The Relationship of Binder Delta Tc (ΔTc) & Other Binder Properties to

Mixture Fatigue and Relaxation

Gerald ReinkeMTE Services, Inc

Binder ETG MeetingMay 10, 2018

Fall River, MA

INTRODUCTION

• Great deal of information has been presented on the ΔTc parameter

• Not going to go into much detail about ΔTc or other parameters because that is not the focus of this presentation

• Goal is to show the relationship between binder and mix relaxation and measured properties of aged binders– R-Value, Glover-Rowe, cross over frequency, ΔTc, Tm-Critical

ΔTc Determination & Sources of Error1. ΔTc=(Ts-critical – Tm-critical) ✓2. To obtain an accurate value for ΔTc the BBR needs to be performed at enough

temperatures so thata. BBR stiffness values < 300 MPa and > 300 MPab. BBR m-values < 0.300 and > 0.300c. Extended aging of binders , high levels of RAP and/or RAS, the use of high levels of

additives such as REOB might require BBR testing at 3 or more temperatures ✓3. If BBR stiffness is less than ≈ 125 MPa when BBR m-value barely exceeds 0.300 then

generally a 3rd BBR test temperature will be required to meet the requirements of 2.a and 2.b ✓

4. If you perform BBR at 2 temperatures where stiffness is <200 MPa so that Tm-criticalwill be <0.300 and >0.300 you can end up with an incorrect Ts-critical ✓

5. Linear extrapolations based on 2 test temperature over 100 to 150°C can result in incorrect predictions. Not all binders are linear (m value) or log linear (S value) with temperature

When a binder exhibits a ΔTc of < -4 or -5 the S critical temperature increases at a substantially slower rate than does the m-critical temperature and this will necessitate the need for a 3rd BBR Test

Ts-Critical ΔTc=-7.2y = -0.5581x - 33.709

Tm-Critical ΔTc=-7.2y = -1.5581x - 33.709

Ts-Critical ΔTc=-2.5y = -1.02x - 28.229

Tm-Critical ΔTc=-2.5y = -2.0482x - 28.234

-40

-35

-30

-25

-20

-15

-10

-8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

S or

m C

ritic

al T

empe

ratu

res

ΔTc

Rate of Change of ΔTc Depending on Binder Composition and Aging Severity

PG 64-22 + 20% Shingle Binder, 5% REOB, SCritical PG 64-22 + 20% Shingle Binder, 5% REOB, m-CriticalPG 64-22 SCritical PG 64-22 m-CriticalLinear (PG 64-22 + 20% Shingle Binder, 5% REOB, SCritical) Linear (PG 64-22 + 20% Shingle Binder, 5% REOB, m-Critical)Linear (PG 64-22 SCritical) Linear (PG 64-22 m-Critical)

Unaged

20 hr. PAV

RTFO

Unaged

RTFO

40 hr. PAV

40 hr. PAV

20 hr. PAV

Slope Ts-Critical for ΔTc of -7 is 50% that of the binder with ΔTc of -2.5

Slope of Tm-Critical for ΔTc of -7.5°C is 75% that of the binder with ΔTc of -2.5

JUST WHAT IS ΔTc?1. Reasons that we should all know are ✓

a. As binders age they become more m-controlled; Tm-critical increases more rapidly than Ts-critical ✓

b. As binders become more m-controlled they are more brittle and lose ability to relax stress ✓

c. As pavements age they are more prone to cracking distress ✓d. As ΔTc becomes more negative pavements become more prone to top down

fatigue cracking ✓e. It may not appear intuitively obvious that a value derived from low temperature

testing should be associated with distresses that are associated with intermediate service temperatures ✓

f. Based on research, some of which goes back 50+ years, research has shown the connections between pavement surface distresses and several parameters the most recent of which is ΔTc ✓

ΔTc can quantify the aging propensity of a binder

WHY IS AN UNDERSTANDING OF ΔTc IMPORTANT?

IN THE FINAL ANALYSIS ΔTc, R-Value, GR COMES DOWN TO 1 THING

IN THE FINAL ANALYSIS ΔTc, R-Value, GR COMES DOWN TO 1 THING

NO! NOT THAT 1 THING

IN THE FINAL ANALYSIS ΔTc, R-Value, GR COMES DOWN TO 1 THING

NO! NOT THAT 1 THINGRELAXATION

IN THE FINAL ANALYSIS ΔTc, R-Value, GR COMES DOWN TO 1 THING

NO! NOT THAT 1 THINGRELAXATIONSPECIFICALLY BINDER

RELAXATION

TIME TO GET SERIOUS• As with most advances in technical research developments are

the result of cumulative increase in knowledge ✓• I will briefly reference the work of three individuals, but

reading their research will show many other contributors along the way

• Prithvi (Ken) Kandhal – Pennsylvania DOT Bituminous Engineer• Dr. Charles Glover—Research Professor Texas Transportation

Institute at Texas A&M• Mike Anderson—Director of Research at the Asphalt Institute

References1. Kandahl, Low Temperature Ductility in Relation to Pavement Performance, ASTM STP 628,

Marek, Ed., 19772. Glover, Charles J, Davison, Richard, Domke, Chris, Ruan, Yonghong, Juristyarini, Pramitha, Knorr,

Daniel, Jung, Sung, “Development Of A New Method For Assessing Asphalt Binder Durability With Field Validation”, FHWA/TX-05/1872-2, August 2005

3. Anderson, R. M, King, G.N., Hanson, D.I., Blankenship, P.B. "Evaluation of the Relationship between Asphalt Binder Properties and Non-Load Related Cracking." Association of Asphalt Paving Technologists, 2011 Volume 80, pp 615-663, 2011

