7/30/2019 SHRP-91-510

1/41

SHRP-A /UWP-91-510

Chemical Properties of Asphalt sand Their Relationship to

Pavement Performance

Raymond E. Robertson

Western Rese arch InstituteLaramie, WY

Strategic Hig hway Researc h Pr ogramNationa l Research Counc il

Washington, D .C. 1991

7/30/2019 SHRP-91-510

2/41

SHRP-A /UWP-91-510Contract A-002A

Product Code 1001, 1007

Program Manager: Edward T . HarriganProject Manager: Jack S . YoutcheffProgram Area Secretary: Juliet Narsiah

March 1991Reprinted November 1993

key words:adhesion

agingasphaltchemical mod efailure mod els

fatigu e cr ackingimer molecularmetalsmoistur e damageoxidation

pavem ent p er for manc eper manent defo r mationpolarit yruttingther mal c r ackingvirgin asphalt

Strategic High way Research ProgramNati onal R esearch C ounci l2101 Constitution Avenue N.W.

Washington, DC 20418

(202) 334-3774

This manual represents the views of the author o nly, and is not necessarily reflective of the views of the

National Research Council, the views of SHRP, or SHRP's sponsor. The results reported here are notnecessarily in agreement with the results of other SHRP research activities. They are reported to stimulatereview and discussion within the research community.

50 /NAP /1193

7/30/2019 SHRP-91-510

3/41

Acknowledgmen ts

The research described herein was supp orted by the Strategic H ighway ResearchProgram (SHRP). SHRP is a unit of the National Research Council that was authorizedby section 128 of the Surface Transportation and Uniform Relocation Assistance Act of1987.

Th is pr oj ect w as conducted by the Weste rn R esearch In stitute in coo perati on wi th thePenn sylvania Tran spo rtation In sti tu te, the Texa s Tran spo rtation In sti tute, and SRIIn ternational. Raymond E. Robe rtson was the p rincipal in vestiga tor. The suppo rt andenco uragemen t of Dr . Edwa rd T. Harr igan, SHRP Asphalt Prog ram Manage r, and Dr.Jac k Youtche ff, SHRP Technical Cont ract Mana ger, are gra tefully ac knowled ged.

The experimental work cited herein was conducted by Dr. J. F. Branthaver , Dr . K.Ensley, Dr. J. Duvall, H. Plancher, P. M. Harnsberger, S. C. Preece, F. A. Reid, J.Tauer, M. Aldrich, A. Gwin, M. Catalfomo, G. Miyake, and J. Wolf. Their support inthe preparation of this report is gratefully acknowledged. The review of this manuscriptby Dr . J. Clain e Petersen, former pr incipal inves tigato r for A-00 2A; Dr . Jan F.Branthaver at W RI; and Dr. David Anderson at PTI, co-principal investigator forA-002A; and Dr. Edward T. Harrigan and Dr. Jack Youtcheff of SHRP is herebyacknowledged.

nl

7/30/2019 SHRP-91-510

4/41

Con ten ts

1.0 Introdu ction ................................................... 11.1 Objectives of the SHRP Asphalt Program ........................ 11.2 Purpose of Report ......................................... 2

2.0 Chemistry .................................................... 32.1 Molecular Level ........................................... 3

2.1 .1 Typic a l Spe cies in Virgin As phalt ........................ 32.1.2 Oxidation to Form New Molecules ....................... 52.1.3 Metals ............................................ 92.1.4 Polarity ........................................... 9

2.2 Intermolecular Level ...................................... 102.3 Chemical Mode of Asphalt .................................. 17

3.0 Spec u lation on Relationship of Chemistry to PavementPerformance ................................................ 203.1 General Comments ....................................... 203.2 Aging ................................................. 213.3 Speculation on the Relationsh ip of Composit ion

to Various Failure Models ................................. 233.3.1 Rutting and Permanent Deformation ..................... 243.3.2 Thermal and Fatigue Cracking ......................... 253.3.3 Adhesion and Moisture Damage ........................ 26

4.0 Summary .................................................... 29

V

7/30/2019 SHRP-91-510

5/41

Li st of St ructu res

1 M olecular Fragment of an Aliphatic Hydr ocarb on ........................ 4

2 M olecular F r agment of an A ro matic Hy dro car bon ........................ 4

3 M olecular Fragmem of an A ro matic Hydr ocarb on ........................ 4

4 M olecular Fragment of an Aliphatic and Ar omatic Hy drocar bon ............. 4

5 M olecular Fragmem of a Pyridine ................................... 5

6 Molecular Fragmem of a Thi ophene .................................. 5

7 M olecular Fragmem Sh owing Benzyl Carb on ........................... 6

8 M olecula r Fragmem of a Ket one .................................... 6

9 M olecular Fragmem of a Carb oxylic Aci d .............................. 6

10 Molecular Fragmem of a S od ium Carb oxylate ........................... 7

11 M olecular Fragment of a Calcium Carb oxylate .......................... 7

12 M olecular F ragment of a Ca r boxylic Anhy dride ......................... 7

13 M olecular Fragment of a Phen ol .................................... 7

14 M olecular Fragment of a H omolog .................................. 8

15 M olecular Fragment of a H omolog .................................. 8

vii

7/30/2019 SHRP-91-510

6/41

16 Molecular Fragment of a Homolog .................................. 8

17 Molecular Fragment of a Quinolone .................................. 8

18 Molecular Fragment of a Sulfoxide .................................. 9

19 Metalloporphyrin ............................................... 9

Vll"""

7/30/2019 SHRP-91-510

7/41

List of Figures

1A Randomized Molecules......... " " " * " " ......... .............. 10

1B Organized Molecules . o o 10

2 Three Dimensional Molecular Matrix .............................. 11

3 S.E.C. Chromatograms ........................................ 14

4 Effects of Temperature on Aging Kinetics of an Asphalt Sensitiveto Molecular Structuring ....................................... 17

