Evaluation and Mix Design of Cement-Treated Base Materials with High RAP Content
Paper Number 11-2742
2011 Annual TRB Meeting
Deren Yuan
Research Engineer
Center for Transportation Infrastructure Systems
The University of Texas at El Paso
500 W. University Ave., El Paso, Texas 79968
Tel: (915)-747-5474, Email: [email protected]
Soheil Nazarian
Professor and Director of
Center for Transportation Infrastructure Systems
The University of Texas at El Paso
500 W. University Ave., El Paso, Texas 79968
Tel: (915)-747-6911, Email: [email protected]
Laureano R. Hoyos
Associate Professor
Department of Civil Engineering
The University of Texas at Arlington
Arlington, Texas 76019
Tel: (817)-272-3879, Email: [email protected]
Anand J. Puppala
Professor
Department of Civil Engineering
The University of Texas at Arlington
Arlington, Texas 76019
Tel: (817)-272-3879, Email: [email protected]
Item Number Words
Abstract 1 250
Text 3850
Table 1 250
Figure 12 3000
Total 7350
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 1
ABSTRACT 1 Reclaimed asphalt pavement (RAP) and granular base materials were collected from stockpiles 2
throughout Texas to evaluate the feasibility of using mixes containing high RAP content for base 3
course applications. Mixes containing 100%, 75% and 50% RAP treated with portland cement 4
of 0%, 2%, 4% and 6% were evaluated in a full-factorial laboratory experiment. For mixes with 5
75% and 50% RAP, both virgin and if available recycled base were used. Experimental results 6
indicates that, besides the cement content, the RAP content and fine content in RAP-granular 7
base mixes significantly affect the properties of the RAP mixes, and that the effects of RAP type 8
and asphalt content in RAP are very limited. 9
The goal of this project was to evaluate the applicability of the current Texas Department 10
of Transportation (TxDOT) mix deign for cement-treated bases with high-RAP-content base 11
mixes, and if necessary, to propose necessary modification to it. Despite many specifications that 12
limit the RAP content to 50%, it is quite feasible to develop and construct high-quality bases 13
with more than 50% RAP. The use of mixes with 100% RAP does not seem to be either 14
economical or perform as well as mixes with up to 75% RAP primarily due to lack of fine-15
grained particles. Economical alternative to achieve the strength and durability by modification 16
of the gradation of the RAP is also provided. 17
To achieve a 300-psi unconfined compressive strength as required by TxDOT, the 18
optimum cement contents are statistically about 4%, 3% and 2% percents for mixes of 100%, 19
75% and 50% RAP, respectively. Since the achievement of any specified strength and/or 20
modulus may not always ensure the durability, a number of other parameters, which may be 21
relevant to performance and long-term durability, were also evaluated through laboratory testing. 22
23
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 2
INTRODUCTION 24 The use of recycled materials in roadway maintenance, rehabilitation and construction has 25
become increasingly more prevalent over the past two decades. Each year, approximately 100 26
million tons of hot-mix asphalt is milled in the U.S. (1). One of the specific goals of the FHWA 27
recycled materials policy includes increasing the percentage of RAP used in the highway 28
projects. The primary use of RAP is currently in their reintegration into new hot-mix, warm-mix 29
or cold-mix pavements. Nonetheless, a large quantity of RAP, especially of marginal quality, 30
remains unused in some areas in the US, such as Texas. The use of RAP in base course has been 31
encouraged to reduce waste, and provide cost effective materials for roadway maintenance, 32
rehabilitation and reconstruction (2, 3, 4, 5). This is particularly true for projects that require 33
long-hauling distances for disposal of RAP or suitable base aggregates are scarce. 34
Given the fact that the quality of base material is one of the most important factors for 35
long-term performance of flexible pavements, the question that remains to be answered is how 36