4. TRB papers in 2010, 2011 and 2012 by Sui and Farrar, et al from Western Research Institutre5. EECongress in Istanbul, 2012, Farrar, et al

In the interest of time I have hidden several background slides which will be available when

this presentation is provided to the ETG members

Ductility and Pavement Condition of 1961 and 1962 Pennsylvania Pavements Reported by Kandhal (Kandhal 1977)

Ductility value at 60°F (15.5°C), 5 cm/min, cm

Pavement Condition Observed

More than 10 Satisfactory8 to 10 Loss of fines (matrix)5 to 8 Raveling3 to 5 Cracking, needs resurfacing

Less than 3 Very poor, extensive crackingSOME COMMENTS REGARDING KANDHAL’S WORK1. At 10 cm ductility there is no cracking reported, however when it takes longer than 3

years to reach 10 cm loss of fines and some raveling is noted2. Regardless of the time it takes to reach less than 5 cm of ductility that ductility value

is associated with the onset of cracking ✓

Extent of binder aging is the key factor and not the time of binder aging

What Can We Infer From This Data?• There is a point in the aging of a binder when cracking begins to

develop• Binder aging rate is not the same for every binder (crude source

impacts performance) or perhaps it is not the same time point for the same binder depending on the conditions of the job– Time of year constructed– % bitumen in the mix– Air voids – Aggregate type and/or gradation– Other factors e.g. RAP, RAS, polymer or ???

• Extent of Binder Aging is the Key Driver• How can we age binders and mixtures sufficiently in the lab to

tell us something useful about long term performance?

Taken from Glover, et al 2005, plot shows1. Linear correlation between

G’/ η’/G’ and 15°C ductility for ductility values < 10 cm✓

2. Based on Kandhal’s data when ductility drops below 10 pavement distress begins ✓

3. Glover used this data to develop relationship between ductility and binder rheology at 15°C ✓

4. Glover used time temperature superposition principles to adjust the DSR test to 44.7°C and 10 rad/sec ✓

NOTE: ALL THESE ARE CONVENTIONAL BINDERS

Moving from Ductility to ΔTc• Mike Anderson, et al AAPT 2011—introduced ΔTc concept ✓• Rheological & ductility of PAV binders and binders recovered from

aged airfield mixtures• Established Relationship of ΔTc to non-load associated distress• Key findings ✓

1) Glover @ Texas A&M had shown ductility @ 15°C & 1 mm/min correlated to long term pavement distress ✓

2) G’/(η’/G’) correlated to ductility @ 15°C & 1 mm/min ✓3) Also showed G’/(η’/G’) correlated to ΔTc (difference between the BBR Tm-

critical – BBR Ts-critical ✓

4) ΔTc of 2.5°C = cracking warning limit, ΔTc = 5°C point where binder durability lost ✓

ΔTc and 4 mm DSR Testing

• Much of the data to be discussed next was generated at MTE using a 4 mm DSR test developed at Western Research Institute (see reference list)

• Requires very little material to perform test ✓• Results correlate well to BBR, but there is a learning curve ✓• Provides a broader temperature range (-36°C to +30°C or

+40°C) of data collection in less time than BBR test at 3 temperatures ✓

The size advantages are obvious for performing tests on field samples and other forensic workWhen the main mixture layer that needs testing is binder recovered from the top ½ inch of a 6 inch diameter core very little binder is obtained and the 4 mm test requires only one core to provide sufficient binder for a 25 mm and 4 mm test

Just How Does ΔTc Relate to Mix Performance?• Need to get back to RELAXATION• As binders age their ability to relax stress diminishes ∴ BBR

result becomes increasingly m-controlled (poor relaxation) ✓• Some binders have inherently poor relaxation properties, BBR

will show this and ΔTc can quantify impact of poor relaxation✓

• Relaxation is not just a low temperature (i.e. sub 0°C) problem– Ductility decreases when binder cannot relax fast enough to prevent

the binder thread from breaking (Kandhal & Glover at 15°C)– The DSR data shows similar behavior (Glover’s DSR vs Ductility Plot

another test performed at 15°C)

Just How Does ΔTc Relate to Binder Relaxation and Ultimately Mix Performance?

• How many of you have really looked at or compared the BBR data plot for different binders?

• BBR test is not just a single data point at 60 seconds• In that plot is the story of how the binder relaxes (or doesn’t) due

to the imposition of load

100.0

1,000.0

10 100 1000

BBR

Stif

fnes

s, S

(t),

& 4

mm

Stif

fnes

s G

(t),

MPa

Reduced Time, sec's

BBR S(t) mastercurve @ -18° Ref Temp, Binder A, ΔTc = -5°C

BBR S(t) mastercurve @ -18° Ref Temp, Binder B, ΔTc= 1°C

Relaxation time = 60 seconds

60

COMPARISON OF BBR MASTERCURVES @ -18°C FOR TWO DIFFERENT BINDERS

If you only look at the 60 second results from the BBR test you are blind to the relaxation behavior of the binder

1. If you only focus on the slope at 60 seconds you will see a difference, but it is just a comparison of 2 numbers

2. When you look at the complete BBR mastercurve you see how much more readily the binder with a ΔTc of 1°C relaxes stress compared to the binder with a ΔTc of -5°C

10.0

100.0

1,000.0

10,000.0

0.001 0.01 0.1 1 10 100 1000 10000 100000 1000000

BBR

Stif

fnes

s, S

(t),

& 4

mm

Stif

fnes

s G

(t),

MPa

Reduced Time, sec's

BBR S(t) mastercurve @ -18° Ref Temp, Binder A, ΔTc = -5°C

BBR S(t) mastercurve @ -18° Ref Temp, Binder B, ΔTc= 1°C

Relaxation time = 60 seconds

COMPARISON OF BBR MASTERCURVES @ -18°C FOR TWO DIFFERENT BINDERSCOMMENTS1. Binder sample A has a ΔTc of

-5°C compared to sample B with a ΔTc of +1.

