Longitudinal joints – bad, ugly and how to make them good

By Bob Humer, P.E.

The longitudinal joints are often the weakest link in an otherwise good performing asphalt pavement. 

Unless creating a hot-joint by paving in echelon, the typical cold joint will have less density than the center of the paving lane. This occurs for dense-graded mixes, and to a lesser extent for SMA mixes and asphalt-rubber mixes. The deterioration of the joint area is caused by it being permeable. The permeability leads to water and air intrusion, resulting in binder oxidation and a scouring of the mix matrix caused by the combination of traffic and water intrusion. This leads roads to early maintenance/repairs of the joint area, and on airports to earlier-than-anticipated overlays to avoid the risk of Foreign Object Damage (FOD) from dislodged mix. These are the costly results of failing joints and the associated reduction in service life. 

Insufficient density and higher permeability are detrimental to the performance of a pavement in general and the longitudinal joints specifically. Therefore, our best opportunity for improvement is to avoid permeability in the joint area. 

Studies have shown that lower density is to be expected in the joint area than in the center of the paving lane; in the best cases, the joint density is only 1-2% lower. This doesn’t mean that we can’t do better. The Asphalt Institute’s cooperative work with the Federal Highway Administration resulted in increased knowledge and training on how the performance of the joints can be improved. Numerous recommendations regarding the mixes and construction methods are now available to reduce permeability and achieve better joint performance. Many of these recommendations will be discussed. 

The basics of permeability and segregation 

First, we need to discuss the basics of permeability and how it relates to the properties of density, TMD, Gmm, air voids, and also physical and thermal segregation.

Mix permeability is directly related to mix density/compaction and segregation.

The mix density is expressed/measured as a percent of Theoretical Maximum Density (% TMD). For this, the Theoretical Maximum Specific Gravity (Gmm) of the mix is determined by means of the AASHTO T 209 test procedure. The TMD is then calculated as: 

TMD = Gmm x 62.43 lb/ft3 

(with 62.43 lb/ft3 = unit weight of water at 4°C) 

Once the TMD of the mix has been established in the lab, the densities of the mat and the joints achieved in construction are expressed as a percent of TMD (% TMD). 

The percent air voids (Pa) in a compacted mix specimen (core or lab compacted mix specimen) is determined with the equation: 

= the Bulk Specific Gravity of the compacted mix, determined with the AASHTO T 166 procedure. 

The terms “% density” and “% air voids” are often used when discussing compaction. 

Percent density is expressed as % of TMD and relates to % air voids as shown in the following equation: 

% Air Voids = 100 – % TMD. 

Thus if the achieved density is 93% of TMD, then the compacted specimen contains 7% air voids (Pa). Higher density corresponds with lower air voids. Lower density corresponds with higher air voids.

Relative compaction and the reference density 

Whenever relative compaction is required, it is important to know the reference density. This can be the TMD of the mix, or maybe the Gmb of the designed mix, or the Gmb of that day’s production mix. Always determine what the agreed-upon reference density is for the project. Because TMD is the least variable property it has become the preferred reference in specifications. 

Specifying an acceptable minimum % density and/or a maximum % air voids (Pa) is a powerful tool to control the permeability and with that the quality and service life of the constructed pavement. This has been a longtime practice for the mat and is more recently being implemented for the joint areas. 

Physical and thermal segregation 

A more difficult condition to measure and specify is the degree of segregation. Segregation can be identified as physical or thermal segregation. For both types of segregation, meeting the specified limits of % density and % air voids will be problematic. 

Physical segregation is where the larger size aggregates separate from the mix, forming rock pockets/more rocky areas. In these rocky areas, the mix is no longer the optimized design mix and will have a significantly higher percentage of air voids. These physically segregated areas tend to be permeable. Other than visual observation, the segregated area can be confirmed by coring and testing for low binder content, high air voids content, and gradation deviating from the specified mix gradation. 

Depending on the type of project and the extent of the segregation, remedial surface treatments such as seal coating may be a solution. For significantly segregated areas, the better solution may be to remove and replace. The joint areas are specifically prone to physical segregation when the augers and tunnels are not operated properly. Proper construction practices go a long way in avoiding or minimizing this problem. 

Thermal segregation is the other type of segregation that can occur. Thermal segregation is not visible but can be detected and measured by infrared scanning. The infrared “gun” is now commonplace among the tools of the inspectors. Significant mix temperature differences can occur in the mat as a result of hiccups at the plant/storage silos, incorrect truck loading and unloading, incorrect operation of the paving machine (running the hopper and/or the augers low), and/or mat repairs by the rakers during paving. The colder areas do not compact as well as the rest of the mat, resulting in localized areas with lower density, higher air voids, and the risk of permeability. 

