ASPHALT MONOMER REACTIONS
ALLEN EDWARD LEI BOURNE III
A DISSRTATION PRFSFNTFD T1 THFF GRN.I1.ATF. COUNCIL Of
THL UNI PL IT I OF FL:.RIl[E
IN PARTIAL L'LFILLIENT i-.F TIIF Ri I(. PIIrEl- NIN h OR THE-
DEGIRFF OCF L(.CII..F OF PHILO'(tPH I
UNIVERSITY OF FLORiDA
Allen Edward Leybourne III
ACKI;OtLE DGME NT
The author wishes to express his sincere appreciation
to Dr. Herbert E. Schueyer who, through his interest, stim-
ulated and guided this research program, to Dr. tack Tyner
and the other members of the Supervisory Committee whose
continued efforts and guidance made this research possible.
The author is indebted to Mr. John U. ealway whose assist-
ance in construction and maintenance of the experimental
equipment was invaluable. The facilities provided for this
research by the University of Florida, especially the Chemi-
cal Engineering Department, have been used with the greatest
possible regard for their value. The financial aid of the
fellowship grant of the Ethyl Corporation and the assistance
and fellowship provided by National Science Foundation Grant
NSF-ul4249 are also acknowledged with gratitude.
TABLE CF CITrTS
ACIHA1W DGENIMT ....................... ii
LISE F TABLES ........................ vi
ITC d6F FIGU3.LB ........................ viii
i. Mi w . . . . . . . . . . 1
II. T . . . . . . . . . . ..
A. Stabilized and/or Trapped Free Radicals Are Present
in Asphalts . . . . . . . . . .
Introduction . . . . . . . . . .
Causes of Free Radical Stability . . . . 5
Tree Radical Types Expected in Asphalts . . 7
Natural and manufactured Materials Having Stable
Free Radicals ................. 8
B. Principles Underlying Free Radical Induced
Polymerization ... . . . . . . . 14
A Theoretical Equation for Thermal Polymerization 18
C. The Influence of Stabilized Free Radicals . . .. 22
A Theoretical Equation for Polymerization Influenced
by Stabilized Free Radicals . . . ... 27
III. IWBm AL . . . . . . . . .. ..... 32
A. *Aerlals Used . . . . . . . .... 32
:i|r-olump Refidues . . . . . . . . 32
romi lidueu............... 32
ii rl .i .ailtee .. . . . . . ..... 31
. . . . . . q . . . 3b
i ...t l gent . . . . . . . 37
WlaScriptin aieW Operation ....... 37
I ld Aq t Description . . . 37
MtiO 2 t Opvrational Procedure .. 38
C. e. hms........... .................. 41
ii tlen Ali of Mrimoer by Hydroeaastion . 41
SMiiil hds of 01efin Analysis . . . .. 48
- .ii .
- iv -
A. Method of Calculating Monomer Concentration . . 54
B. Overall Polymerization Rate Constant Calculations 57
Empirical Determination of the Polymerization
Rate Constant . .. . . . . 57
Correlation of the Temperature Effect . . . 59
An Absolute Basis for the Overall Polymerization
Rate Constant . .. . . . . . 61
Determination of Constants for the Theoretical
Rate Equation .. . . . . . . 64
The flatran Program . . . . . . . 65
C. Calculation of Relative Free Radical Coacentrations 66
D. Estimation of the Absolute Free Radical
Concentration . . . . . . . . .. 67
V. mESBUI S . . . . . . . . . . . 68
A. Introduction . . . . . . . . . . 68
B. Correlation of Polymerization Rates with Monoma
Concentration . . . . . . . . . 69
C. The Arrhenius Correlation . ... .. . . ..... 73
D. Asphaltene Constituents Exhibit a Dominant Effect 75
K. Influence of a Pure Stable Free Radical . . . 82
F. Spectroecopic Evidence of the Existence of Stable
free Radicals . . . . . . . . . 85
VI. DISCUSSIOB (1 RESUIM . .... . . . . . . . 95
A. Introduction . . . . . . . . . . 95
B. The Taeoretical Kinetic Equation .... .. . . 95
Analysis of the Theeretical Enetic Equation
Related to Minmer Concentration . . .9
Analysis of the TWeoratical Kinetic Eqawtion
Related to Stable Free RaM cal Coacnntw lti .
C. A Ceoeriso n of Theory with E ,rMlment; CorwmAlit|||
of .boli.mrizatia Rate# wi;1th:b&ma=m r Comentiralie 1 ||00
D. AggxragatL i aid Solvauti ofa the Abhalteaws . .
VII. CMC M i .Ab . . . . . . . . . . . .
I. cI CSI .. .......... .............. i
1II. SUliWHi .toR W. .C . . . * * * ..
viii. sm .. .............. .......... I11
X.. Isum--W-I I .w . . . . . . . . . .
XI. I MEE ~ C-1 . . . . . . . . . . U9
XII. APa i ICES ... ... .... .. ...... ... 126
A. Polymrization Run Ra Data . . . . . . 127
B. Calculated Values Determined for the Overall
Polymerization Rate Constant . . . . 143
C. Flatran Object Program for the I. B. 14. 650 Computer 145
D. temple Electron Spin BResnance Derivative Absorption
Curves of the Asphalt Residua . . . . .. 149
LIST CF TABLES
1. Free Radicals Stabilized in Pyrolytic Carbons . .. 10
2. Equations Derived from the Assumed Kinetic Mechanism for
Thermal Polymerization and Related Material Balances. 21
3. Equations Derived from the Assumed Kinetic Mechanism in
the Presence of Stabilized Free Radicals and Related
Material Balances . . . . . . . . .. 30
4. Properties of the Petroleum Residues Studied .... 33
5. Hydrogenation Blank Values for Several Asphalts ... . 41
6. Hydrogenation Blank Values for Two Asphaltenes . .. 44
7. A Comparison of Analyses for Styrene by Hydrogenation
with Other IMethods ... . . . . . . . 51
8. Densities of Styrene, S119 and Solutions of These
Materials ... . . . . . . . . . 63
9. A Comparison of Experimentally Determined Rate Constants
to Show Reproducibility . . . . . . . . 72
10. Free Radical Concentration of the Asphalts and
Asphaltenee Used in This Study . . . .... . 86
11. Graphically Determined Overall Polymerization Rate
Constant for Runs with the Gulf Coast Naphthenic
Residuum (8119), Styrene and Xylene . . . . . 87
12. Graphically Determined Overall Polymerization Rate
Constant for Runs with the East Central Texas Residuum
(8120), Styrene and Xylene . . . . . . . 88
13. Graphically Determined Overall Polymerization Rate IIII
Constant for Runs with 5000F. Air Blown S119 (R60-11),
Styrene and Xylene . . . . . . . . . 89
14. Graphicelly Determined Overall Polyamrization Rate Con-
stant for Runs with Various Petroleum Materials . . 9
-!!iii vii -
15. Graphically Determined Overall Polyerizatiom Rate Con-
stant for Runs with Asphaltenes Derived from S119 the
Gulf Coast Naphthenic Residuum (B61-6), styrene and
Xylene . . . . . . . . . . . . 91
16. Graphically Determined Overall Polymrization Rate Con-
stant for Runs with Amphaltenes Derived from 5120 the
Bet Central Texa ReSiduum (S61-9), Styrene and Xylene 92
17. Grlahioally Determined Overall Polymerization Rate Con-
stant for Runs with Asphaltenes Derived fron 5120 the
Iut Central Texas Residuum (S61-9), o(-vinylnaphthalene
and Xylwe ...................... 93
18. GOrphically Determned Overall Polymerization Rate Con-
atent for Runs with l,l-Diphenyl-2-Picrylhydrazyl
(e61-11), Styreme and Xylene . . . . . .. 94
19. OCmtants Determined for the Theoretical Polymerization
Itate Constant Equation . . . . . . . .. 104
20. S~Mltmri4aton Ran Data for RW with the Gulf Coast
4Wllthenic llsiduta (8119), ityrene and Xylene ... 128
21. eJl eritatien Run Bta for Runs with the East Central
Sir w EfMiluum (S120), Styrene and Xylene . . . 131
22. iAi ncMtion Run Data for Runs with 5000F. Air Blown
Sc19 (160-11), StyKer and Xylaee . . . . . 134
23. Iirrt aim. Rmm i ta. far Rus with Various Petroleum
.t .I S..tyrem ................ 136
24. i len IRtun ust. f9r Rume with Asphaltanes Derived
i q i i .te Gulr Gouft Ktthmanic ealAduum (S61-6),
Aiift l. 4 yl06t .. . . . . . . . . 137
29. =ar.=n ~ ft for Rums with Aljaklt~ es aerived
ike~tCs Oients iexa ftetAu (s61-9),
~1i ............... 139
26. B l I h Ri"%.a fr MIWn with ABgaltenes Derived
7t = :::.64 I Ag$ 0gibal T Rmon niag (M61-9),
27. 110:i E &lilit for *ituna Mih 1,1 .4jiwnyl-
.... "i........... .....i. :il jii.e ignl i y.. .. . . . . 1 6
1. Cross Sectional View of an Asphalteae Model . . . 12
2. The Influence of Solvent Ratio on the Yield of Petroleum
Asphaltenes . . . . . . . . . 35
3. Polymerization Equipment . . . . . . . . 39
I. Bomb Assembly . . . . . . . . . . 39
5. Selective Catalytic Micro-bydrogenation Assembly . . 45
6. A Scheatic Diagram of the Hydrogenation Apparatus .. 46
7. A Comparison of the Results of Analysis by Bruiaation
and Hydrogenation ................... 53
8. Vapor Pressure of a-xylene . . . . . . . 58
9. Calculation of Experimntlly Determined Overall
Polywriaztion Rate Constant . . . . .... 60
10. The Effect of Temperature on the Yete of Polymerization 70
11. The Effect of Temperature on the Wate of Polymerization 71
12. Arrhenius TePerature Correlation of Overall Polyeri-
zation Rate Constants . . . . . . . . 7:
13. summry of thbe OfrUa Polymrizatioua ate Constants
gMaized in -Lt Variom-Syeas- Stuadie . . ..
14. The Effect of Vari ion in As#hslt*ae Cmcina in e -t- 1
the Aelyierlatiln Rate Co t . . . . . .
15. Tte effect of Va st&L L in Asphalt Cnceatfst.A on tho. e
Solymeriamian bate feastant . . . .. . . 79
16. A Capwrieft of the lafl iene dt an Air Blown Ayihaltlite
tlSt of tA h M~aH fri r Which It iB Fre .ma- . 81
S ihlflh ii! Si!!!!! lilt -
17. The Effect of Variation in Asphalteme Cwemetration
on the Blyrization Rate Co ta . . . . . 83
18. The EfSect of Variation in the Concentration of A Pure
Free Radical on the Polymrization Rate Coantant . 8.
19. ualitative Apects of the Factors Affecting the
Ovresml Palyriraten Rate Coamtant . . . ... 98
20. A Caimarison r Theory with Experlasmt . . . .. 102
21. A CoWrleon of Theory with Experiment . . . ... 103
22. An Aephalteka Model Showing the Solvatlon effect
Occurring in Relatively Concentrated Asphalt System 109
Asphalt has been used in a variety of ways since ancient times.
Archaeologists discovered that asphalt had been used as an adhesive
for building stones and paving blocks as early as 3800 B C. These
early asphalts were obtained from scattered sources formed by
nature's own refining of petroleum deposits through geologic pro-
ceases (10) The 114-scre Trinidad lake and the 1000-acre Bermudez
deposit in Venezuela are impressive examples of these natural
processes. In the United States, a naturally occurring asphalt-like
material called gilsonite is found in the Uinta River basin of Utah.
Today virtually all asphalt used commercially is manufactured
from materials of petroleum origin. In order to obtain satisfactory
asphalts, early refiners of this century required the use of special
crude sources containing residues yielding high quality asphalts
through simple distillation procedures. This limitation has
Largely been removed in recent years through more sophisticated
procedures employing selective solvent extraction or precipitation,
etc. The patent literature contains many references to the use of
special chemical treatments in attempts to modify asphalt character-
istics.(l) Large quantities of asphalt are made by air-blowing at
temperatures in the neighborhood of 500 700 OF.(4S)
The inherent characteristics of asphalt, together with its
relatively low cost, make it ideally suited for such wide-spread
- 1 -
applications as road building, roofing, w'atcrproofing and a multi-
tude of others. The comple:-x world in which we live todsy is
constantly demanding new and improved mAterials to meet the require-
ments of modern applications. Asphalt is recei.'ing a large sn:rc-
of the research stimulated by these der'nds.