2. The important point is that sample A relaxes the same applied load over the same time period at slower rate than sample B

1. When you incorporate the 4 mm data for the same binders similar ΔTc results are obtained, but you also observe how the relaxation disparity carries over to longer relaxation times

2. Longer relaxation times are a surrogate for relaxation behavior at warmer temperatures

1.0E-05

1.0E-04

1.0E-03

1.0E-02

1.0E-01

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

1.00E-06 1.00E-04 1.00E-02 1.00E+00 1.00E+02 1.00E+04 1.00E+06 1.00E+08 1.00E+10

BBR

Stif

fnes

s, S

(t),

& 4

mm

Stif

fnes

s G

(t),

MPa

Reduced Time, sec's

BBR S(t) mastercurve @ -18° Ref Temp, Binder A, ΔTc = -5°C

BBR S(t) mastercurve @ -18° Ref Temp, Binder B, ΔTc= 1°C

4 mm DSR ,G(t) mastercurve @ -18°C Ref Temp, Binder A, ΔTc= -4.9°C

4 mm DSR ,G(t) mastercurve @ -18°C Ref Temp, Binder B, ΔTc= 0.6°C

Relaxation time = 60 seconds

COMPARISON OF BBR & 4 mm MASTERCURVES @ -18°C FOR TWO DIFFERENT BINDERS

If binders have a relaxation disparity at low temperatures they also have a relaxation disparity at warmer temperatures

An additional benefit of the 4 mm test is the ability to examine the binder’s behavior at temperatures beyond those capable by the BBR

1.0E-05

1.0E-04

1.0E-03

1.0E-02

1.0E-01

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

1.00E-13 1.00E-10 1.00E-07 1.00E-04 1.00E-01 1.00E+02 1.00E+05 1.00E+08 1.00E+11

BBR

Stif

fnes

s, S

(t),

& 4

mm

Stif

fnes

s G

(t),

MPa

Reduced Time, sec's

BBR S(t) mastercurve @ -18° Ref Temp, Binder A, ΔTc = -5°CBBR S(t) mastercurve @ -18° Ref Temp, Binder B, ΔTc= 1°C4 mm DSR ,G(t) mastercurve @ -18°C Ref Temp, Binder A, ΔTc= -4.9°C4 mm DSR ,G(t) mastercurve @ -18°C Ref Temp, Binder B, ΔTc= 0.6°C4 mm DSR ,G(t) mastercurve @ 25°C Ref Temp, Binder A, ΔTc= -4.9°C4 mm DSR ,G(t) mastercurve @ 25°C Ref Temp, Binder B, ΔTc= 0.6°CRelaxation time = 60 seconds

COMPARISON OF BBR & 4 mm MASTERCURVES @ -18°C & 25°C FOR TWO DIFFERENT BINDERS

1.00E+00

1.00E+01

1.00E+02

1.00E+03

1.00E+04

1.00E+05

1.00E+06

1.00E+07

1.00E+08

1.00E+09

1.00E+10

1.00E-04 1.00E-02 1.00E+00 1.00E+02 1.00E+04 1.00E+06 1.00E+08 1.00E+10 1.00E+12 1.00E+14

Com

plex

She

ar M

odul

us, P

a

REDUCED FREQUENCY, rad/sec

Comparison of G* Moduli @ 25°C of 40 hour PAV residues Showing Greater R-Value for Binder That has lower crossover frequency

G* MN1-4, PG 58-28, 40 hr. PAV G* MN1-5, PG 58, 40 hr PAVMN1-4 Crossover Frequency MN1-5 Crossover Frequency

R-Value forMN1-4

R-Value forMN1-5

Illustration of Determination of R-Value (Rheological Index)1. MN1-5 binder

performed the best and has the lowest R-value

2. MN1-4 performed the worst and has the highest R-value

R = Log(Glassy modulus) – Log at G* at the crossover frequency)For practical purposes the Glassy modulus is 1 x 109 PascalsCrossover frequency is where phase angle = 45°As a binder’s ability to relax stress diminishes the binder stiffness must decrease to achieve a phase angle of 45°. As a result the R-value increases

The R-Value is another way to quantify binder relaxation by comparing the shear modulus (G*) mastercurves The method of determining

the R-value from rheological data is summarized at the left

A graphical presentation of R-Value is shown in the difference in length for the 2 sets of brackets

SOME FIELD EXAMPLES• I’ve presented this information at AI and other places such

as ETG meetings, ∴ I will only provide a couple brief comments

COMPARATIVE CRUDE SOURCE STUDY

• CTH 112 Olmsted Cty, MN; 2006 construction• 3 virgin test sections to compare 3 different crude sources of

PG 58-28 binder (MN1-3, MN1-4, MN1-5)• 1 virgin PG 58-34 PMA binder (MN1-2)• Project specified mix of a PG 58-34 + 20% RAP (MN1-1)• Substantial surface cracking began to show up between

years 4 and 5

Mathy Technology & Engineering27

28

MN1-2

MN1-3

MN1-4

MN1-5MN1-2MN1-3

MN1-4MN1-5

y = -54.788x + 151.08R² = 0.9638

y = -4.1875x + 16.721R² = 0.6744

y = -50.601x + 134.36R² = 0.9349

0.0

100.0

200.0

300.0

400.0

500.0

600.0

-7.0 -6.0 -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0

DIS

TR

ESS

DAT

A, 2

014

SUR

VE

Y, m

eter

ΔTc OF BINDER RECOVERED FROM TOP 1/2 INCH OF 2014 CORES

Total Distress = F(ΔTc from Top 1/2''); Transverse Cracks = F(ΔTc from Top 1/2'') & (Total Distress-Transverse Cracks)=F(ΔTc from Top ½’’ Recovered

