.

Roll it at the right temperature and each pass drives air voids out and locks the stones tighter. Roll it too late and you are working against a skeleton that has already set. The drum rides the surface, you get roller marks and maybe a little tightening of the very top, and the air voids deep in the lift stay where they were. That mat reads low on the core and there is no fixing it after the fact.

The failure is delayed and it is expensive. A mat left a couple of points low on density has interconnected air voids, so water gets in, the binder oxidizes, and the surface ravels and cracks years before it should. Nobody sees it the day you pave. They see it on the warranty.

What temperature do you stop rolling asphalt?

You stop rolling for density when the mat cools to its cessation temperature, the point below which the roller no longer reduces air voids. For many dense-graded HMA mixes that floor is commonly cited near 175 to 185°F, but it is not a universal number. The binder grade, the mix design, and the agency spec move it, and polymer-modified and stiffer binders can carry a higher stop temperature.

The window has a top end too. Breakdown rolling, the first and most important pass set, should start while the mat is still hot, commonly above roughly 280°F behind the screed for typical dense-graded mixes, again depending on the mix. Most of your density is won here, in the hot zone, right behind the paver. Wait and you give the heat away.

Treat these as ranges to confirm, not gospel. The reliable move is to pull the numbers for the specific mix from the supplier and the spec, then verify with a temperature check on the mat, not a guess from the truck ticket. Below cessation you can still run a finish roller to take out marks, but understand you are smoothing, not compacting.

What shrinks the compaction window?

What shrinks the compaction window is anything that pulls heat out of the mat faster. The biggest one is lift thickness. A thin lift has little mass and a lot of surface, so it sheds heat fast, and a 1 to 1.5 in surface course can give you a fraction of the working time a 3 in lift does. Thick lifts hold heat and buy you minutes.

Then comes everything the weather does. A cold base or existing pavement pulls heat out the bottom of the mat the instant it is placed. Cold ambient air, and especially wind, strip heat off the top. Wind is the sleeper. A 15 to 20 mph wind on a cool day can close the window faster than the air temperature alone would suggest, because it scrubs the still layer off the surface.

Delivery temperature is the input you can lose before the mat ever hits the ground. A load that sat in traffic, ran uncovered, or left the plant cool starts the clock with less heat in the bank. Check the mat temperature behind the screed, not just the truck ticket, because that is the temperature your window actually starts from.

PaveCool: putting a clock on the mat

PaveCool is a free cooling-rate model from the Minnesota DOT and the University of Minnesota, published through the FHWA, that estimates how long you have to compact before the mat reaches its stop temperature. You enter the date and time, the location, the mix and lift thickness, the delivery temperature, and the conditions: base temperature, air temperature, wind, and sky. It runs a heat-flow model and tells you the minutes available down to about 175°F.

Use it as a planning tool, not an oracle. It is most valuable in the shoulder seasons and at night, when the answer is not obvious and the difference between a 1.5 in and a 2.5 in lift, or a calm afternoon and a windy one, decides whether you make density or chase it. Run it before the paver shows up so the rolling plan matches the time you will actually have.

The number it gives you is a budget. If PaveCool says you have eight minutes, the roller train has to be sized and positioned to put the full pattern on in eight minutes, which usually means tight behind the screed and no wandering.

Field example: a thin overlay on a cool morning

Put numbers on it. Lay a 1.5 in surface course at 300°F delivery on a 55°F morning over an existing pavement at 50°F, with a light 10 mph wind. A thin lift on a cool base under wind is the short-window case, and a cooling estimate for those conditions can leave you only a handful of minutes from screed to the mid-170s°F before the mat is done.

Now change the lift to 2.5 in and hold everything else. The thicker mat has more heat to give and sheds it slower, and the available time can roughly double. Same mix, same morning, very different rolling plan.

That swing is the whole point of running the estimate first. On the thin lift, the breakdown roller lives on the screed's heels and the operator does not get to stop and chat. On the thicker lift there is room to breathe. Guess wrong and you find out from the cores, which is the most expensive way to learn it.

Lift thickness and the t/NMAS ratio

Lift thickness is set by more than the heat it holds. There is a floor based on the stone size in the mix, expressed as the ratio of lift thickness to nominal maximum aggregate size, t/NMAS. The mat has to be thick enough that the largest stones can shift and seat without bridging on each other.

The common rule is a minimum lift of about 3 times the NMAS for fine-graded mixes and about 4 times for coarse-graded mixes, with many agencies and references citing a working range of roughly 3 to 5 times. Confirm the ratio your spec calls for, because they vary and some research argues for going higher. A 1/2 in NMAS surface mix, then, wants a lift on the order of 1.5 to 2 in, not 3/4 in.

