Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Paper Presented to
Transportation Research Board
78th Annual Meeting
by
August 16,1999
Way 2
Phone: 602-712-8085
Fax: 602-712-8138
Email: gway@dot.state.az.us
large scale Asphalt Rubber (AR) test project in Flagstaff, Arizona on the very heavily traffiked
Interstate 40. The purpose of the test project was to detennine whether a relatively thin overlay
with AR could reduce reflective cracking. Asphalt Rubber is a mixture of 80% hot paving grade
asphalt and 20% ground tire rubber. This mixture is also commonly referred to as the asphalt
rubber wet process or McDonald process. The overlay project was built on top of a very badly
cracked concrete pavement, which was in need of reconstruction. The asphalt rubber overlay has
perfonned beyond the original expectation. After nine years of service the overlay is still
virtually crack free, with good ride, virtually no rutting or maintenance and good skid resistance.
The benefits of using asphalt rubber on this project represents about $18 million dollars in
construction savings and four years less construction time. Strategic Highway Research Program
SPS-6 test sections built in conjunction with the project further illustrate the very good
perfonnance of the asphalt rubber. Results of this project have led to widespread use of asphalt
rubber hot mixes throughout Arizona. Based upon this work over 3,333km (2,000 miles) of
KEYWORDS
INTRODUCTION
The Arizona Department of Transportation (ADOT) maintains about 12,333km (7,400 centerline
miles) of highways, of which approximately 500km (300 miles) are Portland Cement Concrete.
A twenty year old, 16.7km (ten mile) section of concrete pavement on Interstate 40 near
Flagstaff, Arizona suffered severe deterioration warranting reconstruction (Figure 1). Flagstaff is
a mountainous area of 2134m (7,000 foot) elevation with an Alpine like climate. High
temperatures in the summer are a pleasant 27C (80F) with winter time lows of -31C (_23 F).
Overall rainfall is 0.7m (28 inches) per year and winter snows average over 2,500mm (100
inches) per year. Interstate 40 cuts through this mountainous area and is built on soils and rock
of generally poor engineering quality. The current traffic is over 20,000 vehicles per day. The
traffic loading has rapidly increased over the years and is quite heavy, with presently over 35
Interchange (Mile Post Marker 195) and the Walnut Canyon Interchange (Mile Post Marker 205).
The concrete pavement originally built in 1969 (I) consisted of two 3.65m (twelve foot)
eastbound and westbound lanes of 200mm (8 inch) thickness with 25mm (3 inch) thick hot mix
asphalt (HMA) shoulders. The total width of the pavement surface is 11.6m (38 feet). The
concrete pavement is non-reinforced except for reinforcing steel that tie the two concrete lanes
together longitudinally. The concrete pavement had sawed and skewed joints that were randomly
spaced 4.6m (15 feet) apart. Joints originally were not sealed. Underneath the concrete
pavement is a 175mm (7 inches) cement treated base (CTB), built out of a frost susceptible
limestone aggregate. Beneath the CTB is an unbound subbase, 175mm (7 inches) thick
composed of the same limestone. Underlying the subbase is a clayey cinder material
representative of the natural ground. The supporting base layers are the same under the shoulders
except that 100mm (four inches) of cinder base was placed between the asphaltic concrete and
the CTB.
The concrete pavement began to fail in 1974, just five years after construction. The failure began
as large comer cracks and progressed to transverse cracks and severe spalling at the transverse
joints. Maintenance sealed the cracks and patched spalls as best as possible and in the process
As the pavement deteriorated the cracking continued to increase. From 1980 to 1989 the percent
cracking increased from one percent to nine percent (Figure I). The ride quality also suffered,
even though maintenance repaired the worst locations. In 1980 the ride was 1500mm!km (100
2220mmlkm (148 inches/mile) is considered objectionable. Twenty percent of the project miles
exceeded the objectionable ride level before overlaying. In general the performance of the
concrete pavement was very poor. In addition the traffic loadings over the course of time
increased dramatically. In 1969 the annual 80kn (18 kip equivalent single axle loads (ESAL's))
was about 120,000. In 1990 it was 1,600,000 and by 1999 it is presently 2,500,000 which is
Design began in 1988 and reconstruction was very strongly considered. The adjacent 6.7km
(four mile) section of 1-40 (MP 191-195) had experienced the same type and degree of failure
and had been reconstructed. The reconstruction (2) involved the building of detours and closing
the interstate in one direction for one year. Thus two reconstruction projects were built over a
two year period in the years of 1985 and 1986. The reconstructed sections were composed of
200-275mm (8-11 inches) of HMA on top of 150mm (6 inches) of permeable asphalt bound base
and 100-175mm (5-7 inches) of drainable aggregate base. The pavement structural section was
placed on top of a geotextile separation fabric to keep the wet clay from pumping into the
aggregate base. A complete edge drain system with slotted pipe was also installed. The total
Way 7
cost of construction of the two projects was about $15 million. In order to reconstruct the
Flagstaff 1-40 (MP 195-205) project it would have been necessary to build the project in four
phases, since detours of more that 8.3km (five miles) were not allowed. In addition the
construction would have taken four years to complete, however it was strongly questioned
whether maintenance could maintain the pavement for that long, given its very poor condition
(Figure I). The overall cost of reconstruction was estimated to be at least $30 million.