5 Performa nce ................................................ 21

ix

7/30/2019 SHRP-91-510

8/41

F, _C_ ZVE S_Y

The che mistry of petrole u m asphalt at th e molecular and

intermolecular levels is discussed herein and a combination of

interpretation and speculation placed on how chemical data may explain

performance characteristics in roadways. At the molecular level

historical studies have shown that there are at least hundreds of

thousands of uniqu e mol e cular species that exist within any particular

asphalt. Asphalts are known to b e comprised of molecular species that

vary widely in polarity and molecular w e ight. Th e major obj e ctiv e of

this portion of th e work is to explain behavioral characteristics of

asphalts in t e rms of ch e mistry. Polarity is a very major contributor to

th e performance characteristics. Polar materials tend to associate

strongly into a matrix which is dispersed in less polar and non-polar

materials. In g e neral, th e mechanical or structural properties of

asphalt ar e r e lat e d to th e inte rm olecular structuring among polar

components. These polar interactions may arise by involvement of any of

numerous different ch e mical species. Th e exact nature of th e chemical

specie is l e ss important than th e overall ass emb lag e of a s e t of polar

materials to form a matrix within th e non-polar continuous medi um . Th e

matrix gives elastic charact e r to th e asphalt while the continuous non-

polar phase gives a viscous component to asphalt. It is speculat e d that

ex cessive structuring leads to brittl e cements which tends to crack

while too little structuring leads to mat e rials which defo rm und e r

stress. Oxidation adds to the ag e hardening and brittleness of asphalt

cement by contributing additional polar materials to the structured

zones within the binder. At very low temp e ratur e it appears that the

xi

7/30/2019 SHRP-91-510

9/41

non-polar mate r ials also tend to organize into a very rigid material,

and that this rigid mat e rial shrinks at low t e mperatur e . It is

sp e culated that this type of shrinkage is largely responsible for low

t e mpe ratur e cracking.

A second major obj e ctiv e is to translat e r e s e arch me thods into t e st

methods. Numerous types of research m e thods have been employed to

differentiate among asphalts with an emphasis on identifying methods

which distinguish among asphalts in the same fashion that they ar e

distinguished by their performance characteristics.

Th e e xp e cted advantages of this work, both technical and economic,

ar e to dev e lop methods for s e l e ction of asphalts that will perform in a

predictable fashion th e reby alleviating the current co n unon problem of

prematur e road failure. Throughout t h e progr am , both us e rs and

produc e rs of p e trol e um asphalt have worked closely with SHRP contractor

personn e l both to give advic e on progr am dir e ction and to maintain a

sense of what is practically implementable in each sector.

As with any new development questions arise as to how much

sp e cialization will be required for implementation of new test methods.

Certainly n e w test methods could become extraordinarily complex, but it

is still another obj e ctiv e to assur e that there is no mor e complexity

than is absolu te ly necessary. Problems of impl em entation, ther e fore,

should be minimiz e d and espec'ially so when one considers that each new

test method will enter a round robin test progr am in both state highway

and user laboratories. The round robin effort, in fact, has already

xii

7/30/2019 SHRP-91-510

10/41

b e gun. F i n a ll y , t his w o rk i s su ppo rt ed by pu bli c f u n d ing and the

r e sults that ar e impl e mented into new t e st m e thodology will be public

domain.

xiii

7/30/2019 SHRP-91-510

11/41

1. 0 INTRODUCTION

i.i Objective s o f the SHRP A s phalt Pr o gram

Th e SHRP asphalt research area is a highly focused endeavor which

ultimately must d e v e lop specifications for binders and asphalt-aggregate

mixtures that relat e to th e p e rformanc e of both binders and mi x tures in

pav eme n t s. To accomplish this th e fund a mental material properties must

b e known first. Then fund am ental properties are to be used to develop

me aningful sp e cifications. The current specifications for petrol e um

asphalt are little mor e than a quality control exe rcis e for th e r e finer's

vacuum crude tow e r operations, hence there is little or no distinction

among asphalts other than a s e t of viscosities at two elevated

t e mp e ratur e s and a mix plant oxidation susceptibility. That is, all of

many asphalts may b e classifi e d as, for e xampl e , AC-20, and all will be

applied as if they are a single material. Yet, their perfo rm ance

charact e ristics in pavements often vary quite significantly. Th e results

of using poorly classified asphalts have been rude and costly surprises

in th e form of various types of early failures with increasing frequency.

It behooves the user, th e refore, to d e t erm ine what properties of asphalt

binders are r e flected in pavement p e rformance and then select binders

which have d e sirabl e p e rformanc e prop e rties. From a practical viewpoint,

no doubt many petrol eum residua will require modification to prepare

asphalts that have desirable p e rformanc e characteristics.

7/30/2019 SHRP-91-510

12/41

1.2 Purpo s e o f Rep o rt

Th e purpose of this report is to d e scrib e th e current status of th e

SHRP chemical studies of petrol e um asphalt. It is intended to b e

instructiv e to th e non-chemist and further is a speculative effort to

correlate some of th e known ch e mical properti e s of asphalts with pavement

performance characteristics. It is generally believed that performance-

bas e d specifications for binders will b e mainly physical property tests.

The obj e ctive of the composition studies is to develop correlations

between th e chemical and physical properties. Further, the composition

studies ar e to define a s e t of practical analytical methods to describe

composition. This then will allow definition of acceptable composition

for a given p e rformanc e which in turn is instructive to the producer in

th e manufacture of asphalt. If any given asphalt has a compositional

deficiency, th e producer can determine how to modify th e asphalt and

remedy th e deficiency. It has become obvious that deficiencies vary from

on e asphalt to another. It is important to note that ma ny ch e mical

properties will not correlate with performance so it is a ma jor goal to

distinguish th e chemical properties that do relate.

Referenc e s ar e made to ongoing research which has been reported to

SHRP, but in many cases has not been published. Many statements are

based on current SHRP research and have no literature referenc e s.

7/30/2019 SHRP-91-510

13/41

2.0 CH EMISTRY

It is convenient to divide the chemistry of petroleum asphalt into

two parts. The first part is the ch e mistry at the molecular level and

th e second part is the chemistry of interaction among all molecular

species in asphalt. Much of the physical nature of asphalt can best b e

describ e d, in t e rms of composition, as an a sse mbly or matrix of molecular

species (building blocks) into large multi-molecular units within the

asphalt. N ume rous techniques exist to e xamin e asphalt at both th e

molecular and int erm ol e cular lev e ls. In the following subsection the

ch e mistry of asphalt at th e m ol e cular l e v e l is r e vi e wed without r e f e r e nc e

to e x p e ri me ntal methods. This is to illustrate what types of "building

blocks" ar e pres e nt. The che mi stry of interaction am ong asphalt

molecules is r e viewed in th e second subsection with reference to som e of

th e t e chniqu e s being used curr e ntly in th e SHRP progr am .