Binder)

Total Distress = F(ΔTc of Binder from Top 1/2'') Total Transverse = F(ΔTc of Binder from Top 1/2'')

(Total Distress-transverse) = F(ΔTc of Top 1/2'' Binder) Linear (Total Distress = F(ΔTc of Binder from Top 1/2''))

Linear (Total Transverse = F(ΔTc of Binder from Top 1/2'')) Linear ((Total Distress-transverse) = F(ΔTc of Top 1/2'' Binder))

Transverse cracking does not correlate well with change in ΔTc, but Total Distress and Total Distress-Transverse cracking are well correlated to ΔTc

Mathy Technology & Engineering

Olmsted County, MN CTH 112, 2014 (8 yrs) COMMENTS1. ΔTc does not

correlate well with transverse cracking

2. transverse cracking level is similar for all mixes, but ΔTc varies widely

3. Substantial difference in top down cracking in the test sections does correlate well with ΔTc

Relationship of Cracking to Binder Relaxation

• For purposes of my objective in this discussion the next few slides are more important than looking at ΔTc plots correlated to cracking

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

1.0E+05

1.0E+06

1.0E+07

1.0E+08

1.0E+09

1.0E+10

1.0E-13 1.0E-11 1.0E-09 1.0E-07 1.0E-05 1.0E-03 1.0E-01 1.0E+01 1.0E+03 1.0E+05

Rela

xatio

n M

odul

us, G

(t),

Log

Scal

ed

Reduced time, Log ScaledG(t) @15°C 1478, 08-27-14-D, MN1-5, 58-28, 40 HR. PAV, 4mm, G(t) @15°C 1478, 08-27-14-E, MN1-3, 58-28, 40 HR. PAV, 4mm

G(t) @15°C MN1-4, 58-28, 07-10-14-D, 40 HR. PAV, 4mm

MN1-5ΔTc=+0.8°C

MN1-3ΔTc= -4.2°C

MN1-4ΔTc= -7.6°C

Reduced Time VS Relaxation Modulus @ 15°C for MN1-3, MN1-4, MN1-5 of 40 hour PAV Residue

COMMENTS1. MN1-3 & MN1-5

have greater relaxation moduli than MN1-4 at short relaxation times

2.HOWEVER3. MN1-4 relaxes

stiffness so slowly that at extended time it intersects MN1-3

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6

Slop

e of

Log

redu

ced

time

vs. R

elax

atio

n M

odul

us d

ata

Log Reduced time

Log Reduced Time VS. Slope (m vlaue) of Relaxation Modulus

slope MN1-5, 40 HR PAV RESIDUE slope MN1-3 40 HR. PAV RESIDUE slope MN1-4 40 HR. PAV RESIDUE

MN1-5ΔTc=+0.8°C

MN1-3ΔTc= -4.2°C

MN1-4ΔTc= -7.6°C

COMMENTS1. The first derivative of

relaxation modulus curves show more clearly what is happening

2. The 1st derivative plot is the same as determining the m-value at every point along the relaxation modulus mastercurve

3. The slope of MN1-3 decreases at a faster rate than the slope of MN1-4 and the slope of MN1-5 decreases at the fastest rate of all.

4. This rate of relaxation emphasizes the interrelation of relaxation slope and level of ΔTc

1.0E+00

1.0E+01

1.0E+02

1.0E+03

1.0E+04

1.0E+05

1.0E+06

1.0E+07

1.0E+08

1.0E+09

1.0E+10

1.0E-08 1.0E-06 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04 1.0E+06 1.0E+08 1.0E+10 1.0E+12

RELA

XATI

ON

MO

DU

LUS,

G(t

), Pa

REDUCED TIME, SECONDS

G(t) @-18°C MN1-2 (PMA), 8 yr core Top ½ in, 4mm G(t) @-18°C MN1-1, 8 yr core Top ½ in, 4mm

G(t) @-18°C MN1-4, 8 yr core Top ½ in, 4mm G(t) @-18°C MN1-3, 8 yr core Top ½ in, 4mm

G(t) @-18°C MN1-5, 8 yr core Top ½ in, 4mm

Reduced Time VS Relaxation Modulus @ -18°C of Recovered Binder from Top ½ inch of8 year Field Cores of MN1-1, MN1-2, MN1-3, MN1-4, MN1-5

COMMENTS1. Plot is of relaxation

moduli of binders recovered from the top ½ inch of 8 year field cores

2. The 3 PG 58-28 binders have relaxation moduli plots that reflect their ΔTc values;

3. The plots of MN1-1 and MN1-2 (PMA binder) appear to have worse relaxation moduli even though they have the 2nd

& 3rd best ΔTc values

ΔTc Binder Recovered from top 1/2'' of 8 Year Old Field

CoresSample

IDBinder Grade ΔTc

MN1-158-34 +20%

RAP -2.5

MN1-2PG 58-34

PMA -1.1MN1-3 PG 58-28 -3.0MN1-4 PG 58-28 -6.4MN1-5 PG 58-28 1.5

MN1-4, ΔTc= -6.4

MN1-3, ΔTc= -3.0

MN1-5, ΔTc= +1.5

-0.5

-0.45

-0.4

-0.35

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

0

-8 -6 -4 -2 0 2 4 6

SLO

PE O

F RE

LAXA

TIO

N M

OD

ULU

S M

AST

ERCU

RVE

LOG of REDUCED TIME

1st Derivative of Relaxation Modulus Mastercurves @ -18°C Versus Log of Reduced Time @ 60 seconds