does the use of high quantities of RAP in a base course affect the mechanical properties of the 37
mix and the performance of a flexible pavement? This paper presents the results from a 38
comprehensive laboratory testing program on cement-treated RAP (C-RAP) mixes as part of a 39
Texas Department of Transportation (TxDOT) research project. Based on these results, statistic 40
relationships between strength/modulus and cement content/RAP content are proposed for mix 41
design consideration. 42
43
MATERIALS 44 RAP and granular base materials were collected from stockpiles in six districts of TxDOT. To 45
ensure that representative materials are considered in this study, a survey was distributed among 46
the 25 districts of TxDOT. Based on the responses of the districts in terms of desire to use C-47
RAP and their geographical location in the State, materials from Childress, El Paso, Fort Worth, 48
Lubbock, Pharr and Waco were gathered for laboratory testing. Table 1 provides a brief 49
description of these materials. RAP samples were collected randomly from each stockpile at six 50
different locations. In addition, representative local base materials were also sampled. Type I/II 51
(ASTM C-150) portland cement was used throughout the testing program. 52
53
Table 1. Description of Collected RAP and Granular Base Materials 54
RAP Base Material District
Ownership Aggregate Type Virgin Salvage
Childress (CHR) State Limestone/ Granite Gravel NA
El Paso (ELP) Private Limestone/Dolomite Hard Limestone NA
Fort Worth (FTW) State Limestone Limestone NA
Lubbock (LBB) State Limestone Limestone Limestone
Pharr (PHR) Private Silica Shard Soft Limestone Gravel/Limestone
Waco (WAC) State Limestone Soft Limestone NA
55
PROPERTIES OF RAP 56 RAP properties are governed by the milling and crushing operation, as well as by the 57
characteristics of the aggregate, binder and age from which the RAP is obtained. Since the 58
quality of virgin aggregate used in asphalt concrete usually exceeds the requirements for granular 59
aggregate used in base courses, there are generally no durability concerns regarding the use of 60
RAP in base courses. The main properties of RAP considered here are the gradation, asphalt 61
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 3
content and sand equivalency value. Sieve analysis, asphalt content test, and sand equivalency 62
test were performed on each of the six randomly sampled RAP materials from each stockpile. 63
The particle size distribution of milled RAP is determined by the characteristics of the 64
milling unit, speed of and temperature during milling operation and the original HMA gradation 65
and binder content. Figure 1a shows the average gradation curves of the RAP from the six 66
stockpiles obtained through dry sieving. For comparison, typical range of particle size 67
distribution of RAP reported by FHWA (6) and the upper-lower limits for Grade 1 (high quality) 68
base materials specified by TxDOT are also shown in this figure. The lack of aggregates passing 69
No. 40 sieve is evident, similar to the findings in a previous study in Texas (7). A significant 70
portion of coarse aggregates visually consisted of a conglomerate of finer particles. Wet sieving 71
was not deemed effective in breaking down these aggregates because of the gluing action of the 72
binder. The differences in the average gradations among all six RAP materials are not very 73
significant with the largest variation of about 15% occurring around No. 4 sieve. The coefficient 74
of variation among the results from the six samples of each RAP stockpile increased as particle 75
size decreased from less 1% for aggregates retained on 1 in. sieve to about 50% for particles 76
passing No. 200 sieve. 77
Asphalt contents of RAP were determined by using NCAT ignition oven method. The 78
RAP samples from each of the six locations within a stockpile were divided into the following 79
three bins: 80
Bin 1: particles retained on 0.5 in. sieve 81
Bin 2: particles passing 0.5 in.-sieve and retained on No. 4 sieve 82
Bin 3: materials passing No. 4 sieve 83
The asphalt contents by bin and by weighted average for all RAP materials are shown in Figure 84
