The Use of Phosphoric Acid to Stiffen Hot Mix Asphalt Binders

CHAPTER 2. ANALYTICAL METHODS

QUANTITATIVE ANALYSIS OF ASPHALT BINDERS FOR PHOSPHORIC ACID

X-ray fluorescence spectroscopy (XRF) is an analytical technique by which all the elements in the periodic table from sodium to uranium can be quantitatively and rapidly detected with minimal sample preparation. Test samples are irradiated with an X-ray beam, and the resulting spectrum can be used to provide quantitative information on each element present.

The use of XRF to quantitatively determine the amount of phosphoric acid contained in asphalt binders was developed by Puzic et al.⁽³⁾ The method has been refined by Reinke et al.⁽⁴⁾

Samples are placed in cups consisting of two concentric polypropylene rings over which a thin plastic film is stretched like a drum skin. The X rays are able penetrate the film with no attenuation of the beam. Initially, 6-micron Mylar® polyester film was chosen for its strength. It was discovered that it contains the equivalent 0.1-percent phosphoric acid and was discarded in favor of polypropylene, which contains none. Pictures of the inverted cups are shown in figure 1 and figure 2. A drop of water has been placed on the film of the empty cup in figure 1 to make the film visible. Warm asphalt is poured into the empty cup while it is sitting on a ¼-inch thick aluminum plate. The plate acts as a heat sink and prevents the heat of the asphalt from melting the plastic film. The asphalt temperature is not critical; the asphalt just needs to be molten. Typical pouring temperatures are 140 °C.

The plastic cup, filled with asphalt, is then placed inside a stainless steel cup holder (figure 3), which is placed inside the spectrometer (figure 4). Each sample takes 20 to 25 min to run. The program runs automatically, and the spectrometer is capable of analyzing up to 52 samples unattended.

Figure 1. Photo. XRF cup (inverted) with plastic membrane.

Figure 2. Photo. XRF cup (inverted) filled with asphalt.

Figure 3. Photo. Steel XRF cup holder.

Figure 4. Photo. Interior of XRF spectrometer.

All asphalt binders contain a significant amount of sulfur; they do not contain phosphorus. Because sulfur and phosphorus are next to each other in the periodic table they have very similar XRF energies. The major Kα peak energy for phosphorus is 2.013 KeV and for sulfur 2.307 KeV. This proximity causes the peaks in the XRF spectrum from these two elements to overlap. Because the amount of sulfur in asphalt is very much higher than the amounts of phosphoric acid typically used for modification, the sulfur peak is very much larger. It overwhelms the phosphorus peak. This negatively affects the accuracy of the analysis. The XRF spectrometer software "sees" a phosphorus peak when none may be present. This phenomenon is clearly shown in figure 6, the XRF spectrum of asphalt AAB-1. At the energy level of approximately 2.0 KeV on the x-axis, the software has labeled the spectrum P-Ka1 indicating a phosphorus Kα peak when none is present. The software is using the intensity of the first part of the sulfur peak and interprets it incorrectly. Ninety samples of unmodified asphalt binders showed a phosphoric acid level of 0 to 0.2 percent when we know that none was actually present. Compare this with figure 5, the XRF spectrum of the same asphalt, AAB-1, modified with 1 percent of 105-percent phosphoric acid where the phosphorus peak can be clearly seen. There is no fixed detection limit. The results may also depend on the spectrometer used; however, these results suggest that XRF analyses indicating the presence of low levels of approximately 0.2 percent or less might be misleading.

The accuracy of the phosphoric acid analysis was improved markedly by entering into the spectrometer software phosphoric acid calibration standards, the known sulfur content of the binder used. The sulfur levels in the binders were determined using XRF. Accuracy was improved further by using the published sulfur contents of the SHRP reference binders.⁽⁵⁾

Figure 5. Chart. XRF spectrum of asphalt AAB-1 modified with 1-percent of 105 percent superphosphoric acid.

Figure 6. Chart. XRF spectrum of asphalt AAB-1.

Reference standards for the XRF analyses were prepared by blending 105-percent phosphoric acid in asphalt at 165 °C while stirring briskly with a propeller stirrer under air for 30 min. The hot asphalt was then poured into XRF cups for analysis. Addition levels of superphosphoric acid used were 0.25, 0.5, 1.0, 2.0, and 3.0 percent. The correlation chart showing phosphoric acid concentration plotted against the XRF intensity (measured in counts per second per milliamp of current), taken from the spectrometer is shown in figure 7. The R² correlation is 0.9973.