Slope G(t) @-18°C MN1-4, 8 yr core Top ½ in, 4mm Slope G(t) @-18°C MN1-2 (PMA), 8 yr core Top ½ in, 4mm

Slope G(t) @-18°C MN1-1, 8 yr core Top ½ in, 4mm Slope G(t) @-18°C MN1-3, 8 yr core Top ½ in, 4mm

Slope G(t) @-18°C MN1-5, 8 yr core Top ½ in, 4mm LOG OF 60 SECONDS

TEST SECTION

Slope_at_-18°C & 60

sec

MN1-1 -0.2541MN1-2 -0.2957MN1-3 -0.2634MN1-4 -0.24845MN1-5 -0.2911

COMMENTS1. This is a zoomed plot of the

slope of the relaxation modulus mastercurve vs log of reduced time for all 5 CTH 112 binders

2. MN1-1 starts out at a slightly lower relaxation modulus than MN1-5, but relaxes more slowly and by 60 seconds is relaxing at a slower rate than MN1-2

3. MN1-4 which has the lowest relaxation modulus at short times relaxes so slowly that it eventually crosses over all of the other binders and has the worst slope of all materials

Relationship of Tm-Critical to Several Parameters

MN1-1

MN1-2

MN1-3

MN1-4

MN1-5

y = 108.57x + 2.3376R² = 0.9829

-34

-33

-32

-31

-30

-29

-28

-27

-26

-25

-24

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tm-C

ritic

al, °

C

Slope of Relaxation Modulus at 60 seonds, -18°C

Tm-critical = F(slope @ -18°C), for top 1/2'' Recovered Binder

Low temp grade is controlled by Tm-critical for all bindersIt is logical that the slope of the binder relaxation modulus at 60 seconds is strongly corelated to the low temperature binder grade because the Tm-

critical value is derived from the 60 second slope values.However, it is not as obvious that relaxation modulus slopes for PMA, RAP and plain binders would all plot on the same correlation line versus Tm-criticalespecially when ΔTc and Tm-Critical are not strongly correlated for the same binders

MN1-1

MN1-2

MN1-3MN1-4

MN1-5

y = -0.6866x - 28.679R² = 0.6456

-34

-32

-30

-28

-26

-24

-22

-20

-7 -6 -5 -4 -3 -2 -1 0 1 2

Tm-C

ritic

al, °

C

ΔTc of the Binder

Tm-critical = F(ΔTc), for top 1/2'' Recovered Binder

ΔTc does not strongly correlate to the binder low temperature grade

MN1-158-34

20% RAPCanadian

MN1-258-34PMA

Canadian

MN1-358-28

Canadian

MN1-458-28

Mid East + REOB

MN1-558-28

Venezuelan

-32

-30

-28

-26

-24

-22

-20

0 50 100 150 200 250

T m-c

ritic

al

Glover-Rowe

Tm-critical=F(Glover-Rowe)

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5y = -4.1659x + 16.279R² = 0.6582

0

10

20

30

40

50

60

-7 -6 -5 -4 -3 -2 -1 0 1 2

Tran

sver

se C

rack

s (m

)

ΔTc

Transverse Cracks = F(ΔTc )

Transverse Cracks = F(ΔTc) Olmsted CTH 112 Crude Source Study

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5y = -4.1659x + 16.279R² = 0.6582

0

10

20

30

40

50

60

-7 -6 -5 -4 -3 -2 -1 0 1 2

Tran

sver

se C

rack

s (m

)

ΔTc

Transverse Cracks = F(ΔTc ) Even if you eliminate the polymer only (MN1-2) & polymer + RAP (MN1-1) data there is still not a linear correlation of transverse cracking with ΔTc

R2 was not determined, but it is easy to discern that there is no good correlation

Transverse Cracks = F(ΔTc) Olmsted CTH 112 Crude Source Study

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 488.81x + 158.4R² = 0.5519

0

10

20

30

40

50

60

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)

There is not a linear relationship between the slope of the binder relaxation modulusand the level oftransverse cracking

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 488.81x + 158.4R² = 0.5519

0

10

20

30

40

50

60

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 488.81x + 158.4R² = 0.5519

0

10

20

30

40

50

60

70

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)

Y=14.013+7.15E16*Exp(-X/(-0.00706)

The exponential relationship fits the data, but I suspect that this functionreally fits the physical reality of transverse cracking as a function of binder relaxation modulus

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 488.81x + 158.4R² = 0.5519

0

10

20

30

40

50

60

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 488.81x + 158.4R² = 0.5519

0

10

20

30

40

50

60

70

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)

Y=14.013+7.15E16*Exp(-X/(-0.00706)

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 488.81x + 158.4R² = 0.5519

y = 195.71x + 73.702R² = 0.7215

0

10

20

30

40

50

60

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)For the non-REOB binders the linear correlation between slope of the binder relaxation modulus and transverse cracking is reasonable. Keep in mind thatMN1-2 was a virginPMA PG 58-34 mix and MN1-1 was a 20% RAP mix with PG 58-34 binder. MN1-5 and MN1-3 were virgin PG 58-28 mixes. I think the most wecan conclude from thisdata is that binderrelaxation plays a rolein transverse cracking, but is certainly not the whole story

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 505.36x + 162.66R² = 0.5542

0

10

20

30

40

50

60

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24

Tran

sver

se C

rack

ing,

m

Slope of Relaxation Modulus, -18C, 60 sec

Transverse Cracking (m) = F(slope of Binder Relaxation modulus, -18°C, 60 sec)

Even if you eliminate the polymer only (MN1-2) & polymer + RAP (MN1-1) data there is still not a linear correlation of transverse cracking with Relaxation Modulus slope

R2 was not determined, but it is easy to discern that there is no good correlation