1b. The lowest asphalt contents were usually observed for Bin 2. The unusually high asphalt 85
content in Lubbock RAP can be attributed to a significant amount of recycled surface treatment 86
that manifested as disk-shape aggregates. 87
Average sand equivalency values (ASTM D 2419) for six samples of each stockpile 88
varied from 50% to 91%. The standard variation among individual stockpiles varied from 1.7% 89
to 5.6%. The high sand equivalency values can be attributed to the asphalt binders coating finer 90
aggregates. 91
92
LABORATORY EVALUATION OF RAP MIXES 93 Since the main goal of this study is to examine the feasibility of using mixes with high RAP 94
contents, only mixes containing 100%, 75% and 50% RAP were utilized. For mixes with 75% 95
and 50% RAP, both virgin and recycled base materials, when available, were added to the RAP. 96
Four levels of cement content, 0%, 2%, 4% and 6%, by dry weight were used for all 97
mixes except for 100% RAP. Stable specimens could not be made from the untreated (0% 98
cement) mixes of 100% RAP. As listed in Table 1, since six RAP materials and eight granular 99
base materials were used in this study, there are 18 (3 x 6) mixes for 100% RAP, 32 (4 x 8) 100
mixes for 75% RAP and 32 (4 x 8) mixes for 50% RAP. 101
102
Evaluation Protocol and Methodology 103 Achievement of a specified strength/stiffness does not always ensure the durability of chemically 104
stabilized mixes. To develop a realistic mix design procedure for RAP mixes, a number of 105
parameters were considered. Some of these parameters were necessary for a comprehensive 106
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 4
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
a) Average Gradations of RAP 130 131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
b) Asphalt Content by Bin and by Weighted Average 149
Figure 1. Characteristics of RAP Materials Used in This Study 150
0
10
20
30
40
50
60
70
80
90
100
0.010.1110100
Sieve Size, mm
Pe
rce
nt
Pa
ss
ing
Typical FHWA Range Limits of TxDOT Grade 1
Childress El PasoFt. Worth Lubbock
Pharr Waco
#4 #40 #200
9.4
4.7 4.7
11.0
3.94.6
3.85.3
3.4
5.7
3.14.3
7.45.8 5.8
8.8
6.5 6.7
6.5
5.34.7
7.9
4.85.2
0
2
4
6
8
10
12
14
Childress El Paso Ft. Worth Lubbock Pharr Waco
RAP Source
As
ph
alt
Co
nte
nt,
%
Bin 1: > 0.5"Bin 2: 0.5" > but > 0.187" (No. 4)
Bin 3: < 0.187" (No. 4)Weighted Average
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 5
evaluation but may not lend themselves to the day-to-day mix design and operation for most 151
state highway agencies. These parameters and their related tests are briefly described below. 152
153
Gradation. The RAP aggregates consisting of crushed rock or gravel covered by asphalt binder 154
exhibit mechanical properties that may be much different from the aggregates in typical granular 155
base materials. As will be seen later, the strength and stiffness of a compacted cement-treated 156
RAP mixes are mainly determined by the gradation of the mix and the cement hydration rather 157
than the properties and shape of aggregate (especially for mixes with high RAP content. 158
159
Moisture-Density Characteristics. Based on this study, the optimum moisture contents (OMC) 160
and maximum dry densities (MDD) can be determined in a simplified manner. For each RAP 161
content, moisture-density curves were developed for the mixes with 0% and 6% cement contents. 162
The OMC and MDD values for mixes with 2% and 4% cement contents were estimated through 163
linear interpolation of the OMC and MDD values obtained from the mixes with 0% and 6% 164
cement contents. This was deemed justified since the variations in these two parameters between 165
0% and 6% cement contents are reasonably small. 166
167
Strength. Unconfined compressive strength (UCS, ASTM D 1633 or Tex-120-E) has been used 168
by many highway agencies in their mix design for cement-stabilized base layers as the only 169
quantitative requirement for engineering properties. A target 7-day UCS in the range of 200 psi 170
to 600 psi or more is currently adopted by different highway agencies. A minimum 7-day UCS 171
