Pavement Management
Pavement_Management_250

De-Bonding of HMA Pavements/Layer Bonding

Print this pageEmail this to someoneShare on FacebookTweet about this on TwitterShare on Google+Share on LinkedInShare on RedditDigg this

Major Themes From Past Tack Coat Studies

This section attempts to identify some of the emerging themes associated with tack coats based on a qualitative meta-analysis of past research. Ideas that are corroborated amongst multiple studies are briefly discussed under individual headings while ideas emerging from single studies that have yet to be corroborated by other studies are listed next.

No Consensus Exists for Best Bond Testing Technique

In most cases interlayer bonding is tested as resistance to shear.This means that physical properties of the layer materials (e.g., gradation, maximum aggregate size, surface roughness) as well as tack coat adhesion are significant. Testing apparatus of this nature include the Superpave Shear Tester (SST), torque bond test, wedge-splitting test or other tests with a special shear box or other loading devices attached to a shear collar or box (e.g., Leutner test, Swiss LPDS tester, ASTRA test device, Nottingham shear box, Florida Department of Transportation shear tester). A majority of these tests load specimens in a strain controlled mode although stress controlled modes are also used on some. In other cases an attempt to isolate tack coat adhesion is made by using tests that pull apart a sample (e.g., ATTACKer™ or the UTEP pull-off test).

Each test has shown promise but none have been widely adopted. Because of the wide variety of test methods it is difficult to compare specific test values from study-to-study. Although field testing is often desired (in order to determine the quality of layer bonding) there are few tests (ATTACKer™ and the UTEP pull-off test) designed for quick field use.

Layer Bond Strength is Inversely Proportional to Temperature

Laboratory studies that varied test temperature (Sholar et al., 2002[1]; Deysarkar and Tandon, 2005[2]; Canestrari et al., 2005[3]; West et al., 2005[4]) have all concluded that as test temperature increases layer bond strength decreases. West et al. (2005[4]) found that, “On average, bond strengths were 2.3 times greater at 50ºF compared to 77ºF; and the bond strengths at 140ºF were about one sixth of the bond strength at 77ºF.” Most conclude that at higher temperatures tack coat adhesion becomes relatively insignificant and most measured shear resistance comes from layer surface roughness. This implies shear resistance at layer interfaces in the field are likely to be lowest during hot days.

Layer Bond Strength is Proportional to Normal Stress

Laboratory studies that varied the normal pressure applied to a sample (Uzan, et al., 1978[5]; West et al., 2005[4]) have all concluded that as normal pressure increases layer bond strength increases. This implies that although a heavier load is more likely to produce higher horizontal stresses making slippage failure more likely, it is also likely to provide a higher normal stress, which increases resistance to slippage failure.

Layer Surface Roughness is a Larger Contributor than Tack Coat Adhesion in Resisting Shear

Studies that compared tack coated surfaces to uncoated surface (Canestrari et al., 2005[3]; Mohammad et al., 2005[6]) generally found that tack coat improved bond strength somewhat. Mohammad et al. (2005[6]) tended to show that at 77ºF (25ºC) tack coats increased bond strength by no more than about 1/3 and in some cases decreased it (Table 1). At 131ºF (55ºC) tack coat had either no effect or a negative effect on bond strength; the exceptions being the two tack coats that were latex modified (Table 2).

 

Table 1: Summary of Selected Results taken from Mohammad et al. (2005[6])for Bond Strength Tests at 77ºF (25ºC).

 

Table 2: Summary of Selected Results taken from Mohammad et al. (2005[6])for Bond Strength Tests at 131ºF (55ºC).

 

Additionally, Tashman et al. (2006[11]) found that for a milled surface “the absence of tack coat did not significantly affect the bond strength at the interface”, which suggests that a rough milled surface provides significantly more shear resistance than a tack coat can add. Findings from Cooley (1999[7]) and Sholar et al. (2002[1]) support this view.

