Affinity of Hydrophilic Aggregate for Asphaltic ... - ACS Publications

the bottom up of porous oil mats; and the stripping of the bitumen in open-typepavements, which causes loss of the cohesion part of the stability of t...
0 downloads 0 Views 1MB Size
Affinity of Hydrophilic Aggregate HANS F. WINTERKORN University of Missouri and Missouri State Highway Department, Columbia,Mo.

MACHINE FOR MIXINGSOIL WITH BITUMINOUS MATERIALS

1. Those which show a greater natural affinity for bitumen than for water (hydrophobic). 2. Those which show a greater natural affinity for water than for bitumen, but whose affinitycan be beneficially (or detrimentally) altered by base exchange (hydrophilic). 3. Those, like quartz, which do not possess base-exchange capacity and which, therefore, cannot be altered by base exchange (hydro hilic). 4. Those wiich possess a chemical nature similar to quartz, but which, because of a porous structure, ermit the intrusion of asphaltic plugs and thus result in a type o f mechanical adhesion; or the pores of which can be filled by mixed hydroxides and oxides of higher valent metals (iron, aluminum, lead) resulting from the effect of the atmosphere on their soluble salts. Consequently, a treatment of this type of materials with metal salts may result in a partial change of their affinity for water and bitumen, respectively.

The action of bituminous materials varies to a great extent with their chemical composition. Although very little is known about the chemical composition of asphaltic bitumens, HE problem of the affinity of mineral aggregate for it appears to be established that best adhesion is obtained if bituminous coatings is foremost in the mind of many the asphalts possess a sufficient amount of active components highway engineers. Although numerous road failures, which are preferentially adsorbed by the mineral and, poswhich have been blamed on improper adhesion, were actually sibly, form water- and oil-insoluble compounds with comcaused by insufficient base support or by faulty design of the ponents of the mineral. These active compounds may be bituminous mix, a distinct field exists in which adhesion is of very stable-for example, the acids found in many natural primary importance in bituminous road construction. This asphalts. On the other hand, they may be unstable if profield includes the surface stripping, scaling, and pitting of duced by excessive heating. Compounds showing special relatively dense bituminous pavements; the stripping from affinity for the hydrophilic, quartzlike type of aggregate are the bottom up of porous oil mats; and the stripping of the usually of the latter type; although they exhibit desirable bitumen in open-type pavements, which causes loss of the adhesion properties, they usually possess an undesirable cohesion part of the stability of the pavement. A special weathering susceptibility and have, therefore, fallen into nuisance and a source of embarrassment is the stripping of disfavor in the conventional semipermeable types of bitumicoated aggregate between the time of mixing and that of nous road construction. total compaction on the road. The mix is most vulnerable, I n the light of present knowledge, it appears most desirable and financial and time loss is most apt to occur a t this time. to use a hydrophobic aggregate in connection with an asAn attempt is made in this paper to point out the factors phaltic material with good weathering resistance. If a hydrowhich are of greatest importance in the adhesion problem phobic aggregate is not economically available, an attempt and to propose a cure for at least some of the ills caused by should be made to reduce the detrimental effects of stripping improper affinity of the aggregate for the bitumen. by the adoption of a proper design, such as the use of a dense grading and of more viscous bituminous materials. The Theoretical Discussion utilization of a more active bitumen may also be considered; but if the greater activity has to be paid for with a decreased The theory of the adhesion interrelation of mineral surweathering resistance, the wisdom of such procedure appears faces and bituminous materials is well advanced. A recent doubtful. I n cases where the methods noted above are not survey (8) on the present status of the adhesiveness problem sufficiently effective, an attempt should be made to change in bituminous road construction demonstrated conclusively the surface character of the aggregate, and to choose the that the resistance to stripping of bitumen-coated aggregate bitumen on the basis of its weather-resistant properties. is dependent upon (a) the surface chemical character of the The main problem in the field of adhesiveness, then, is to mineral, (a) the chemical composition of the bitumen, and change the surface properties of the hydrophilic mineral (c) the time and condition of contact between bitumen and materials which are not susceptible to improvement by base mineral. exchange or which do not possess the advantages of porous According to their surface chemical character, minerals structure. may be divided into four classes: 1362