of 300 psi is used as the main criterion for cement content selection in this study, since this value 172
is used recommended by TxDOT. Requirements for other parameters such as modulus, moisture 173
susceptibility and long-term durability are established on the basis of this UCS requirement. 174
The indirect tensile strength (ITS, ASTM D 6931) could be a primary parameter for 175
evaluation and qualification of asphalt emulsion-treated or dual-stabilized (asphalts emulsion 176
plus calcium-based additive) base mixes including those containing RAP up to 80% (8). To 177
study the significance of this parameter for cement-treated RAP mixes, ITS tests were performed 178
on specimens prepared from all mixes. Each specimen for ITS testing is 6 in. in diameter and 179
about 4.5 in. in height. 180
181
Modulus. Moduli of specimens were measured through the resilient modulus (RM, AASHTO T 182
307) test and the free-free resonant column (FFRC, ASTM C 215) test. RM test is widely 183
accepted since it attempts to simulate vehicular loading conditions on pavement structures even 184
though the test is time consuming, complicated and less accurate, in particular, for stabilized 185
materials. As part of a comprehensive evaluation program in this study, RM tests were limited to 186
those mixes with the optimum cement contents and are not presented here due to space 187
limitation. 188
The FFRC test is nondestructive, rapid, reliable and easy to perform for stabilized 189
materials. The principle of the FFRC method is based on the determination of the fundamental 190
resonant frequencies of vibration of a cylindrical specimen. From the longitudinal frequency, 191
Young’s modulus (called FFRC modulus here) of the specimen can be calculated. Over the years, 192 the test setup and data reduction of this method have been significantly simplified and enhanced for 193 day-to-day use (Figure 2 and draft Tex-148-E). FFRC tests were performed on all UCS specimens 194
to see the feasibility for establishing the relationships between the FFRC modulus (instead of 195
RM) and the strength parameters from the standard tests. 196
197
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 6
198
199
200
201
202
203
204
205
206
207
208
209
210
a) FFRC Test Setup b) Records from a FFRC Test 211
Figure 2. Free-Free Resonant Column (FFRC) Test 212 213
Moisture Susceptibility. Moisture susceptibility represents the potential of a soil to lose 214
strength/stiffness by absorbing water under capillary conditions. Moisture susceptibility is 215
particularly of major concern with the RAP blends in pavement base courses due to potential for 216
stripping. This parameter was evaluated by using tube-suction test (TST, 9). Outputs from TST 217
on a moisture-conditioned specimen include the retained strength, retained modulus and 218
dielectric constant. The retained strength is here defined as a ratio of the UCS measured after 2-219
day oven (140o F) cure and 8-day capillary soaking to the standard 7-day UCS (ASTM D 1633 220
or Tex-120-E). The current acceptance criterion for retained strength is 80%. It may be 221
reasonable to propose the same number for the retained ITS and FFRC modulus. 222
223 Long Term Durability. This parameter was studied by following the procedure of ASTM D 224
559. The procedure involves a cyclic wet-dry process that simulates rainfall events in a 225
reasonably short time period. The test was carried out by immersing samples in water at room 226
temperature for 5 hrs and then oven-drying at 160º F for 48 hrs to complete one cycle. This 227
process was repeated for up to 14 cycles. Figure 3a shows the setup of wetting-drying process. 228
After removal from the oven, the specimen was subjected to volume change and moisture 229
content measurements. After 3, 7, and 14 cycles, the specimens were subjected to UCS tests. The 230
results obtained provide adequate information whether the cement treated RAP materials are 231
durable or fail prematurely. This test was applied to those mixes with the optimum cement 232
contents. 233
234
Leachate. Due to the coarse nature of RAP materials, cement-treated RAP mixes should be 235