Importantly, there may be some difference between shear resistance as measured in the laboratory and effective layer bonding in the field. Typically, samples prepared in the laboratory with no tack coat show substantial shear resistance (Uzan et al., 1978[5]; Mohammad et al., 2005[6]; Kruntcheva et al., 2006[8]). However, experiments using field cores (Tayebali et al., 2004[9]; West et al., 2005[4]) found that layers without tack tended to de-bond and thus could not even be tested for shear resistance. To speculate, samples taken from the field may have been subjected to additional variables that helped cause de-bonding such as compaction with construction equipment, non-uniform application rate, and torsional/normal forces created by the core drilling machine. If this speculation is true then laboratory prepared samples may not be adequately reproducing a key component of bond failures in the field.

Gradation of Surrounding Layers Influences Bond Strength

Coarse gradations provide more shear resistance than fine gradations, however smaller nominal maximum aggregate size (NMAS) mixes benefit more, on a percentage basis, from tack coat application. West et al. (2005[4]) and Sholar et al. (2002[1]) both reached these general conclusions.

Tack Coat Application Rate is Somewhat Related to Bond Strength

Based on an evaluation of the studies that varied application rate and/or included a sample with no tack coat applied (Uzan et al., 1978[5]; Buchanan and Woods, 2004[10]; Mohammad et al., 2005[6]; Kruntcheva et al., 2006[8]) the following conclusions can be drawn:

  • Straight asphalt (e.g., PG 64-22) appears to be relatively insensitive to application rate within reason. There is no maximum bond strength but rather bond strength remains relatively constant over a wider range of application rates.
  • Some emulsions tend to have an optimum application rate. Mohammad et al. (2005[6]) reports this as around 0.09 L/m2 (0.03 gal/yd2 – residual application rate of about 0.02 gal/yd2) for the CRS-2P examined.
  • Some emulsions tend to be no better or even worse than no tack coat at all.
  • Emulsions containing polymer modified asphalts tend to have higher bond strengths than those that do not when applied at the optimum rate.
  • Excess tack coat (high application rates) usually produces weaker bonds. A generalization of “high application rate” might be any rate greater than 0.10 gal/yd2 (about a 0.06 gal/yd2 residual rate).

These general ideas seem to be consistent with individual study findings that may, on initial impression, seem to go against them. For instance, Kruntcheva et al. (2006[8]), when using a 0.33 L/m2 (0.10 gal/yd2) application rate of K 1-40 tack coat (British) concluded that “A dry and clean surface with no tack coat has similar properties to the same interface with a standard quantity of tack coat.” However, If results from Mohammad et al. (2005) for 25ºC (77ºF) are interpolated between tested application rates of 0.23 L/m2 (0.07 gal/yd2) and 0.45 L/m2 (0.14 gal/yd2) the results are similar: PG 76-22M, CRS-2L, CRS-2P, SS-1, CSS-1, SS-1h and SS-1L all showed no improvement over no tack coat and PG 64-22 showed only marginal improvement at 0.23 L/m2 (0.07 gal/yd2).

Influence of Curing Time is Not Well Corroborated

Some studies (West et al., 2005[4]; Tashman et al., 2006[11]) suggest that paving over unbroken tack coat (an emulsion that still contains water and has not cured) does not adversely affect bond strength while other studies (Hachiya and Sato, 1997[12]; Buchanan and Woods, 2004[10]) suggest that longer cure times improve bond strength. Additionally, Shahin et al. (2002[13]) found that after paving bond strengths tended to increase over time. This area needs more investigation. Of note, both studies that found paving over unbroken tack coat to be okay used field samples while both studies that found longer curing time improves bond strength used laboratory samples. Finally, it may be that bond strength increases with time regardless of when the actual paving occurs.