T

for Asphaltic Bitumen Use of Furfural and Its Resinous Derivatives for Improving Affinity Early Experiments Some time ago it was conceived that a change of the surface properties of aggregates from hydrophilic to hydrophobic could be achieved by synthesizing very thin films of artificial resins on the surfaces of the minerals (3). Experiments along this line, undertaken a t the author's laboratory and also by V. B. Saville (1) of the Missouri State Highway Department, showed conclusively that the surface properties of gravel can be permanently changed from hydrophilic to hydrophobic by such a procedure. Especially good results were obtained with a furfural-phenol type of .resinification product. For the reaction to give the best results with this type of material, somewhat elevated temperatures (up to 100" C.> are desirable. This feature of heating appears to make the procedure comparable to attempts to bake thin tar films on hydrophilic aggregates in order to change their sur-

SOIL-BITUMEN MIXTURE BEFORE COMPACTING

I

I

I

face properties. The main differences between the two methods are: (a) a lower temperature is needed to produce the artificial resin films. which are formed at a lower drvinp temperature than the maximum used with the higher t&e ovf asphaltic pavements; and (b) the aggregate is wetted better by the components of the artificial resins, and this makes for a better adhering and more uniform hydrophobic film. The main objection to the use of baked tar films appears to be the time loss incurred by heating the material well above the ordinary drying and mixing temperature of bitumen and aggregate, and cooling it to the desired working temperature. Resinous films, on the other hand, can be synthesized during the ordinary process of drying the aggregate. The successful changing of the hydrophilic properties of quartzite aggregate by resinous precoating was an inducement to try the effect of admixtures of resin-forming substances, such as phenolfurfural and aniline furfural, with bituminous materials prior to the coating of the aggregate. Interesting results were obtained from tests performed on Ottawa sand which was coated with different bituminous materials containing various percentages of admixed tar plus furfural, aniline plus furfural, phenol plus furfural, and similar materials. Certain percentages of some of the admixtures improve the adhesiveness of certain bituminous materials. Others decreased it or had no observable effect. It was apparent that in some cases the admixtures had reacted with the bitumen and thus changed its physical and chemical properties; in other cases the solvent action of the admixtures had changed the natural equilibrium of the bituminous materials. The results of these tests were described in detail in a recent publication (3). Because of the complexity of the bituminous systems, no definite rules can be set up a t present for the effect of such admixtures on bituminous materials. I n certain cases the admixtures are beneficial, in others they are detrimental. Therefore, a contemplated use on a larger scale should always be preceded by laboratory experiments to ascertain the compatibility of the admixture with the bitumen. The question remained as to whether the precoating of hydrophilic aggregate with furfural and its resinous derivatives would permanently increase its affinity for bitumen even in a cold mix. To clarify this question, a number of tests were performed on materials submitted by the Alabama State Highway Department and by the Eagle Picher Lead Company. The following paragraphs contain data of general importance in the adhesion problem, and data which bear directly on the improvement of adhesiveness by precoating with furfural and its resinous derivatives. The physical properties of the bituminous materials are given in Table I. The following mineral materials were used :

Alabama 1, Montgomery, Ala. (No. 1739); gravel and sand screened into five fractions: a. Retained on No. 2 sieve b. Passing No. 2 sieve, retained on No. 4 c. Passing No. 4 sieve, retained on No. 10 d. Passing No. 10 sieve, retained on No. 20 e. Passing No. 20 sieve Alabama 2, Selma, Ala.; gravel 3/s to 1 inch (Southern Sand and Gravel Company), coated with brownish impurities Alabama 3, Plantersville, Ala.; gravel 3/8 to 1 inch (American Sand and Gravel Company) Alabama 4, Tallassee, Ala.; gravel S/sto 1 inch (Southeastern Sand and Gravel Company), coated with brownish impurities Joplin chat (chert gangue of tristate zinc district); passing 1-inch sieve, retained on 8/4-inch 1363

INDUSTRIAL AND ENGINEERING CHEMISTRY

1364

VOL. 30, NO. 12

TABLEI. PHYSICAL PROPERTIES OF BITUMINOUS MATERIALS" Identification No. Type of bitumen

1 2 10 Medium- cur- Medium-cur- Slow-curing in# cutback ing cutback road oil

Saybolt-Furol viscosity, 8ec. Sp. gr., 6O0/?O0 F. Flash point, F. Oliensis test Residue,. % Penetration of residue at 77" F.c Ductility of residue at 77' F., om. LOBSon heating 50 rams 5 hr. at 325' F . , Vol. 7 distilled off at 68OooF.

k

0

b

11 Slow-curing road oil

13 Slow-curing road oil

17 Crude oil

16b 18b 21b Pa. asphalt Pa. asphalt Trinidad asphalt (penetra- (penetra(penetration tion 237C) tion 55C) 108c)

327 at 140" F . 0 9771 160 Nee.