A considerable amount of effort has been e:.pended tu improve
asphalt properties by blending or reaction in "situ" vitn other
materials, particularly polyreric substances, both natural and
synthetic.(4, 16, 33, 2, 7,67, 68, '., 9,) Since 1930, interest h23
been shown in the effect of the addition of small amounts of natural
rubber. I-calorub and Pulvatex were the firs- special rubber Additives
developed principally for this purpose Todsy interest in ruDber
additives for paving asphalt 13 very nigh.(101) A number of vinyl
compounds have been reacted rith asphalts in attempts to impart
better qualities.(60) Recently developed epo.xv resin formulations
in asphalts, having improved solvent and heat resistance, have been
used quite successfully to pave jet aircrAft riuna:,s and for pro-
tective coatings.(2, 3, 58)
Even though a large amount of effort has gone into the study
of asphalt systems, relatively little is yet known because of the
complexity of its structure. There has been recognition of distinct
constituent groups present such as asphaltous acids and arn-iydrides,
asphaltenes, asphaltic resins, petroleum resins, petroleum oils,
carbenes, carboids and inorganic material.(69) WMny attempts have
been made by previous investigators to characterize these groups as
completely as possible with the u2thods available This research has
been under-t.kerl to establish a better undcrstandirtg of the fundamental
aspect. affecting the runner in which v'inyl monomers will behave in
asphalt reaction s',stems. This study has been approacrned throuEg- an
anil.;is of factor affecting the kinetics or vinyl polymerization in
aspnaltic media. It is by correlation of the knowledge gained by
plreioLs investigators with the results obtained in this study, that
the object of tnis investigation has been, in part, achieved.
....... ,. ..........
A. Stabilized and/or Trapped Free Radicals
Are Present In ..sphaltc
With the recent advent of techniques such as electron para-
ugr.etic resonance and nuclear magnetic resonance spectroscopy, a
considei-ble amount of effort his beenr focused on the study of
stabilized free radicals; an area that heretofore has suffered
from neglect. This neglect has stemmed mainly from the inability
to make suitable measurements for proper analysis Steacie in his
review of free radical mechanrusms uhich appeared prior to the
ability to apply these new techniques; contributed much to our under-
standing of active free radical processes, but did not present
information of the reactive character of stable free radicals, or
even of their widespread e:dstence.(90) Since the preparation of
the stable free radicals, pentaphenylcyclopentadien'l and 1,1-diplenyl-
2-picrythydrazyl, the ability of free radicals to remain dissociated
has been recognized.(52, 102)
Electron paramagnetic resonance absorption probably offers the
most direct means of evaluating free radical concentrations and the
effects of structural configurations. A striking feature of the
paramagnetic resonance spectra of stabilized free radicals is the
sharpness of the absorption exhibited.(28, 59) It is sufficient to
. .... ..... ..... ..... ..... ..... ..... ...... ..... ..... ... .. ..... .....
say at this point that there appears to be little doubt as to the
validity of the conclusions drawn, at least qualitatively from these
measurements. (28, 47, 51, 52, 59)
Since the early discovery of stablilzed free radicals, the
List of compounds wrich exist as free radicals, per se, which form
free radicals through dissociation in solution (in some cases strong
electrolytes), or which may yield stable free radicals through mild
oxidation with chemical or electrical means, has continued to grow.
Also, the evidence for the existence of trapped and/or stabilized
free radicals in many natural materials has continued to increase.
At cryogenic temperatures, experiments nave been devised to trap many
of the more reactive free radicals .(52)
Causes of Free Padical Stability
It is obvious that a more thorough understanding of the cause
of free radical stability is needed. Apparently, free radicals can
be trapped through what might appear to be mechanical means. Poly-
merizing systems in which the polymer precipitates have been shown to
occlude or otherwise trap free radicals.(53) For example, Bamford has
determined that a polymer of acrylonitrile prepared at 25 "C by
photopolymerization was found to have about 8 x 10-5 gn. mole radical
per liter. This represented about 13 percent of all radicals initiated
photochemically during the experiment.(13) In some materials such as
dicarboxylic acids, for example, free radicals appear to be trapped
by maintaining hydrogen bonds with the lattice molecules after
irradiation with x-rays or gamma-rays have produced them.(5)
mi ....... .ii .ii .i
Most of the free radicals which are obtained in pure forn or in
solution appear to be stabilized through what is normally termed
It is convenient at this point to introduce the terms singlet,
triplet, and bi-radical to describe concepts which will be discussed
in slightly different context later in this work An olefin or con-
jugated double bond system is capable of existence in a state such
that all unsatur-tion electron pairs have the orbital spins paired
This is termed the singlet state, with only one possible conmination
of angular momentum vectors, because the electron spin vectors cancel.
A bi-radical is visualized as the situation arising when a pair of
unsaturation electrons' orbital spins become unpaired, and the
electrons become distributed so that they are predominately at
opposite ends of the previous olefin system. If these electrons ar-
not physically, widely separated, then the orbital spins which are
no longer paired can interact, giving rise to three possible com-
binations of the angular momentum vectors, with the so-called triplet
state resulting. Hence, the difference between a bi-radical and
triplet is one of degree. Rex.road and Gordy have illustrated these
phenomena with the compound, p,p-biphenylene-bis-diphenymnethyl. (73)
C = C C-
S Not:l diianagnetic state Bi-radical state
(singlet state) (triplet state)
- 7 -
The singlet, bi-radical and triplet states, are but a few of
the possible electronic configurations that can exist for molecules
that have complex systems of conjugated double Fonds, such as
condensed ring aromatic: hydrocarbon s:,stems. Indeed, r.ost of the
stable fr-e radicals which have teen cited in the recent literature
have as a part of their structure multiple phenyl groups. Tne
possibilities for resonance i: tnese types of compounds is evident.
Free Radlcal Types Expected In Asphalts
The evidence for the existence of hijgly .'oraplicated aroimtic
structures; and in addition, structures contcinin& nitrogen, oxygen
and sulfur, is abunaant for asphalts This seems to be particularly
true for the nigher molecular weigni fractions such as tne asphaltcnes.
(.3L, "6, 95, 105)
Perinapntnslene, whose structure is sho'W, belov yields free
r-dicals in a very interesting mannrer.(89)
..H H A H
H1 rr I -CCl r\ + H
In carbon tetrachloride solutions, the central hydrogen atom is
removed leaving an unsatisfied valence at this carbon atom.
These conclusions uere reached by examination of the fine line
splitting of the electron paramagnetic resonance spectra. This
solution also becomes highly colored upon standing. This type of
compound is almost certain to exist in asphalt systems,
- 8 -
particularly the high molecular we'iht. portion.(29, 31, 105)
Aromxtic free radicals can be produced in solution by the dis-
sociation of structures which are not tightly bound. An e:amprlc of
this is the dibenzoyi tetrazanes, whose generic structure is as
follows: R R
1,1,4, h-te t re.-p- ~'kyI phenryl-2-
The dissociation occurs between the central two nitrogen atoms,
yielding a compound of the hydrazyl type.(87) It maj be possible
that analogous structures, capable of dissociating into fr.-e radicals,
exist in asphalts. This has not been demonstrated to tre knowledge
of the author to date. Although no direct reference has been made
to condensed ring aromatic compounds, otler than perinaphtbalene, a
large number of condensed ring aromatic systems have been demon-
strated to yield stable free radicals.(52)
Natural and Manufactured Materials Hav-ing Stable Free Radicals
The spectra of a large group of pyrolytic carbons, as deter-
mined by electron spin resonance (ESR), have been examined by
numerous investigators. With this group of pyrolytic carbons, the
high molecular weight constituents of petroleum, its residues, and
other similar naturally occurring bitumens, including those derived
from coal, shall also be discussed.
These materials generally e-nibit sharp ESR absorption bands.
Estimates of tne free radical content of materials representative
of this group of pyrolytic carbons has been collected in Table 1.
From the data collected in Table 1, it may be seen that
stabilized trapped free radicals are, indeed, characteristic of
pyrolytic carbons produced either naturally or artificially. It is
clear from these data that the content of free radicals is a
function of the thermal history and total carbon content. Pyrolytic
carbons formed slightly below 6C0 C. give the stronger ESR
absorption. X-ray examination has shown tnat it is at the point of
pronounced free radical content thst the formation of condensed ring
clusters bcgin; both with regard to the variation in formation
temperature and carbon content.(12) Investigations by Austen, Ingram
and Tapley have led them to suggest that:
"The essential mechanism in the trapping and stabilization
of unpaired electrons [in the complex carbons being dis-
cussed7 is the existence of ring clusters containing more
than a certain number of carbon atoms. Such ring clusters
will possess a high degree of resonance energy available
for stabilization of the electrons. . The decrease in
radical concentration above the 90 percent carbon content
or 550 C. carbonizing temperature is explained by the
gradual formation of graphitic sheets, the individual
carbon clusters joining and thus saturating their broken
In the above hypothesis, it was assumed that the radicals have been
formed by the breakage of bonds around the edge of carbon clusters.
It was also pointed out that some radicals may occur in packing de-
fects producing internal trivalent carbon atoms. The ideas proposed
by Austen, et al., have been generally agreed upon by other working
in the field.
. i....... ... ..
i= .....- 10 -
TABLE 1, FREE RADICALS STABILIZED II FYROLYTIC CARBONS
A Pittsburgh Coal
Coal Hydrogenation Asphaltene
SIn CSg Solution)
In Dioxane Solution)
Dextrose Clar (in vacuo)
wotleum. Reidue Char
14 5 C 0)
(12, 17, 71, 96)
1 8 10.9 (44)
20-30 9-10 30)
0.5-1.0 7.6-8.2 30)
ail 6.0-6.2 (30)
q.' iUW i si, IAeBllly that any charred
3l4til'^iiw 6aOe., give5 s streg ESSR
f'MMBr kAin aMiolar to Oat of
:MiE, .S t : e 'I!1t'm P g
- 1 -
An interpretation of the bond breakage at the edge of the con-
densed carbon rings by Ingr=m, is that the electron released by the
bond breakage is stabilized by the large amount of resonance
associated with aromatic rings, and as a result, nove in highly de-
localized r -orbitals over the system. The presence of such de-
localized radicals has been suggested to be the cause of carbon blacks
actively participating in the reinforcement of rubber.(52)
Petroleum products have been placed in the category with pyro-
lytic carbons because of the analogy that can be drawn between their
probable origins and similar constitutions.(35) In an inter-sting
paper by Yen, Erdman and Pollack, it was determined that the charac-
teristic X-ray diffraction of the high molecular petroleum fraction
(asphaltenes) from nany. crudes could be reproduced by blending
samples of polyethylene and carbon blacks. Also the X-ray diffraction
pattern of the aromatic portion of these asphaltenes compares with
that of a blend of condensed aromatic compounds of known structure.
Figure 1 is a model constructed from the data they obtained.(105)
The analogy to pyrolytic carbons from other sources seems very
The paramagnetic resonance of crude oils is apparently con-
centrated in the high molecular weight colloidal portion. This
portion has been labled "asphaltenes" from a comparison of these
materials with the so-called asphaltenes obtained freomasphalts by
precipitation with light paraffin hydrocarbon solvents. Outowaky,
ot al., observed that the ESR absorption of a crude oil ik pear
L U VCi 43
*L h Q'0 U
43 -i .- 0I
after extensive centrifigation at 80,000 G's.(25, 44) Some concern
has b.-cn expressed as to the possibility of the pwarsmiTnetic resonance
of petroleum stemzu;inU firon other than frne radical sources. Vanazdium
in the +4 valence state as vanr dyl readily forC s chelate compounds
which exist frequently in nature as porpnyrins.(T0) T.is -iny be
true for other metals as well, and therefore, the possibility of LSR
absorption from this source should not be overlooked. Eldib, Dunninc
and Bolen, in a study of the colloidal rr.terials in petroleum,
obtained by ultracentrifugation, nave determined that vanadiur. and
rickel are present in this colloidal material in concentrations less
than 25 and 8 p.p.m. respectively. (34) These uere the only ettals
detcnmined to be present in appreciable concentration. O'Reilly has
compared the ESR spectrum of a solid asphaltene %ritl that of vanadyl
etioporphyrin I (VEPI) dissolved in a high viscosity petroleum oil.