Critical air void levels 

Past studies show for different mixes that there are critical air voids levels at which permeability of the compacted mix starts. 

For 9.5 mm mixes Critical air voids level
E. Zube – California Dept. of Highways, 1962 8.0%
L. Cooley, B. Prowell, R. Brown – NCAT, 2002 7.7%
R. Mallick, et al – NCAT, 2003 (fine graded) 8.5%
For 12.5 mm mixes
B. Choubane, et al – Florida DOT, 1998 7%
J. Westerman – Arkansas HTD, 1998 6%
R. Mallick, et al – NCAT, 2003 (coarse graded) 7%

In other words, the larger aggregate size mixes will have even a lower critical air voids level. And that makes sense, because where there is a large aggregate piece there is no air void, forcing the air voids to be concentrated in the rest of the mix. Therefore the AI/ FHWA study concluded that selecting smaller aggregate mixes is preferred. They are less prone to be permeable. In addition, these smaller aggregate size mixes do not segregate as easily, typically have a higher binder content, and are more workable and compactable. These benefits are especially important for joint areas. 

How air voids relate to pavement service life 

The fact that a higher percentage of air voids results in a reduced pavement service life has been concluded in numerous studies. R. Linden, Joe Mahoney, and N. Jackson concluded from a nationwide study of 48 DOTs that for every 1% increase in air voids (over the base air-void level of 7 percent) it tends to result in about a 10% loss in pavement life (as reported in TRR 1217). More recently the National Center for Asphalt Technology (NCAT) has confirmed this conclusion in their publications. 

The increasingly rapid reduction in pavement service life is shown by WA DOT in the following graphic based on many years of their Pavement Management System (PMS) data. 

The percent reduction in pavement service life can be equated to a similar increase in infrastructure cost.

The three types of longitudinal joints 

Before getting into the recommendations for improving the joint density by mix choice, and construction practices, consider that there are three types of cold joints. 

The butt joint is the original type of cold joint, where nothing special was done to help improve the joint density. The second paving pass is butted up against the first pass to close the joint. 

Improvements can be made to the butt joint by milling or cutting back and removing the 3-4” lesser compacted material from the edge and then placing the second pass against the better-compacted edge material. 

The “notched wedge” joint was developed more recently, first as a safe traffic transition between lanes when the second lane is placed later as a better performing joint design. There are concerns with the level of compaction achievable in the bottom wedge. However, the bottom wedge does provide some sideway confinement of the first pass. Tacking and rolling should be applied as with the other two joint types. 

For both the butt joint and the milled or cut back joint a number of options regarding treating the joint area are recommended. This can be placing a void-reducing asphalt membrane (VRAM) under the joint area, treating the edge with tack or joint adhesive (JA), and when later needed a surface treatment such as over-banding with joint sealant. 

Void-reducing asphalt membrane (VRAM) 

A more recent technology is that of placing a joint enrichment underneath the joint area by utilizing VRAM material. It is a localized rich application of asphalt material, which prior to rolling, migrates upwards about 1” to fill the voids. In this manner, a waterproof joint is created. Its non-tracking property is important during construction and its resistance to flow avoids bleeding later on. At the right application rate, it imparts crack resistance at the joint. 

Compaction, roll-down and overlap 

To achieve proper compaction of dense-graded mixes, about 25% over-height of the specified lift thickness needs to be placed by the paving machine. This allows the rollers to achieve density at the specified lift thickness. Special attention to this over-height needs to be given to the joint area. Past practice of raking the joint, by scalping and starving the hot side of the joint of material resulted in low joint densities. A minimum amount of overlap is recommended of the first pass edge, about 1” for the butt joint and about ½” for the trimmed back joint and 1” for the notched wedge joint. This requires the edge of the first pass to be straight, so that joint raking can be avoided. The AI/FHWA study recommends this minimum amount of overlap to remain in place and for the rollers to just crush it down. Weather and traffic will wear off any crushed aggregate material from the joint area. 

Another interesting recommendation from the study is to place the first pass along the shoulder and successive passes towards the pavement crown, with each pass approximately 1/8th to 1/10th of an inch higher than the previous pass. This avoids bridging of the rollers across the joint, assuring good compaction at the joint, and positive and unobstructed cross drainage of surface/rainwater. 