studied for any possible leaching of cement stabilizer due to moisture flow. Figure 3b shows the 236
apparatus for leachate test (10). The test utilizes flexible wall molds housing specimens. The 237
specimens were first cured in a moisture room for 7 days before subjecting to leachate testing. 238
This test was applied to those mixes with the optimum cement contents. Leachate was collected 239
after 3, 5, 7, 11, and 14 cycles of leaching, while the UCS tests were conducted at the end of 14 240
cycles of leaching. The leachate collected was tested for pH value change and the amount of 241
calcium present after the corresponding cycles. Results were analyzed to address the loss of 242
cement due to leaching. 243
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 7
244
a) Wetting b) Drying 245
a) Setup for Wetting/Drying Process 246
247
b) Apparatus for Leachate Test 248
249
Figure 3. Test Setups for Long-Term Durability and Leachate Tests 250 251
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 8
Presentation and Discussion of Results 252
Moisture-Density Characteristics 253 Figure 4a shows the variations in OMC and MDD with different RAP contents for the untreated 254
mixes containing virgin base materials. In general, both OMC and MDD decrease as RAP 255
content increases except for OMC with Ft. Worth mixes. Figure 4b compares the OMC and 256
MDD for mixes of 100% RAP without and with 6% cement. The OMC and MDD of all mixes 257
are greater with 6% cement than those without treatment. The average differences are about 258
1.2% for OMC and 3.5 pcf for MDD. These two numbers may have significance for roadway 259
maintenance and minor rehabilitation projects as well as quality assurance if only the moisture-260
density curve is developed for untreated mixes in the laboratory. 261
262
263
264
265
266
267
268
269
270
271
272
273
274
a) Variations of OMC and MDD of Untreated Mixes with RAP Content 275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
b) OMC and MDD of 100% RAP Mixes without Treatment and Treated with 6% Cement 290 291
Figure 4. Moisture-Density Characteristics of RAP Mixes 292
293
Effect of Asphalt Content 294 As shown in Figure 2b, the asphalt contents of RAP materials collected from the six stockpiles 295
cover a quite large range from 4.7% to 7.9%. The effects of asphalt content on strength and 296
6.6
6.1
4.8
7.2 7
.7
7.6
6.5
5.8
5.2
6.6 7
.2
6.9
5.4
5.2
6.0 6.4
5.2
6.2
0
2
4
6
8
10
12
CHR ELP FTW LBB PHR WAC
Material Source
OM
C, %
50% RAP 75% RAP 100% RAP
13
3.2
13
1.8
12
5.6
12
8.8
11
9.8
11
7.3
12
5.5
121
.7
12
4.8
11
9.6
11
6.3
12
3.2
116
.0
115
.9 121
.7
119
.3
10
8.5
13
0.0
100
110
120
130
140
150
CHR ELP FTW LBB PHR WAC
Material Source
MD
D, p
cf
50% RAP 75% RAP 100% RAP
5.4 5.26.0
6.4
5.2
6.26.5 6.47.1
8.2
6.3
7.2
0
2
4
6
8
10
12
CHR ELP FTW LBB PHR WAC
RAP Source
OM
C, %
0% Cement 6% CementOMC
123
.2
118
.0
11
6.9
11
9.7
119
.3
11
1.5
12
6.7
120
.8
12
2.3
12
2.9
12
2.4
11
5.3
90
100
110
120
130
140
150
CHR ELP FTW LBB PHR WAC
RAP Source
MD
D, p
cf
0% Cement 6% CementMDD
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 9
modulus for all mixes with 100% RAP are shown in Figure 5. Strength and modulus are perhaps 297
independent of asphalt content for different levels of cement treatment, even for the RAP from 298
Lubbock stockpile with 7.9% asphalt content. 299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
Figure 5. Variations of UCS with Asphalt Content in Mixes of 100% RAP 337
338
Effect of Fine-Grained Aggregate 339 The lack of particles passing No. 40 sieve is apparent in all RAP materials collected as shown in 340
Figure 1. In particular, the fines contents (particles passing No. 200 sieve) are about 1% or less. 341
This occurs because the fines in RAP manifest themselves as larger particle sizes in the presence 342
0
50
100
150
200
250
300
3 4 5 6 7 8 9
Asphalt Content, %
UC
S, p
si
2% Cement
150
0
100
200
300
400
500
600
3 4 5 6 7 8 9
Asphalt Content, %
UC
S, p
si
4% Cement
290
0
200
400
600
800
3 4 5 6 7 8 9Asphalt Content, %
UC
S, p
si
6% Cement
440
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 10