The spectra of both samples showed the same general characteristics.
The ESR absorption demonstrated by the porphyriuns present in the
solid asphaltenes however, is quite small compared to the band
attributable to free radicale.(70) Only about 0.1 of the spin con-
centration observed, as determined by elemental analysis for nickel
and vanadium, was attributable to these materials in asphaltenes
investigated by Gutovskys, et al.(44)
In addition to the evidence cited for the presence of free
radicals in asphalt, that has been essentially based upon measure-
mcnts made by ESR; Poindexter has observed the Overhauser Effect
in asphalt solutions. The details will not be discussed here,
- 14 -
However, it can be said that the contribution to proton relaxation
from the asphalt is a linear function of asphalt concentration.
This contribution to proton relaxation arises through the interaction
of the electron spins of the free radicals with the nuclear spin of
the hydrogen atoms present.(Y5) It mws stated tnat there was no in-
dication of any unusual quality of the asphalt radicals observed.
Saraceno and Coggeshall conclude from their measurements, that
the fire radical species in petroleum oils are stable, and are not
created through dissociation of diamagnetic molecules upon dilution
with solvents.(83) Tnis is consistent with the linear contribution
to proton relaxation with change in concentration as mentioned above.
This however, is in contrast to the results reported for coal hydro-
genation asphaltenes; see Table 1. Ho effects on the free radical
concentration were observed to be a function of either sunlight,
ultraviolet light or bubbling molecular oxgen into a petroleum oil.
B. Principles Underlying Free Radical
That there are polymerization processes which are induced by
the presence of free radicals does not seem to be in dispute.(14, 20,
26, 39, 77, 99) The strongest evidence for this appears to be that
acceleration of polymerization rates of man;, vinyl compounds occurs
in the presence of substances which are known to give free radicals
upon thermal decomposition, after exposure to ultraviolet light, or
bombardment from radioactive sources. In some cases, polymerization
has been known to occur in the absence of light or free radical
forming materials. Tnis naz become known a- tnermal polymeriza-
tion, but can still be explained on the basis of a mechanism
rich proceeds through the action of free r-ailcals, once polymeri-
zation is initiated. These free rid-cals have generally been
considered to be quite reactive, in that their life time is rela-
Attempts to correlate the kinetics of some of the more simple
systems have been reasorLnbly successful A very good review of
this subject has been given by Bamford and Barb (14) It appears
tiat the cuse of homogeneous solution polymeri-ation, is the
simplest to treat kinetically, particularly at relatively low
concentrations of nonomer. Several simplifying assumptions have
usually been introduced by various authors, in the derivation of
correlation equations. The main justification of these assumptions,
of course, lies in the fact that they yield correlations which are
The overall polyueri-ation may be described in terms of
three phases which are distinct conceptually.(14, 37) These
phases are initiation, propagation (in some treatments which
chain transfer) and termination. During the initiation phase,
several methods of opening an initial double bond have been dis-
cussed. During thermal polymerization, the assumption has often
been made that a bi-molecular reaction between monomer molecules
yields active free radicals. When an active free radical is
formed through the action of radiation, bombardment by high
..... iiiiiiiiii i
- 16 -
velocity particles or ther al decomposition of an unstable compound,
initiation is usually assumed to occur through a bi-molecular reac-
tion with the radical forced, and a monomer atom. During the propa-
gation phase, the active free Irdizals previously formed ad.d to the
double bond of the styrene molecules in a chain fashion, with the
rate of each new addition being governed by a bi-molecular reaction
between the previously formed active free radical, and tne new mono-
mer molecule. Since each of the steps in this chain reaction
occurs between molecular species which are very much alike chemically,
even though different from one another in the abLolute sense, the
rate constant for each of the propagation steps is assumed to be
identical. During the final or termination phase, the assumption
is made that termination occurs through d icproportionation and combi-
nation between any tro of the existing active free radical species.
A fairly subtle effect which merits our attention is that of
chain transfer with the solvent, polymer, and monomer molecules.
This occurs when an active polymer radical abstracts a hydrogen atom
or other atom from one of these molecules to satisfy its unsaturated
valence, leaving the molecule which was attacked a free radical.
If this newly formed free radical is sufficiently reactive, the
effect on the overall rate of polymerization will be slight, since
the propagation of the oonomer reaction by this radical continues
more or less uninterruptedly. Breitenbach and aschin, also Mayo,
have effectively shown this to occur when using carbon tetrachloride,
a solvent ideally suited for chain transfer. By assay for chlorine
in the resultirG polyimr, it vas found that approximately four chlorine
atoms per poly:rmr molecule were incorporated, and that the effect on
the overall polylernZ ation rate was onl;, slight. The najor effect
noted for this process is a substantial decrease in the polymr molec-
ular oeight.(22, 62)
Some solvents may have a retarding effect on the polyraorizntion
rate through the action of chain transfer. This has been e::plainod
by Price and Durham on the basis of formation of stabilized solvent
free radicals with, insufficient reactivity to propagate continued
polym.er grrrth.(76, 93) Of the solvents studied, nitroben2cne,
pht nols and 2,i-dinitrochlorobe;izene ha'e been shown to have a
pronounced retarding effect. In some instances as hiChl an one dini-
trochlorophenyl group per polymer cnain uas found dhen polymerization
was conducted in this solvent. In each of these cases the possibil-
ities for resonance stabilization of the free radicals is apparent.
A concept which is useful in the development of kinetic rela-
tions is the so-called "steadjy state".(l4, 41, 90) The steady
state condition can be visualized as a condition in vhich the rate
of formation and removal of certain reactive species arc approximately
the same, so that a seeming equilibrium situation is manifest. Even
thour. this is only an approximation to what is actually occurring
*very useful relations have been developed uben the magnitude of the
error introduced is small. In polymer kinetics, the assumption of
'steady state" Generally implies that the concentrations of the
free radicals, intermediate to the polymerization process, attain
,- 18 -
equilibrium values. From this it follows that tihi rate of change
of a given free radical specie, with time, is nearly zero.
A Theoretical Equation for Thermal Polymerization
An example use of the principles cited is the treatment of the
thermal polymerization of styrene in solution.(lL) Experimentally,
it has been shown that styrene (when quite pure) has a solution
polyn rization rate which is second order in terms of monomer
concentration, if the reaction is not carried on to too great an
extent.(6, 14, 20, 27, L3, 45, 97) This, houe.er is not the case
when polymerization proceeds in bulk, i.e., pure styrene alone.
When the effect of change in the activity is accounted for, the
polymerization rate is second order in terms of monomer activity.
Walling, Briggs and Mayo measured the vapor pressures above various
solutions of styrene and polystyrene to obtain the data required for
Bamford and Barb, in their review of polymerization processes,
have presented the following mechanism for the thermal polymeriza-
tion of styrene.
The Assumed Kinetic mechanisms m for Thermal Pol'yerization
Equations Governing Rate Rate Constant
The Initiation Phase
(1) M + ----- 2Rc* kc
(2) + M ------ R1* kpc
The Propagation Phase
(3) R* + M ---- R* k
- 19 -
(6) Rn-l*+ ---
The Termination Phase
With Active Free Radicals
(7) R + Ra*----
W\he r .
R* Free radical
k Rate constants identified by
the proper subscripts
1,2,3, ..,n,n-l denotes the number
of monomer units affixed
c Denotes a radical formed in the
s Denotes a stabilized free radical
a Denotes all active free radicals
If the polymer chain length is relatively great, the overall
rate will be governed by the rate of monomer reacting in the propo-
gation steps to a fair approximation. From the preceding statement
of the assumed mechanism it can be shown that the overall rate of
polymerization may be expressed analytically as:
-d(M)/dT =k P(R *)A + k kpi(Hi*)M
Utilizing the assumption that the kp's are identical.,Lnd notf.i
- 320 -
i = n
R-* = Re* + Ri* by definition, 2
i = 1
-l(!l)/dT = kp(Ra*)1 3.
The rate of change of concentration of any free radical species
may be stated in general as:
Fate of Formation Rate Propagation Rate Termination Rate
Change of Species R1* to Species Pil* of Species FRi
For purposes of clarity, it is shown in Table 2 through the use
of material balances derived from the preceding cmchanism, how these
relations may be solved to obtain an equation explicit in R *. Com-
bining equation 7 from Table 2 and equation 3 yields the overall rate
in terms of H.
-d()/dT =k.p (2k/ / ta- ..
The correlation thus obtained predicts that the thermal polymer-
ization of styrene should be second order in terms of monomer concen-
tration. To be completely rigorous, this entire analysis should have
been made on the basis of thermodynamic activity, as indicated previous-
ly, when it was stated that a correction for this must be applied in
bulk polymerization. However, in dilute solution, say 10 percent
monomer or less, the concentration expressed in units proportional to
the number of molecules per unit volume should adequately approximate
the activity. If the density does not change appreciably as polynmer-
ization proceeds in this dilute solution, a further approximation -an
be made; i.e., the concentration expressed in weight percent will very
nearly be proportional to the concentration expressed as above.
- 71 -
r-1 (C4 r'1 -4
.r.. Da n
I C cl]
* .4 - .
'4 e. 4-
r re CA
*^ I ^
C .. II
i 0. -
cfl ?- >f)
j^. 4 i --
I C 1q-4
i- 0 4(41
I iL .T)
4-. I I--
r-4 T f *l
I TO(UC -
C ,i .4.
11 .c 01 PN.'
TC t441C ^
r 1T-. vi 11
^' (-, 1 1
4o < ( '
1 I .4
+ J (t'
6 I- L
.4b 14 Ii
S +. *1
'-4 i'4 I'j
- E i-
I' 1 ;L
; 1 1
.C- C Cl 1-.
l'-4 '14 <
' L. 4. C-
44 C C
.4-I 0 -
r. 43 .-
0) C-. LU
G. -o I
f~ ri 1
8 lo .ah
(1 -r C
Itl . I^
.C .c -
44 .4- (439
C, r- *
64 (1 -. ; in
*ri Gd 0
Integration of the rate equation allows it to be shown that a plot
of reciprocal monomer concentration versus time should yield a
straight line whose slope is the overall rate constant. The details
of this are shown on page 59.
C. The Influence of Stabilized Free Padicals
Previous discussion has pointed out that stabilized free radical
and/or trapped free radicals are demonstrably present irn U. types
of materials being studied. Several investigators have noted catalytic
and inhibiting effects in natural materials whicr are analogous to
these materials under consideration.
Very recently Wright has made a careful study of the inhibiting
effects Of a number of pure synthetic stable free radicals and several
natural materials containing free radicals (104) During the polymeri-
zation of styrene catalyzed by bcnzoyl peroxide, it was found that tlhe
stable free radicals, l,l-diphenyl-2-picryl.hydrazyl and Banfield's
free radical (which is prepared from Banfield's hydroxide (18) ),
exhibit an inhibiting effect which increased with increasing concen-
tration. The same effect was observed for the asphaltene fraction
of gilsonite and a Wilmington crude oil.(104)
Heexaphenylethane, when in solution, has been shown to dissociate
into the triphenylmethyl radical.(214, 102) Mayo and Gregg, in a study
of styrene polymerization, used hexapbenylethane over a considerable
concentration range.(61) From material balance calculations based
on the polyipr molecular weight and the extent of reaction, they
estlaBte that about 77 molecules of hexaphenylethano disappear for
each molecule of polystyrene that would have been formed in its ab-
sence. This work lead to the following conclusion:
"Any free radical may start or terminate the polynerization
of a styrene chain. Whether a source of free radicals will
behave as a catalyst or an inhibitor depends on the balance
between the rate of addition of these radicals to monomers,
the rate of interaction of radicals and the rate of growth
of the polyner radicals at the chosen temperature. If the
radicals do not add rapidly or if they are supplied too fast,
then a high radical concentration results and chain growth
is restricted. If the radicals add very rapidly, or are
supplied slowl; enough, polymerization will result. These
statements mean simply that the dividing line between cata-
lysts and inhibitors is not clear-cut; the differences between
them are quantitative rather than qualitative."