Without going into great detail, the many recommendations from the study are as follows: 

In the planning and specifications writing stage consider the following: 

• Echelon paving: to have “hot joints” instead of “cold joints” 

• “Mill and Fill”: one lane at a time to eliminate unconfined edges 

• Cutting back: the unconfined edge (3” to 4”) to remove lesser compacted material 

• Stagger the joints: (at least 6”) in successive lifts, for better pavement strength in the joint area 

• Plan the joints: in the surface lift to avoid wheel paths, recessed pavement markings, and striping 

• Have well-defined specifications: regarding the longitudinal joints and the QA requirements for the joints 

• Lift thickness: should be at least 4x NMAS (for course mixes) and 3x NMAS (for fine mixes) 

• Use less permeable mixes: by selecting; 1) the smallest NMAS (that will not rut), 2) a finer gradation, and 3) lower design Air Voids 

• Warm Mix Asphalt: (WMA) technology can assist as a compaction aid 

• Consider the “notched wedge” joint in 1.5” or thicker lifts for increased joint density 

• Pay for tack as a separate pay item, to assure getting sufficient application of the residue 

• The pre-construction meeting should include discussion of the longitudinal joints 

• Pave from low to high: A 1/10” to 1/8” higher second lane creates a “shingle” effect and avoids roller bridging at the joint and impeding water run-off 

• Apply “joint enrichment”: such as VRAM under the joint, and/or a surface seal or over-banding of the joint area later to increase joint impermeability and joint longevity. 

In laydown consider the following: 

• Follow best practices to avoid mix segregation

• Use a string-line to produce a straight mat edge, allowing for a consistent overlap with the second pass 

• Apply adequate tack coat uniformly to the full width and preferable 4” to 6” wider than the paving lane 

• Use paver automation to get sufficient depth of material on the hot side of the joint 

• A “joint matcher” provides the best opportunity to place the correct depth of material on the hot side 

• Maintain a uniform head of material at the augers and extend the augers and tunnels to within 12” to 18” from the end gate to avoid segregation of the mix at the joint area. 

• Firmly seat the end gate on the existing pavement to control the proper overlap width 

• When closing a butt or notched-wedge joint, overlap by 1” (+/- ½”). At cut-back joints overlap approximately ½”. 

• Place enough material on the hot side so that after rolling the surface is about 1/10” to 1/8” higher than the cold side creating a “shingle” effect. (This avoids roller bridging and allows for unimpeded water run-off across the joint.) 

• Avoid luting or raking the overlap, and never “scalp and broadcast” excess material across the mat 

• Treating the cold side joint face: Consider the use of infrared joint heaters, which can improve joint density by 1 to 2% 

• Tack the face of the joint with JA, a PG grade binder, or a double application of tack coat. 

Avoiding segregation 

If the paving machine is operated correctly much of the physical and thermal segregation can be avoided. The two main areas of focus should be the receiving hopper and handling the mix at the augers. 

The hopper should at all times contain no less than 25% of its capacity filled with the mix. If the hopper does run low, do not fold in the hopper wings after every truckload. And if for some reason there are no trucks, do not slow down the paving or try to run all the mix out. It is better to make a hard stop with sufficient mix in the hopper and at the augers to retain the temperature while waiting for the next truck to arrive. If a much longer wait is expected, then finish up as you would do at the end of the day. 

How the mix is handled at the augers can significantly affect the quality of the joints. Keep the mix in front of the screed within +/- 1” of the height of the auger axle, so that half the auger blade is in the mix and half is above the mix. At this height, there is the least amount of segregation occurring. Starving the augers of mix is a guarantee to cause segregation in the mat and especially at the joints. 

When extending the screed to pave wider, have the auger extensions (whether bolted on or automatic) extended and the confining tunnel extended out to approximately 12” to 18” from the endplate. This will allow the mix to be properly spread in front of the screed and reach the joint area without segregating. If the augers and tunnel are not extended properly the mix is pushed up in front of the screed and will tend to cascade forward with the larger aggregate pieces rolling to the bottom of what is being paved. 

Rolling the longitudinal joints 

The final step in getting joints with good density is the rolling process. Options vary for both the unconfined edge of the first paving pass and the confined joint situation after closing up the joint with the second paving pass. The choice of joint rolling method will be based on how the joint area behaves. 

In the past, mixes tended to move laterally at an unconfined edge under the breakdown roller. With the more stable mixes used today this lateral movement is occurring less, depending on lift thickness and sufficient tack coat application. 

For the unconfined edge, the recommendation is to make the first roller pass with the roller drum overhanging the edge by 4” to 6”. If lateral yielding of the mix does occur, the recommendation is to adjust the breakdown roller pattern, placing it approximately 6” away from the edge. The second roller pass should then overhang the unconfined edge 4” to 6”. 

For the supported edge situation, the preferred method is to stay 6” to 12” away from the joint with the first pass. Then roll down the joint area with the second roller pass with about 6″ of roller overhang onto the cold material. This procedure has the risk of inducing a stress/shear crack induced by the edge of the roller drum during its first pass. If this happens, adjust the roller pattern to have the first pass overhang the joint onto the cold material by 4″ to 6″. 

