Curing agent dosage decides two things at once: how strong the finished part becomes, and how long the processor has to shape it. In a reclaimed thermoset binder, going from 10 to 50 percent reactive system raised tensile strength from 1.2 to 3.0 MPa while elongation at break peaked near 200 percent and then fell back to 90 percent.
Why an amount of curative is a scheduling decision
A reactive binder does not stop reacting when mixing stops. Anyone who has watched a batch stiffen in the pot knows this. It is also why the reactive fraction of a formulation can never be picked on strength alone. Whatever the dosage does to the final network, it does the same thing to the clock.
So the real question is not how strong the mix can get. It is how strong it can get while there is still time to pour it, compact it, or coat it. A study on a reclaimed thermoset binder ran the whole chain of measurements, strength, viscosity, stiffness, relaxation, phase morphology, and compaction quality, on the same set of formulations. The pattern that came out of it travels well beyond the road material it was measured on.
How the experiment was organized
The reactive system was an epoxy resin with an amine hardener at a 1:3 mass ratio, accelerated by 0.4 percent of a tertiary amine catalyst. Dosage was changed by swapping virgin asphalt mass for reactive system, which held the aged fraction and the total binder content steady.
| Variable | Levels | Notes |
|---|---|---|
| Reactive system content | 10, 20, 30, 40, 50 percent | Reference formulation contained 0 percent |
| Cure temperature | 110, 120, 130 degrees Celsius | Isothermal hold after mixing |
| Cure time | 0, 20, 40, 60, 80, 100, 120 minutes | Zero minutes means compacted immediately |
| Tensile test | 500 mm per minute at 23 degrees Celsius | Specimens precured 16 hours at 120 degrees Celsius |
| Viscosity test | Brookfield spindle at 50 rpm | Continuous recording during cure |
| Rheology | 25 mm plate, 1 mm gap, 1 percent strain, 10 rad per second | Temperature sweep from 58 to 82 degrees Celsius |
| Low-temperature test | Bending beam, 980 mN load, 60 seconds | Tested at minus 6, minus 12, and minus 18 degrees Celsius |
| Mixture | 60 percent reclaimed material, 5.12 percent total binder | Air void content and indirect tensile strength |
Running both scales was the point of the design. The binder tells you what the chemistry is doing. The mixture tells you what that chemistry costs on the line.
How curing agent dosage shifts the strength and ductility balance
More reactive system bought more strength, cleanly and without a plateau. Ductility behaved differently. It climbed to a peak and then fell off a cliff.
| Reactive system | Tensile strength (MPa) | Elongation at break |
|---|---|---|
| 10 percent | 1.2 | about 90 percent |
| 20 percent | intermediate | about 130 percent |
| 30 percent | intermediate | about 165 percent |
| 40 percent | intermediate | about 200 percent |
| 50 percent | 3.0 | about 90 percent |
The specification applied in the study required at least 100 percent elongation at break. Two formulations missed it, at 10 percent and at 50 percent, and for opposite reasons. One had too little network to carry load. The other had too much to bend. Only the 20 to 40 percent band cleared both requirements at the same time.
This is where a dosage sweep stops being a single number. Pick the strongest compound in the table and you have picked the one that fails the ductility clause.
The viscosity curve has two regimes
Brookfield viscosity was logged continuously at three hold temperatures. Every reactive formulation traced the same shape: a slow early climb, then a sharp surge. The unreactive reference hardly moved over the full 120 minutes.
At 120 degrees Celsius the 40 percent formulation began 215 mPa s below the 20 percent material and 85 mPa s below the 30 percent material. It ended the hold 1393 and 1019 mPa s above them. Held at 130 degrees Celsius, the same formulation hit 4485 mPa s at around 105 minutes.
Two things pull the surge forward: more reactive system, and a hotter hold. Both squeeze the window from the same side. Raising temperature to improve wetting is not free, because it is paid for out of processing time.
High-temperature stiffening is real but one-sided
Cure had a big effect on stiffness at elevated temperature. At 58 degrees Celsius the complex shear modulus climbed throughout the hold, and the rutting factor, the complex modulus divided by the sine of the phase angle, followed it up.
| Binder | Rutting factor at 0 minutes (kPa) | Rutting factor at 120 minutes (kPa) | Change |
|---|---|---|---|
| Unreactive reference | 5.45 | 5.79 | plus 6.1 percent |
| 20 percent reactive | 9.38 | 20.08 | plus 114.0 percent |
| 30 percent reactive | 13.23 | 33.00 | plus 149.5 percent |
| 40 percent reactive | 18.06 | 44.57 | plus 146.8 percent |
The control moved by single digits over the same period. Anything above 100 percent here belongs to the curing reaction and not to the thermal history.