It was finally concluded that the project could not be reconstructed. In addition, due to money
and time constraints the project would have to be overlaid within a tight budget and work
completed in one construction season of about six months. Various overlay strategies were
considered including many different overlay thicknesses, use of a fabric interlayer, asphalt rubber
interlayer, various mixes, edge drains and cracking and seating. Each alternative was discussed
and reviewed at both the central office and the district office. In addition the project was also
selected (3) as a Strategic Highway Research Project (SHRP) Specific Pavement Studies (SPS-6)
to test various overlay and rehabilitation strategies on concrete pavement. With an asphalt rubber
binder the selected project design strategy represented a test of whether a relatively thin
pavement overlay could control reflective cracking. Although the design was for ten years
virtually everyone involved in the project considered this to be at best a six year design given the
thin overlay design section and the very poor condition of the concrete. After much internal
discussion and debate the final pavement design section included edge drains, crack and seat of
Way 8
the concrete pavement, a l25mm (five inch) overlay composed of a 75rnm (3 inch) conventional
dense hot mix asphalt (HMA) and 50rnm (2 inch gap graded) asphalt rubber mix (AR-AC). An
asphalt rubber open graded friction course (AR-OGFC) l2.5mm (one-half inch) thick was placed
as the final wearing surface on the two travel lanes (Figure 2). The asphalt rubber used on the
project was specified to be 80%, AC-lO asphalt binder, hot reacted with 20% ground tire rubber.
The overlay thickness and layer placement was discussed right up to the final days before the bid
advertisement. The discussion centered around whether the AR-AC or the HMA should be
placed directly on top of the broken concrete. Previous experience with asphalt rubber
interlayers indicated that the AR-AC should be placed on top of the cracked surface before
overlay. The other position of placing the AC on top of the broken concrete seemed more in
keeping with its role as a leveling and structural layer which would probably crack very soon
after construction (first winter). The top overlay of AR-AC and AR-OGFC would then perform
not only as the leveling and structural layer but also as the final flexible layer capable of resisting
reflection cracking. The project was designed in this manner, however, a test section was built
with the 50rnm (two inch) AR-AC on top of the broken concrete pavement and a 50rnm (two
The use of the AR-OGFC as the final wearing course had been previously tried on a concrete
pavement in Tucson, Arizona. Its performance in Tucson on Interstate 19 has been very good
and it was always considered as the most appropriate wearing course. Typically in 1990 OGFC's
Way 9
in Arizona were placed with conventional AC-20 (PG-64-16) at a six percent ,maximum binder
content which is normally the maximum the rock can hold before an excess drains off.
Asphalt rubber is over ten times more viscous than AC-20 (PG-64-16) at hot mlXlng
temperatures of 177C (350F) and thus can be applied to an OGFC rock gradation at a rate of
nine to ten percent by weight of the mix. This extra coating thickness increases durability and
slows down aging. In addition the thick rubbery coating helps to retard reflection cracking.
In addition to the material related design issues, constructability issues were addressed in the
design by meeting with district construction personnel. It was agreed early on that the project
construction phasing should be such that the project could be completed in one summer paving
season. To do this the specifications required the contractor to begin with the edge drain. After
sufficient edge drain was completed the crack and seat would start. Crack and seat operations
had to be done at night to avoid interfering with the edge drain installation and to be in sync with
the overlay operation. The crack and seat drop height and spacing was checked by deflection
testing and cores to verifY the quality of the work. In addition an incentive of$15,000 per day to
finish paving the overlay ahead of schedule was included. The maximum incentive was set at
$450,000 with a due date of October I, 1990. With all the numerous design issues described in
plans and special provision specifications the project was bid in April, 1990.