Symons has studied the interaction between acrylonitrile and
amorphous carbons and has found that o:.ygen-free carbons initiate
polymerization in the dark.(91) With no carbon present, thermal poly-
merization did not occur at the temperature used, 600C. Activation of
the carbons at 000C.-5500C. over moist air causes the initiation
effect to be destroyed. This was explained on the basis of the forma-
tion of surface quinone groups which act as inhibitors. It was suggest-
ed by this author that two types of free radicals, i.e., trapped free
radicals and resonance stabilized free radicals may be the active
initiator. Also, the possibility of mechanically isolating some of
these free radicals from the polymerization process may occur.
Breitenbach and Preuesler aVarently confirm the work of Syms re-
garding the inhibiting effect of activated carbons by their ,tudy
in styrena polymrization. (3) Boiever, the inhibiting effect on
styrene which does have a definite theraml polneyrizatioa rate
lasted for only a limited time after which polymeriatim ocacumed at
at least the naeaal rate. Thi is : corAistent with the .eiiomt -i
- 24 -
inhibition by surface quinone groups since it is known that the effect
of inhibition by quinones is gradually lessened through free radical
reactions with quinones to give non-reactive by-products.(14, 61)
Kraus, Gruver and Rollmann studied the effect of carbon blacks
on the free radical thermal polymeri2tion of styrene at 50 oC.(55)
These carbon blacks were treated in the following ways: a) reduction
over platinum with hydrogen or by sodium borohydride, b) heat treatment
at 1000 2000 C., c) and outgassing at 300 oC. under vacuum. Each
of these carbons causes a substantial induction period to occur.
(ote: The pure styrene used gave no induction period.) After the
induction period, acceleration of up to five times the normal polymer-
ization rate was obtained. Hydrogenation, presumably of the surface
quinone groups present in carbon blacks, drastically reduced the
l r th of the induction periods observed, but caused no measurable
change in the acceleration effect. High temperature beat treatment
destroyed both the induction period and acceleration effect. This is
in accord with the known phenaemean that high temperature heat treat-
ent causes graphitization of carbon blacks by allovLng structural
Iavri Vgemnts that destroy stabilized free radicals.(12, 17)
Outgassing of adsorbed oxygen caused no observable changes. Free
a Sia. coaeentration in these carbon blacks was measured before and
after polyflrtzation. Little, if any, change in concentration
(tch.o:,ii,.of the order of 10*18 unpaired spins per gram) occurred.
.."his apparently negativee observation prohibits it being stated
ae lusitBvy that free radicals are the cause of the acceleration
. a m.............
- 25 -
observed. The possibility of this effect being produced by surface
peroxide formation, however, must probably be discarded, since the
thermal history with treatments up to 300 OC. had no effect.
Consideration of the ability of charcoal and other pyrolytic
carbons, as well as paraagnetic substances such as free radicals,
to catalyze the conversion of para-hydrogen to orth--hydregen pofmts
out the ability of these stabilized free radicals to interact with
other systems.(82, 90)
Hulbert, et al. have considered a theoretical mechanism of
vinyl polymerization from the view point of the electronic states
of olefins, particularly ethylene. Their calculations indicate that
ten of the twelve valence electrons in ethylene are localized by
pairs taking up positions between the carbon and hydrogen nuclei.
The remaining two electrons which are essentially distributed over
the entire molecule, have been given the name "unsaturation electrons".
(50) In the lowest energy state, these two electrons are in the same
bonding orbital "N" and by the Pauli exclusion principle must then
have antiparallel spine. Since these electron spin vectors cascel,
there is only one possible combination of angular mwe ntum vectors,
and the so-called singlet state is the result. It is pointed out
that promotiar of an electron frex the bonding orbital to f aon-boitt
orbital should be rare in the absence of an external anpsiwic tibia,
because of the necessary unaouplpg of the W'gi via eachen n a
Wa.m tic energy, :aen though ufltleient easWy-yw be sBOnlable in
the molecule itself. Whainver an "oad electron na6e1cal', i.e., a
free radical, cmr within the kinetic theeay rqdiuB, ti.iiunagjiU 1
.... ...... O
- 26 -
becemes more probable. The result of this uncoupling is the triplet
state in which the electron spins are parallel. When parallel the
spins are uipaired, in other words, what is commonly termed a
diradical has resulted. This diradical may or may not have bonded
with the free radical which aided in the uncoupling of the electron
spine. Active diradicals once formed have been reported to be
capable of propagating a chain reaction from each free radical site
in the molecule.(li, 20, 26) It is not known whether cr not the
stabilized free radical will form a bond during the uncoupling
process; however, even if the bond is formed monmntarily, subsequent
dissociation of this bond might occur. This possible dissociation
of, or failure to form a bond could possibly account for the failure
by Kraus, et al., as previously mentioned, to observe a change in
thi total concentration of unpaired spins before and after polymer-
The total experience gained by the various investigators cited
tintd to igbicate the plausibility of an acceleration effect caused
by the interaction of stabilized free radicals with vinyl monomers.
Tie ability of stabilized free radicals to remain dissociated over
lA:. perteds of time does not neassarily imply that they are
ia~ambtly unreactive. For exaepl, dipbenylpicrylhydrazyl is known
to retst with u:6e reactive free radicals even though alone it
.ieP disaoeiated for long periasl of time. From the ability of
styrene to urMrgo thermal polyjerilation, it is clear that formation
o0' Sat. free radicals iS paeetble even in the absence of catalytic
substances. The assumption of the initiation of styrene polymeriza-
tion through a bi-molecular collision of styrene moleucles has
successfully correlated the kinetics of the thermal polymerization
of styrene, it may be recalled.
fealizing that even a slight increase in the probability of
forming the diradical state of styrene would increase the overall
polymerization rate, and that a stabilized free radical can poten-
tially increase this probability, leads one to the following specu-
lation: may not the presence of a stabilize& free radical, near a
monomer molecule, at the time of interaction with a thlrd wolecule
(which possesses a definite probability for reacting), increase this
probability through the mechanism described by Hfulbe-,t', An
alternative question -Is-, may there perha-ps be irsufficienIt energy
normally available for the stabilized free radicals prouent in
aphalts to react alone with L monomer or active frae z-Ldical
A TonAclEL~to o oyrzto rfunc ySaiie
A nechaniam -wbich embodies thbe anE;umptiorw and principles
utilizeid previously for the derivation of a rate equatid% correla-
ting the tlienws polymorizatien, bas been develOP'ed for poly1**riZa-
tiom influencL-d by the pmesence. of stabilized free rwdicala In
addition to the ezssumptions opoa Previously, it hac a140 bien
Assumed that (1) during, the intiation rbatt, the ftorm-ti96 of mctlve
frEe radioils vey occur thrmoig a. tor-oleculax reftttiom involIvia
the wtabt31Zd free radcual WVd two WWWma moleCU34a aot (2) during
- 28 -
the termination phase, reaction ceases through a ter-molecular
reaction involving an active free radical, a stabilized free radical
and a monomer molecule. The assumed mechanism follows.
Tne Assumed Kinetic Mechanism for Polymerization Influenced by Stable
Equations G verning Rate Rate Constant
The Initiation Phase
(1) + M ---- 2R* k
(2) Rs* + + M ---- M + Re ks()
(3) Re* + -- R" kpc
The Propagation Phase
(4) R1* + M ---- R2* kpl
(5) R2* + 1 --- R- R: p2
(6) R3* + --- R* h3
(7) n-1* + M ---- kp(n-1)
The Termination Phase
With active free radicals
(8) Rn* + R ---- Dead Polymer kta
With stabilized free radicals
(9) R* + Rs* + M ---- Dead Polymer kts
(a) Me cAnfaet~A tian dependence of the rate in this step is assumed to
be sia-linMr d: Rg*. Therefore, the forward rate equation for this
step:. i arbit arily written as. 2
(Rc*/farbard = k, *) (M)
TI eaFSaBrt D ji..,been introduced to represent a simple noa-linear
- 29 -
For purposes of clarity, it is shown in Table 3 through the use of
material balances derived from the preceding mechanism how these rela-
tions may be solved to obtain an equation explicit in Ra*. Coa ining
equation 7, Table 3 with equation 3 previously shown in the discussion
of thermal polymerization for the overall rate in terms of Ra* and M
-d(M)/dT = Ipkts/2ktas 5.
S 2 4k ta D .] 2
(Re*) + (2kc + ks (RA) )) R M
Setting the term in front of H2 equal to K yields an overall rate
constant that is analogous to that obtained for the theril polymeriza-
tion process, in that it is a correlation constant for an equation
second order in mononmer concentration.
The overall rate constant, K, is therefore, actually a function
of the part jeter Rs*. This of course means that K will not be a
constant unless the concentration of RH* remains essentially constant
during a given polymerization run. This will be approxdnately true
for polymerization that represents a small quantity of S leer poly-
merized As before, the rate equation may then be integrated to show
that a plot of reciprocal concentration versus time yields a strain t-
line curve whose slope is the rate constant K. At a 4HPlBitftfi
of HR* 0, the rate comaiet, K, obvtissiy MWiuiU llb* l *fl k
thereal polymerization rate ceauSit.
It is believed that the preseeiC Ailscemain cM .
reassoabtle justificatima fbr the 1a1Aiu H that hg .wiiBWi||[ J.gg At
-i 30 -
II 3 I II I
C' CD ul D'
* P-1 +W.
4/-. *I S 4 --
O a 4 11
N N- P'
*"'""""""" E-. 0- F
m *-H co
o C. w
S.-, IA C *, -
L- L 0-- ,
-4J C .*
Ir c r- c
C -I -l C
* .1 0.-, *
04 Cn C 4)3 3
IL cc L .; Z
450 0 Thl U
r-I .' np v r
N Z i J rI
a Cr 1 -1 t4- r 0
4 v C C iGt -,
Y cF CZ -
I C iT + O C H
4P91 + m 45%Li*
H A: W c
'-1 +21B N -r^l.D ui
16n 93 + 30 0' 4C
C Di 0 C N
n .!r 4t 0j "-
4 5 + TCJ .
U CLIi n
IC 10 ~rO
..i .... a::
this point, the theoretical development of a rate equation has been
completed which correlates the variation in monomer and stabilized
free radical concentrations.
A. Materials Used
The four asphaltic residues used in this investigation vere
supplied through the courtesy of the Texas Company. These residues
represent different types of asphalt sources that yield products
having different properties. Of these asphalts, several were
selected for modification by the processes of air-blowing and
selective solvent removal of the asphaltene fraction.
The air-blown asphalt (R60-11) was prepared by bubbling air
through a 5 kilogram batch of the Gulf Coast haphthenic Residuum
(ii19) for a period of li hours at 500 5 OF. in a stainless steel
heated ai vented vessel.
The asphaltene free asphalts were prepared from the filtrates,
debcribed in the following section discussing asphaltene prepara-
Sisp, by stripping the solvent (n-pentame) to a final temperature
qf 'bout 350 O at 1 em. Hg absolute pressure.
The properties which have been determined for these materials
a presenedA in Table 4.
isUe, a product of Chesebrough-Pbads, Inc., obtained from
*S Iqil phtbacetical supplier was used.
.9 .4 E
11 .* 4
ca F -
_li r' a
4 S -l
*^ 'ge C)
O 0Cj r a 0"
-, 1. 1- '-I e
(34B) 9 O
a-A s 4
^co nnr- I
- 33 -
c ( e
- 34 -
Petrole-a AsMt lteses
The asphaltenes were prepared froa the stock indicated by the
a) Stake a mLture of the petroleum residue and n-pentane
(1:1 weight ratio) until the solution is judged to be
haememeeus. This was actually accomplished by allowing
several days Idxing in a bottle supported on moving
b) Dilate a portion of this mixture to give a final solvent
to asphalt weight ratio of 20:1 in one-gallon batches.
c) Filter the solution on a large dense filter paper, retain-
ing the filtrate adn filtered material.
d) All the filtered material to dry, free of pentane, by
reading the filter paper out flat, exposed to the air.
e) Callime the filtrate with previous portions obtained
for future stripping of the solvent under vacuum conditions.
I 5ttglMt was r.Ae to wauh the asphaltenee thus obtained free of
"u "iii Wiotaas, since the purpose here is to obtain a large
qlity of the .asphaltser fraction from each material rather than
tl taiii u a strictly quantitative separation as is possible In one
a:? theb ,,iansaadized ~aasytical procedures.