Joint density specifications 

Whereas for many decades the Federal Aviation Administration (FAA) has specified the minimum required joint density in its P-401 Specification, highway agencies more recently have started to focus on the quality of the joints as well. The FAA P-401 specification is statistics-based and uses Percent Within Limits (PWL).

AASHTO recently surveyed the state DOTs on joint specs implementation, with the following result for the question if they have a longitudinal joints specification and if it is a density or method specification. 

Summary
Responses 40
Yes 27
No 6
No, but considering 7
Density 23
Method 6
Density Satisfaction (1 to 5: 5 is best) 3.4
Method Satisfaction (1 to 5: 5 is best) 3.6

At least 14 states that did not have a longitudinal joint (LJ) spec in 2009 now have, or are in the process of implementing an LJ spec. The survey shows improvement compared to the AI/FHWA finding ten years earlier that only 17 DOTs had a longitudinal joint density requirement. 

Suggestions for implementing a joint specification 

If an agency aims to improve the performance of the longitudinal joints, it is necessary to start testing the joint densities, then consider gradually implementing an incentive specification focused on getting better joints. Incentive provisions in the project specifications are a great way to stimulate contractor innovation on improving the density of the joints. 

Testing the joint density is done preferably with 6” cores, in such a manner that the core represents a 50/50 split of material from each side of the joint. For butt joints and trimmed back joints, the visible joint line should be in the center of the core. For notched wedge joints, plan to take the core at the halfway point on the width of the wedge to assure a 50/50 split of material from the top and the bottom of the wedge. Nuclear gauges are not preferred for joint density testing. They need a level surface for reliable test results. On a notched wedge joint, they give more weight to the density of the top wedge than to the bottom wedge material. By beginning to test the joint densities, a baseline is created for comparison and measuring future improvements. 

Once this baseline is established, the agency and industry can work together on what is achievable. For each agency the approach to implement an incentive joint density specification may be different. 

The AI/FHWA did come up with the following suggested approach: 

• Have a multi-year plan (versus all at once) 

• Encourage agency and industry participation and cooperation 

• Conduct joint industry-agency training (best practices, possible alternatives, etc.) 

• Establish a baseline of existing joint densities (randomly selecting projects to test) 

• Make incremental changes (trying different techniques, products, or specs.) 

• Evaluate progress by measuring densities to compare to baseline, and monitor performance 

When implementing the joint density incentive specification, take incremental steps: 

• First year require “report only” (calculate any bonus/ penalty without adding/subtracting dollars) 

• Start bonuses and penalties in the second year of implementation 

• Gradually increase density requirement to reach 90%, or possibly higher when it can be shown to be accomplished on a regular basis 

• Evaluate progress by comparing densities to the baseline 

Use a longitudinal joint density pay scale, such as; 

• 90% of TMD or better earns 100% pay, 

• 92% of TMD or better earns maximum bonus, 

• In between 90% and 92% of TMD gets a pro-rated bonus, consider sealing of the joint area 

• Less than 90% of TMD results in reduced payment and requires sealing of the joint area. 

Implementation example from Penn DOT 

Good examples are specification implementations by the DOTs of Pennsylvania, Maine and Connecticut, where over the span of several years the joint densities were improved. It took Penn DOT seven years to gradually implement their incentive specs for the joint density. They increased their statewide average joint density from 87.8% TMD in 2007 to 92.3% TMD in 2014.

The Penn DOT joint density specifications highlights are: 

• Both types of LJs are allowed (butt or notch wedge) 

• Joint Lot = 12,500’. One core every 2,500’, with 5 cores per lot. 

• Core location: for the butt joint directly over the visible joint, and for the notched wedge joint in the middle of the wedge. 

• Using statistics based on Percent within Limits (PWL). 

• Incentive starts at 80% PWL Disincentive starts at <50% PWL 

• Lower Specification Limit 

° For the years 2010-2013: 89% TMD 

° In 2014: 90% TMD; (In 2020 Penn DOT Section 405 now has the min. Spec limit at 91.0%) 

° Corrective action for < 88% TMD (typically over-banding or other surface treatment.) 

How to get from the ugly and the bad to the good joints 

States not specifying the quality of the longitudinal joints, typically get joint densities 2% to 5% lower than the mat densities. For the asphalt mat and the joints to be impermeable, the in-place air voids need to be less than 7-8%. Yet, good joint construction practices are typically achieved between 8-10% in-place air voids. 

Using a combination of the improvement options mentioned in this article, uniform impermeable longitudinal joints with densities of at least 92%TMD should be in our reach.

For further information regarding longitudinal joints visit asphaltinstitute.org/ engineering/longitudinal-joint-information 

Humer is a Senior Regional Engineer based in California