Way - 10
CONSTRUCTION
The project was awarded in May 1990. The low bid was $10,783,486 which was $190,000
(17%) lower than the state estimate. The quantity of asphalt rubber binder was 3.9Mkg (4,316
tons) the largest amount ever bid by ADOT in a single contract at that time. The unit cost of the
asphalt rubber mix was $0.05/kg($45 per ton) and the HMA was $0.025/kg ($23 per ton).
Construction began in June, 1990 with the installation of the edge drain and paving of some of
the ramps and crossroads with HMA. In late June the crack and seat night operation began.
Cracking of the pavement was accomplished with a guillotine hammer. Several drop heights
from 0.6m (two feet) to 1.2m (four feet) were tried. It originally appeared from cores that 1m
(40 inch) drop height was adequate although ultimately a l.2m (four feet) drop height was
adopted to be certain that the slab was cracked full depth. A 0.6m (two feet) spacing was
originally set and later reduced to 0.5m (18 inches). Seating was accomplished with a 45,359kg
(50 ton) rubber tired roller. Generally one to two roller passes was adequate to seat the concrete.
Paving of each nights crack and seat followed the next day.
The 75mm (three inch) HMA overlay consisted of a dense graded HMA with 4.6 percent asphalt
by weight of mix of an AC-20 grade (Table I). No paving problems occurred, although later
after traffic was applied about a dozen small areas failed. Each area was dug out and replaced
with new HMA. During removal it was noted that the concrete had been previously in very poor
condition. Apparently the crack and seat operation totally destroyed the concrete pavements
After the HMA overlay was placed and opened to traffic some minor roughness was noticed.
This roughness appeared to be related to the seating process, since as more traffic was applied the
pavement became noticeably smoother. Following the overlay the 50mm (two inch) AR-AC was
placed. AR-AC was a slightly finer gap graded mix with the largest retained aggregate of 9.4mm
(3/8 inch) (Table I), whereas the AC overlay was 19mm (3/4 inch) mix. For the AR-AC mineral
aggregate the minimum required sand equivalent was 55 percent and the crushed faces 70
percent. Both of these values are higher than typically required for AC mineral aggregate. The
AR-AC contained 6.5 percent asphalt rubber by weight of mix. The asphalt rubber consisted of
about 20 percent vulcanized ground rubber sized from 100 percent pass the 2.00mm (number 10
sieve), to no more than five percent passing the 75!lm (200 sieve). The rubber was mixed and
reacted with an AC-l0 asphalt binder and kept at a 177C (350F) temperature for one hour, such
that the mixture viscosity was between 1.5-4.0 pascal seconds (1,500 to 4,000 centipoise). The
resultant asphalt rubber was pumped into the drum mixer hot plant at about 177C (350F). The
AR-AC mix in the field had to be at least 135C (275F) at time of compacting. Compaction
was achieved with steel rollers. Rolling was not allowed after the mix cooled to 104C (220F).