Emulsion Dilution Tends to Reduce Bond Strength

When emulsion tack coats are diluted their bond strength decreased. Deysarkar and Tandon (2005[14]) found an “Increase in dilution reduced tack coat strength even though the similar residual amounts were applied”. Buchanon and Woods (2004[10]) found that dilution also reduced bond strength, however they measured application rates and not residual rates so their diluted samples actually contained less residue, which they believe contributed to the lower bond strengths. These results, especially those from Deysarker and Tandon (2005[14]), could infer that diluting tack coat emulsions is a poor practice because it tends to weaken bond strength for a given residual rate.

Field Performance May Not Be Adequately Modeled In The Laboratory

Studies that examined field samples (West et al., 2005[4]; Kulkarni et al., 2005[15]; Deysarkar and Tandon, 2005[14]; Rodrigo et al., 2005[16]; Canestrari et al., 2005[3]) found widely varying bond strengths in the field and, in the case of West et al. (2005) found significantly lower bond strengths in the field.

Construction Factors Have a profound effect on tack coat

Actual tack coat application rates (versus target rates), construction vehicle tire pickup, weather, surface cleanliness and application uniformity are all construction issues that can affect tack coat performance. Few, if any, of these factors are measured and archived, making construction a difficult influence to quantify. At best, some basics are known. For instance, Hayachi and Sato (1997[12]) and Collop et al. (2003[17]) found that, in general, dirty surfaces result in lower bond strengths.

Actual application rates can be measured, however, accuracy of these measurements is questionable. While West et al. (2005[4]) found ASTM D 2995 an effective measurement method, Tashman et al. (2006[11]) report measured residual application rates that were fairly consistent regardless of the target residual rate using the same ASTM D 2995. Further, West et al. (2005) reported application rates significantly different than target rates on 3 of 6 field projects measured by ASTM D 2995. Coincidentally, all 3 of these projects were CRS-2 emulsion tack coats, while those that most closely match target rates were straight paving grade asphalt tack coats (2 projects) and 1 special heavy application of a polymer modified emulsion. Tashman et al.’s (2006[11]) project used a CSS-1 emulsion. One possible cause is that these emulsions lost water weight before they were weighed making the measurement inaccurate. It remains to be seen whether the mismatch between target and actual application rates is caused by improper/variable tack truck application rates, ASTM D 2995 measurement inadequacies.

Mechanistic models show reduced bond strength can lead to early fatigue failure

Studies using mechanistic models usually employ a type of layered elastic model (e.g., BISAR, Everstress) and vary the slip parameter between layers. In general, studies by Shahin et al. (1986[18]) and Willis and Timm (2006[19]) suggest that loss of bond results in reduced fatigue life; an expected result from layered elastic theory. Willis and Timm (2006[20]) present substantial evidence showing that structural sections at the National Center for Asphalt Technology (NCAT) test track de-bonded, which they speculate led to early cracking. Willis and Timm’s analysis with WESLEA corresponds reasonably well with Shahin et al.’s (1986[18]) analysis with BISAR. In addition, Willis and Timm (2006[20]) showed strain gauge data that correlated well with a loss of bond WESLEA model.

There is no information on minimum adequate bond strength

No study offers any compelling evidence or speculates on what constitutes adequate bond strength to prevent or at least minimize the chances of de-bonding. Efforts by West et al. (2005[4]) produced the best indication of typical field bond strengths. When tested in a shear collar device at a 2 inch/min shear rate they found a distribution of bond strengths with a mean of about 100 psi. From this distribution they suggest a bond strength less than about 50 psi could be considered poor, while one above 100 psi could be considered good. The actual level at which de-bonding becomes likely is still unknown.

Tack Coat Themes Substantiated By One Study

Water on broken tack reduces bond strength.Sholar et al. (2002[1]) found that water (in the form of simulated rainwater) reduces bond strength. The tack coats used were RS-1 (55 percent minimum residual and 60 minimum penetration), RS-2 (63 percent minimum residual and 100-200 minimum penetration range) as well as two other projects where the tack coat types were not stated.