Transverse Cracks = F(Slope of Binder Relaxation Modulus) Olmsted CTH 112 Crude Source Study

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 5710x + 1761.4R² = 0.7915

y = 8523.2x + 2518.7R² = 0.9939

0

50

100

150

200

250

300

350

400

450

-0.3 -0.29 -0.28 -0.27 -0.26 -0.25 -0.24Tota

l Dis

tres

s = F

atig

ue a

rea

+ lo

ngitu

dina

l + T

rans

vers

e

Slope of Relaxation Modulus Mastercurve @ -18℃ & 60 sec

Total Distress (Excludes CL cracks)=F(Slope of Binder Relaxation Modulus, -18°C, 60 sec)

Total Distress (Non CL)=F(Slope of Relaxation Modulus, -18°C, 60 sec) all Binders

Total Distress (excludes CL)=F(slope of Binder Relaxation Modulus @ -18C, 60 sec PG 58-28 only)

Total Distress = F(Slope Binder Relaxation Modulus) Olmsted CTH 112 Crude Source Study

The relationship between total fatigue distress and the slope of the binder relaxation modulus master curve has a R2 value of 0.79for all binders, including the REOB binder mix MN1-4. Considering that this relationship includes the virgin PMA mix (MN1-2) and the 20% RAP containing PMA mix (MN1-1) this is a good result. When the PMA mixes are removed the relationship is nearly perfect. Both relationships indicate that binder relaxation plays a greater role in fatigue cracking than in transverse cracking

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 335.57x - 603.22R² = 0.6106

0

50

100

150

200

250

300

350

400

450

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3Tota

l Dis

tres

s =

Fatig

ue a

rea

+ lo

ngitu

dina

l + T

rans

vers

e

R-Value Binder from top 1/2 inch of 8 year Field cores

Total Distress (Excluding Center Line Cracks) =F(R-Value of Binder from Top ½ inch of 8 year Field Core

There is not a good correlation of Total Distress as a function of Binder R-Value when the data for all mixes are evaluated

Total Distress = F(R-Value) Olmsted CTH 112 Crude Source Study

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = 335.57x - 603.22R² = 0.6106

y = 393.42x - 705.98R² = 0.9052

0

50

100

150

200

250

300

350

400

450

1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6 2.8 3Tota

l Dis

tres

s =

Fatig

ue a

rea

+ lo

ngitu

dina

l + T

rans

vers

e

R-Value Binder from top 1/2 inch of 8 year Field cores

Total Distress (Excluding Center Line Cracks) =F(R-Value of Binder from Top ½ inch of 8 year Field Core

There is not a good correlation of Total Distress as a function of Binder R-Value; however when the polymer mix (MN1-2) and the polymer + RAP mix (MN1-1) data are removed there is a linear correlation with the non modified PG 58-28. Binder R-Values differ when polymer and/or reclaimed binders are included in the mix.The base binder for MN1-1 and MN1-2 are from the same crude source as MN1-3

Total Distress = F(R-Value) Olmsted CTH 112 Crude Source Study

MN1-1

MN 1-2

MN 1-3

MN 1-4

MN 1-5

y = -47.508x + 106.49R² = 0.9578

0

50

100

150

200

250

300

350

400

450

-7 -6 -5 -4 -3 -2 -1 0 1 2

Tota

l Dis

tres

s =

Fatig

ue a

rea

+ lo

ngitu

dina

l + T

rans

vers

e

ΔTc

Unlike relaxation modulus or other parameters such as R-Value, crossover frequency and Glover Rowe which are impacted by binder additives or crude source and therefore do not correlate well with pavement distress the ΔTc parameter appears to be blind to the presence of polymer or RAP when looking at the correlation to pavement performance.ΔTc may not correlate this well for all mixtures with a wide variety of binder types, but it appears it will always correlate better than other parameters.

Total Distress = F(ΔTc) Olmsted CTH 112 Crude Source Study

MnROAD TEST OF 3 BINDERS

1. CONSTRUCTED IN SEPT 19992. 3 BINDERS

a. PG 58-28b. PG 58-34c. PG 58-40

3. TRAFFICED UNTIL APRIL 20074. ANNUAL OR NEARLY ANNUAL PAVEMENT DISTRESS

SURVEYS CONDUCTED

Mathy Technology & Engineering

58-28

58-34

58-40

PG 58-28

PG 58-34

PG 58-40

0

200

400

600

800

1000

1200

1400

1600

1800

-9.0 -8.0 -7.0 -6.0 -5.0 -4.0 -3.0 -2.0 -1.0 0.0

Tota

l Cra

cks (

Non

CL

), fe

et

ΔTc, °C

4 Year Total Cracks (Non CL)= F(ΔTc @ 40 hr. PAV) 5.5 Year Total Cracks (Non CL) = F(ΔTc @ 40 hr.)

7.5 Year Total Cracks (Non CL) = F(ΔTc @ 40 hr. PAV)

Mathy Technology & Engineering

Total Crack Length (Non CL) @ years 4, 5.5 & 7.5 =F(ΔTc 40 hr PAV)

COMMENTS1. Between years 4 and 5.5 a

substantial increase in cracking took place for the PG 58-40 section. While the increases for the other 2 sections were not as severe they also showed an increase after 5.5 years

2. Regardless of the years in service, the cracking trended with the ΔTc of the 40 hour PAV residue.

3. No binder was recovered from field cores over the course of the project.

58-40REOB & PMA

58-34PMA

58-28Straight

0

200

400

600

800

1000

1200

1400

1600

1800

0 50 100 150 200 250

Crac

king

at 7

.5 y

ears

, Fee

t

Glover-Rowe, kPa

Cracking @ 7.5 yrs VS F(Glover-Rowe 40 hr. PAV)