Low-temperature relaxation pays for the gain
The bending beam results are the invoice. At minus 18 degrees Celsius, creep stiffness went up while the creep rate, which says how fast the material sheds stress, went down.
| Formulation | Stiffness at 0 minutes (MPa) | Stiffness at 120 minutes (MPa) | Creep rate at 0 minutes | Creep rate at 120 minutes |
|---|---|---|---|---|
| 20 percent | 113.05 | 164.20 | 0.314 | 0.271 |
| 30 percent | 137.68 | 215.83 | 0.297 | 0.248 |
| 40 percent | 172.04 | 278.11 | 0.281 | 0.227 |
Stiffness rose by 45.2, 56.8, and 61.7 percent. All three formulations stayed under the 300 MPa stiffness limit. None of them stayed above the 0.300 relaxation threshold after a full cure. At minus 6 degrees Celsius the 40 percent binder crossed below that threshold after 80 minutes, while the other two held above it for the whole run.
Dosage and cure intensity therefore buy one property with another across the entire service temperature range. No dosage improves both ends simultaneously.
Phase morphology confirms the mechanism
Fluorescence microscopy made the growth of the reactive phase visible. Bright regions mark the fluorescent reactive phase, and the area they cover was measured at each cure time.
| Formulation | Area fraction at 0 minutes | Area fraction at 120 minutes |
|---|---|---|
| 20 percent | 2.35 percent | 22.12 percent |
| 30 percent | 3.12 percent | 25.79 percent |
| 40 percent | 5.19 percent | 32.62 percent |
The reactive phase both appeared and merged as cure advanced. At the start the bright areas were scattered dots. Two hours later they had grown and joined. That is the physical face of the viscosity surge and the stiffness gain, and it is why neither is reversible.
At the mixture scale the optimum arrives and then leaves
The mixture data carry the clearest operational message of the whole study. Air void content dropped to a minimum and then climbed back. Indirect tensile strength mirrored it exactly, peaking and then declining.
A 40 percent formulation held at 130 degrees Celsius reached about 2.5 percent air voids at 40 minutes, then drifted back up to about 5.2 percent at 120 minutes. Its indirect tensile strength peaked near 3.0 MPa somewhere between 20 and 40 minutes and sagged to about 1.35 MPa by the end of the hold.
The unreactive control mixture barely moved on either measure. That contrast is what ties the hump to the curing reaction rather than to the compaction procedure.
The counterintuitive result: the lowest viscosity does not compact best
Follow viscosity and you would compact at zero minutes, when the binder is thinnest. That is not what the data show. Air void content at zero minutes sat above its minimum, and the best densification came after a hold.
Two processes are competing, according to the study. Early on the binder is still fluid enough not to block particle rearrangement, and the continued soak helps the virgin and aged fractions interdiffuse, which closes voids. Later the network has grown enough that flow resistance takes over and voids reopen. Compaction quality is the product of flow and interdiffusion together, not flow alone.
That changes how curing agent dosage should be chosen. The best hold time is not the point of lowest viscosity, and it is not a fixed number of minutes copied from a datasheet. It has to come off a measurement on the real formulation at the real temperature.
The three factors are not independent
A three-way factorial analysis of variance treated dosage, temperature, and hold time as factors. Every main effect, every two-way interaction, and the three-way interaction came back significant at p below 0.01.
| Source of variation | Degrees of freedom | F value for air voids | F value for indirect tensile strength |
|---|---|---|---|
| Formulation composition | 3 | 668.270 | 783.610 |
| Cure temperature | 2 | 12.150 | 22.978 |
| Cure time | 6 | 396.762 | 281.600 |
| Composition x temperature | 6 | 227.695 | 51.522 |
| Composition x time | 18 | 130.618 | 74.085 |
| Temperature x time | 12 | 59.925 | 33.850 |
| Composition x temperature x time | 36 | 7.615 | 5.470 |
Formulation composition owns by far the largest F value for both responses. The interaction terms matter just as much for planning, because they say the effect of a dosage change depends on the temperature and the hold, and the other way round. No one-factor-at-a-time dosage sweep can locate the operating point.
What to do with this in a formulation lab
- Fix the specification before the strength target. The strongest formulation in this set was the one that failed the elongation clause.
- Measure the workable window instead of reading it off a datasheet. Run the viscosity sweep at the temperature the process actually reaches, because a 10 degree shift moved the surge onset appreciably.
- Do not count on a hotter process to buy time. Higher temperature starts thinner but surges earlier, so the safe interval gets shorter, not longer.
- Expect the compaction or shaping optimum in the middle of the hold. Take data at several hold times before the procedure is locked.
- Screen with a factorial design once the space is defined. Every interaction here was significant, which rules out varying one factor at a time.
- Check low-temperature relaxation before pushing dosage up to hit a stiffness target. The creep rate fell below the reference threshold in every fully cured formulation at the coldest test temperature.
- If the cold end governs, cut dosage and extend cure rather than raising both.