141 at 140' F. 171 at 122O F. 327 at 122' F, 317 at 122' F. 33 at 122" F. 1.0285 1.0402 0.9766 0.9775 0 9286 170 240 240 260 190 Pos. Pos. Neg. Neg. Neg. ...... 71.8 65.1 73.2 49.8

85d

56d

150f

1504-

......

...... 21.5

......

86

93

101 9.92

1 :Oios 615 Neg

1 .'0349 620 Neg.

1.'ib 430 1

197

458

85 e

...

...

...

85

113

150f

150 f

77

118f

160 -I-

150 f

12.25

8.36

10.84

0.004

0 00

1.08

20.8

Tested according t o the specifications of the Missouri State Highway Commission. In actual use these asphalts were changed into naphtha cutbacks. 100 grams, 5 seconds.

The influence of the following variables on stripping resistance in the wash test was studied: (a) aggregate, ( b ) bitumen, (c) curing for different periods in air and water, respectively, (d) washing the aggregate with distilled water, hydrochloric acid, and aqua regia, respectively, (e) precoating the aggregate with furfural, aniline, phenol, and their resinous derivatives. WASHING. I n the washing with hydrochloric acid, the aggregate was left for 12 hours in dilute (5 per cent) acid; it was then repeatedly rinsed with distilled water until no chlorine ion was detectable. I n the washing with aqua regia, the samples remained from 5 to 6 hours in the acid; the acid was completely removed by rinsing with distilled water, and the gravel was dried on a steam-heated plate. For water washing, the samples were soaked overnight in distilled water and then thoroughly rinsed with fresh distilled water. COATING. The larger size gravel and chat was coated by dipping each individual piece into asphalt heated to 60" C. Samples of fractions b of Alabama 1 (passing screen No. 2 and retained on No. 4) were coated by dipping a thimbleful into asphalt heated to 60" C. Gravel retained on No. 10 screen was mixed with 1per cent by weight of asphalt (60" C.). Gravel retained on NO. 20 screen was mixed with 1.5 per cent by weight of asphalt (60' C.). Gravel and sand passing No. 20 screen were mixed with 2 per cent by weight of asphalt (60" C.). Asphalts 16, 18, and 21 were cut with 25 per cent by volume of naphtha to render them sufficiently fluid. The samples which were to be soaked in water were permitted to lose the larger part of the cutting liquid by overnight airing. Since differences in atmospheric conditions may influence the amount of cutting agent lost and also the resistance of the film to stripping, only such series of data were used as were run a t the same time. As priming materials, furfural, PRIMING OR PRECOATING. aniline, phenol, aniline-furfural (3 :2 mixture), and phenolfurfural (2:3 mixture) were employed. The aggregate was dipped into the priming liquid and allowed to stand 20 minutes in air before it was coated with asphalt. Tests were also made with gravel exposed 48 hours to the vapors of a mixture of aniline-furfural (3:2) as a priming treatment. TESTING.The washing machine used for the tests was described previously (3). It possesses the following features: A metal disk, ap roximately 10 inches in diameter and fitted with clamps to h o l i Erlenmeyer flasks or half-pint jars on both sides, is fastened to a shaft which runs through the center of the disk and holds it in the geometrical center of a water bath. The shaft rests in a cap on one side of the bath and penetratas the opposite side, where it is connected with an electric motor by means of a gear system and a belt. The gear system is adjusted to permit sixty-five rotations of the shaft per minute. Each bituminous sample is placed in a 125-cc. Erlenmeyer flask containing 50 cc. of water if the aggregate is fine-grained (filler,

Standard distillation. 0 After heating 5 hours at 325O f Residue of 114 penetration. d

F.

sand), or in half-pint jars if the aggregate consists of gravel and crushed stone. Each flask is fitted with a perforated rubber stopper in which a glass capillary has been placed to release the r s u r e caused by raising the bath temperature. Then the asks are fastened to the disk with their bottoms towards the latter, and the disk is rotated for 30 minutes at a bath temperature of 30" C. The bottles are taken out, examined, reinserted in the clamps of the disk for 30-minute tests at 4 5 O , 60°,and 70" C. The bath temperature can be kept constant either by relay thermostatic control or by a controlled influx of steam into the bath. Because of the churning action of the moving disk, no other stirring device is necessary.