of asphalt binder. As shown in Figure 6, the higher the fines content is, the higher the UCS 343
seems to be. Similar effects were also found for ITS and FFEC modulus. 344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
Figure 6. Effect of Percent Aggregates Passing No. 40 Sieve on UCS for 100% RAP Mixes 363
364 The fines contents (particles passing No. 200 sieve) of the granular base materials 365
collected for this study are significantly higher (from 1% to 6%). Thus, for mixes with different 366
RAP contents, the total finer contents in these mixes are different. Figure 7 shows an example of 367
the positive impact of fines content on the UCS for El Paso materials. Again, similar effects were 368
also found for ITS and FFRC modulus. Of course, other material properties may also affect the 369
strength and modulus of a cement-treated mix. However, the effect of materials passing No. 40 370
sieve on strength and stiffness seems to be significant for cement-treated RAP mixes 371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
Figure 7. Effect of Fines Content on UCS 388
0
200
400
600
800
1000
1200
0% 2% 4% 6%
Cement Content
UC
S, p
si
100% RAP (0.3% Fines)
75% RAP (1.5% Fines)
50% RAP (2.7% Fines)
0
100
200
300
400
500
5%
(WAC)
6%
(ELP)
7%
(FTW)
11%
(LBB)
11%
(CHR)
12%
(PHR)
Aggregates Passing No. 40 Sieve in RAP
UC
S, p
si
2% Cement
4% Cement
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 11
Effect of Coarser Aggregate 389 Figure 8 shows the gradation curves of RAP, recycled base and virgin base collected 390
from Pharr stockpiles. The gradations of the two bases are very different except for the particles 391
passing No. 100 sieve. However, with 2% cement treatment, the two mixes, 50% RAP plus 50% 392
recycled base and 50% RAP plus 50% virgin base, provide quite similar strength and modulus 393
values. That is, 396 psi vs. 359 psi for UCS, 51 psi vs. 46 psi for ITS, and 1350 ksi vs. 1233 ksi 394
for FFRC modulus. This example indicates that the particle size distribution of coarse aggregate 395
has a lesser role on strength and modulus of cement-treated RAP mixes. The significance of this 396
phenomenon is that under cement treatment more RAP resources can be used in base course 397
applications. 398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
Figure 8. Gradations of RAP, Recycled Base and Virgin Base 418 419
Moisture Susceptibility 420 The average retained UCS, ITS and FFRC moduli for all mixes involved in tube suction tests are 421
shown in Figure 9. The bar on each value corresponds to ± 1 standard deviation. All three 422
parameters meet or closely meet the requirement of 80% retained values. The smaller retained 423
ITS can be attributed to the shorter specimen heights relative to those used for the UCS and 424
modulus tests (about 4.5 in. vs. 8 in.), and thus the moisture due to capillary has more effect on 425
the properties of ITS specimens. The longer error bars for the retained ITS can be attributed to 426
the uncertainty in indirect tensile testing. Dielectric constant measurement was applied to all 427
specimens involved in the moisture susceptibility study. All dielectric values were significantly 428
less than 10 (a recommended value by TxDOT draft Tex-144-E). Thus, dielectric constant 429
seems to have less meaning for cement-treated materials. 430
431
432
433
0
10
20
30
40
50
60
70
80
90
100
0.010.1110100
Sieve Size, mm
Pe
rce
nt
Pa
ss
ing
TxDOT Limits for Grade1
RAP
Recycled Base
Virgin Base
#4 #40 #200
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 12
Figure 9. Average Retained Strengths and Modulus 434 435
Long Term Durability and Leachate 436 Figure 10 presents the results from wet-dry tests on the mixes of Childress and Ft. Worth 437
materials to assess the long-term durability of different mixes. Term “original” in the figure 438
stands for the UCS measured on specimens subjected to the 7-day standard cure. All three Ft. 439
Worth mixes exhibit high retained UCS strengths even after 14 cycles of wetting-drying. The 440
retained strengths of mixes from Childress exhibit different patterns and the retained strength of 441
the mix of 100% RAP decreases significantly. The adequate strength of a RAP mix from 442