Go thinner than the ratio allows and you pay for it. The mat will not compact to density, it turns more permeable, it is prone to segregation, and the roller can fracture the aggregate instead of seating it. That is the trap on a thin scratch course over a rough surface. You spread it too thin to hold density and it never seals. Match the lift to the stone, and match the stone to the lift you need.

NMAS

Nominal maximum aggregate size, a measure of the largest stone in the mix, read from the gradation

t/NMAS

Ratio of lift thickness to NMAS; a minimum near 3 (fine) to 4 (coarse) is commonly specified

The roller train: breakdown, intermediate, and finish

The roller train is the sequence of rollers working the mat in order, each with a job. Breakdown is the first and the one that matters most. It runs as tight behind the paver as it safely can, in the hot zone, and it builds most of the density. Lose the breakdown pass to a cold mat and no downstream roller buys it back.

Intermediate rolling follows and adds density while the mat is still workable, often with a pneumatic tire roller, whose kneading action chases voids the steel drum leaves and seals the surface. Finish rolling comes last, usually a static steel drum, and its job is to take out the roller marks and leave a tight, smooth surface, not to add density.

On wide mats and high production you see echelon rolling, two breakdown rollers running side by side to cover the full width in the time the heat allows, instead of one roller making twice the passes and losing the window doing it. The principle under all of it is the same. The pattern is built around the clock the mat gives you, and the breakdown roller stays on the screed's heels.

How many roller passes does it take to hit density?

The pass count is whatever it takes to reach target density inside the window, and you find it by running a test strip, not by reading a number off a chart. As a rough starting point for a typical 2 in dense-graded lift, breakdown might run on the order of 3 to 5 coverages, intermediate a few more, and finish a couple of light passes, but treat that as where the test strip begins, not as a spec.

The test strip, sometimes called a control strip, is how a real crew sets the pattern. You pave a section, roll it in a known pattern, and check density as passes accumulate until the gauge stops climbing. That coverage count, in that order, at that speed, becomes the pattern for the day's mix and conditions. Change the mix, the lift, or the weather and you reset it.

A pass and a coverage are not the same thing, and conflating them is a common error. A pass is one trip of the drum over a point. A coverage is the full mat width getting that pass. Operators count coverages. The roller has to hold a steady speed too, commonly in the low single-digit mph range, because a roller that races leaves voids and one that crawls in the tender zone shoves the mat.

Static, vibratory, and pneumatic rollers

Three roller types show up on a mat, and each compacts differently. A static steel drum compacts by dead weight alone. A vibratory steel drum adds a dynamic force on top of its weight, so it gets more density per pass and is the usual breakdown choice, though the amplitude and frequency have to suit the lift. Run high amplitude on a thin lift and you can crush aggregate or bounce the drum and mark the mat, so thin lifts call for low amplitude and high frequency, or static.

The pneumatic tire roller is the odd one and the underused one. Its rubber tires knead the mat, working voids out and sealing the surface in a way a steel drum does not, and it is the tool that gets a stubborn mat through the tender zone, because rubber does not shove the mix the way a steel drum does. Many density problems on tender or fine-graded mixes come down to nobody putting a pneumatic in the train.

Match the tool to the mix and the lift. The drum is not one setting for every mat, and the operator who knows when to drop the vibration and when to bring up the pneumatic is the one who makes density without chasing it.

The tender zone: when the mix shoves under the drum

Some mixes have a tender zone, a mid-temperature band where the mat is too unstable to support a steel drum and shoves, cracks, and checks under it instead of compacting. It commonly falls somewhere around 240°F down to about 190°F, depending on the mix, and you see it as a bow wave of material crawling ahead of the drum and hairline checking left behind.

Roll hard in the tender zone with a steel drum and you do damage, not density. The mix moves sideways instead of densifying, and the surface checks. There are two ways through it. One is to get your density above the zone, hot behind the screed, and below it, so the steel drum never works the mat in the unstable band. The other is to roll the band with a pneumatic tire roller, which kneads the mix without shoving it the way a steel drum does.

Not every mix is tender, and you find out which is on the test strip. Fine-graded Superpave mixes and certain sand contents are the usual suspects. When the mat starts crawling and checking under the breakdown roller, that is the zone talking, and the fix is to change tool or timing, not to add more steel.

Longitudinal joint compaction

The longitudinal joint is where density goes to die. It is the seam between two adjacent paving passes, and the chronic failure is low density along that seam, which lets water in and starts the joint raveling and cracking before the rest of the mat shows any age. More pavements fail from the joint in than from the field of the mat out.