At each inspection-i. e., for each temperature-the observation is listed as no stripping, slight stripping (less than 25 per cent uncoated), or bad stripping (more than 25 per cent uncoated). I n previous work the data were published in the form as listed but proved rather difficult to visualize. Therefore, the following scheme was adopted in the present paper. The numbers obtained and listed for each specimen during the testing from 30" to 70" C. are added. The resulting sum is subtracted from the highest possible sum, indicating worst measurable adhesiveness. The difference is divided by the difference of the greatest and smallest possible sum. The resulting quotient is multiplied by 100 and indicates the quality of stripping resistance on a scale from 0 (lowest measurable) to 100 (best measurable) adhesiveness. This procedure is explained by the following example: a.

b. c.

Test data Worst pogsible Best possible

Quality =

b --' X b -c

100

30°C. 1 3

-

1

45OC. 1 3 1

5 X 100 8

60°C. 2 3 1

70OC. 3 3 1

Sum 7 12 4

= 62.5

I n this form the results of the present experimental investigation are given in Tables I1 to VII.

Reproducibility of Data I n accordance with the method by which they are obtained, the data are of a semiquantitative nature. They are completely reproducible within the defined limits, if the method of coating is kept the same and if the aggregate has been exposed to the same influences. With the same aggregate, variations in adhesiveness have a t times been observed, if exposed to a very humid or very dry atmosphere before coating. I n the present case the adopted precaution of drying the aggregate on a steam-heated plate before it was coated or primed ensured uniformity of the results. The absolute data in themselves are not so important as the functional interrelations of these data. The data in Tables I1 and I11 further substantiate the facts that the resistance to stripping of bitumen-coated

DECEMBER, 1938

INDUSTRIAL AND ENGINEERING CHEMISTRY

-

TABLE11. EFFECTOF CURINGON TESTRESULTS Curing Gravel No. Bitumen

NO.

Av

.

-

42.2

General av.

I n Air 2 Weeks

I n Air 1 Week 2

---.

3

4

2

I n Water 1 Week

-. 2

4

3

3

Av.

4

Rating

50

19

56.2

-

54.7

37

37.5

49

9.4

34.4

0

33.2

I _ _

14.6

TABLE111. RATINGOF BITUMINOUS MATERIALS I n Air 2 1 21 11 62.5 58.3 54 54 0 1 2 10 16 11 79.2 79.2 70.9 41.7 37.5 I n Water 10 2 1 16 21 11 Bitumen No. 33.3 25 21 21 16.6 0 Rating after 1 week Av. Rating for All Methods of Curing 10 2 21 1 16 11 Bitumen No. Rating 55.5 54.2 52.7 51.4 43.2 12.5 --Av. Effect of Curing 7-Av. Effect of Curing I n water ---In Ak-Gravel 3 Time 1 week 1 week 2 weeks Rating 46 Rating 14.6 37 49

Bitumen No. Rating after 1 week Bitumen No. Rating after 2 weeks

16 66.7 21 87.5

10

-

13 0 13 0

17 0 17 0

13 0

17 0

13 17 0 0 Gravel2 4 36 19

1365

the materials reported in Table IV, washing with hydrochloric acid resulted in moderate permanent improvement, whereas washing with aqua regia resulted in great permanent improvement of the stripping resistance. 5. The adhesive quality of the bituminous materials decreases in the series 2 > 10 > 21 > 16, in accordance with the behavior of these materials with gravels 2, 3, and 4, reported in Tables 11, 111, and V. The data in Table V conform, in general, with the conclusions from the preceding ones. They show again that, although washing with hydrochloric acid and aqua regia is, in general, beneficial, the specific effect depends upon the gravel, the bitumen, and the history of the bitumen-coated mineral before the test. It is worthy of emphasis that, with previous soaking, bitumen No. 21 is the most satisfactory of all the bitumens used with unwashed gravel No. 4, and is only half as satisfactory with the same gravel if washed with either hydrochloric acid or aqua regia. This is in contrast to the general trend of the effect of acid washing and

aggregate depends upon the properties of the bitumen, the properties of the aggregate, and the time and condition of contact between bitumen and aggregate. Although all the gravels are silicic, gravels 2 and 3 show much more affinity for bitumen than gravel 4. The data in Table IV show the following interesting facts: 1. The increase of the stripping resistance with the time of contact of bitumen and mineral in a dry atmosphere is, at least partially, reversible, as shown by the tests on air-cured specimens which were subsequently submerged in water, before being tested. 2. The different sizes of material coming from the same gravel pit show similar adhesiveness data if coated in the unwashed and air-cured state before being tested. If they are soaked in water before being tested, differences in the materials are noticeable in the test data. The reality of these differences is substantiated by similar ones occurring if the gravel portions have been washed with hydrochloric acid and with aqua regia, respectively. 3. Combined soaking and washing represents a more severe but also a more discriminating test procedure. 4. The effect of washing the aggregate with acids before it is coated with bitumen depends upon the character of the aggregate, the type of the acid, and the type of bitumen employed. With