2.1 M o lecular Level

Extensi ve research on petrole um composition at th e molecular l e v e l

has be e n done by many workers for many years. At this point, much of it

is e ssentially textbook material b u t worth r e vi e wing h e r e .

2.1.1 Typical Species in Virqin Asphalt. At the molecular level

most of th e total mass of neat (tank) asphalt is a mixture of a wide

variety of high boiling hydrocarbons. Some are aliphatic (waxy

mat e rials), some ar e aromatic (more like air-blown asphalts) and some

molecules hav e both aliphatic and aromatic carbon. E xam pl e s of such

molecular fragm e nts ar e shown in Structures 1 (aliphatic type) and 2

(aromatic type).

7/30/2019 SHRP-91-510

14/41

H H HI I I

H_ C _"C" C_"C "c /C_ C /

_CH2_ ICH2 _x_CH3_ I I II IHIC-_c /C_c /C _c _C_H

x Typically = 15 or More Carbons I I IH H H

Structure 1 Structure 2

Structure 2 is more co mmonly represented by Structure 3. Both s tructures

1 and 3 are only representative molecular fragments of larger molecules

that make up asphalt s . Structure 1 is s hown as a straight chain of

carbon atoms, but typically in asphalt there ar e numerous co mbinations of

aliphatic carbon chains which have one or more branches.

Structure 3

Structure 4 shows a molecule which i s a mixture of aliphatic and ar o matic

carbon.

2

____C H 2 (CH)x- CH3

Structure 4

While the major mass of asphalt is hydrocarbon, a large proportion of

molecules also contain one or more heteroato ms; nitrogen, sulfur, oxygen

4

7/30/2019 SHRP-91-510

15/41

and metals. Nitrogen, in the form of a pyridin e is shown in Structure 5.

CH2-(CH2)x" - CH 3

Structure 5

Sulfur, as a benzothiophene, is shown in Structure 6. Structure 6 on the

right is a shorthand notat i on of the one to the le f t. These are

identical molecular fragments.

HI

--H or

/

Structure 6

Bo th su l fur and ni tr o ge n ma y a ppe a r i n a ny o f a vari e ty o f sites withi n

molecules. Hence, it is easy to see that tens of thou s ands of di ff erent

molecul a r species may be present in a s phalt con s idering that every

different arrangement of elements constitutes a different molecule. It

also follows that any definition of properties ba s ed on specific

molecular species would be a mon u mental task. It is much more e f fective

to classi f y the chemistry in ter ms of the molecular type s .

2.1.2 Oxidation to Form New Molecules. Certain types of carbon in

asphalt are susceptible to o x idation. An aliphatic carbon ne x t to a n

7/30/2019 SHRP-91-510

16/41

aromatic ring is known a s a b e nzyl carbon and is an example of a r e adily

o x idizab l e site. Structure 7 is an e x ample of a mol e cular fragm e nt

showing a b e nzyl carbon in bold type (adjacent to th e aromatic ring).

CH2-(CH2 )x

Structure 7

Sites such as these oxidize to form ketones as shown in Structure 8.

C_ (c H2)x--

Structure 8

More severe oxidation m ay result in formation of carb ox ylic acid s and

lo ss o f part of the molecule such as shown by both versi o ns of

Structure 9.

rStructure 9

Carbox y l ic acids, whether present in th e origina l crude or for me d up o n

o x idati o n, may be c o nv e rt e d to sodium (Na) salts (Structur e 1 0) o r

7/30/2019 SHRP-91-510

17/41

calcium (Ca) salts (Structure 11 ) by appropriate reaction with sodi u m or

calci u m inorganic compounds.

O O OII II II

Struct u re i0 Structure 11

Carboxy l ic acid anhydrid e s (typica ll y call e d anhydrides) may be formed

upon o x idation wh e n two b e nzyl carbons are pr e sent on adjac e nt aromatic

rings. An ex ampl e is Structur e 12.

0% c/O_c // 0

(CH2)xCH3

Structure 12

Anothe r type of o x yg e n containing m o l e cule which may b e present in

asphalt is a class known as ph e nols wh e r e o x ygen is attach e d dir e ctly to

an aromatic ring. This class is illustrat e d by Structure 1 3.

OH

CH3-(CH2) x'- CH2,_ CH 3

Structure 13

7

7/30/2019 SHRP-91-510

18/41

Again n u merous variations, or iso me rs, containing t h e basic phenol unit

may e x ist. I n addition to isomeric combinations o f all mo le cular types

shown above, homologs of each also typically exi s t. An example of a

homologous s e ries is shown by Structures 1 4, 15, and 1 6.

0II

C_ cH2 (CH2) x CH3

-- -- Structure 14

OII

C_c H2--(CH2)x+I-CH3 Structure 15

OII

C_ cH2--(CH2) x+2--C H3 Structure 16

In the above case , each varies only by one aliphatic carbon. Each is a

unique molecule, although the properties a mo ng all of th e s e will be quite

similar.

Another important class of c o mpo und s typically found in aged

asphalts is quin o lones a s shown in Structure 1 7.

Structure 17

7/30/2019 SHRP-91-510

19/41

Many s ulfur comp ou n ds are al so s u s cept ib le t o o xidation and

typically f o rm sulfoxides. A sulfoxide containing fr a gment is shown in

Structure 18.

OII

-(CH2)xCH2 -S -CH2 -CH 2 -

structure 18

2.1.3 Metals. There are also metals present in asphalts, again in

varying amounts and distributions. The most common metals are vanadium,

nickel, and iron although other metals may also be present. Typically

metals are present as organo-metallic materials, specifically as

porphyrins. An ex a mple is shown by Structure 19.

CH3CH 2 H CH3

CH3@__ CH2CH3

3NZ'N (

cH cH2 H3CH3 H CH2CH3

Str ucture 19

By now it is obvious that hundreds of thousands of unique molecules

may be found in any given asphalt. Further, the second, third and so

forth asphalts will contain hundreds of thousands of different molecules.

2.1.4 Polarit y. All of the naturally occurring heteroatoms,

nitrogen, sulfur, oxygen, and metals contribute to polarity within these

7/30/2019 SHRP-91-510

20/41

molecules. Likewise, o xidation products formed up o n aging are p o lar and

further contribute to th e polarity of the entire s ystem. Polarity, which

is th e separation of charge within a molecule, can be seen by the

following exampl e . Th e dipole moment (separati o n of charge) of pyridine

(C5H5 N) is 2. 1 9 d e byes (in th e gas phase) wherea s the dipole moment o f

benzene (C6H 6) i s zero. B e nzene is th e all carb o n anal o g of pyridine.