MnRoad Comparative Binder Study

0.0518369258-40 REOB & PMA

0.886362258-28 Straight

0.166357558-34 PMA

0

200

400

600

800

1000

1200

1400

1600

1800

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Crac

king

Dist

ress

, Fee

t

Crossover Frequency

Cracking @ 7.5 yrs VS (Crossover Freq @15°C 40 hr. PAV)

MnRoad Comparative Binder Study

Evaluation of Relaxation of Mixes Aged for 10 and 20 Days @ 85°C

1. Six Mixturesa) PG 52-34 + 5% RAS, PG 52-34 + 5% ADD#1+5% RAS, PG 52-34 + 5%

ADD#1, 2.5% ADD#2 +5% RASb) PG 58-28 +5% RAS, PG 58-28 + 5% ADD#2 +5% RAS, PG 58-28 + 5%

ADD#3 +5% RAS

2. Binders recovered from aged mixes and characterized3. Relaxation modulus determined for mixes and binders4. Relationship between mixes and binders evaluated

1.00E+05

1.00E+06

1.00E+07

1.00E+08

1.00E+09

1.00E+10

1.00E+11

1.00E-13 1.00E-11 1.00E-09 1.00E-07 1.00E-05 1.00E-03 1.00E-01 1.00E+01 1.00E+03

RELA

XATI

ON

MO

DULU

S, G

(t),

Pa

REDUCED TIME, sec's

Log Relaxation Modulus (G(t)) vs Log Reduced Time for 10 day aged mix

MODEL: G(t) @25°C Summary 1531, 07-05-16-BA, 58-28 Straight, 4, 10d85MODEL: G(t) @25°C Summary 1531, 07-05-16-AU, 58-28 5% RS 1100 Arizona, 3, 10d85, RSSMODEL: G(t) @25°C Summary 1531, 07-05-16-AX, 58-28 5% Cargill 1103, 3, 10d85, RSSMODEL: G(t) @25°C Summary 1531, 07-05-16-AO, 52-34 Straight, 3, 10d85, RSSMODEL: G(t) @25°C Summary 1531, 07-05-16-AL, 52-34 w5% Sterol, 3, 10d85, RSS, HR3-3 (2)-3MODEL: G(t) @25°C Summary 1531, 07-05-16-AR, 52-34 5% Sterol 2.5 Cargill 1103, 3, 10d85, RSS

AO & AP PG 52-34 + 5% RAS

AL & AMPG 52-34 + 5% ADD#1+5% RAS

AR & ASPG 52-34 + 5% ADD#1, 2.5% ADD#2 +5% RAS

BA & BB PG 58-28 +5% RAS

AX & AYPG 58-28 + 5% ADD#2 +5% RAS

AU & AVPG 58-28 + 5% ADD#3 +5% RAS

These are relaxation moduli for the mixture NOT the binderNOTE the moduli at low values of time are > 1E10 Pa

1.00E+05

1.00E+06

1.00E+07

1.00E+08

1.00E+09

1.00E+10

1.00E+11

1.00E-14 1.00E-12 1.00E-10 1.00E-08 1.00E-06 1.00E-04 1.00E-02 1.00E+00 1.00E+02 1.00E+04 1.00E+06

Rela

xatio

n M

odul

us fo

r 20

day,

85°

C ag

ed m

ix T

orsi

on B

ars

Reduced Time, Seconds

Relaxation Modulus of Commpacted Mix aged 20 days @ 85°C all mixes contained 5% RAS, different Binders and Additives were employed

G(t) @+25°C 1531, 07-05-16-BB 58-28, 5% RAS 20D aged @ 85°C

G(t) @25°C 1531, 07-05-16-AV, 58-28,5% RAS, R% BO#2, 20D aged @ 85°C.

G(t) @25°C Summary 1531, 07-05-16-AY, 58-28,5% Cargill 1103, 20d85, RSS.

G(t) 1531 @25°C Summary 07-05-16-AP 52-34 + 5% RAS

G(t) @+25°C 1531, 07-05-16-AM. 5% RAS + 5% EP#1, 20 D aged @ 85°C

G(t) @25°C, 07-05-16-AS, 52-34, 5% RAS +5% EP#1, 2.5% BO#2, 20 D aged @ 85°C

Modulus results obtained using Torsion Bars tested on Dynamic Shear Rheometer

AO & AP PG 52-34 + 5% RAS

AL & AMPG 52-34 + 5% ADD#1+5% RAS

AR & ASPG 52-34 + 5% ADD#1, 2.5% ADD#2 +5% RAS

BA & BB PG 58-28 +5% RAS

AX & AYPG 58-28 + 5% ADD#2 +5% RAS

AU & AVPG 58-28 + 5% ADD#3 +5% RAS

These are relaxation moduli for the mixture NOT the binderNOTE the moduli at low values of time are > 1E10 Pa

1.00E+03

1.00E+04

1.00E+05

1.00E+06

1.00E+07

1.00E+08

1.00E+09

1.00E+10

1.00E-13 1.00E-11 1.00E-09 1.00E-07 1.00E-05 1.00E-03 1.00E-01 1.00E+01 1.00E+03

RELA

XATI

ON

MO

DU

LUS,

G(t

), Pa

REDUCED TIME, SECONDS

Relaxation Modulus of Binder Recovered from 20 day, 85°C Compacted Mix with 5% RAS and Different Binders

G(t) @25°C 1531, 07-05-16-BB, 58-28 straight, 2, 20d85, rec ac, 4mm, hr3-2G(t) @25°C 1531, 07-05-16-AV, 58-28, 5% BO#2 20d85, Rec AC, 4mm, HR3-2G(t) @25°C 1531, 07-05-16-AY, 58-28, 5%, BO#1, 20d85, Rec AC, 4mm, HR3-2G(t) @25°C 1531, 07-05-16-AP, 52-34 20d85, rec ac, 4mm, hr3-2G(t) @25°C 1531, 07-05-16-AM, 52-34 w5% EP#1, 20d85, rec acG(t) @25°C 1531, 07-05-16-AS, 52-34 w5% EP#1, 2.5% BO#1 20d85, rec ac