The root cause is the unconfined edge. When the first lane is laid, the outside edge of that pass has nothing beside it to push against, so the roller spreads it sideways instead of compacting it, and that edge ends up lean. The confined side, where fresh hot mix is laid against an existing compacted lane, densifies much better. The hot side of a joint commonly comes out a couple of points denser than the cold, unconfined side.

The fixes are about confining or restoring that edge. An edge-restraining or pinch device on the breakdown roller presses the unconfined edge in on the first pass so it cannot spread. Rolling the unconfined edge with the steel drum hanging a few inches over it, in vibratory, helps. Keep the pneumatic off the very edge, because near an unconfined edge it spreads the mix and loses density rather than gaining it. And do not leave the joint for last and cold. Joint density usually carries its own spec line, because everyone knows it is the weak point.

What density do you need to pass?

In-place density is the headline acceptance number, and it is measured as a percent of Gmm, the theoretical maximum specific gravity of the mix, sometimes called the Rice value. The common field target on a dense-graded mat is about 92 to 93 percent of Gmm, which is the same thing as 7 to 8 percent air voids. Many specs set an acceptance band, often somewhere in the range of roughly 92 to 96 percent of Gmm with a number near 94 to 95 percent as the pay target, but the agency spec sets the actual limits and pay schedule, so confirm them for the project.

The floor matters more than the number sounds. Below about 92 percent of Gmm, the air voids stop being isolated bubbles and connect into channels, and a connected void structure lets water and air move through the mat. That is when premature raveling, stripping, and cracking start. A study by the Asphalt Institute and the Kentucky Transportation Center found that lifting density from 92 to 93 percent of Gmm improved fatigue life on the order of 10 percent, which is why agencies put real money behind density.

The joint gets its own, lower target, because everyone knows the unconfined edge runs lean. Minimum joint density requirements commonly land somewhere around 88 to 93 percent of Gmm depending on the agency. A well-built joint runs only 1 to 2 percent below the surrounding mat. A bad one runs 5 to 10 percent low, and that is the one that opens up in two winters.

How in-place density is measured

Density is a percentage, so you need two numbers: the maximum density the mix could reach and the density you actually got. The maximum is Gmm, the theoretical maximum specific gravity, run in the lab by the Rice method under AASHTO T209 or ASTM D2041. That is your 100 percent reference. Get the Gmm wrong and every density on the project is wrong with it.

The in-place number comes three ways. Cores are the referee. You cut a plug, measure its bulk specific gravity in the lab under AASHTO T166, the saturated-surface-dry method, or AASHTO T331 vacuum sealing for coarse or open mixes that drain, and compare it to Gmm. Cores are slow and they leave holes, but they are what settles a dispute. The nuclear density gauge, run per ASTM D2950, reads density in the field in a minute or two and is how crews track the rolling pattern in real time, but it has to be correlated to cores to be trusted for acceptance. Non-nuclear electromagnetic gauges, the PQI type, do the same field tracking without the licensing and storage that come with a nuclear source, and they also need correlation.

The practical workflow is gauge for control, cores for acceptance. You roll to a gauge reading you have correlated to cores, then the cores confirm it. When a gauge number and a core number disagree, the core wins, and the usual reason the gauge was off is a bad correlation or a Gmm that drifted with the mix.

Gmm (Rice value)

Theoretical maximum specific gravity of the loose mix, the 100 percent density reference (AASHTO T209 / ASTM D2041)

Gmb

Bulk specific gravity of the compacted mat or core, the in-place number (AASHTO T166 or T331)

% Gmm

In-place density as a percentage of Gmm; 100 percent minus % Gmm is roughly the air-void content

Can you pave asphalt in cold weather?

You can pave in cold weather, but the colder it is the shorter your window, and below a point the spec stops letting you. Many agencies set a minimum ambient and surface temperature for paving, with something around 50°F and rising as a common line for thinner surface courses and lower minimums allowed for thicker lifts, but the number is agency- and lift-specific, so check the spec. The hard part is not the air temperature on the thermometer. It is what the cold does to your compaction time.

Cold base steals heat from the bottom of the mat the moment it is placed, and cold air and wind strip it off the top. The same mix that gives you a comfortable window in July gives you a few minutes in October. The defenses are real but limited. Lay it thicker where the design allows, raise the delivery temperature within what the binder tolerates, keep trucks covered and moving, and shrink the haul. Run PaveCool so the rolling plan matches the cold-weather clock.

Night paving is its own case. Crews go to nights when daytime lane closures are not allowed, and summer nights in the 60s°F can compact fine. The trap is the early-morning hours when the air and the ground keep dropping. A mat laid at 2 a.m. on falling temperatures has a shorter window than the same mat at 10 p.m., and the rolling has to be planned around it.

What do I do if the mat cooled too fast?