O .aW a aHi, the results of a series of deterBinatimau based
i aliC r qisi ts" ofr IattriaJl for the asphalt 5119. It can be
*ii lai ~l'tuthis figure that at a solvent to asphalt ratio of 20 ml.
.......l e i'pei. . ashalt, a definitely larl decrease in the yield
..... S...i.. I
I~II al' E~iiii
- 35 -
o I I
I I These determinations made on
- the Gulf Coast Naphthenic
SResiduum, S 119.
0 I I-- I I I
0 20 O0 60 80 100
Solvent Ratio, ML. n-pentane Per Gn. Asphalt
Figure 2, The influence of solvent ratio on the yield
of petroleum asphaltenes.
of asphaltenes occurs. For this reason, the ratio of 20:1 on a weight
basis, which is obviously somewhat greater than the above value, was
used in the asphaltene preparative procedure, as a compromise between
the loss of quantitativeness in yield and the necessity of handling
extremely large quantities of solution.
The monmers used were obtained from the Borden Chemical Company,
Honpmer-Polymer laboratories at intervals to assure that fresh
materials were being used. Generally, the inhibitor hydroquinonee)
was removed by washing twice with two percent aCHB solution and
rinsing thrice with distilled water to a neutral point indicated by
phenolphthalein indicator. Drying and removal of trace polar mater-
ials (oxygenated impurities, etc.) is accomplished by percolation
through a column containing a section of granulated calcium chloride
aSd 80-200 mesh activated alumina.
To determine if additional purification of the moomer by
di&tillation would change the thermal polymerization rates, simple
vacuum adstillations at 120 OF. and 1 ca. Hg absolute pressure, using
a ii...fied vigreaux column, were twice performed on a styrene sample.
:iio change .sa: ..observable.
:iThie mo rs used have been styrene, vinyl-2-ethylbexanoate and
II 'viy'prhe t&l aAonn of polymer from the styBree used, with anhydrous
aiiitiimel, .it :was Jdsa that less than 0.5 percent polymer was present.
- 36 -
- 37 -
In all cases, coamercially obtainable reagent grade chemicals have
been used. It is considered that the following materials warrant
a) Xy/lene (mixed o, p, and m isomers) was found to be only
suitable in the best grade (Baker's White Label Analyzed
Reagent) obtainable, because of the presence of trace
quantities of materials which are slowly hydrogenatable
in the inferior grades.
b) The palladium hydrogenation catalyst, Grade 937, was
furnished by the Davison Chemical Company, Baltimore,
Maryland. This catalyst comprises 5 percent palladium
on a carbon support.
c) The methyl alcohol used in determination of polystyrene
was supplied anhydrous in one-pint, sealed bottles to
assure that the presence of water was minimized.
d) Adsorption alumina, 80-200 mesh, was purchased from
Fisher Scientific Ca pany.
B. Apparatus Description and Operation
Polymerization Equipment Description
Reactions of asphaltic aqd eammaeric materialss are catnitted
in a sealed autoclave which rocks to agitate the reactants. This
type of apparatus provides a mmans of reactltd in the dark, i.e.,
the effect of the catalytic action of light is eliaiated. Thtls
reactor is equipped with a temperature controller ( 12 a-d
- 38 -
recorder which allows reactions to be conducted over long periods
of time with a minimum of personal attention necessary from the
operator; however, this should not to construed to indicate that
neglect may be permitted. The apparatus is shown in Figure 3. It
may be noted that a nitrogen cylinder equipped with pressure
reducer is used to maintain pressure on the system. A list of the
principal equipment follows:
1) American Instrument Coapany super atmospheric
pressure rocking autoclave, -23/'8 inch series,
equipped with 3 liter and 1.5 liter bombs.
2) Brown Pulse Pyrovane temperature controller,
0-800 OF. range, Model No. 105C4PS-25, Serial
3) Bristol Company temperature recorder,
0-1000 OF. range, Model No. 12PG560-21, Serial
Auxiliary equipment is provided which allows introduction of onomer
and rae al of reaction sales whenever desired. Measurement and
Ciltrol of temperature is obtained by means of two thermocouples
is~eted into':a nll provided in the center of the autoclave. A
cakll water-cooled heat exchanger is provided on the outlet line
to. coolmii. i ap3!iii: Ldately as they are withdrawn from the reactor.
ftijsWzlattios SigEI mn t operating Procedure
To- chae r the Vi", it is removed from the rocker, opened and
satU" on ead ith the opening upward. The desired amount of asphalt
- 39 -
Figure 3, Polymerization Equipmnt
riire 4, Almaly
- 40 -
and solvent is placed in the bomb, the bomb is then closed and the
compression bolts tightened. Particular care must be exercised in
seating the cylinder gasket to avoid scoring or otherwise damaging
it. The compression ring should always be turned with the same
side facing the bolts so that the opposite side will remain smooth.
This may be accomplished by noting which side of this ring has
been previously scarred by the bolts. The bolts should be tightened
evenly by pulling thea down in a "criss-cross" fashion to a torque of
about 20 ft.-lbs. which is ample to seal 6000 p.s.i., as stated by
the mnaufacturer. The bomb assembly should be put together as
shown in Figure h. It is necessary that the retaining bolt be
fastened securely after placing the bomb in the rocker.
After completely assembling the apparatus and all auxiliary
equipment, nitrogen is allowed to flow through for about five minutes
to cqpletely purge air from within. Then the heater, controller,
shaking amchanism, etc. are turned on to bring the apparatus to
the desired temperature and allowed to reach themal equilibrium.
During the run, occasional adjustment of the controller set point
Sby be necessary. After the reactor has reached equilibrium temper-
ature, the n-apamer is added via the sample outlet by forcing it
into the reietor with slight pressure from the nitrogen tank.
Finally, the pressure is adjusted to 100 p.s.l.g. by the reducing
valve at the nitiaen tank. Samples are taken at intervals during
the run ant stewed in a refrigerator to prevent further polymeriza-
tiE..& It e. neeasary to lower the otUlet end of the shaking.
I;vl i ir e ......
C. Alytical Methods
Determination of oaoer by yd rogatioa
A considerable amount of effort was expected to establish a
suitable means of detersning the eoccentration of the reactants,
i.e., vinyl-moamers in the complex asphalt syj'ttm encountered.
Quantitative selective hydra~gantion has been found to be a reliable
method of analysis. Because of the nature of aaphalts the more
common methods of olefin analysis were discarded for at least o Q or
The following is a list of possible methods of analysis which
might appear to be feasible, but which were discarded by at Ilsat the
a) Spectroscaey Aiphealt are esasa.tally opaque to the
region of the spectrum which are~,a plicabe to thbe
arthaUi with the mtertfleeo Nued. (15, 20, 31, 36, 65, .I)
b) halogsnatius Either ijhiggterim iauthiietl 4ith
hlydro-hl .in cur. ifienty is g .
~~~~~~.... iiiiiiii ..................iiiiiiii
- L2 -
c) Aqueous phase addition reactions Apparently all aqueous
systems must be discarded be cause of the occurrence of
the problem of mass transfer in the heterogeneous system
resulting from the precipitation of the asphalts into a
d) Dilatometer techniques The expense of manufacturing
high-pressurentllatmeters capable of operation at a precise
temperature over the wide range of temperature desired is
prohibitive and the ability to obtain precise results
with systems containing only 1-10 percent monomer is
questionable.(20, 79, 10h)
e) Viscosity methods These methods would not be applicable
because no direct conparicon could be iade between the
various asphalts encountered, which is part of the object
of this research.(20, 32)
f) Removal of polymer by precipitation It has not been
possible to discover a solvent which vill precipitate the
polymer from solution without precipitating the asphalts
The above list is included principally to justify use of the
procedure of analysis adopted. This procedure, namely a selective
nicro-hydragenation, is actually quite reliable but very time
consuming and tedious to perform.
Selective hydrogenation of olefins over palladium has been used
by several investigators.(19, 64, 100) However, materials containing
- 43 -
araatic carbaoyl, nitrile and nitro groups nay exhibit interference.
(6, 64) The procedure described here is essentially a modification
of that described by Mitchell.(64) Low blank values have been found
when hydrogenating the various asphalts in the absence of vinyl-mono-
mers, indicating that the procedure followed Ir qte -w eleMtfe Onr
olefins in the systems encounters. Tbbles 5 a-i 6 iiUcate that of aoL
materials studied, only the asphafM; l any be aIs cfl f interfere
in this analysis. In all cases, the asphaltenes have been present
(when removed from the asphalt) in less than 20 percent concentration.
Therefore, approximately 3 x 10-6 g. holes hydrogen per gm. sample
will be used by the asphaltenes in the mxi;a situation.
Refer to Figure 5, a photograph and Figure 6, a scheatic diagram
of the equipment used to obtain details of the equipment construction.
The analytical procedure used is as follows:
a) Weigh accurately a quantity of saamle which will take up
fr m 75 to 100 El. of hydrogen into the appropriate con-
tainer. This is the sample boat shown in figure 6 in the
ose of asphalt samples amr the Victor-Okler type bulb
in the cme where calibration vith pure olefin is desija.
The use of the ma boat allow s. m ewC ioaaIm ama la:,
of the vaiaile caWtitweats of the i ple. ~ai can
ein1*medo by a4ltitid f:l.ip--dt snes with 5 to 6 h.. dr the
sualent xylea- after they Iberl been weighed.
b) mlns1e 50 :1. of 31W:., e a. ipni.ly 0.05 as q.i tbq; il
l daat4i. m
iiiiiiiiiiiiii iiii iii iiiiiiiiii iiiii i iiii
=.....- il- .
TABLE 5, HYDROGENATIOII BLAIK VALUES FOR SEVERAL ASPHALTS
Test Conditions: Hjdrogenation over palladium catalyst at roomc
temperature and pressure for at least eight hours.
nG. holess Equivalence
Hydrogen Per in Weignt
Desigiation Description Gm. Sample % Sty:vne
S117 East Texas Asphalt 7.5 x 10'- 0.078
511. South Texas Heav. 5.3 x 10- 0.055
Asphalt Base Residuum
5119 Gulf Coast Naphthenic 10 C 0 010
S120 East Central Texas 6.6 x 1o- 0.068
561-1 Asphaltene free 5119 < 10-" 0.010
861-12 Asphaltene free S120 < 10 0.010
R60-ll Air Blown 5119 < 10" 0.010
TABLE 6, HYDROGENATIOw BLANK VALUES FOR TWO ASPHALTENES
Test Coidetions: Hydrogenation over palladium catalyst at room
temperature and pressure for at least eight hours.
Gm. Moles Equivalence
Hydrogen Per in Weight
Designation Description Ga. Sample ( Styrene
s61-6 Asphaltenes from 5119 1.2 x 10-5 0.125
861-9 Asphaltenes from S120 9.9 x 10" 0.100
- 45 -
Figure 5, Selective catalytic micro-
bar, and the sample co i lner into the reactiep anel
as sho&n in Figure 6 aamd. aireile the apparatus.
c) Purge the burette of air and fill it with hydrogn. Turn
on the stirrer andt ter, and adjust the valvew o that
hydrogen will flaw thigh the ractii vessel ta pAlI
it of air. About 30,iMlliIA. hb been tentml ail" to:.
reduce any oxide oa the caaWysat .t ,siiihs thIe eltieon
with hyda4ggan a tblalit a Miier up-taug at, W3aRdMi
- 146 -
- 47 -
occurs. A longer time should be avoided because of sample
d) Vent the ballast tube to the atmosphere, fill the burette
with hydrogen, adjust the valves to connect the burette
with the reaction vessel and obtain an initial burette
reading; checking to make sure that the pressure in the
ballast tube is balanced by the pressure in the burette
via the manometer. Close both vents and empty the sample
containers into the solvent by rotating the ground-glass
joint which supports the sample boat. The Victor-Meyer
bulb, when used, may be broken against the open gaaple
boat by gently shaking the reaction vessel which is
connected to the burette by a flexible coupling. Obtain
the atmospheric pressure and water temperature.
e) Occasional adjustment of the level of the mercury in the
burette is necessary to avoid large pressure differentials
which would otherwise cause leaks to develop. After con-
pletion of the hydrogenation (this is judged by the end
of hydrogen up-take), obtain the final burette reading
after adjusting the pressure to that in the ballast tibe
by means of the leveling bulb. Usually abe"t two to three
hours is required for cqpletiar of the analysis after
complete solution of the saiae is ache-ed. Became of
the shape of the s aple bd.t, the hydiiAz~ptlqp doiStt
tends to gradually ac'ac4late in it. It is thedlfore
Other Methods of Olefin Analysis
Several other test methods were devised for analysis of olefirs,
particularly styrene, and used to check the accuracy of the hydro-
genation procedure. A monWmer volatility test, a test employing
quantitative precipitation of the polymer formed and a bromination
test will be discussed.