AO & AP PG 52-34 + 5% RAS

AL & AMPG 52-34 + 5% ADD#1+5% RAS

AR & ASPG 52-34 + 5% ADD#1, 2.5% ADD#2 +5% RAS

BA & BB PG 58-28 +5% RAS

AX & AYPG 58-28 + 5% ADD#2 +5% RAS

AU & AVPG 58-28 + 5% ADD#3 +5% RAS

These are relaxation moduli for binder recovered from aged mixNOTE the moduli at low values of time are > 1E9 Pa

BA

AUAX AO

AL

AR

BB

AVAY

APAM

AS

y = -0.0218x - 0.496R² = 0.8572

-0.5500

-0.5000

-0.4500

-0.4000

-0.3500

-0.3000

-0.2500

-0.2000

-9.00 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00

Slop

e of

Tor

sion

Bar

Rel

axat

ion

Mod

ulus

at 2

5°C

@ 1

seco

nd

ΔTc of Recovered Binder from 10 and 20 Day Aged Compacted Mix @ 85°C

Slope of Torsion Bar Relaxation Modulus @ 25 °C & 1 sec = F(ΔTc of recovered Binder)

All Samples 10 day aged compacted mix20 day aged compacted mix Linear (All Samples)

AO & AP PG 52-34 + 5% RAS

AL & AMPG 52-34 + 5% ADD#1+5% RAS

AR & ASPG 52-34 + 5% ADD#1, 2.5% ADD#2 +5% RAS

BA & BB PG 58-28 +5% RAS

AX & AYPG 58-28 + 5% ADD#2 +5% RAS

AU & AVPG 58-28 + 5% ADD#3 +5% RAS

Recovered Binder ΔTc correlates well with the slope of the mixture relaxation modulus

BA

AP

AXAO

AL

AR

BB

AV

AYAPAM

AS

y = 0.0109x - 0.1058R² = 0.84

-0.60

-0.50

-0.40

-0.30

-0.20

-0.10

0.00

-38 -36 -34 -32 -30 -28 -26 -24 -22 -20

Slop

e of

Tor

sion

Bar

Rel

axat

ion

Mod

ulus

at 2

5°C

@ 1

seco

nd

Tm-critical of Recovered Binder

Slope of Torsion Bar Relaxation Modulus @ 25 °C & 1 sec = F(Tm-critical)

10 Day aged mixes 20 Day Aged Mixes

AO & AP PG 52-34 + 5% RAS

AL & AMPG 52-34 + 5% ADD#1+5% RAS

AR & ASPG 52-34 + 5% ADD#1, 2.5% ADD#2 +5% RAS

BA & BB PG 58-28 +5% RAS

AX & AYPG 58-28 + 5% ADD#2 +5% RAS

AU & AVPG 58-28 + 5% ADD#3 +5% RAS

In addition the lowtemperature Tm-Critical value of the recovered binder also correlates well with the slope of the mixture relaxation modulus

BAAPAX

AO

ALAR

BB

AV

AY

AP

AM

AS

y = -0.4855x - 17.511R² = 0.92

-9.00

-8.00

-7.00

-6.00

-5.00

-4.00

-3.00

-2.00

-1.00

0.00

-38 -36 -34 -32 -30 -28 -26 -24 -22 -20

ΔTc

Tm-critical of Recovered Binders

ΔTc = F(Tm-critical)

10 Day aged mixes 20 Day Aged Mixes

AO & AP PG 52-34 + 5% RAS

AL & AMPG 52-34 + 5% ADD#1+5% RAS

AR & ASPG 52-34 + 5% ADD#1, 2.5% ADD#2 +5% RAS

BA & BB PG 58-28 +5% RAS

AX & AYPG 58-28 + 5% ADD#2 +5% RAS

AU & AVPG 58-28 + 5% ADD#3 +5% RAS

SUMMARY COMMENTS• Parameters such as ΔTc, Glover-Rowe, R-Value, crossover

frequency are manifestations of binder relaxation• Binder relaxation largely drives mix relaxation for the aged

mixes we studied• Tm-Critical and ΔTc of recovered binders correlated to mix

relaxation• Slope of relaxation modulus mastercurves appear to

correlate well with ΔTc for a variety of binders • Slope of relaxation modulus did not correlate well with

transverse cracking on the Olmsted CTH 112 project

SUMMARY COMMENTS

• ΔTc did not correlate well with transverse cracking on CTH 112, but did correlate well with total cracking

• Slope of binder relaxation modulus at -18C correlated reasonably well (R2 =0.79) with total cracking on CTH 112 for all 5 test sections including virgin PMA (MN1-2) and PG 58-34 + 20% RAP (MN1-1)

• ΔTc correlated well with the project cracking even when modified binders were used

• Glover-Rowe, crossover frequency and R-value did not correlate well when evaluating mixtures produced with straight run and modified binders

WI STH 33 @ 4 years of age

Mathy Technology & Engineering

WI STH 33 @ 8 years of age

Mathy Technology & Engineering

At 8 years cracking has started, some transverse, some wheel path. This is more consistent with the onset of distress than the pervasive deterioration seen on some sections of CTH 112 and MnROAD

Top 1/2 inch of core extracted and recovered

Core Time after construction

S critical,°C

m critical, °C ΔTc, °C

4 year -30.2 -30.9 0.78 year -28.9 -26.6 -2.3