Limitations
The window covered 110 to 130 degrees Celsius and holds up to 120 minutes. The aged fraction was simulated in the laboratory rather than recovered from the field, so interdiffusion may behave differently in a production stream. Rheology was run at one angular frequency, leaving frequency dependence unresolved. One reactive chemistry was tested, so the numbers are not universal constants. Service-life performance and long-term durability were not evaluated.
Bottom line
Curing agent dosage is a scheduling variable as much as a strength variable. In this data set the usable range was 20 to 40 percent of the binder, with strength still climbing at 50 percent while ductility collapsed there. Temperature is not a free accelerator, because it advances the viscosity surge. And the mixture compacts better after a short hold than immediately, so the earliest compaction is not the best one.
FAQ
What does curing agent dosage control in a reactive binder
Curing agent dosage controls both the final crosslink density and the duration of the workable window. Raising the reactive fraction in this study increased tensile strength from 1.2 to 3.0 MPa, raised the rutting factor by more than 110 percent, and shifted the onset of the rapid viscosity surge to earlier times. The same change reduced elongation at break and lowered low-temperature stress relaxation.
How much reactive system gave the best balance of strength and elongation
Between 20 and 40 percent of the binder. In that band tensile strength was well above the unreactive reference while elongation at break stayed at or above 100 percent, which was the specification limit applied in the study. At 10 percent the material was too weak, and at 50 percent elongation had fallen back to about 90 percent and failed the requirement.
Why does elongation at break fall again at high dosage
Because the network becomes too dense to deform. Elongation at break rose steadily to about 200 percent at 40 percent reactive content and then dropped to about 90 percent when the reactive fraction reached 50 percent. The excess crosslink density raises stiffness and strength but removes the chain mobility that lets the material stretch before it fractures. This is a common trade-off in thermoset formulations.
What does the viscosity curve of a curing binder look like
It has two regimes. Viscosity rises slowly for a period and then surges steeply. In this study the unreactive reference barely changed over 120 minutes while every reactive formulation showed the two-stage pattern, and the surge was attributed to progressive growth of the crosslinked phase rather than to simple thermal thinning or thickening.
How does cure temperature change the viscosity surge
A higher hold temperature lowers the starting viscosity but brings the surge forward. At 130 degrees Celsius the 40 percent formulation reached about 4485 mPa s at roughly 105 minutes, and rising temperature generally advanced the onset of rapid growth. That combination means the safe processing interval shortens as the process runs hotter.
How much did the rutting factor increase after curing
Between 114 and 149.5 percent, depending on dosage. At 58 degrees Celsius the 20 percent formulation went from 9.38 to 20.08 kPa over a 120 minute cure, the 30 percent formulation from 13.23 to 33.00 kPa, and the 40 percent formulation from 18.06 to 44.57 kPa. The unreactive control rose by only 6.1 percent over the same period.
What happened to low-temperature relaxation
It weakened, and it became the limiting property. At minus 18 degrees Celsius stiffness rose between 45.2 and 61.7 percent while the creep rate fell below the 0.300 threshold in every formulation after a full cure. Even though stiffness stayed under the 300 MPa limit, the ability to relax stress was compromised, which points to a higher risk of low-temperature cracking.
What did fluorescence microscopy show during curing
The reactive phase grew and coalesced. The fluorescent area fraction rose from 2.35 to 22.12 percent for the 20 percent formulation and from 5.19 to 32.62 percent for the 40 percent formulation over 120 minutes. Sparse bright dots at the start became larger merged regions. This morphology change is the physical counterpart of the viscosity surge and the stiffness gain.
Why did air void content rise again after reaching a minimum
Because the binder eventually becomes too stiff to compact. The mixture densified while the binder was fluid enough not to block particle rearrangement and the thermal soak helped virgin and aged fractions interdiffuse. Once flow resistance from the growing network dominated, voids reopened. A 40 percent formulation at 130 degrees Celsius reached about 2.5 percent air voids at 40 minutes and about 5.2 percent at 120 minutes.
Which factor dominated mixture compaction quality
Formulation composition. In the three-way analysis of variance its F value was 668.270 for air void content and 783.610 for indirect tensile strength, far above the values for temperature and time. Every interaction term was nevertheless significant at p below 0.01, including the three-way interaction, so dosage cannot be optimized without fixing the temperature and hold time.
Can a longer hold time replace a higher cure temperature
Not directly. Temperature and time interact significantly, meaning their combined effect is not the sum of their separate effects. Raising temperature lowers initial viscosity but accelerates network growth, while extending the hold at a lower temperature gives more shaping time before the surge. Which is better depends on the specification, so both variables belong in the same design.
What is the practical limit of this study
The temperature window was 110 to 130 degrees Celsius and the hold was capped at 120 minutes. The aged fraction was laboratory-simulated rather than field-recovered, only one reactive chemistry was tested, and rheology was measured at a single frequency. Field cooling during transport and paving was not reproduced, so construction windows still need site validation.