standard cure may not always ensure its long-term durability even though the wet-dry testing 443
simulates an extreme situation of moisture damage. Results from leachate tests indicate that the 444
loss or leaching of cement for most treated mixes is not significant (less than 50 ppm in terms of 445
calcium ion concentration), and pH values almost keep constant after 14 leaching cycles. Rich 446
limestone in granular base materials may have an impact on leachate testing. 447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
Figure 10. Variation of UCS with the Number of Wetting-Drying Cycles 462
0
20
40
60
80
100
120
140
100% RAP 75% RAP 50% RAP
Mix
Reta
ined
Str
en
gth
/ M
od
ulu
s, %
UCS ITS FFRC Modulus
0
100
200
300
400
500
600
700
100% RAP
(4% Cement)
75% RAP
(4% Cement)
50% RAP
(2% Cement)
Nnumber of Wet-Dry Cycles
UC
S, p
si
Oginal
3 Cycles7 Cycles
14 Cycles
a) Childress
0
100
200
300
400
500
600
700
100% RAP
(4% Cement)
75% RAP
(4% Cement)
50% RAP
(2% Cement)
Nnumber of Wet-Dry Cycles
UC
S, p
si
Oginal
3 Cycles7 Cycles
14 Cycles
b) Ft. Worth
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 13
Strength and Modulus vs. Cement Content and RAP Content 463 The UCS, ITS and FFRC modulus values for all mixes with different cement contents and 464
different RAP contents are summarized in Figure 11. Once again, the error bars corresponds to ± 465
one standard deviation. The effects of cement content and RAP content on strength and modulus 466
are evident. The large standard deviations can be attributed to the variability in RAP and 467
granular base materials from different sources. The larger variations for ITS may also be 468
attributed to the mechanism of indirect tensile testing. 469
Figure 11. Variations of Average UCS, ITS and FFRC Modulus with Cement Content 470
471
MODELS FOR MIX DESIGN 472 One of the main objectives of this study is to provide an easy way to determine the optimum 473
cement content for cement-treated RAP mixes. As shown in Figure 12, linear relationships 474
represent well the dependence of average UCS, ITS and FFRC modulus values on cement 475
content for all mixes involved in this study. The results shown Figure 12 indicate that for a 300-476
psi UCS, the optimum cement contents are about 4%, 3% and 2% for mixes with 100% RAP, 477
75% RAP and 50% RAP, respectively. Also, corresponding to a 300-psi UCS, the ITS would be 478
about 40 psi and FFRC modulus about 1000 ksi. In addition, in terms of per 1% cement, the 479
average rates of increase in UCS (the slopes of the lines) are 73 psi for 100% RAP content, 115 480
psi for 75% RAP content, and 145 psi for 50% RAP content. Since the relationships were 481
developed based on the RAP and granular base materials from only six stockpiles in Texas, they 482
need to be refined when a larger database becomes available. The large standard deviation 483
reelected as error bars in Figure 11 clearly indicates that the typical cement contents proposed 484
here should be validated through laboratory testing for a given mix before they can be used in 485
construction. 486
0
300
600
900
1200
0% 2% 4% 6%
Cement Content
UC
S,
psi
100% RAP
75% RAP
50% RAP
0
500
1000
1500
2000
2500
0% 2% 4% 6%
Cement Content
FF
RC
Mo
du
lus,
ksi 100% RAP
75% RAP
50% RAP
0
20
40
60
80
100
120
140
0% 2% 4% 6%
Cement Content
ITS
, p
si
100% RAP
75% RAP
50% RAP
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 14
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
Figure 12. Relationships between Average UCS, ITS and FFRC Modulus and 531
Cement Content 532
R2 = 0.9992
R2 = 1R
2 = 0.9995
0
300
600
900
1200
0 1 2 3 4 5 6
Cement Content, %
UC
S, p
si
100% RAP
75% RAP
50% RAP
300 psi
R2 = 0.9999
R2 = 0.9963
R2 = 0.9976
0
20
40
60
80
100
120
140
0 1 2 3 4 5 6
Cement Content, %
IDT
S, p
si
100% RAP
75% RAP
50% RAP
R2 = 0.9896
R2 = 0.9728R
2 = 0.9357
0
500
1000
1500
2000
2500
0 1 2 3 4 5 6
Cement Content, %
FF
RC
Mo
du
lus
, k
si 100% RAP
75% RAP
50% RAP
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 15
CONCLUSIONS 533 Eighty untreated and cement-treated mixes consisting of the RAP from 6 stockpiles and the 534