The Volatility Test
The purpose of this test is to evaluate the amount of unreacted
mamLer remaining in the samples of nonomer-polymer-asphalt mixtures
resulting from reactions occurring in monomer-asphalt systems. It
is uggarally to be assumed that only the unreacted portion of the
apnmuir is determined. However, this is necessarily only an approxi-
.:jteas because of inherent test limitations. The principle involved
is that of the removal by vacua distillation of the volatile portion
(lii mmer) from a suitably arrange sample at an appropriate tempera-
ue- ad-a pressure.
A kanutant teWmerature oven large enough to contain a large
siZedl li dessicator provided with shelves or a high vacuum oven
i inireqM tfor beating the samples. The ssples themselves are con-
tabat xai the bgftt half of 10 cm. diameter petri dishes covered
wiWthA 'Iith b-] s turam. with the coneavity downward. This allows
any ;~ae Eipte rag flow back into the petri dish rather than
u th s.sib as vs would otherwise be the case.
- 49 -
Using reagent grade chemicals, a solution of the concentration
of 1.00 gram per liter of p-t-butyl.:atcchol dissolved in benzene is
prepared to serve as an inhibitor during the test to prevent further
unwarranted polymerization of the momxmer.
Weigh accurately (t 0.001 gmn.) an approximate 5 gi. sample
into a petri dish and record the total gross weight of sample, petri
dish and watch glass used for the sale. Imandiately add 10 al.
of the solution of p-t-butylcatechol and allow the sample to dissolve.
After solution of the sample, gently tilt the dish several times to
mix the inhibitor. Place the samples in the oven at aseepheric
pressure, adjusting the watch glass so as to caly cover about two-
thirds of the petri dish. This allows more rapid reamval of the
When the samples appear nearly dry, remove them fras the oven
and allow them to cool to room teoaerature (overnight is permiesble
since further polymerization will be inhibited). Cover cooletely
and again place the sales in the oven, but this time uaer vacuum
for a period of eight hours before removing san reweighih the
sales. The owvn ast be cold at the start to avoid sample spat-
tering. It my be necessary to vary the taiirature and ji~ieasure
when removing various .aemwr. However, it as..beaen iaut Miti:
*ucceesful remaeal of styane :piger can be ac~ir litl= la
5-6 am. U pressure sad 200 OP .
- 50 -
The weight percent styrene is determined as follows:
% Styrene= 100(Gross ut. loss + Blank. sample wt. gain) (a)
The Quantitative Precipitation of Polymer (20)
To 100 milliliters of anhydrous methanol in a 250 milliliter
beaker add 25 Pm. of sample. Stir well and place on a steam bath for
five minutes to coagulate the sample. Allow the solution to cool
and filter on a tared Gooch crucible. Dry under an infrared lamp
for five minutes, then allow the polymer to dry to constant weight in
an oven at 70 oC. Cool in a dessicator and weigh.
The weight percent styrene is determined as follows:
% Styrene = % Styrene Charged (Wt. polmer)(0lO)
It must be noted that this procedure will work only when asphalt
is not present since methanol will also precipitate a portion of the
The Broadnation Procedure
Refer to the American Society of Testing Materials standard test
D 1158 57T. (11) It must be noted that this method will work only
when the asphalt is present in a small amount since it precipitates
and otherwise interferes in samples which contain an appreciable amount.
A c.anaXrson of the results obtained by these methods with
those obtained by hydrogenation is given in Table 7. The hydrogenation
(a) 'Te blank is determined on an asphalt sample adding only the
p-t-butylcatechol solution. A value of 0.017 gm. appears to be satis-
factory as a result of the average value obtained using several
samples for blank determination. Further attempt at increased
.acuracy is not felt justified in view of the overall accuracy
"of the method .
- 51 -
4 N N
,j r- CN
o n", a
, 1,- r'
N 0 .2
I 0 '
i. ) )
0 0 .
r- 1 14r
c- r- r-
fl \0 C%
no a VI
I ba o
I'laB 0 t
0 Nr C
0 0' 0
0 h I
w0 o' 0
... ... ...
0-I CY H
- 52 -
procedure consistently gives results which are slightly low. This
can be attributed to a slight evaporation of the monomer from
samples during the purging phase of the hydrogenation procedure.
This has also been found to be true for hydrogenation of samples
which have been blended to give a known composition. The error
from this source appears to be less thar 5.0 percent of the total
quantity of monomer present. The results obtained by hydrogenation
and bromination are plotted in Figure 7 in the manner employed for
the use that these data are intended. It can be seen that very
reasonable agreement is obtained, even though the hydrogenation
values are somewhat low.
o o N
0-J 0aTA~ro 0ragZoiriP
A. Method of Calculating Monomer Concentration
Berivetiom of an Equation for Calculating Monomer Concentration From
Jacket te merature
System pressure (by reference to
the atmospheric pressure)
From the ideal gas law,
V = nRT
Where: n Ga. males of gas
R Ieal gas law constant
T Absolute temperature
P/T = n=/V
(P/T)baUllt = (nR/V}ballast = a constant
(f.i is true for a given determination since
n aw V are coatant.)
[bam ter reatUng (P /T)
- Jae|t tfbemaature initial = b
- 54 -
h r iiiii ...
Where! b Denotes burette
j Denotes the jacket
i Denotes initial conditions
Note: Under the conditions in which thick test is
performed, use of the ideal gas law is adequately
Since the ballast tube provides a means of reference to a
constant (P/T) ratio, minor fluctuations in the temperature of
the water jacket are admissible as long as the entire apparatus is
all at the same temperature. Continuously flowing laboratory tap
water through the apparatus accomplishes this vith no more than a
20F. variation in temperature from beginning to end of the deter-
mination. The amount of gas disappearing may then be calculated
h b 10.
Vf = V + f
hp r b 11.
Where: r Denotes reaction flask
b Denotes burette
h Denotes hydrogenation system
f Dc.notes final conditions
From the ideal gas law and substituting from the preceding
n = Vh PI/RTi = (Vr + Vb )Pi/Ri 12.
n= V P /RT = (Vr Vf 13
nf Yf f r b Pf f 13.
Subtracting equation 13 from equation 12 and substituting the
constant ratio of (P/T) from equation 9, gives,
% = i nf = (V Vb )P/J 14.
Where: o Denotes the change from initial to
final caEnditiao -
- 56 -
However, since the gas is saturated with solvent vapor, a
correction for this error must be applied. The proper correction
factor may be determined as follows,
nH /ln = pH /'p (i.e., partial pressure law for 15.
2 2 ideal gases)
Where: B2 Denotes molecular hydrogen
S Denotes solvent vapor
p Partial pressure
The partial pressure of hydrogen is the total pressure less
the solvent vapor pressure: this gives,
=2 b 5 16.
Rearranging equation 15 and substituting from equation 6l, yields,
Pb PS 17.
From a material balance on the gas disappearing,
no= ng + nS 18.
Note: The above is approximately true if the small
fluctuation in temperature does not change the
solvent vapor pressure materially.
Substituting equation 17 in equation 18 gives,
n = n.2 x S + 1] 19.
Equating equations 19 and 14, and solving for n.2 gives,
n" = (Vb V) PS)/RTJ 20.
2 1 f
The monmemr concentration may then be calculated as follows,
% Mnoaer = 100 No m/W = 100 m(Vbi pb sb PS)/RTj 21.
Where: a itameer molecular eight
W Weight of sample hydrogenated
Determination of Solvent Vapor Pressure
In order to use the preceding equation, it is necessary to
estimate the solvent vapor pressure. Several factors which will
affect the exactness with which it Is practical to determine this
a) The presence of sample as solute in the system. (72)
b) The x lene used is a mixture of the o, p and m isomers
which tend to vary in concentration from batch to bath.
c) There is a sUight variation in temperature over the
extent of the determination.
The last two isomers of xylene noted above predominate in the mixed
system.(80) For this work it has been considered adequate to use the
data presented in Figure 8, which is the vapor pressure of m-sylene.
Xylene was picked as a solvent because of its ability to
dissolve both the asphalts used and the polymers formed, its low
vapor pressure and ready availability In a grade which has no
observable hydrogenation blank value. (Xylene supplier mentioned
B. Overall Polymerization Rate Constant Calculations
Empirical Determination of the Polymerization Rate Constant
By integration of the equation developed for the polymeriza-
tion process under study, an equation results which aay be rearranged
to give a linear plot when reciprocal concentration is the ordinate
and tim the abscissa.
: 10 30 50
FigurTe 8, apor of -lene
figure 8, Vapor psmsure of m-xylene
This equation is,
-(M)/dT = K(M)2 22.
Where K has been set equal to: (See page 29)
2 4k D f I
K =(kp k /2k ) ((R ) + ta(2 + k (R *)) -R 23.
p ts ta 2 Ce s a
By separation of the variables and integrating between the limits,
T = 0, T, and M = M, M, the following equation may be obtained,
1/M = KT + 1/M 24.
Where: T Polymerization time, minutes
M Monomer concentration, eight percent
R Stabilized free radical concentration,
No. free rad. per unit weight
By plotting 1/M versus T experimentally determined during a
given polymerization run, a straight line vill be obtained whose
slope is identically equal to the overall rate constant. As a check
upon the experimental data, the intercept of this line with the
ordinate should equal the reciprocal of the miassner concentration
initially charged. This calculatiqp is illustrated in Figure 9,
using the experimental data for Run 60-9.
Correlation of the Tedwerature Effect
Determination of the Activation Energy, H
Previous: to the develop pEt I:: the E yrg theory of slp olut
usd. to .e.v.aluate the effect of theatre up. th.e m i.....
velocity c...e.t. This c&u iis been be Siiiiii;. iiii ta......
to a e.. ...f. a la u.r .f
h.. ..........(.......I.....................1...... ..
a........... ................ ..... ........
\ t- -
I IH ,-
10 0 ,
.., ...... .... ,,,i
- 61 -
The Arrhenius equation for correlation of the overall polyeri-
zation rate constant K is given below,
K = A exp (-H/RT) 25.
Where: A Arrhenius frequency factor
H Energy of activation
Taking logarithms of the Arrhenius equation and rearranging
loglo K = log10 A (0.4343H/R)(1/T) 26.
Therefore, the slope of a plot of log10 K as ordinate versus
(l/T) as abscissa yields a plot whose slope N is,
N = (0.4313H/R) 27.
If R, the ideal gas constant, is taken to equal 1.104 cal.
per gm. mole x OR. and T is in OR., the energy of activation, H, may
H = 1.104 N/0.4343 = 2.542 N, cal. per gm. mole 28.
The numerical value of the slope, N, is determined graphically
in a manner analogous to that employed to estimate the value of the
overall polymerisation rate constant as illustrated in Figure 9.
Since the details of this calculation would be repetitious, they are
An Absolute BasiL for ~te Overall Polymertition Iate Coywtmat
Theoretically to be used properly, the K of the. A.rrbhmAi.Mt -
tion must be expressed in :*fM< ,tt of unit which is dlnetly
proportional to tih& number of tom: or sele.lels SAwcting.lpr uait
volume of resotion systt. Fros uiiia0l tuei it, hsqql bqp.i,
this any not be necessary and may even be misleading. Conversion
of the data as obtained experimentally may involve making assumptions
of quantities such as densities, etc., which introduce unnecessary
error into a correlation that might be satisfactory for many purposes
without applying the corrections.
The overall polyaerization rate constant, K ($-1 x min." ), may
be converted to K' in units with an absolute basis (liters per gn.
mole-win.) as follows,
K' = Cf K 29.
Where: cf = 0.1 monomer molecular weight
solution density in gm./cc.
This any be demonstrated in the following mnner,
100 ~. solution gm. monomer cc. solution liter
Am. Ma-omer-ain. om. mole monomer gm. solution cc.
liter per gS. mole-min. 30.