If the mat cooled below its stop temperature before you reached density, the honest answer is that you cannot roll density back into it. Once the binder has set, more passes polish the surface and seat nothing. The air voids you did not close are locked in. Chasing a cold mat with a heavy vibratory roller does more harm than good, marking the surface and sometimes fracturing aggregate.

What you do is stop, get a density check, and find out how far short you are. If the mat is close to spec, the agency may accept it, dock it on the pay schedule, or want a confirmation core. If it is badly low, the fix is removal and replacement of that section, which is exactly the conversation you do not want, so the real answer lives upstream.

The recovery is in the next load, not the cold one. Figure out what closed the window early. Thin lift, cold base, wind, a cool delivery, a roller train that fell behind. Then change it before the next truck. Tighten the breakdown roller onto the screed, add a roller, raise the delivery temperature, or stop paving if conditions have moved outside what the mix and the spec allow. The cheapest cold mat is the one you prevent on the load after the one that taught you.

What to document

The record is what defends the mat when a core comes back light months later. Density disputes are won and lost on whether anyone wrote down the temperatures and the rolling, and the crew that documents is the crew that gets paid for marginal lots instead of penalized.

Capture the mix and lift, the delivery temperature read off the mat behind the screed, the mat temperatures at breakdown, intermediate, and finish, the ambient and base temperatures and the wind, the rolling pattern actually used, coverages by roller, and the density results with which gauge or cores produced them. Note any delay and why, because a stalled truck or a paver breakdown is exactly the gap that shows up later as a soft spot, and a note made at the time is worth more than a memory at the dispute meeting.

Field checklist

Common mistakes

• Rolling too late, letting the mat fall below cessation temperature before the breakdown pass is finished.
• Leaving the longitudinal joint for last when it is cold, then wondering why it ravels first.
• Working a tender mix with a steel drum inside the tender zone instead of a pneumatic or a changed pattern.
• Over-rolling a cold mat, which marks the surface and can fracture aggregate without adding density.
• Reading the truck ticket temperature instead of probing the mat behind the screed.
• Running a nuclear or PQI gauge for acceptance without correlating it to cores.
• Spreading a lift thinner than the t/NMAS minimum, so it never seals or reaches density.
• Setting one rolling pattern and never resetting it when the mix, lift, or weather changes.

Standards and references

The Asphalt Institute's MS-22, Construction of Quality Asphalt Pavements, is the practical reference for placement and compaction, and the chapters on rolling and acceptance are where this material is laid out in depth. For mix design and volumetrics, the Superpave gyratory compactor procedure is AASHTO T312.

The test methods behind the density numbers come from AASHTO and ASTM. Theoretical maximum specific gravity, Gmm or the Rice value, is AASHTO T209 and ASTM D2041. Bulk specific gravity of compacted specimens and cores is AASHTO T166, the saturated-surface-dry method, with AASHTO T331 vacuum sealing for mixes that drain. In-place nuclear density measurement follows ASTM D2950. Non-nuclear electromagnetic gauges have their own procedures and still require correlation to cores.

Acceptance itself, the density target, the pay schedule, the minimum ambient temperatures, and the joint requirement lives in the state DOT or owner specification, and it varies by agency. The temperature ranges, the t/NMAS ratio, and the density targets in this guide are the commonly cited figures. The mix design, the binder grade, and the controlling spec set the numbers you actually build to, so confirm them against the project documents and the adopted specification before you rely on any single value.

Units, terms, and conversions

The same mat gets described in a few unit systems and a few names, so the numbers read differently across a mix design sheet, a spec, and a gauge readout.

Temperature is in °F on most U.S. paving jobs, with the cessation floor commonly near 175 to 185°F, which is about 80 to 85°C. Lift thickness is in inches and aggregate size in inches or millimeters, so a 1/2 in NMAS is 12.5 mm. Density is a percent of Gmm, the theoretical maximum specific gravity, and the inverse of density is air voids, so 93 percent of Gmm is 7 percent air voids. Gmm is the Rice value, Gmb is the bulk specific gravity of the compacted sample, and the two together give you the percent.

Compaction window

The time and temperature range in which the mat can be rolled to density

Cessation temperature

The mat temperature below which rolling no longer reduces air voids

Breakdown rolling

The first, hottest rolling pass set, which builds most of the density

Tender zone

A mid-temperature band where some mixes shove or check under a steel drum

% Gmm / air voids

In-place density as a percentage of Gmm; the remainder is roughly the air-void content

NMAS / t/NMAS

Nominal maximum aggregate size, and the lift-thickness-to-NMAS ratio that sets minimum lift

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• Why is asphalt density important?
• Why does asphalt density matter?
• What is the minimum temperature to pave asphalt?
• What is warm mix asphalt?
• What is intelligent compaction?