granular materials from 8 stockpiles in Texas were evaluated to develop a realistic mix design 535
procedure for high-RAP-content mixes used in roadway base course construction. The results 536
from the study can be concluded as follows: 537
538
• The RAP content in a mix strongly impacts strength, modulus and durability of the mix. 539
• For a UCS of 300-psi, the average optimum cement contents are about 4%, 3% and 2% for 540
mixes of 100% RAP, 75% RAP and 50% RAP, respectively. 541
• The results from UCS, ITS and FFRC modulus tests are quite consistent. Corresponding to a 542
300-psi UCS, ITS and FFRC modulus are about 40 psi and 1000 ksi, respectively. 543
• For the mixes that meet the 300-psi UCS requirement, the average retained UCS, ITS and 544
FFRC modulus from tube suction tests meet or closely meet the recommended value of 80% , 545
and the average retained UCS values from wet-dry testing are similar. 546
• Dielectric values are significantly less than 10 for all cement-treated RAP mixes. 547
• Percentage of particles passing No. 40 sieve in general, and passing No. 200 sieve in 548
particular, in a RAP mix significantly impact its strength and modulus. Since the lack of 549
these particles in Texas RAP is common, RAP mixed with granular base (including recycled 550
base) materials with higher fines content can improve the quality of the mixes. 551
• Asphalt content in RAP does not seem to have a considerable impact on strength and 552
modulus of cement-treated RAP mixes. 553
• Particle size distribution of coarse aggregate only has a minor impact on strength and 554
modulus of cement-treated RAP mixes. 555
556
ACKNOWLEDGEMENTS 557 The authors wish to gratefully acknowledge the financial support and guidance of the Texas 558
Department of Transportation. Special thank you is extended to Dr. Jimmy Si of TxDOT for 559
directing this project. 560
561
REFERENCES 562
1. Recycling of Asphalt Pavement. Missouri Asphalt Pavement Association (MAPA), 2007, 563
(www.moasphalt.org/facts/environmental/recycling.htm) 564
2. Maher, M. H., and Popp, W., Jr. Recycled Asphalt Pavement as A Base and Subbase 565
Material. Testing Soil Mixed with Waste or Recycled Materials, ASTM STP 1275, 1997, pp.1 566
– 10. 567
3. Taha, R., A. Al-Harthy, K. Al-Shamsi and M. Al-Zubeidi. Stabilization of Reclaimed 568
Asphalt Pavement Aggregate for Road Bases and Subbases. Journal of Materials in Civil 569
Engineering, Vol. 14, 2002, pp. 239 – 245. 570
4. Guthrie, W. S., A. V. Brown, and D. L. Eggett. Cement Stabilization of Aggregate Base 571
Material Blended with Reclaimed Asphalt Pavement. Transportation Research Record, No. 572
2026, 2007, pp 47-53. 573
5. Saeed, A. Performance-Related Tests of Recycled Aggregates for Use in Unbound Pavement 574
Layers. NCHRP REPORT 598, 2008. 575
6. Chesner, W., R. Collins, M. MacKay and J. Emery. User Guidelines for Waste and 576
Byproduct Materials in Pavement Construction. FHWA Report FHWA-RD-97-148, 577
Federal Highway Administration, 1998. 578
TRB 2011 Annual Meeting Paper revised from original submittal.
Yuan et al. Paper No. 11-2742 16
7. Rathje, E. M., A. F. Rauch, K. J. Folliard, C. Viyanant, M. Ogalla, D. Trejo, D. Little and 579
M. Esfeller. Recycled Asphalt Pavement and Crushed Concrete Backfill: Results from Initial 580
Durability and Geotechnical Tests. Center for Transportation Research Report 4177-2, The 581
University of Texas at Austin, 2002. 582
8. Franco, S., P. Moss, D. Yuan and S. Nazarian. (2009), Design, Constructability Review and 583
Performance of Dual Base Stabilizer Applications. FHWA/TX-09/0-5797-1, Center for 584
Transportation Infrastructure Systems, University of Texas at El Paso, 2009. 585
9. Barbu, B. and T. Scullion. Repeatability and Reproducibility Study for Tube Suction Test. 586
FHWA/TX-06/5-4114-01-1, Texas Transportation Institute, The Texas A&M University 587
System, 2006. 588
10. Chittoori, B. C. S. Clay Mineralogy Effects on Long-Term Performance of Chemically 589
Treated Expansive Clays. Dissertation, University of Texas at Arlington, 2008. 590
TRB 2011 Annual Meeting Paper revised from original submittal.
Top Related