Figure 7. Chart. XRF Calibration chart for phosphoric acid.

Figure 8 shows that the XRF signals of the four SHRP reference asphalts at zero phosphoric acid addition differ slightly.

Figure 9 shows the same data corrected for the difference in zero acid addition.

Figure 8. Chart. Plot of XRF signal of SHRP reference asphalts modified with superphosphoric acid content.

The XRF up to 1-percent acid modification is the same for the four asphalts. However, at higher modification levels, the curves diverge, indicating some asphalt dependency. Test results were found to be less accurate at higher modification levels as shown in figure 10.

Figure 9. Chart. X-ray fluorescence of SHRP reference asphalts modified with superphosphoric acid corrected for baseline fluorescence.

Figure 10. Chart. Accuracy of PPA analysis using XRF.

A SIMPLE QUALITATIVE TEST TO DETECT THE PRESENCE OF PHOSPHORIC ACID IN ASPHALT BINDERS

Because not all State agencies have access to XRF spectrometers, a simple procedure, the "Susan P. Needham Test," was developed. It is a very simple technique; it requires no special equipment-just the use of a few simple chemicals. The test is very sensitive, and care must be taken that the equipment used and chemical used do not contain phosphates. This is particularly true for the use of metal cans that contain a phosphate film on the surface because they will give a positive result. The test has been submitted to AASHTO and is published as provisional test method TP 78-09, "Detecting the Presence of Phosphorus in Asphalt Binder." The following describes the reagents and procedures used in the test method.

Reagents

• Antimonyl Tartrate/Ammonium Molybdate Solution: Dissolve 0.13 g of potassium antimonyl tartrate hydrate [C₈H₄K₂O₁₂Sb₂·H₂O] in 50 mL of distilled water. Add 5.6 g of ammonium molybdate [(NH₄)₆Mo₇O₂₄·4 H₂O] and swirl until dissolved.
• 1N Sulfuric Acid Solution (H₂SO₄): This can be purchased in 1-L polyethylene bottles.
• Stock Solution Mixture: Mix (solution 1) and approximately 950 mL of (solution 2) above. This can be done by adding solution 1 to the 1 L of solution 2 if there is sufficient space in the bottle. The exact amount of solution 2 is not critical. This stock solution is stable for 1 year.
• Ascorbic Acid Color Reagent: Dissolve 0.50 g of L-Ascorbic Acid [C₆H₈O₆] in 100 mL of solution 3. This reagent is stable for a week if stored at 4 °C; otherwise, prepare the reagent fresh daily or as needed.
• Iso Butanol ((CH₃)₂CHCH₂OH)(n-butanol can also be used).

Procedure

Description:

• Heat the asphalt and pour 1 to 2 g into a small container, glass beaker, or test tube.
• Place the container in an oven for 10 min to ensure the asphalt is fluid.
• Remove the container and immediately add 2 mL of solution 5 while swirling the container.
• Continue to swirl the container and add 2 mL of distilled water.
• While still swirling the container, add 2 mL of solution 4.

Identification:

• If phosphoric acid is present in the asphalt, a blue color (figure 11) will develop at the bottom of the tube within 5 to10 min. (Decant into a second 1 oz can/glass tube if unable to see color.)

Figure 11. Photo. Phosphoric acid detected by the blue color developed in the Susan P. Needham test.

SATURATE, AROMATIC, RESIN, AND ASPHALTENE ANALYSIS OF ASPHALT BINDERS MODIFIED WITH PHOSPHORIC ACID

Solvent separation of asphalt binders (saturate, aromatic, resin, and asphaltene (SARA) analysis) was accomplished using Chromarods® (thin layer chromatography) run on the Iatroscan® TH-10 hydrocarbon analyzer. This method results in the separation of the four operationally defined fractions inherently present in all petroleum-derived asphalt and asphaltic residuals, namely are asphaltenes, resins, aromatics, and saturates.

Asphaltenes are the viscosity builders in asphalt. They are black amorphous solids that contain the bulk of the heteroatoms (nitrogen, sulfur, and oxygen) found in asphalt. Trace elements, such as nickel and vanadium, are also present. Asphaltenes are highly polar aromatic materials of aggregated molecular weights of 750 (number average), and constitute approximately 5 to 25 percent of the weight of asphalt.