Polarity also exists in all other heteroat o m containing species.

Polarity is imp o rtant in asphalt because it tends to cause m o lecule s t o

organize themselves into preferred orientati o ns. Historically, the s e

have been referred to a s formations of micelles, coll o ids, etc., alth o ugh

th e s e t e rms have been mi s us e d. A more current understanding of molecular

orientation within asphalts is given in t h e following Subsecti o n.

2.2 Interm o l e cul ar Lev e l

At the int erm ol e cular level, polar m o lecule s including tho s e in

a s phalt, have an o ther behavioral characteri s tic. Thi s is attraction of

o n e polar molecule for another as a result of their s eparated charges, or

dip o les. Figur e 1 illustrates this sch e matically.

(- + )C -

Figure 1A Figure IB

1 0

7/30/2019 SHRP-91-510

21/41

In part A the polar mol e cul e s are randomized, but in part B th e molecules

ar e well orient e d with respect to each other. Part B r e pres e nts a mor e

stable th e rmodynamic state. It is important to note here that it makes

little difference which of the man y polar molecules shown earlier is

involved. Any one of th e m any types of polar mol e cu le s may fil l th e

mo le cular sch em atics shown in Figure 1 A or B. The primary require me nt is

that some sort of charge separation is pres e nt in the molecules. It is

obvious that a mu l ti- molecular structure may form as illustrated

sche matica l ly in Figure 2, although the individua l m o le cular co mponents

will vary from one to the ne x t so that no sp e cific reg u larity e x ists

within the organized zone (see note I).

G 0C+ -) C+ -) C+ - 7

0 0F igure 2(Note 1 ) For si mp l icity, this schematic shows positive and negative

charg e s at the ends of molecules. The associations among molecules ar e

combinations of e l e ctrostatic and other short range forces. Th e actual

charges are b e st d e fined in term s of asymm e tric e lectron den s ity and ar e

not true plus a nd minus charges as would be the case with ions. Neither

are charges neces s arily distribut e d e nd to end. F ig u r e 2 is only

intended as a convenient illustration.

1 1

7/30/2019 SHRP-91-510

22/41

Du r i ng t h e o r g an i zat io n , so me a moun t of thr e e -d i men sional, i n t e rmolecul a r

structure forms. Historically, this has be e n called the mic e ll e , or

colloid. While it is not a true mic e lle, it is an organized set of

mol e cul e s. Th e set does hav e som e pr e f e rr e d organizational structure as

compared to that shown in Figure IA where it is only randomized

molecules. Th e structur e is held together by el e ctrostatic and other

short rang e forc e s which are weak compar e d to covalent ch emi cal bonds.

Sh o rt rang e ( n on - coval e nt) forc e s rang e from about 3 to I0 Kcal / mol e ,

whil e coval e nt bonds are much strong e r. For comparison, carbon-carbon

covalent bonds, the bonds that hold organic molecules together, are 80 or

great e r Kcal / mol e and carbon-hydrogen covalent bonds are typically i 00

Kcal / mo le . It follows th e n that the o rganized (intermolecular) structur e

may be subject to rearrangement or may be scrambl e d either from physical

stress or by raising its temperature. However, this will occur without

changing mol e cular composition. All of th e molecular speci e s remain th e

sam e . On th e other hand, t h e physical properti e s will be different.

Whe n th e mol e cul e s ar e rando mi z e d, th e y can mov e about with r e spect to

e ach other more easily than when they ar e more organiz e d. Th e mor e

highly organiz e d structur e has mor e resistance to motion or d e formation.

Said differently, structured asphalt is more of a springlik e material,

more viscous, and is stiffer. Th e ability to form an organized, or self-

asse mbl e d, structure d e p e nds upon th e strengths of th e attractions and

upon th e n u mb e r of sites wh e r e int e rmol e cular attractions occur.

Ox idation has a pronounc e d e ff e ct on th e organiz e d structure. As

o x idation occurs, new sites that ar e gr e at e r polarity ar e fo rme d and ar e

formed in larg e r amounts than in the virgin asphalt. So th e propensity

to self-as sociat e will increase. Also, the rate of association depends

1 2

7/30/2019 SHRP-91-510

23/41

upon the nu mb e r o f s it e s an d the magnit ud e s o f the a tt r act i on s . The

stronger the attractions and the more there are, the greater the driving

force to associate. But, association is inhibited by the high viscosity

of a sphalt. Hence, the overall process is slow. For example, in one

experiment, virgin asphalt was observed to double in viscosity over a

period of a few years. The sample was protected so that oxidation did

not contribute anything. After this extended storage, the sample was

heated to mix plant temperature while protecting it from oxidation. The

viscosity returned to near its original value. The rates and magnitudes

of stiffening of asphalt after oxidation, with and without aggregate

present, and at various temperatures, are being studied at the present

time in the SHRP asphalt progr a m.

The degree of association varies from one asphalt to another and

several methods can be used to dete r mine type and magnitude of

association. One method being used at the present ti me is size exclusion

chromatography. It is illustrated in Figure 3 which shows three size

exclusion chromatography (SEC) experiments, all plotted on the same axes.

Asphalts AA G, AAK, and AAM (hereafter noted as G, K, and M) are all

similarly classified petroleum asphalts. All three were separated by S EC

in the same fashion. The S EC process separates materials (in this case,

asphalts) into components according to size at the molecular or multi-

molecular level. It matters not whether the size excluded entity is a

single molecule or an associated group of molecules. In the current

work, a system was chosen that causes the least possible disturbance to

the association. The SEC separates asphalt by apparent molecular size so

that if there are associated groups of molecules in the whole asphalt

they will also exist and therefore be separated as an associated group by

SEC. The SEC profiles sh o wn in Figure 3 are plots of the fraction of the

13

7/30/2019 SHRP-91-510

24/41

SEC CHROMATOGRAMS

3 0 AAM /_\

-- \ ' _AAG25 /

" == _ //1 _

\

20 II

.,. , \15: \

*_ 1 0 f'\ \/ _ /(J I _a \

I \u_ I

\0

0 20 40 60 80 100 120

Large SmallMolecular Molecular

Size Size

Fi 9ure 3

14

7/30/2019 SHRP-91-510

25/41

wh ol e a s pha l t o n t he y -ax is and t he mo l e cula r or mul t i - mo l e cular si ze on