Latex modification improves high temperature bond strength but is not significant at lower temperature.Mohammad et al. (2005[6]) showed the highest percentage of bond strength gain for tests at 55ºC (131 ºF) with the two tack coats containing latex modified asphalt cement. However, for tests at 25ºC (77 ºF) the emulsions containing latex modified asphalt cements were not significantly different than the same emulsion without the latex.

More tack does not overcome dirty surface. Collop et al. (2003[17]) found that for dirty surfaces on specific HMA mixes “…extra tack coat did not compensate and the interface shear strengths were significantly reduced.” This implies that surface cleanliness affects bond strength.

Discussion

Of the general ideas listed previously, early fatigue failure mechanisms and construction factors warrant further discussion.

Fatigue Failure: De-bonding cracking

Layered elastic models (Shahin et al., 1986[18]; Willis and Timm, 2006[20]) and evidence (Willis and Timm, 2006[20]) suggest de-bonding that leads to early fatigue cracking (termed “de-bonding cracking”) can and does occur. Slippage cracking, the underlying concern in many studies, is not an abundant HMA pavement distress outside of runways, taxiways, intersections and other braking/accelerating areas. A majority of the HMA placed is not subject to excessive braking, acceleration and turning and thus is generally free of slippage cracking. Therefore, de-bonding cracking, which could occur anywhere, could potentially be much more prevalent on highways and thus represent a far greater concern. Since the current trend is to build thick HMA pavements (i.e., perpetual pavements) with many layers it seems that perhaps the critical item in pavement design and construction has shifted from ensuring overall adequate thickness (drainage and subgrade concerns notwithstanding) to ensuring adequate bonding between layers so that the pavement performs as a whole.

It is likely that evidence of de-bonding cracking and its extent already exists in pavement management system records. However, such cracking is probably indistinguishable from other forms of surface cracking (i.e., top-down cracking and classical bottom-up fatigue cracking) and therefore cannot be recorded separately. It could be that an investigation on the order of Wills and Timm’s (2006[20]) is necessary to identify de-bonding cracking, which would make it impractical to identify in the field. It could also be that a simpler indicator exists but has yet to be discovered.

Construction/field issues dominate bond performance

While variables such as target application rate and tack coat type can be important, it appears the literature is converging on acceptable answers (at least in a laboratory setting). The overriding variables are likely construction-related: actual application rate from the distributor truck vs. target rate, residual rate, cleanliness of site and weather. While these general ideas are known to be important associated data is usually not collected as thoroughly or systematically as other paving data (e.g., density, gradation, asphalt content). Without such data, knowledge of construction impacts comes from speculation, informed opinion or anecdotal evidence.

Additional References

Canestrari, F.; Ferrotti, G.; Parti, M.N. and Santagata, E. (2005). Advanced testing and characterization of interlayer shear resistance. In Transportation Research Record No. 1929, pp. 69-78.

Gomba, S.; Liddle, J. and Mehta, Y. Evaluation of Interlayer Bonding in Hot Mix Asphalt Pavements. International Journal of Pavement Engineering, v. 4, no. 1-2, pp. 13-24.

Kennedy, C.K. and Lister, N.W. (1979). Experimental studies of slippage. In The Performance of Rolled Asphalt Road Surfacing. ICE, London, pp. 31-56.

Miro Recasens, R.; Martinez, A. and Pee Jiminez, F. (2005). Assessing heat-adhesive emulsions for tack coats. Proceedings of the Institution of Civil Engineers: Transport, v. 158, no. 1, pp. 45-51.

Mohammad, L.N. Raquib, M.A. and Huang, B. (2002). Influence of asphalt tack coat materials on interface shear strength. In Transportation Research Record No. 1789, pp. 56-65.

Mrawira, D. and Yin, D. (2005). Field Evaluation of the Effectiveness of Tack Coats in Hot Mix Asphalt Paving. Paper presented at the 2006 Transportation Research Board Annual Meeting, Washington, D.C.

Peattie, K.R. (1979). The incidence and investigation of slippage failures. In The Performance of Rolled Asphalt Road Surfacing. ICE, London, 1979, pp. 3–15.

