Unmasking the problem

Last year, BP Amoco introduced a new product, low fuming asphalt (PG grades 58-28, 58-22 and 64-22), which can help hot-mix plant operators and paving contractors minimize fumes and odors from mix plants and paving job sites.

Amoco became involved with asphalt fuming issues about five years ago while working on a problem with flux that was used by a customer in the production of roofing asphalt. After processing the flux in an air oxidizer, the customer would sporadically see excessively high temperatures in their fumes incinerator. BP Amoco determined that these sporadic problems were related to highly volatile materials that occasionally showed up in the flux, and began experimenting with different test methods. The company eventually identified a test that would detect these volatiles and control the problem.

At about the same time, asphalt paving customers reported visible fumes or odors from hot-mix plants. BP Amoco found that the same test used to control flux also was sensitive to asphalt fumes, and began monitoring asphalt production on a routine basis using this test, the Pensky-Martens Closed Cup Flash Point. The experience with this and other tests to measure the fuming tendency of asphalt played a key role in the development of low fuming asphalt.

Tending to a tendency

In 1999, shortly after Amoco merged with BP, a decision was made to introduce low fuming asphalt. By that time, some additional information had become available which played a role in helping define test methods and specifications. In 1998, the Australian Asphalt Paving Association published a "Code of Practice" for SBS modified asphalt which included three specifications to control the fuming tendency of asphalt. These included the Initial Boiling Point as measured by Gas Chromatography (GC), the Pensky-Martens Closed Cup flash point and a new mass loss test that was not yet fully developed.

Three criteria were identified as being important for the development of the specification. First, the specification should restrict asphalt binders that cause a fuming problem. Second, and just as important, the specification should not needlessly restrict the use of good asphalts that do not cause a fuming problem. Finally, the test should be one that a refinery laboratory can run on a routine basis.

The Boiling Point by GC predicts the distillation curve of an asphalt by analyzing its hydrocarbon composition. Figure 1 shows the results of a predicted distillation on one of BP Amoco’s asphalts. The initial boiling point as defined by this test method is the temperature at which the lightest 1/2% is predicted to distill off, which for this asphalt is just under 800û F. The final boiling point, which predicts the temperature needed to distill off the heaviest part of the asphalt, is over 1400û F. The limit required in the Australian Code of Practice is NMT 482.

The low fuming asphalt met the limit. Since this test is only able to detect the lightest 1/2% of an asphalt, BP Amoco did not expect it to be sensitive enough to detect problem asphalts. In earlier work on asphalt fuming, workers were unsuccessful in using this test to identify asphalts with a fuming problem.

Despite low expectations, BP Amoco initially adopted this test as part of its low fuming asphalt specification. However, when the refinery lab reported that running this test on asphalts on a routine basis was fouling GC columns, routine use of the test was dropped.

The Australian Asphalt Paving Association had not yet finalized a mass loss test but had tentatively set a limit of 0.6% mass loss for the standard Rolling Thin Film Oven Test (RTFOT). Although it was thought this test might be more sensitive than the GC initial boiling point, past work on fumes at hot-mix plants and the fumes incinerator did not show this test to be sensitive to high levels of volatiles. Also, a 0.6% limit with this test appeared to be overly restrictive for the softest asphalt grades such as PG 52-28 and PG 46-34. BP Amoco adopted the 0.6% limit on this test, but only for PG 58-22 and harder grades.

The flash point tests—the Cleveland Open Cup and the Pensky-Martens Closed Cup—are similar in that while a sample is heated it is periodically checked with an ignition source to determine the lowest temperature at which the sample will flash.

The two tests differ as their names suggest. The Cleveland Open Cup is not covered and the lightest fumes are allowed to escape as the sample is being heated. To protect against fires in hot-mix plants, which are open to the environment, the Cleveland Open Cup has historically been specified for asphalts.

Although this test may be good from a fire safety perspective, past experience indicated that it was not sensitive enough to detect asphalts that caused fume or odor problems.

In the Pensky-Martens test the closed cup prevents the fumes from escaping. This usually results in a 50?-200? F lower flash temperature than the Cleveland Open Cup and thus the Pensky-Martens test is much more sensitive in its ability to detect the lightest fumes.

This was the only test that was sensitive enough to detect high levels of the volatiles that had been causing customers problems in the past. It also is one that is easily run on a routine basis at BP Amoco’s refinery laboratory. As part of the low fuming asphalt control specification, the company adopted the Australian limit of NMT 392û F for PG 58-22 and harder grades.

Here’s an example to illustrate the sensitivity of the Pensky-Martens Flash Point test. A typical asphalt might have a flash point of 500û F. Contamination with just a 1/4% of diesel fuel (with a flash point of 150? F) is enough to drop the flash point to 390? F, which would fail the low fuming asphalt spec. This amount of contamination is equivalent to about 1/2 gal of diesel in a 15-ton truckload of hot mix. The example drives home the importance of avoiding contamination of the asphalt binder whether at the refinery, the hot-mix plant or the paving jobsite.

More recently, BP Amoco has looked at another test to evaluate the fuming tendency of asphalts. Thermogravimetric Analysis appears to have some significant advantages over the Rolling Thin Film Oven Test. It allows continous and precise measurement of temperature and weight and it can be run with a nitrogen purge thus avoiding oxidation reactions. Oxidation reactions can be a problem in the standard RTFOT because they can result in a mass gain and interfere with the ability to measure loss of volatiles.

As shown in Figure 2, Thermogravimetric Analysis can easily be run at different temperatures. Here you see mass loss plotted versus temperature for a variety of commercial asphalts. As expected, note that there is a very significant increase in mass loss as temperature increases and that for some asphalts this effect is even more pronounced.

Although more work needs to be done with this test to determine how well it predicts fuming during real-world conditions, these results show that fuming can be minimized by keeping asphalt temperatures to a minimum at the hot-mix plant and the jobsite.

Disrupt the flow

At the refinery, the key to producing a low fuming asphalt is to keep the higher volatility refinery streams out of the asphalt blend stocks. Asphalt by nature is a very low volatility or low fuming product, and most other refinery streams are high volatility in comparison.

The flash point example showed that just a small amount of a more volatile component can have a dramatic impact on flash point and fuming tendency.

Within the refinery there are numerous opportunities for more volatile streams to get into the asphalt. If a valve is opened or closed at the wrong time it could be an opportunity for a contaminant to get into the asphalt. A leaky valve or an improper line flush also are opportunities for contaminants to get into the asphalt. Another point of entry for contaminants is a leak in a heat exchanger. The asphalt coming from a vacuum distillation unit is used to heat incoming crude oil. Because the asphalt is under vacuum, even a small leak in the heat exchanger will draw the much higher volatility crude oil into the asphalt and contaminate it. Although a small amount of volatiles contamination may not be detected in the standard asphalt PG tests, it can be detected by the Pensky-Martens flash point test and problems can be corrected before the asphalt leaves the refinery.

Last year, BP Amoco’s Whiting, Ind., refinery was able to meet the tighter specifications on all shipments of the low fuming asphalt grades.