No placement problems occurred although at such elevated temperatures some smoke does
occur. Asphalt rubber content was checked with a nuclear asphalt content gauge. After some
Following the AR-AC placement, an AR-OGFC was applied as the final wearing course on the
two travel lanes. The aggregate grading for the AR-OGFC (Table 1) is very similar to the
standard OGFC with 100 percent passing the 9.4mm (3/8 inch) sieve and no more than 9 percent
passing the 2.36mm (number 8) sieve (Table 1). A maximum of only 2.5 percent could pass the
751lm (200 sieve). The aggregate properties had to conform to the same quality criteria as the
AR-AC. The asphalt rubber was the same as for the AR-AC. The percent asphalt rubber was 9.0
percent by weight of the mix, typically an OGFC has only 6.0 percent binder. The AR-OGFC
was mixed, placed and compacted in the same manner as the AR-AC except compaction effort
ceased at 93C (200F) instead of 104C (220F). After cooling the AR-OGFC was opened to
traffic with no problems. Virtually all the paving was completed by the end of September. By
working Saturday and Sunday as well as 10 and 12 hour shifts the contractor was able to finish
early enough to receive an incentive payment of well over $400,000. In general the construction
was rushed and hectic, however, the final product appeared to be of very good quality. In
addition to the routine project work, 4.6km (2.76 miles) of SHRP experimental sections and a
PERFORMANCE
Since paving was completed in October 1990 nine years of very good service have occurred, far
more than expected. Each year data for each mile was recorded by the ADOT Pavement
Management Section (PMS). In addition this section also retrieves the traffic loading data and
Way 13
the pavement surface maintenance data from other computer files. Table 2 shows the average
yearly summary of pavement performance data for the 33.3km (20 miles) of pavement. In
addition the data is also shown representative of the pavement condition before overlay. As can
be seen virtually no reflective cracking has occurred since construction. The ride is expressed in
inches per mile units and is indicative of a very smooth riding surface. The rut depth is virtually
not existent. Skid resistance as measured by the Mu meter is good. Maintenance costs have been
For comparative purposes the adjoining section of Interstate 40 referred to as the Riordan project
(milepost marker 191 to 195) was reconstructed in 1985 and 1986 under two separate
construction projects. The old very badly cracked concrete pavement was removed and crushed
for aggregate base. The existing cement treated base had turned to a mushy poor quality material
and was wasted. The wet clay sub grade was regraded and an edge drain system constructed. A
geotextile separation layer was placed followed by 100-175mm (four to seven inches) of
aggregate base. The aggregate base was then covered with 150mm (six inches) of a asphalt
permeable base. The final HMA was 200-275mm (eight to eleven inches) in thickness (Figure
3). All of the HMA was built with an AC-20 binder. The two construction projects required
detours and two years of construction. The total cost for about four miles of new pavement was
$14 million. For comparative purposes Table 3 shows the distress history of these two Riordan
reconstruction projects as measured by the ADOT PMS. Cracking of the 200-275mm (eight to
eleven inch) pavements began in the second year and proceeded to a degree requiring
VVay 14
rehabilitation by the tenth year. Other pavement performance measures indicated that the
Another comparison of how well the ADOT AR section have performed can be seen from
reviewing Strategic Highway Research Project test sections built as part of the Flagstaff overlay
project. The Flagstaff project was selected as an SPS-6 concrete rehabilitation test project.
Table 4 shows the layout of the test sections. ADOT decided to build several of its own test
section to be included in the study. The test section are all located in the east bound direction of
Interstate 40 from mile post marker 202.16 to 204.26. SHRP test section 040608 is a 200mm
(eight inch) overlay of the concrete pavement. Test section 040607, 040606 and 040659 are all
100mm (four inch) overlays of the concrete pavement. The 200mm & 100mm (eight and four
inch) overlay test sections were all built using the ADOT HMA base mix with 4.6% asphalt with
AC-20 binder. Test section 040661 was an ADOT test section composed of a two inch overlay
of ADOT HMA base mix followed by a 50mm (two inch) overlay with the AR gap graded mix
with 6.5% binder. Table 4 shows the cracking by year in lineal feet of cracking (5). These crack
measurements were estimated from the LTPP file and from the crack maps and/or ADOT photos
and field reviews. Again this table shows how well the AR sections are performing compared to
the other conventional overlays. Figure 4 shows a photo taken from 19.8m (65 feet) above the
test section by using an ADOT cherry picker (high boom) truck. It clearly shows the degree of
cracking in the 100mm (four inch) overlay and the lack of cracking in the AR test section.
Way 15
CONCLUSION
Based upon the outstanding performance of this project as well as many others, ADOT routinely
uses AR gap graded and open graded mixes throughout the state. Over 3,333km (2,000 miles) of
AR projects have been built since 1990. Figure 5 is a map showing those ADOT system miles
paved with AR. In addition the performance of all AR projects has been tracked through the
ADOT PMS, Figure 6 shows the percent cracking for those overlay projects built with and
without AR since 1988 (6). As shown the percent cracking is much less over time than for
conventional overlays without AR. In addition Figure 7 shows less routine maintenance for AR
projects than for conventional overlays. It should also be noted that the average thickness of the
AR projects is 42.5mm (1.7 inches) and for conventional mixes it is 137.5mm (5.5 inches). Thus
even though AR mixes cost about twice as much per ton, they are generally placed half as thick
and crack at a rate less than one fourth than that of conventional mix. The overall AR
performance seems to be in agreement with findings in 1977 (7) that AR in cold weather can
withstand about five times the strain before rupture than can asphalt.