the x-a x is. The three materials illustrated are virgin asphalts. Note

that asphalt G has a very small fraction that is in the large molecular

size, and a significantly larger fraction that is smaller molecular size

(the right side of the plot). Asphalt K is almost the reverse. Further,

these two are bimod a l indicating that there is a relatively small amount

of intermediate size materi a l in either. Asphalt M, however, is more of

a continu u m. That is, it is not bimodal. Asphalt M does have a

significant amount of large molecular species present as does K. Upon

further examinations of M and K by vapor-phase-osmometry molecular weight

determination in pyridine, it was discovered that the large molecular

species in K are comprised of many smaller molecules where a s in Asphalt M

the largest fraction (left of plot) is comprised of truly large

molecules. The l a rge molecules in M will not dissociate, for example,

with an increase in temperature, whereas the large ones in K will

dissociate. If nothing else is apparent, note that these three asphalts

a ppear to be very different from e a ch other when exa mined by SEC. Yet

conventional tests would indic a te that all three are si milar.

Another effect within asphalts is the behavior of the molecules

which have very little association. These are not necessarily smaller

molecules, but are the less polar and therefore less associated portions

of the asph a lt (to the right of the SEC plot). Ion exchange

chromatography (IEC) and supercritical fluid chromatography (SFC) h a ve

both been used to show that these molecules a lso vary from one asphalt to

another. Often called a maltene phase, this less-associated material

behaves as a "solvent" for the polar materials and will behave as a

dispersing agent. It will tend to reduce association of the polars.

15

7/30/2019 SHRP-91-510

26/41

At this point, it is worth considering a more integrated description

of asphalt. It appears to be a material containing polar molecules that

associate strongly into organized units that are dispersed in a less

polar and continuous phase. The association of polars appears to depend

upon composition of polars and upon th e rmal history. Some speculation

can be made on the effects of the differing chemistry within the

continuous n e utral phase. First, it is apparent from the SEC curves that

asphalt G has a large fraction of small molecular species which implies

that th e l e ss polar and non-associat e d materials dominate its behavior.

It would b e expected that G would be able to acco mmodate additional

amounts of polars with minor changes in its physical properties. For

e x ampl e , asphalt G should be able to tolerate significant oxidation

b e for e any major viscosity change occurs and this is what is observed.

Asphalt K should be and is th e reverse. Oxidation of K should increase

th e alr e ady predominant polar effect and raise viscosity rapidly with

oxidation. This likewise is observed.

The propensity to oxidize has been observed by another method being

us e d in the SHRP asphalt studies. Oxidation in a pressurized vessel is

b e ing stu d i e d as a m e thod to simulat e long-te rm aging of asphalt in

pav e ment. After 400 hours of oxidati o n at 60C (140F) und e r pre s sure,

asphalt G had an aging index of 17 while K had an index of 23. When th e

temperature was raised to 11 3C (235F) and both oxidized at atmospheric

pressure for 72 hours, asphalt G had an aging index (measured at 6 0C) of

only 1 8 while K jump e d to 530 under th e same conditions. At the elevated

t e mp e ratur e , both asphalts ar e mor e dissociated and oxidize rapidly.

Upon cooling both to 6 0C, G can acco mmodate its own oxidation products

whereas th e newly formed polars in K strongly dominate i t s behavior. The

16

7/30/2019 SHRP-91-510

27/41

po lar s i n aspha l t K a s s o c i ate v er y s t rongly a nd rais e the v i sc osit y qu it e

substa n tially. The same trends have been observed with other asphalts.

While on the subject of differing rates of o x idation of any given asphalt

at different temperatures, note in Figure 4 how the viscosity ch a nge

inc r e a ses with incre a sing oxid a tion temper a ture. This indic a tes that

maximum ro a d service temperatures must be taken into account in the

selection process for asphalts.

EFFECTS OF TEMPERATURE ON AGINGKINETICS OF AN ASPHALT SENSITIVE

TO MOLECULAR STRUCTURING

106 -

60C POV Aging

105

"_ 1 0 4

103 I I I I0 100 200 300 400

Oxidati o n Time, hr

Figure 4

2.3 Chemical Mo d e l o f A s p h alt

Th e model that is emergin g f rom this w ork is b u il t up o n e a rl ier

mo de l s and ha s b een r e fin e d du r ing th e SHRP program. I n this cas e mod e l

means nothi n g mor e than a cl e ar und e rstandi n g of th e b e havioral

charact e ristic s of a s phalt. Th e mod e l must b e abl e to e x plain all of th e

obs er v e d b e havioral charact e ristics a s we ll as all of th e variations i n

b e havior fr o m on e a s phalt t o an o th e r. He r e , it i s wo rth whil e to r ev i e w

17

7/30/2019 SHRP-91-510

28/41

the manu fa ctu r e of as ph a l t. P et r o l e u m asp ha l t i s typ ically a high

boiling vacuum distillation residuu m which is prepared f rom nu merous

petroleum stocks. Some asphalts are produced by alternate methods but

generally very similar portions of the crude oil ends up as asphalt. The

chemistry and physical properties, therefore, vary quite significantly

from one asphalt to a nother and each reflects the nature of the crude oil

used to prepare it. The most consistent description, or model, of

petroleu m a sphalt is as follows. Asphalt is a collection o f pol a r and

non-polar molecules. The polar molecules tend to associate str o ngly to

form organized structures throughout the continuous phase of the non-

polar materials. Nuclear magnetic reson a nce data and the rm odyna mic data

indicate that the associations are not more than about 40 molecules, but

some have smaller asse mblies and a gain it varies f rom one asphalt to

another. Some show very little association. The non-polar phase, on the

other hand, has the a bility to dissociate the organized structure, but

again it varies from one asphalt to another. As temperature is raised,

the associations of polar molecules decre a ses and the materi a l becomes

more dissociated and therefore less viscous. As temperature is reduced,

the opposite occurs. Recent observations indicate that the non-polar

phase also organizes, but at very low te mperature, temperatures below

0C. Further, asphalt is susceptible to oxidation which increases both

the amount of polarity and the n u mber of polar sites pre s ent among

asphalt molecules. This further contributes to the ability of an asphalt

to organize, but again varies from one asphalt to another.