Resins (polar aromatics) are dark-colored, solid or semi-solid, very adhesive fractions of relatively high molecular weight present in the maltenes. They are the dispersing agents or peptizers for the asphaltenes, and the ratio of resins to asphaltenes governs, to a degree, the sol- or gel-type character of asphalts. Resins separated from bitumen are found to have molecular weights of 800 to 2,000 (number average) but there is a wide molecular distribution. This component constitutes 15 to 25 percent of the weight of asphalts.

Cyclics (naphthenic aromatics) comprise the compounds of lowest molecular weight in bitumen and represent the major portion of the dispersion medium for the peptized asphaltenes. They constitute 45 to 60 percent by weight of the total asphalt and are dark viscous liquids. They are compounds with aromatic and naphthenic aromatic nuclei with side chain constituents and have molecular weights of 500 to 900 (number average).

Saturates comprise predominately the straight-and-branched-chain aliphatic hydrocarbons present in bitumen, together with alkyl naphthenes and some alkyl aromatics. The average molecular weight range is approximately similar to that of the cyclics, and the components include the waxy and non-waxy saturates. This fraction forms 5 to 20 percent of the weight of asphalts.

The test method used was kindly provided by Dr. Gaylon Baumgardner of Ergon® Inc. A copy of the method is provided in the appendix.

The four test asphalts, AAD-1, AAK-1, AAM-1, and ABM-1 were dosed with the equivalent of 1-percent acid. For phosphorus pentoxide, the stoichiometry calculates to 0.75 percent. The samples were conditioned overnight at 165 °C. Separation was carried out according the aforementioned procedure. The results are shown in figure 12 to figure 15 .

Figure 12. Chart. Results of SARA fractionation of AAD-1 modified with 1 percent of 105-percent superphosphoric acid or 0.75-percent phosphorus pentoxide.

Figure 13. Chart. Results of SARA fractionation of asphalt AAK-1 modified with 1 percent of 105-percent superphosphoric acid or 0.75-percent phosphorus pentoxide.

Figure 14. Chart. Results of SARA fractionation of asphalt AAM-1 modified with 1 percent of 115-percent PPA, 1 percent of 105-percent superphosphoric acid, or 0.75-percent phosphorus pentoxide.

Figure 15. Chart. Results of SARA fractionation of asphalt ABM-1 modified with 1 percent of 115-percent PPA or 0.75-percent phosphorus pentoxide.

The Iatroscan® technique is very sensitive to temperature, humidity, and time. Test results in the Turner Fairbank Highway Research Center (TFHRC) chemistry laboratory showed a high degree of variability because, at the time of this research, the building temperature and humidity were not well controlled. This variability precluded the detection of any trend at low levels of acid modification in the components separated by the technique. The tests were repeated, with higher levels of acid than would be expected to be used in practice, to see whether a definite trend could be determined.

Samples of the four SHRP test asphalts were dosed with 115-percent phosphoric acid at acid levels from 1 to 4 percent (based on 100-percent acid) and the samples aged overnight at 165 °C. The variability of the technique is evident from the shapes of the curves shown in the following four charts (figure 16 through figure 20). The data do, however, illustrate trends.

Figure 16. Chart. Results of SARA separation of asphalt AAD-1 modified with 115-percent phosphoric acid.

Figure 17. Chart. Results of SARA separation of asphalt AAK-1 modified with 115-percent phosphoric acid.

Figure 18. Chart. Results SARA separation of asphalt AAM-1 modified with 115-percent phosphoric acid.

Figure 19. Chart. Results of SARA separation of asphalt ABM-1 modified with 115-percent phosphoric acid.

Figure 20. Chart. Results of SARA separation of B6317 Venezuelan asphalt modified with 115-percent phosphoric acid.

Findings

The following findings resulted from the SARA analysis:

1. Phosphorus pentoxide has less effect on the SARA fractions than does phosphoric acid.
2. The heptane insoluble fractions (labeled as asphaltenes), of all five asphalt samples increased with increasing acid content.
3. The increase in the heptane insoluble fraction with increasing acid content was accompanied by a decrease in one or more of the other fractions:
   1. AAD-1: The increase in the heptane insoluble fraction from 19.5 to 31 percent was accompanied by an almost equal decrease in the level of the resin fraction from 31 to 19.4 percent. The cyclic fraction varied a little but was basically unchanged while the saturate fraction level did not change at all.
   2. AAK-1: The increase in the heptane insoluble fraction from 17.5 to 33.6 percent was accompanied by a decrease in the resin content up to the 3-percent acid level when the resin content then increased up to the 4-percent acid level. This increase in the resin fraction was accompanied by a decrease in the cyclic fraction, which was constant up to the 3-percent acid level and then declined. The amount of saturates was unchanged.
   3. AAM-1: The heptane insoluble fraction showed an increase from 4.6 to 21.8 percent. This was accompanied by an overall decrease in resin content from 35.4 to 26.6 percent although the curve shows an inflexion point at 2- to 3-percent acid. The content of cyclics shows a steady decline from 53.7 to 44.3 percent. There is some fluctuation in the level of saturates but overall, these remain almost unchanged.
   4. ABM-1: There is a 14.85-percent straight line increase in the heptane insoluble fraction from 4.3 to 19.15 percent. This is accompanied by a straight line decrease of 13.9 percent in the resin content from 38.6 to 24.7 percent. The levels of saturates and cyclics remain unchanged.
   5. B6317 Venezuelan Asphalt: The heptane insoluble fraction shows a steady increase from 12.8 to 24.9 percent. The resin content declined from 33.8 to 19.6 percent and showed a small inflection point at the 2-percent acid level.
   6. Four of the charts show a positive inflection point in the resin content, and this is accompanied by a negative inflection in the level of cyclics. Our experience with the Iatroscan® chromatographic technique has shown that the separation of saturates is straightforward and very reproducible. The separation of the cyclics and resins is much more difficult and subject to variation. Each data point in the charts is the average result of reading 10 Iatroscan® rods and so could reasonably be expected to show the true picture. Although four of the five charts show a decline in the level of cyclics with increasing acid, it cannot be assumed that they do actually decline. It may be some quirk with the technique. The level of the heptane insoluble fraction is not affected because it is not determined with the Iatroscan®. This fraction is removed before the Iatroscan® step and determined gravimetrically.
4. The increase in the amount of heptane insoluble fraction is not necessarily accompanied by an equivalent increase in viscosity. ABM-1, which shows no increase in stiffness when modified with up to 1-percent of 115 percent phosphoric acid, also shows the same rate of increase in heptane insoluble fraction as the other asphalts but, up to the 1-percent acid modification level, at least this was found to be actually accompanied by a slight decrease in stiffness (figure 15 and chapter 3).
5. The change in stiffness of the four SHRP reference binders when modified with phosphoric acid is shown in chapter 3. The sensitivity of the stiffness change to phosphoric acid addition found was AAK-1 > AAM-1 > AAD-1 > ABM-1. No correlation could be found between this phosphoric acid/stiffness sensitivity to any of the chemical characteristics published in SHRP-A-645 "SHRP Materials Reference Library: Asphalt Cements: A Concise Data Compilation."⁽⁵⁾

HOW DOES THE PHOSPHORIC ACID REACT WITH THE BINDER?

A small amount of the asphaltenes and heptane-insoluble fractions from the initial solvent separations was analyzed using the energy dispersive spectrometry attachment to the Amray scanning electron microscope. This is purely a qualitative test. It showed that the heptane-insoluble fraction contained phosphorus while the heptane-soluble maltene fraction contained none.

This was confirmed by Liquid State ³¹P nuclear magnetic resonance (NMR) spectra. The spectrum for the heptane-insoluble fraction (asphaltenes) shown in figure 21 clearly shows the peak due to the presence of phosphorus. In figure 22, the NMR spectrum for the heptane-soluble fraction (maltenes) has no phosphorus peak.

Figure 21. Chart. NMR spectrum of heptane-insoluble fraction of phosphoric acid-modified asphalt.

Figure 22. Chart. NMR spectrum of heptane soluble fraction of phosphoric acid-modified asphalt.

MAJOR CONCLUSIONS FROM CHAPTER 2, ANALYTICAL METHODS

• With proper calibration, the phosphoric acid content of asphalt binders can be readily measured using XRF spectroscopy.
• The Susan P. Needham test, which requires no specialized equipment, can be used to detect the presence of phosphoric acid in asphalt binders.
• Addition of phosphoric acid to asphalt binders causes an increase in the heptane-insoluble fraction, which is not necessarily accompanied by a corresponding increase in binder stiffness.
• The phosphorus from the acid all ends up in the heptane-insoluble phase.
• The increase in the heptane-insoluble fraction is generally accompanied by a decrease in the resin fraction. With some binders, there may have been a change in the level of cyclics although variability in the method makes this uncertain. The level of saturates is unaffected by the use of phosphoric acid.