The order of aagnitude of the error introduced in the energy of
activation calculation as a result of neglecting the correction for
density variation with temperature my be estimated. By the method
previously described, I may be determined for a given system in the
f-11 ol iiw aMer,
l 400P lo10K00 31.
thids ', dmste a corrected value of N and K, the following
W Glogbto. -logWWoF. 32.
(.-I F.-(/)20. 200 .
- 63 -
Sample Calculation for Runs With S 119
For the system, 10 percent styrene and 90 percent S 119, the
solution densities are: 0.911 gin. per cc. at 200 OF. and 0.851 gm.
per cc. at 0 OF. (Refer to Table 8) Therefore (styrene molecular
weight is 104.1),
KiooOF. = 0.1(1lO.l)(K20OOF.)/0.91' 33
,00oF. = 0.1(104 .1)(K4ooOF.)/0.851
TABLE 8, DE4BITIES OF STYRENE, Sl19 AID
SOLUfIONS OF THESE MATERIALS
at 77 F., at 200 F.,
Sgn. er cc. .a. per ce
+ 90o s119 0.914(e)
at 400 F.,
I.F. per cc.
.. .... .... . .. . ... ......
a) Cited in reference (7).
b) Extrapolation of data cited in reference (7).
c) Measured value.
d) Estimated frca the cubical coefficient of expansion cited in
(e) Calculated assuming additivity of densities of the Materala
comprising the solution on a -weight basis.
By reference to Figure 32, pWSe 714, it may be anPd tha",
ilog o K400. glo0 K t = 2.85 3ll
(1/)4 F. (1/)200 = 0.000354 .. .-1
..... .......................-..:: .....
........ .................... .............. ....EEE= ....::::::: .....
Substituting these values and the values for density gives,
log10K.OW F. lolO200F. + log14(0.914/0.851)
= (2.35 + 0.0306)/-0.00035h = -8,140 oR. 37.
The value for N determined without applying this correction is
-8,060. This represents an error of,
Percent Error = -081 O) x 100=1% 38.
The energy of activation determined may be expected, therefore, to
be low by about one percent.
Determination of Constants for the Theoretical Rate Equation
The I. B. M. 650 digital computer has been used to calculate
values of K as a function of Rs*. By systematically varying the
parameters it has been possible to obtain calculated curves that
agree approximately with experimentally determined data. This
discussion will be presented later in greater detail.
The program for these calculations has been written for the
I. B. 4. 650 by utilizing the "Flatran" system. This system, which
has been developed at the University of Florida, translates the
proeao'g as written from simplified symbols analogous to those used
in ordinary algebraic calculations, to the more complex language
utilized by the machine carputer.
The equation for the rate constant, K, has been reduced to a
simpler form by substitution of symbols as follows: (See eq. 23, p. 59),
K = A [(R2 + BR + C)2 R] 39
A = k ts/2k; B 2= kt k,/s; C = 2kta/s; D = D;
R = *.
The Flatran Program
The Flatran program has been written in accordance with pro-
cedures defined in the Flatran Ianual by Peterson.(73) This program
is designed to calculate values of K in term of R over a range of
R = 10' to 9 x 10" in increments which vary with the magnitude of
R. The Flatran program follows:
0 0001 0 READ,A,B,C,D
0 0002 0 K#A*%SR$R*R&B*$R-*
0 0002 1 DMSCO-RC
0 0000 0 xf#
0 0000 0 1.0000001
0 0000 0 PtN~l ,R,K,A,B,C,D
o 0006 0 *&W
0 0000 0 V2wF
0 0000 0 DOSBEF,G,H
0 0007 0 KrA' As0nQBraW&B*
0 0007 1 DoCa-Ro
0 0008 0 PUCH,R,K,A,B,C,D
0 0000 0 flWOF
0 0000 0 Ir*-o.00o,1,1,1
0 0009 0 STOP
0 0000 0 ED
The m~hlne language program caciled via the flatran pragmw I
included in the Appendix.
C. Calculation of Re~st&ve Free radical C aemtlp
Tia experiantally dete-Otuad data for the varne ouitilt
system has been plotted in the ..~x te g wttOi in t i iMa l ili
i.e., the polyasrization iate a ti-h., K, verla .,b P;i......I.i
.or ethe lti....... s.m Se iii
Sa in tsip -d~ac~iw i-of ren~sa Bflttum~;l ,dlu s;:|iii pi
- 66 -
concentration can be converted to a basis such that the stabilized
free radical concentration can be compared. Experimental data has
been determined to evaluate the free radical concentration in the
various materials on a weight basis. Using these data, which are
given Table 10, page 86, the desired comparison may be made in the
Define the ratio of free radicals in the various asphalts on a
weight basis as,
FR ratio of A:B = Free radicals per g. A 4.
Free radicals per gm. B
D. Estimation of the Absolute Free
The estimated dlm of the absolute free radical concentration
of the various materials have been given in Table 10, page 86. If
it is assmed that the free radical concentration in an asphalt or
asphaltene solvent system is proprotional to the asphalt concentration
R* = qAU( Species A in the system), 41.
in .g. males free radicals per kgm. of system of A
Where: qA The specie qwtient, a quantity that will
cause the uiats of Ra* to be, ga. moles of
free radicals per kilogram of system of
Species A. Species A may be an asphalt
In ofder to obtain the concentration of R,* in the desired units,
it is neapamary to sate the conversion of units from free radicals per
.-~~t0o s L holes free radical per kh. This may be accomplished by
the uee of equation 42, which follows,
Pree L dioal Cone., in bg. males
.R. per kgm. esa4le2.
Free Radical Cec., in x 1i 1 le
F. R. per p. kpi. u.02Y10 cue
Solving equation 41 for qA gies,
qA Specie A in the in t 93.
The nmerieal value Of qA My be obtained by aing the value
of %R carvepoolndn to 100 percent Of Species A, i. e., for the
coaseam tiea of the asphalt or aspbaltene with no solvent aided.
The specie qmlients neiiwl for systems sauAd ae calculated
(NO: The nimt Of qA a m WIs 74. jar .m )
% taieas A
(1.1 x 1016 x 1)/6.Oe x 1023
-t* 00 1.83 x 10-7 4.
(4.4 x 1016 x D?3)/6.02 x 1023
q9120 a IM0 i 7-31 x 10-7 45.
(5.7 x 1016 x 103)/6.02 x 1023
l6oo 0 9.17 K 10.7 7 46.
(11 .x 1016 x 103)/6.02 x 1023
qs61.1 a 10 2 18.3 x 10-7 T.
All of the polymrizations ade in this study have been conducted
is the tinaco autoclave described in the experimental section. This
stuy hba been principally coacermd with the maler in vhich asphalt
a-sati are capbie of interacting in reactions vith vinyl aemomers.
Vinylnreanae ha been the .amaer used in the ajer part of this study.
(tari., imiam rs n eed were o -vinylamphtbalene and vinyl-2-ethylhexanoate.
Swenal typical petmramua residue, representing a vide range of types
2jlily ~countentj Wze the asphalt material used in this study.
".h. :E parties of these saphalts were given in 'able 4, page 33.
h Golf Coast apekthenic residuum, S 119, sad the East Ceotral
"!M ,aditMe, S 120, which represent residue of lor and high
13asi i. am eqest, resoactively, have been studied extesuively.
.. matI fati s bew ales been made in which the asphaltenes removed
iiiww -f hiltse by n-pentame precipitation were used. It will
l .li i nmi"ama tly that the asphaltema fraction of these asphalt
iiqi i# e I wiltfly at the ingredients which are the key to the
il3A "in tbhse results is the awaMUr in which the polymri-
g U| iiiii in i af' te *pltia: studied may be correla ted empirically
vhlbU aam wiatl ia _aI r cceatrats s eat reaction tempePature.
ig SiwillHiii this wat is such that a comaidemable mount of
W 4l Ellt all 4,,is:l|it ~a s igatred to ceaduct a single polymerizatirn
- 68 -
run. This run yields a set of data from which may be obtained a
single value, namely K, the overall polymerization rate constant.
With this single value, if the initial concentrations are known,
it is possible to again construct the entire curve representing
the results of this run. Therefore, the quantity of data cited
in this section any appear to be small. The numerical values of
these empirically determined rate constants are tabulated at the
end of this section. The original rau data from the polyerlzation
experiments are given in the appendix.
B. Correlation of Polylerination Rates
With tanomer Comentration
Experimental data are obtained in a tabulated form with the
unreaated maemr concentration (weight percent) measured as a
function of the time elapsed (minutes) from the start of polymer-
ization. It has been fouta that straight lines are obtained if
these data are plotted with reciprocal macear concentratiea (weigh
percent)-l as ordinate, versus the time elapsed (.mnuses) from the
start of polymeri~ati~ o as abscissa. Piguar 10 a d 11 incllpadt in
this section, and Figure 9 of the calculation section and figure 7
of the experimental section have been plot-sd n this awijair Sj
straight line dApendance as shaba in these figu-re is. tyi il of
the data obtained in all of the polyam~Pai ti.As preu-S .
ft~j~f QO f O.... ...... .. .. ....... .1 :::: ... ^" :::: i
This type of data ay be carzdated I-ams- aof-t Sad qar
rate eqttion in teamm of ampimr comoentntten. (Refr ii the .
calculati"a uan.ti.on ) "te ....... l.l.
A *. :
- 70 -
o 0 0 0
P I iJ d
\ \ s
- y 1 -
I I D
x o ( D
^ 43 !
T% 'uouox^J4U90UOQ i^Eea^^s rdfoet
- 72 -
If the units are applied consistently and the data plotted as just
mntiQaed, K will be identically equal to the slope of the line
abtaiind and have the units, (min.)-l(percent mo omer) Here K
has been used as the empirically determined overall polymerization
Check nrms Qn the same system under the same conditions have
agreae within about 10 percent. Data from the runs which have been
Wme in duplicate in this study are presented in Table 9.
TABLE 9, A CiWARISQI OF EXPERMENTALLY DETERMINE PATE
CWdTMITAS TO SaHW REPMIUCIBILITY
.. ... .. .. ... ..... ......... . .. .. ... .
Iagzmijatian Run Coaditims: These runs were charged with the
g~peMt asphalt as indicated below plus a quantity of Styrene
approaching 10% with xylene comprising the difference.
Teaierature, CCe ant&a
a bB. Aspbalt j Asphalt "F. min. x %L
R61-4 meae -- 300 4.08 x 10-5
R61-5 oea 300 4.12 x 10-5
B61-16 name -- 300 3.96 x 10-5
IjF-19 S 120 90.0 300 10.85 x 10-5
8as2 120 90.0 300 10.09 x 10-5
ai- s 8119 90.0 4oo 1.17 x 10-3
u S 119 90.0 400 1.22 x 10-3
i.... . . ............ ...... . .. . .... .. I I I U
Sk eWalt. of pmaipitfl*e Of the p&e3Q r during the pelyeeriza-
,gAQl^N baris McoaptSAu as a priiDble maruce of erroemsuws mrults.
- 73 -
Petrolatun, a solvent incompatible with the polystyrene, was used to
determine the nag itude of this effect. The rate constant determined
for this run is plotted in Figure 12 for the purpose of comparison
with the data obtained for asphalt systems. It is obvious that the
possible source of error is potentially quite large. The possible
occurrence of a polymer precipitate on the interior of the autoclave
vould be readily observable. This did not ever occur, however, for
polymerization in any of the other solvent systems used.
C. The Arrhenius Correlation
It is apparent from the data given in Figures 10 and 11 that
temperature is a very important variable. This is in accordance with
the effect observed for many reaction system that are essentially
irreversible. As might be expected from the verk of others in the
polymer field, it has been possible to correlate this temerature
dependence by mans of the Arrhenius equation. The sme data shown
in Figures 10 and 11, for the Gulf Coast naphthenic re idua (S U19)
have been used to calculate the overall polymeriation rate constant,
K, and plotted in Figure 12. The second line shown in this figure
is for the air-blown asphalt R60-11.
The slopes of these lines are about -8,060 oR.
The energy of activation, H, is theefore,
H = -1.104 (-8,060)/0.4343 =+20,500 cal. per goi. ple 49.
If a L&. percent peitive. correction is applied, a. ::*i5iid in
the calculatia section, A will be fer thope aptm.:
3 =1.01 (20,500) = 20,700 cal. per ag... Slie. IH0
< 1,000 -
Temperature, OR."lx 103
yigae 12, Arrheniue temperature correlation of overall
gigiiuipliuigiti kli rte constrnts.