It is important here to point out that the variations in behavioral

characteristics of asphalts must be measurable. And ag a in the objective

of the SHRP program is to elucidate methods which can distingui s h among

18

7/30/2019 SHRP-91-510

29/41

asphalts and ev e ntually reduc e thes e methods to highway laboratory

practice. Whil e th e quantitation of all important aspects of th e model

has not b ee n completed, it is int e r e sting to speculate on how the

variations in chemical properties may be reflected in the pavement

p e rform a nc e .

1 9

7/30/2019 SHRP-91-510

30/41

3. 0 SP ECULATION ON R ELATI ONSHIP OF CH EMISTRY T O PAVEM ENT P ERFO RMANCE

3.1 General C o_tn ent s

Whil e it is unusual to d e scrib e p e rformanc e in terms of failure,

nev e rtheless i t is instructive to do so here since it is g e nerally

und e rstood that all pav em ents will eventually fail unless th e y ar e

r e built periodically. Performanc e must b e d e fin e d as sufficient time to

failure or to r e constructio n to justify th e us e of any particular

met h odology an d any g iv e n set o f ma teri a ls. S uff i ci e n t s er v i c e li f e is

determined by cost of construction, traffic density, harshness of the

environ ment, soil (support) conditions, and nu merous other factors. The

determination of acceptable service life is not the subject of this work.

However, service lifetimes of a few months to a few years, which are

experienced all too frequently, are unacceptable, whereas i0 to 20 years

of service gives more acceptable life cycle costs. Ne i ther is

methodology the subject of this work, so it will be assumed that

construction methodology is both consistent and adequ a te. The focus will

be upon variations in the quality of construction materials.

The s i gnificant fa ilure modes in asphalt p a vement that may be

related to m a terials are g eneral l y a g reed to be

I. Permanent deformation

2. Rutting

3. F a tigue crack i ng

4. Low temperature cracking

5. Moisture damage

6. Total loss of adhesion

2O

7/30/2019 SHRP-91-510

31/41

Some speculations on the relationships of these failure m o des t o the

chemical properties of asphalts ar e given in the s e ctions following the

section on aging.

3.2 Aging

Aging is often included in the above lis t, but aging is actually a

conditioning st e p which may be ben e ficial o r detrimental. It may b e

detrimental when excessive hardening or stiffening is ob s erved in an

already adequate pavement. On the other hand, aging may be beneficial

when a soft mi x ture hardens into an adequate pavement. A simplified

view of performance in terms of m e chanical properties is that a pavement

may be too soft and therefore rut and def o rm, or ma y be too stiff and

brittl e and therefore subject to cracking, either under traffic l o ad or

under th erm al s tress. Figure 5 illustrate s these points schematically

and also include s the effects of temp e rature, time, and oxidation.

PERFORMANCE

Little Well HighlyAss o ciati o n Balanced Associated

I ISOFT BRIT TLE

Perma n e n t Fatig u e and Low

Deformatio n , Temperat u reRutting Cracki n g

Time, Aging, Cooli n g

Moist u re, Heati n g

Figure 5

21

7/30/2019 SHRP-91-510

32/41

Oxidation, or o x idativ e hardening, imparts permanent hard e ning in

asphalt whil e the hardening from reduced temperature and from molecular

organization ar e r e versibl e . That is, as an asphalt cement oxidizes,

whether in bulk or in a mixtur e , the cement becomes and stays stiffer at

any given s e t of conditions. However, warming a cold asphalt in the

absence of any other effects will soften it to its original value and

h e ating a c e ment to an elevated t e mp e rature such as a mi x -plant

temperature will reverse the effects of organizational hardening.

It is important to note that hard e ning resulting from mol e cular

organization can be reversed periodically by recycling pavement.

However, the recycling process removes only the organizational

hardening, but since pavement suffers oxidative hardening also, this

portion is not removed during the recycled process.

Th e e v e nts that lead to aging in a pavement are very slow because

th e driving forc e s for orientation (dipole of each molecule) ar e sm a ll

and th e whole m e dium is quite viscous at road service temperature.

Nonetheless, slowly th e molecules will shuffle about and eventually find

the b e st orientation, known as the th e rmodynamic stable state or

eq u ilibrium, and in so ori e nting themselves become a better pack e d and

bound together system. Aggregates, no doubt, have a distinct effect on

this orientation but also differ in their ability to cause orientation.

The result is an increasingly stiffer (more rigid) material until

th erm odynamic equilibrium is achi e ved.

In a roadway, achievement of the rm odynamic e quilibri u m is a moving

targ e t ! There ar e constant changes in composition which r e sult from

oxidation of asphalt. During oxidation polarity changes, hence, the

"b e st packing" changes. Further, te mp e ratur e changes constantly in

22

7/30/2019 SHRP-91-510

33/41

pavement and th e rmodynamic equilibrium varies with temperatur e . Also,

traffic t e nds to disorient molecular species, especially under heavy

traffic loads and this further effects achievement of the rm odynamic

e quilibri um. It is not clear wheth e r traffic loads speed up orientation

by providing additional energy for molecules to move or if orientation

is slowed by k ee ping the syst e m "stirred."

Oxidative aging has on e interesting and saving feature to it. While

some asphalts, such as asphalt G, may oxidize and acco mmodate their own

oxidation products without major changes in viscosity, other asphalts do

show s ubstantial increases in viscosity upon oxidation. As shown in

Figure 4 e arli e r, virtually all asphalts eventually quench their own

oxidation and in so doing qu e nch th e ir own increase in viscosity as a

result of oxidation. Also, as was noted earlier in Figure 4, this

ph e nomenon varies with th e ma x imum t em peratur e of oxidat ion.

Compositional variations among asphalts dictate th e amount of viscosity

increase that is observed as a result of oxidation at different

t e mpe ratur e s. While th e chemistry of quenching is not fully understood,

it is a measurable chemical property of an asphalt and does relate to

th e visco e lastic prop e rti e s, and th e r e fore relates to th e p e rformanc e

charact e ristics of th e asphalt.