Administration and the Strategic Highway Research Program have designed and constructed an
extremely valuable research project. Field performance in nine years of service is providing
experience indicating the effectiveness of a thin flexible overlay and wearing course composed of
asphalt rubber. Results of this work are being incorporated into paving projects, throughout the
Way 16
state. Continued field performance measurements provide the data necessary to develop rational
ACKNOWLEDGMENT
The author is indebted to the good work of Jackie Hostetler in typing the paper. Also, Ali Zareh
REFERENCES
1. Way, George "FlagstaffI-40 Overlay Project", De., 11-13, 1991 4R Conference & Road
Design and Research Plan for experiment SPS-6: Rehabilitation of Jointed Portland
-;.
Way 17
379-11.
5. ATRC Research Notes, "Concrete Pavement Rehabilitation Update", ADOT, Dec., 1994,
SPS-6 Note.
6. Way, George "Asphalt Rubber - The Arizona Experience", Feb., 3,1999 Asphalt Rubber
7. Green, E. L., Tolonen, William J., "The Chemical and Physical Properties of Asphalt
LIST OF FIGURES
FIGURE 4. Overhead cherry picker view of overlay with and without asphalt rubber.
FIGURE 5. Arizona State Highway Map showing asphalt rubber overlays built since 1988.
FIGURE 6. Percent cracking over a ten year period with and without asphalt rubber.
LIST OF TABLES
TABLE I. 1990 mix design, HMA, AR gap graded and AR open graded.
Typical Sectioll
1-40 Flagstaffproject
38' ..
Typical Section
1- 40 RIORDAN Project
38' 20'
16' ......
I 22'
Centerline ----..f
Drain
Tack Coats
3" AC (3/4")
~~~~!=-~3~" AC (3/4")
t 3" AC ( 3 /4 " )
6" BTB
Prime Coat
4" AB
Engineering
Fabric Subgrade
Way - 22
FIGURE 4. Overhead cherry picker view of overlay with and withont asphalt rubber.
1-40 SHRP test sections 4" conventional AC (left) and 2" Asphalt-Rubber Hot Mix placed in
FIGURE 5. Arizona State Highway Map showing asphalt rubber overlays built ~ince 1988.
tlJf,ij
!JTAfE flltk>fiIiINi ;'{STf?.t
d
MUlIJ/I",
Way 24
FIGURE 6. Percent cracking over a ten year period with and withont asphalt rubber.
14
12
10
""
=
0
"..
0:;
8
U
~
~~
.
~
I'.
6
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Years
1400
1200
.: 1000
~
=
...l
~
..'"
..!!
800
'0
Q 600
400
200
0
0 1 2 3 4 5 6 7 8 9 10
Years
-o-Overlays/lulays Neat Asphalt - 0 - AR-ACFC
~ ~
Way " 26
Test Section Begin Mile End Mile Test Section Linear Feet Cracking Test Section Description
# Post Post Type 1994 1999
040660* 202.16 202.25 STATE 91 784 8" HMA Overlay, Rubberized Concrete
040663 202.29 202.38 STATE 10" PCC Overlay, 2" HMA Overlay, C&S
040608* 202.43 202.53 SHRP 259 658 8" HMA Overlay, Crack & Seat (C&S)
040607* 202.58 202.68 SHRP 617 990 4" HMA Overlay, C&S
040606* 202.72 202.82 SHRP 726 934 4" HMA Overlay, Max., Restoration PCC
040659* 202.88 202.99 STATE 521 1198 2" HMA Overlay, Fabric, 2" HMA, C&S
040661* 203.02 203.10 STATE 30 291 2" ARAC Surface, 2" HMA, C&S
040604* 203.13 203.23 SHRP 504 554 4" HMA Overlay with Saw & Seal
040662* 203.32 203.42 STATE 546 1171 2" HMA Surface, 2" ARAC, C&S
040603* 203.54 203.64 SHRP 770 1339 4" HMA, Minimum Restoration PCC
040605 203.68 203.87 SHRP Out of Service PCC Section
040602 204.02 204.13 SHRP Out of Service PCC Section
040601 204.16 204.26 SHRP Out of Service PCC Section
* A 0.50" open graded (ACFC) surface placed after overlay to improve skid resistance. An AC-20 binder was used, as the binder.
Note some sections have a dual number. TS #040659 was 040610, TS #040660 was 040609, TS #040661 was 040611 and TS