TABLE IV. WashineGravel 1 Passing sieve No. Retained on sieve No. Bitumen No. 2 10 16 21 Av.

a

-

2

2

4

50 25 37.5 50 40.7

50 25 25 50 37.5

e 20

-

37.5 87.5 87.5 25 25 37.5 25 12.5 12.5 50 25 25 34.4 37.5 40.7 J

General av.

Av. General av.

37.5 37.5 0 0

18.7

100 25

2

50 50

5

.

-With a

HCI b c d e 2 4 10 20 2 4 10 20 Rating after 2 Weeks in Air 87.5 100 50 50 87.5 100 100 50 50 50 37.5 100 25 25 37.5 25 100 37.5 3 7 . 5 62.5 62.5 100 40.7 4 0 . 7 6 9 . 4

.

d

7 -

a 2 100 100 37.5 100 84.4

60.6

25 37.5 0

Rating after 3 Weeks i n Air and 1 Week in Water 37.5 37.5 62.5 37.5 25 37.5 50 50 62.5 0 25 100 50 62.5 0 25 50 0 0 12.5 37.5 0 12.5 100 50 75 50 37.5 28 28 53 28 3 1 . 3 37.5 65.6

50

28 33.1

75 62.5 0 25 40.6

35.6

With Actua Reaiab C d 2 4 10 4 10 20 100 100 100 100 100

100 50 50 50 62.5

e 20

.

100 25 0 25 37.5

100 25 25 62.5 53.3

100 62.5 0 50 53

62.5 25 0 12.5 25

Average 79 52.5 36.7 53.3

.

38 r

2 10 16 21

IN

EFFECTOF WASHINGGRAVEL 1 BEFORE COATING

None c d 4 10 10 20

b

ADDINGCHEMICAL ACTIVATORS TO OIL SUPPLYTANK

67.5 87.5 100 62.5 100 87.5

100 87.5 62.5 87.5 84.4 63

58.4 52.5 18.6 46.7 44.0 44.0

INDUSTRIAL AND ENGINEERING CHEMISTRY

1366

VOL. 30, NO. 12

Effect of Pracoating the Aggregates .

also of the average effectiveness of the bituminous materials. This behavior again seems to emphasize that, although general trendscan be formulated in the stripping phenomenon, each aggregate and each bituminous material representsa specific problem.

~~

The Of the Priming experiments on grave' are shown in Table VI. Although bitumen No. 13 resisted improvement by any treatment, the effect of the primers on the stripping resistance of the other bitumens was remarkable. ANILINE. Aired Specimens. Aniline decreased TABLE V. EFFECTOF WASHINQ OF GRAVELS 2, 3, AND 4 BEFORE cO&TlNG the stripping resistance somewhat in three out Weahin& . None-With HCI-With Aqua Regisof nine bitumen-gravel combinations and inC.reve1 0 2 3 4 2 3 4 2 3 ~ A I . creased this resistance in four combinations. Bitamen NO. Rating sflei 2 Weeks in Air Soaked Spe&mens. The aniline was detri2 75 100 5 2 . 5 37.5 inn io0 87.5 87.6 75 70.8 mental in three and beneficial in five out of nine 10 75 1W 50 62.5 87.5 50 100 37.5 100 69.5 50 25 37.5 37.5 87.5 37.5 6 2 . 5 16 50 50 48.6 combinations. 3 7 . 5 87.5 6 2 . 5 75 1W 6 2 . 5 25 25 55.5 21 100 ANILINE-FURFURAL (VAPOR). The amount of AV. 86.6 81.3 52.5 37.5 59.4 90.6 59.4 81.2 62.6 61.0 aniline-furfural vapor absorbed by the gravel Qsneral LIY. 69'. 8 $2 68 was so small that hardly any discoloration of ----Rating after 3 Weeks in Air and 1 Week in Waterthe aggregate was observable. The aniline-fur50 25 36 25 87.5 n 87.5 25 f u r d resin itself has a dark red color. 25 0 2 62.5 75 50 2s 12.5 o 25 5n 33.4 u in Aired Specimens. The priming resulted in u 75 n 26.4 u 87.5 12.5 37.5 75 o 16 n 25 n 25 25 is9 0 60 0 o 21 an adhesion rating of 100 in seven and a slight 5.3 3 1 28 1 31 3 22 53 22 55 3 25 27 AV. decrease in two out of nine cases.