3.3 Speculati o n on the Relatio n s hip o f C om p os iti o n t o Vari o u s Failu r e

Mo de s

Th e me asurable chemical properties that ar e believed to relate to

th e me chanical or structural strength of a pavement ar e several. Again

from th e model, it is b e li e ved that asphalt is a set of "hardcore"

agglomerat e s (structured u nits) that consist of polars disp e rs e d in a

23

7/30/2019 SHRP-91-510

34/41

less polar to nonpolar phase. The r e sult is a mat e rial which has an

e lastic behavior as a r e sult of the network form e d by the polar

molecules, but it is also a material with a viscous behavior that can

flow or cr ee p becaus e various parts of the network can move with respect

to each other under prolonged stress. The relative contributions of th e

e lastic and viscous behavior vary with composition.

3.3.1 Ruttinq and Permanent Deformation. Again, consider asphalts

G, K, and M. Th e ir SEC pl o ts ar e shown in Figure 3. Asphalt G is

la r g e ly th e disp e rsing phas e , not much e lastic in character. On e would

ex p e ct it to b e v e ry compatibl e as is o b se rv e d in oxidation and

compatibility index studies. Asphalt G is somewhat ins e nsitiv e to

oxidation. It does not harden well which makes it a mat e rial that will

rut or deform, especially at higher temperatur e s. It has little of th e

compositional feature of large molecular size to give it elasticity.

Asphalt G is mostly th e non-associated material which do e s tend to

organize and harden at low temperature. Th e refore G would b e expected

to be v e ry stiff (susc e ptible to cracking) at very low temperature. In

fact, th e se ar e observed behaviors for G in roadways. Contrast this

with th e S EC plot for asphalt K. Asphalt K has a large amount of

agglo me rat e d mat e rial and would b e e xp e c te d t o hav e a greater elastic

modulus at high temperature than asphalt G and not have t he pr o pensity

to rut; and this is also what is obs e rved. Asphalt M, with its truly

high molecular weight, should and do e s behave somewhat like asphalt K at

high temperature. Asphalt M is also a more homogeneous material than K

or G. This feature should impart a relatively low temperature

susceptibi l ity to M, and this likewise is observed.

24

7/30/2019 SHRP-91-510

35/41

If heretofore th e relationship of composition to performanc e

properties has n ot b ee n obvious, consider some recent results from cross

blending of disp e rs e d and dispersing phas e s of asphalts. Th e s e

separated phases were generated by S EC. Major changes in properties

were observed in an asphalt prepared by blending fractions of different

asphalts. In one case, the dispersed phase of one asphalt was mixed

with th e dispersing phase of a second asphalt. The resulting asphalt

was more than i000 times th e viscosity of either original mat e rial ! The

dispersing phase of the s e cond asphalt is not a good "solvent" for the

dispersed phase of th e first. By appropriate cross bl e nding, one can,

within r e ason, achi e v e a wide variety of properti e s. Viscosities can be

pushed up or down at will. Tan deltas (a measure of th e relative

contribution of th e elastic and viscous moduli) can be varied

significantly. This area is not yet fully understood and is under

intensiv e study at the present time. Obviously, a clear understanding

of this phenomenon is needed for specification purposes, since many

asphalts are sold as mixtures from different crude oils.

3.3.2 Thermal and Fatigue Cracking. Cracking is another serious

failure in roadways, and again can be related to binder composition. If

th e mol e cular network (agglomerate, micell e , colloid, or what e v e r term

is pr e f e rr e d) becomes too rigid, th e ability of an asphalt to d e form

elastically will be lost. Instead, the asphalt fractur e s and likely

will b e s e parated to a point that healing cannot occur. Th e constant

working of very rigid matri x will ev e ntually suffer fatigue and crack.

Th e potential to crack is compounded by yet another organizational

feature. At lo w temperature, th e more neutral materials b e gin to

organize into a more structured form as can b e seen by diff e r e ntial

25

7/30/2019 SHRP-91-510

36/41

scanning calorimetry (DSC). Now, th e asphalt is quite brittle and

subject to cracking under str e ss. To mak e matters worse, some cements

shrink significantly at low temperature as a result of the organization

of low polarity and nonpolar components. This aggravates cracking.

Again, it varies with composition from on e asphalt to another. All

other things being equal, th e more linear aliphatic materials show the

most pronounced tendency to shrink with decreasing temperature. Hence,

cracking at low t e mp e ratur e would b e expected if shrinkage occurs, and

is e x pect e d to b e most clos e ly related to the compositional feature of

ali p hatic / aro matic ratio when all other characteristics are equal.

Aliphatic / aro matic ratios can be det e rmined by NMR, but it is probably

more practical to either measure shrinkage directly or predict it from

very rapid DSC measurements.

3.3.3 Adhesion and Moisture Damaqe. Adhesion and moisture damage

go together only to a point. While loss of adhesion certainly is a

s e rious moisture damag e problem, other forms of moisture damage may also

occur. At this point, consider only adhesion. By definition, it must

involve both asphalt and aggregate. While SHRP has co mmissioned studies

to investigate th e int e raction of asphalts and aggr e gat e s, the

e x am ination of adhesion, p e r se, is not within the scope of the binder

composition studies. Nev e rth e less, som e interesting observations can be

made considering only th e binder.

Adhesion of components in asphalt to aggregate appears to be

governed as much at the molecular level as at th e inter-molecular level.

Sp e cific functional groups (molecular types) seem to be very important.

Ce rtainly overall polarity, that is, separation of charge within the

organic molecules, promotes attraction of polar asphalt components to

26

7/30/2019 SHRP-91-510

37/41

the also polar surface of aggregate. Several workers within the

Strat e gic Highway Program have shown that aggregates vary quite

substantially, and in some cases aggregate behavior varies with

e nvironm e nt. Some aggregates hav e positive sites, some negative, and

some show variation in polarity with moisture content, temperature, etc.

For this discussion, aggregate will b e considered to be simply a highly

polar surfac e . Adhesion arises because of th e interaction of th e polars

in asphalt with th e polar surface of an aggregate. But polarity alone

in asphalt may not b e sufficient to achieve good adhesion in pavem e nt

b e caus e asphalt is affected by its environment. Asphalt has th e

capability of incorporating and transporting water. More on this in the

moisture d am ag e section. Absorption of water, like all other behavior,

vari e s with asphalt composition including changes in composition as a

result of oxidation. Incorporation of wat e r is m e asurabl e as are th e

e ffects of the invasion of water into th e asphalt aggr e gat e mixtures.