Romanoschi, S.A. and Metcalf, J.B. (2001). Characterization of asphalt concretelayer interfaces. In Transportation Research Record No. 1778, pp. 56-65.



Footnotes    (↵ returns to text)
  1. Preliminary Investigation of a Test Method to Evaluate Bond Strength of Bituminous Tack Coats. Research Report FL/DOT/SMO/02-459. Florida Department of Transportation, Gainesville, FL.
  2. Field Evaluation of Tack Coat Quality Measurement Equipments. International Journal of Pavement Engineering, v. 4, no. 1-2, pp. 25-37.
  3. Temperature effects on the shear behaviour of tack coat emulsions used in flexible pavements. International Journal of Pavement Engineering, v. 6, no. 1, pp. 39-46.
  4. Evaluation of Bond Strength Between Pavement Layers. NCAT Report 05-08. National Center for Asphalt Technology, Auburn, AL.
  5. Investigation of adhesion properties between asphalt-concrete layers. Asphalt Paving Technology, vol. 47, pp. 495–521.
  6. Investigation of the Behavior of Asphalt Tack Interface Layer. FHWA/LA.04/394. Louisiana Transportation Research Center, Baton Rouge, LA.
  7. Evaluation of the Influence of Tack Coat Construction Factors o the Bond Strength Between Pavement Layers. WA-RD 645.1, Washington State Department of Transportation, Olympia, WA.
  8. No-Tack Inlay on Miled (sic) Surface: Project Report. NCAT Report 99-02. National Center for Asphalt Technology, Auburn, AL.
  9. Properties of asphalt concrete layer interfaces. Journal of Materials in Civil Engineering, v. 18, no. 3, pp. 467-471.
  10. A Mechanistic Approach to Evaluate Contribution of Prime and Tack Coat in Composite Asphalt Pavements. FHWA/NC/2004-05. North Carolina Department of Transportation, Raleigh, NC.
  11. Field Tack Coat Evaluator (ATACKer™). Report No. FHWA/MS-DOT-RD-04-168. Mississippi Department of Transportation, Jackson, MS.
  12. Effect of tack coat on bonding characteristics at interface between asphalt concrete layers. Proceedings of the 8th International Conference on Asphalt Pavements, v. 1, pp. 349-362.
  13. Shahin, M.Y.; Kirchner, K.; Blackmon, E.W. and Tomita, H. (1986). Effect of Layer Slippage on Performance of Asphalt Concrete Pavements, In Transportation Research Record, No. 1095, pp. 79–85.
  14. Deysarkar, I. and Tandon, V. (2005). Field Evaluation of Tack Coat Quality Measurement Equipments. International Journal of Pavement Engineering, v. 4, no. 1-2, pp. 25-37.
  15. Evaluation of tack coat bond strength for mixtures containing baghouse fines. International Journal of Pavement Engineering, v. 6, no. 3, pp. 147-162.
  16. Recasens, Rodrigo Miró (Technical University of Catalonia, Jordi Girona 1-3, Modulo B1, 08034 Barcelona, Spain); Martínez, Adriana; Jiménez, Félix Pérez Source: Transportation Research Record, n 1970, p 64-70, 2006, Bituminous Paving Mixtures
  17. Assessment of bond condition using the Leutner shear test. Proceedings of the Institution of Civil Engineers: Transport, v. 156, no. 4, pp. 211-217.
  18. Effect of Layer Slippage on Performance of Asphalt Concrete Pavements, In Transportation Research Record, No. 1095, pp. 79–85.
  19. Willis, J.R. and Timm, D.H. (2006). Forensic Investigation of a Rich-Bottom Pavement. NCAT Report 06-04. National Center for Asphalt Technology, Auburn, AL.
  20. Forensic Investigation of a Rich-Bottom Pavement. NCAT Report 06-04. National Center for Asphalt Technology, Auburn, AL.

 

 

 

close citation

Thanks for sharing De-Bonding of HMA Pavements/Layer Bonding.