- 74 -
- 75 -
Bamford and Barb, in their review of polymerization processes,
cite values of H, the activation energy of the thermal polymeriza-
tion constant of styrene ranging from 19,200 to 23,200 cal. per gm.
mole.(14) These values were determined over the teamerature range
of 0 to 120 C., which is somewhat lover than the temperature range
of this study. It may be noted that the value of B obtained here
is within the range determined by other investigators.
Although the data obtained are insufficient to allow an accurate
calculation of the activation energies (only one or two points
obtained per system), the data of Figure 13 show qualitative agree-
ment with those obtained for the more definitive data presented in
D. Asgalttene Constituents Exhibit a
Data have been obtained to show that cMiements concentrated
in the asphaltene fraction are largely raespasible for the influeneen
that the asphalt studied have had on the poly3er kintics.
Reference to Figure 13 indicates that asphalts having a high
asphaltene content have caused the greatest acce: -satt e of the
polyeirizatioa rate. At a reciprocal ts;dieitare cories.Iafim. .nliti
300 3P., the slte cahd"ite obtained Tor pol.taSi action in, ylewn i la
shown for coeqarisen. This single point represents time results dI
three uns.wme one the sepasaie batches of styre!:im u : ini the rIun
with the aasfplt mys te. tiCe "A" ga 5 ;i hasfliiiibJtiiUI UP.ue;
afl*tlenemi while the "B" (sp has d -i 1.5-2.6 percent j
- 76 -
As Indicated 10% Styrene
o Asphalt S 117
* Asphalt S 118
o Asphalt S 119
* Asphalt S 120
A Asphalt R 60-11
(R60-11 is air-blown S119)
Reciprocal Absolute Temperature, OR .-x 10
tigur 13, S~ary of the overall polymerisation
rte coataints obtained in the various systems
A = High Asphaltene Content
B = Low Asphaltene Content
- 7 -
It is obvious that the grouping of these data in this manner is
permissible; however, it must also be noted that the sulfur content
(see Table 4, page 33) of the asphalts has a similar relationship.
The asphalt S 119, having a low asphaltene content, and the
asphalt S 120, having a high asphaltene content, were selected for
further investigation of the asphaltene content variable. By
concentration of the asphaltenes in one fraction through n-pentane
precipitation, and removal from the remaining asphalt by filtration,
it was possible to study the two fractions separately. If
asphaltene concentration is an important variable, it should be
possible to observe its effect by dilution of the asphaltene with
a suitable inert solvent, or solvent having known effects. Xylene
was chosen because it is mutually compatible with styrene,
asphaltenes and polymer, and is known to have no acceleration
influence of its own. The results of these experiments are shown
in Figure 14. It can be seen that at low asphaltene coaeentratima,
acceleration of the polymeristlen is paemosmoed. As the caames-
tration is increased, this acceleratieni gradually disappears. Also,
it can be seen that differeacee in the character of the asphalteee
themmnlves have a netionable effect on the rate, since the
asplaltenes fra S 120 gave a substantially greater accelontims
than did the aephal tes from 8 119.
epeating these apewiments with the basw aaphalta S 119 an
S 120 yielAld aults ealupmse to they obhflmed for the agiaalatenes.
TMMe resla 0s. abmn in piguwe 15. An iWaoflant dsiffueamse tald
- 78 -
ete U3 00
V CII I
n 11 r
_ _______ I _
I n n*--
4 H r
o "i nl
c-. XT- 'u'AR 50OT x 1 ietsuoD e0B uoyTIZ-laLod
x x e-Q----- 0
-I----- I- i I 0
r-- ,-< al
-. I I I o
c( X o
O v co
r p. &. ci. c i r *' '4-
0 '0S ('!
- --------- t ------ 38 !
- 80 -
be noted for these plots since the concentration variable is extended
considerably in Figure 15. These data indicate the diluting effect
of the inert asphalt fractions. Polyerizations at the concentration
of asphalt yielding the maislm rates, were mde, using the asphaltene
free residues (ceataining less than 0.3 wt. percent). The results
ef these pelmarizabtieo ruas are shown in Figure 15. The acceler-
ation of polySerization noted for the original asphalts, at the
coneEtrati n of asphalt yielding the maxiimu rates, was greatly
reduced. This streagly demnstratat that the asphaltene fraction and
constituents contalned therein are a key factor of this phenomenon.
Mmiftse should be made of phenomena which are annifest for the
asphalt systems, not observed for studies of the asphaltene fractions.
As relatively high asphalt comcentratinas are reached, the polyMeriza-
1tiL is again accelerated. Clarification of the cause of this vill
be covered in the discussion of results.
Earlier in this section it was noted that asphalts with high
aigalt~bl e content had the greatest acceleration influence. Among
this gre a ws the air-blown asphalt R60-11, prepared from S 119.
Ina 0F1ie 16, the eencemtBtion variable for S 119 and R60-11 is
Awil d. S M leans ti a obtained for the air-blown asphalt is
'- partbslarly Odmif sbhed by increasing asphalt concentration, as
it the den for the midim S 119.
Altheigh it ha not been explicitly stated, styrene has been
ap oLy inmWpr ftr which %gilmlt have been cited up to this point.
imAiwbr to ua alisze this sti y, other mmemrs were included.
- 81 -
8 a X X *t o
|J x_'"uTI *Tpr x 2 I7uiUtUo 3 esa uOalis FMCitEo
- 82 -
Alpha-vinylnaphthalene and vinyl-2-ethylheanoate were polymerized
in xylene and in sylene containing asphaltenes obtained from the
asphalt S 120. Vinyl-2-ethylhexanoate which had no measurable thermal
polymerization of its own was not caused to polyerize by the addition
of the asphaltene. The presence of asphaltawes did affect the
polymerization rate of o(-vinylnaphthelene.
Since no polyaerization occurred for vinyl-2-ethylhexanoate, the
data obtained are not included. A control test containing one percent
benzoyl peroxide in this monomer caused gelling to occur at 200 OF.
in 2.5 hours.
The data detained for o -vinylruphthalene is plotted in Figure 17.
Because very few points were Cbtaiaed for this system, the curve drava
B iee e w e dehed. Data w r limited because of the high cost of
this ce gout. 2The is qualitative agreement of these data with those
baMin for the styrene system. The magnitude of the observed
effect is not large, being of the order of the anticipated exprl-
Smental error for these poly% risation experlimls.
I5. aInaee of a Pure Stable Fuse Radioal
Sgb~lifed free incals were supeoted of being the comtituent
przaet in the asplalteae fractm .vhich is rsponible for the
li;Ej iewsd-. l2afdments usi ,l-d 1,1-phbel-2-picrylhydrazyl
yiM~idet: the esls ple*sed in figue 18. Althlgh the data are
fhi ,st aebrieli a, there is at least an izdica tla of an acceleration
defeat fou.tstm- by a., giM Il aieemm of this effect as the fae
..l4mala ca .atp is incnrawed. It is believed that the
WA* 11.41 x
- B4 -
I -- -
S1-1 r- .
0 N -t C. -
0 % -1J c
u..TX 501 xL~O' X Luiouoo e 4uH uoT.,4vz-pxIriS"[o
CI. 0 -
C t 0i.
- 85 -
acceleration noted here is definite, based on the reproducibility
of the experiments conducted in pure xylene. The spread in the
values obtained at the maximum is probably caused by the delicate
balance of factors responsible for the occurrence of this maximum
F. Spectroscopic Evidence of be Existence
of Stable Free nt-
The four asphalts used in this study, and the two asphaltenes
prepared from these asphalts were submitted for analysis by ESR,
on a contract basis, to the:
Ridgefield Instrument Group
P. 0. Box 337
The determinations were made on a relative basis by comparison at
the sane instrument gain. Examination was umade at a field-strength
of 3,380 gauss, with a micro-wave frequency of 9.47 kileamgacycles.
The spectroscopic splitting factor, "g", was determined to be
2.0021 for these materials. This is quite close to the theoretical
value which can be calculated for the free-electron, i.e.,
- 86 -
TABLE 19, IE RADICAL COCENTRATIO CO THE ASRHALTS
AND ASPrA LMEES USED IN THIS STUDY
Sample Relative Signal
5.1 x 101
1.1 x 101
5.7 x 1016
11 x 1016
(a) These values are estimated to be accurate to within about
10 percent of the value determined. They are the average
"f determinations made In duplicate.
(b) Deeradined by comparison with a 0.002 M aqueous solution
of ME tal-.
(c) easureennts were made by Schlutberger Corporation using a
Strand Labs modell So. 600 spoctrcmeter at 9.47 kae. and
3380 gane. Sbaplea were prepared by means of scooping
1.5 m. dieter pyrex capillary tubes.
TABLE 11, GRAPHICALT.Y DETEPKTNED C'VERALL PCLYMERIZATITC RATE CONSTANT FGR
.LIRUS ITH THE GULF COAST NAPHTENIC R ESIDUUM (S119), STYREVE AND .TLENE
Polymerization Run Conditions: Theae runs were charged with the percent
S119 as indicated below, plus a quantity cf ?tyrene approaching I04,
with 'Xylene comFrisrng the difference.
Pun No.. 119 (Ln.-1 x -
R61- 4 0.0 300 4.08 x 10-5
F61- 5 0.0 300 4.12 x 10-5
R61-16 0.0 300 3.96 x 10-5
P61-48 2.91 300 9.32 x 10-5
R61- 3 4.97 300 9.92 x 10-5
R61- 6 4.99 300 10.57 x 10-5
R61- 7 14.9 300 6.54 x 10-5
R61- 2 29.8 300 500 .x 10-5
R61- 8 36.9 300 4.38 x 10-5
R61- 1 6c.6 300 4.75 x 10-5
F6C- 5 90.0 300 5.87 x 10-5
R60-1C 89.2 200 1.07 x 10-6
R60- 9 89.8 250 7.22 x 10-6
P60- 4 90.0 400 1.17 x 10-3
R60- 3 90.0 400 1.22 x 10-3
The foll.-wing run was made with 5119 after removal of asphaltenes by
preciFitation with n-pentane. This material designated .61-1.
Run No. % 61-1 OF. (min.-1 x %1)
300 5.79 x 10-5
6l1- 9 4.83
TABLE 12, GRAPHICALLY rETEPRMIJET CVoRALL PKLYTEPRIZATTCI RATE CCrJSTANT FCR
RUTS ~rWTH THE EA3T CE-TRAL TEXAS RESTPUUM (3120), STYRELE A!D T EE
Polymerizaticn Run Conditions: These runs were charged with the percent
S120 as indicated bel-w, plus a quantity cf ?tyrene approaching 10C',
with Xylere comprising the difference.
Run lo. 9 "120 OF.
(min.-l x t-1)
The following run was made with 3120 after removal of asphaltenes by pre-
ei station with n-pentane. This material designated !61-12.
Run No. % S61-12 OF.
(min.-1 x -1)
0.837 300 4.96 x 10-5
T.AbBTE 13, GRAPHICS'Y DETERi'N!ET," :'TAl[ PR'YIERZATI:; RATE CONSTANT
F.R RUT.S .*ITH 500'F. AIR 5BL:T 5119 (R60-11), nTYFP.E 4b ViTEliE
Polymerizaticr, r'un Conditions: 'hese runs were charge with the percent
60-11 as indicate below, plus a quantity of Ct:rene approaching 10f,
with yylene comprising the difference.
Wur. lIurr.er RPC-r! F.
(min.-I x 1-1)
- 90 -
TAPE 14, GR4FHICAUY D.TFFURMIT:ED C'TERATL P.:ClTiERIATIC? RArTE Ci-:JTAf:T
FCR RUI'S WITH 'JP.RI.:US PLTR';EU MATERIALS
Pclymerisation FRn Conditions: These runs were charged with the material
as indicated below, plus a quantity nf Ftyrene aprr-aching 10..
Run Number Material
(S60-1 is petrolaturm)
ll7 300 :
('l17 is the East Texas
Asphalt Base Resitwum)
(S118 is the South Texas
Heavy Asphalt Base
(min.-l x f-1-
1.L-9 x 10-5
eP.6 x 10-5
5.8c x 10-5
(&)A dhinit& precipitation of polymer occurred on the interior of
the "u'toclave during this run.