At the molecular level in asphalt it has been observed that basic

nitrogen compounds (pyridines) tend to adhere to aggregate surfaces

tenaciously. Carboxylic acid salts, while quite polar, tend to b e

r e mov e d from aggregate more easily, but this varies with the type of

salt. Monovalent cation salts, such as sodi u m or potassi um , of acids

tend to b e removed from aggregate quite easily. Calci u m or other

divalent salts of acids ar e much more resistant to the action of water.

From a practical viewpoint, it would b e hoove the user to assure th e

acids in asphalts are not in th e form of monovalent salts. The e xam pl e s

of pyridin e s and sodi u m salts are som e what th e extremes and it should b e

obvious that th e r e is a spectrum of tenacity of adhesion am ong th e

organic mol e cul e s in asphalt. Th e e valuation of this spectr um is being

don e els e wh e re within the progr am .

27

7/30/2019 SHRP-91-510

38/41

Moisture damage without apparent loss of adhesion is another co mmon

problem in pavement. Certainly, highway design is a major factor in

reducing the availability of water, but to b e clear, water is

omnipresent in roadways, so the e ffects of moisture are unavoidable.

For this discussion, consider that design i s adequate to remove most

water a nd also to resist its invasion in to pavement. Still, some water

will come in contact with asphalt in pavement and will affect its

p e rformanc e when the water soaks into th e concrete. Water, like

aggregate, is a highly polar material and to some extent is transported

into th e asphalt by virtue of attraction of polar water molecules to

polar asphalt components. Upon invasion into the asphalt, water will

effect the mechanical properties, typically softening it. From a

chemical viewpoint, the action of water is somewhat like the dilution of

asphalt with a low molecular weight solvent. This typically results in

r e duc e d strength and further results in rutting or other deformation.

Aged or oxidized asphalts, which have greater amounts of polars

(oxidation products), tend to incorporate water to a greater extent than

n e w asphalts. This would be expected from polarity considerations. The

probability of moisture invasion, from a ch e mical viewpoint, increases

with pavement ag e . But aged pavements are also harder than their

counterpart n e w pavements, so the effects of moisture and oxidation ar e

somewhat counter to each other. The co mbined effects are better

measur e d in t erm s of the mechanical properties of the mi x. As usual,

th e behavior varies with composition.

28

7/30/2019 SHRP-91-510

39/41

4. 0 SU _a 4 ARY

The chem is try o f pe tr o l eu m a s ph alt a t th e mol ec ular and

intermolecular levels has been discussed. It is clear that hundreds of

thousands of molecular species exist within any particular asphalt. The

polarity a mong asphalt molecules varies widely and the physical

properties are governed by the balance of polars and nonpolars

components. Polars tend to associate while less polar and nonpolars

cause dissociation. Several specul a tions on the effects of chemical

composition on pavement performance have been offered. In general, the

mechanical or structural properties of asphalt are related to the

intermolecular structuring among the polars. These interactions may

arise by involvement of any of numerous different chemical species. The

exact nature of the chemical specie is less important than the

distribution of charge within the specific molecule. Excessive

structuring leads to brittle cements which tend to crack, while too

little structuring leads to materials which deform under stress.

Oxidation adds to age hardening and brittleness of asphalt cement by

contributing to the structures zones within binder.

Finally, the word asphalt should be used in the sa me sense a s the

word glue. As much as there are significant differences a mong glues,

e.g., carpenters glue, airplane glue and rubber cement perform very

differently from each other, so are there tremendous differences among

asphalts. The differences among asphalts are as great as the diversity

of the crude oils used to manuf a cture them. The objective of developing

new specifications to describe and achieve consistent behavior, and

therefore consistent performance, is of high priority from the user's

29

7/30/2019 SHRP-91-510

40/41

cost standpoint. But th e user must also realize that the cost-effective

supply line (crud e oil) consists of a wide variety of mat e rials, and

there is little chance of significantly limiting the source of supply.

That is, th e s e l e ction of a very limited s e t of crude oils to

manufacture asphalts is im possibl e . It is therefore obvious that

achievement of consistent performanc e with asphalts frequently will

require modification of materials that are produced today. A principal

valu e of th e composition studies is to develop an understanding of what

compositional features ar e needed to produce ma terial with the desired

properti e s. Then this information can be used to select and / or modify

asphalts to obtain binders that will p e rform in a cost-effective manner.

3O

7/30/2019 SHRP-91-510

41/41

Aspha lt Adviso ry Comm ittee George WestShell Oil Co mpany

Thomas D. Moreland, chairmanMoreland Altobelli Associa tes, lnc . Liai sons

Gale C. Page, vice chai rman Avery D. AdcockFlorida Department of Transportation United States Air Force

Peter A. Bellin Ted FerragutNieder sachsi sches Landesa mt Federal H ighway Ad ministration

fiir Strass7enbauDonald G. Fohs

Dale Decker Federal Highway Administ ration

National As phalt Pa ving AssociationFredrick D. Hejl

Jose ph L. Goodrich Transpor ta tion Research BoardChevron Research Company

Aston McLaughlin

Eric Harm Federal Aviation Ad minis tration

Illinois De partment of Tra nsportationBill Weseman

Charles Hug hes Federal Highw ay Adminis tration

Virgin ia H ighway & Transpor tation Research Council

R obert G . J enkins Exp ert Task GroupUniversity of Cincinnati

Ernest Bas tian, Jr.Anthony J. Kriech Federal Highw ay Administr ati onHe ritage Group Co mpany

Wayne BruleRichard Langlois New York State Depar tment of Tra nsportationUn ivers ite La rval

Joseph L. G oodrichRichard C. Meininger Chevron Res earch Co mpanyNational Aggregates Association

Woody Hals te adNicholas Nahas ConsultantEXXON Chemical Co .

Gayle KingCharles F. Por ts Bitu minous Materials Co ., lnc .APAC, lnc .

Robert F. LaForce

Ron Re ese Colorado Department of TransportationCalifornia Department of Tra nsportation

Mark Plummer

Donald E. Shaw Marathon Oil CompanyGeorgia -Pac ific Corporation

Raymond PavlovichScott Shuler Exxon Che mical Co mpa nyThe Asp halt Institute

Ron Reese

Harold E. Smith Californ ia Depar tment of TransportationCi ty of Des Moines

Scott Shuler

Thomas J. Snyder Colorado Paving AssociationMarathon Oil Company

Richard H. Sullivan

Minnesota Department of Transportation

A. Hal eem Tahir

American Association of State Highw ay andTransportation Officials