--

_I

\__I

35

12

TIBLEVI. Bitumen No. O r a d No. Primer

Nolle AdIirzS Aniliobfvrlural (vapor) Phenolfurfural Aniuobfurfure.1 (dip1

F"d"d

Nons Anifin* Ahnilinbf.fudurel (vapor) Phsoal-furfural Aniline-furfur4 (dip)

Furfural

-~

34

EIFECT OF PRIMlNO THE AGOKEGATE BEFORE COATWQ WlTH BITVXEN

2

1 3

4

62.5

75

25

an iw IW

100

loo 0

12.5

25 50 87.6

ion

iw

Inn

IM

ion iw 62.5 12 5

inn 100 62.5 IOU

-13-, 3 4 2 3 Rating after 1 Week in Aii

- 1

iw mo iuo 1uu 100

u o

12.6 87.6 37.5

1W

2

87.5

87.5

100

100 100 100

12.5

ino

12.6 100

roo

37.5

inn 82.5 iw

100

Ion

inn

Ion

iuo

n

75

12.5

100

o 0 n oo n

n 0

u u 12.5 (I

Rating after 1 Week in Water 87.5 0 0 0

37.5

12.5 100

inn

Ipo

87.5

25 12.5 100

100

0

0

0 0

0

n

0

0

n

n

- 1

4

2

3

4

0

60

loo

50

n n 12.5 o

o

0 0

100 100 100 100

IOU 0

37.5

0 0

100 100

0

1w

n

inn

ivn

87.5 100 100

loo

52.5 50

75 100 100

100

,

37.5 37.5

100 100

87.5 0

37.6

100 50 87.5

1w

WASHINGMACHINEPOR DETERMIN~NG RESISTANCE OF BlTVXINoUS COATINGS TO %RlPPING

Aversus 45 59.4

68.7 08.7

70

74

18

31 38.5 58.3 50.5 70

DECEMBER, 1938

INDUSTRIAL AND ENGINEERING CHEMISTRY

1367

The adhesion of bitumen 18 was increased somewhat by washing the chat with aqua regia Curing 1Day in Air--2 Days in Wateras compared with that to the untreated chat, Bitumen No. 10 18 21 10 18 21 and was raised to optimum by all other treatTreatment Rating . Average ments. The adhesion' of bitumen 21 to the None 100 75 50 100 50 62.5 73 100 100 Washed with distd. water 100 75 100 62.5 90 chat was raised by all treatments; aniline-fur87.5 75 100 37.5 100 Washed with aqua regia 75 79 fural and phenol-furfural precoating brought the 100 100 62.5 100 37.5 81 Primed with lead acetate 87.5 100 100 100 87.5 92 Primed with furfural 75 87.5 stripping resistance up to optimum rating. 100 100 Primed with aniline 75 100 75 90 87.5 100 100 50 75 Primed with phenol 50 79 100 Soaked Specimens. Bitumen 10 showed opti100 100 100 100 Primed with aniline-furfural 100 94 62.5 mum rating in all cases except in that where phenol 100 100 62.5 100 Primed with phenol-furfural 100 94 100 was used as a primer. 5 Aired 5 hours after coating was applied. The stripping resistance of bitumen 18 was increased somewhat by washing with distilled water and by priming with phenol-furfural, was increased more by priming with aniline and with furfural, Soaked Specimens. The priming was beneficial in seven, respectively, and reached optimum rating by priming with detrimental in one, and neutral in one out of nine cases. an aniline-furfural mixture. The stripping resistance is d e PHENOL-FURFURAL. Aired Specimens. The priming was creased by washing the chat with aqua regia and by priming beneficial in seven (six improved to optimum rating) and it with lead acetate before coating. The stripping resistance neutral in two cases which already showed optimum adhesion. of bitumen 21 was improved in all cases but one; optimum Soaked Specimens. The priming was beneficial in all nine adhesion was achieved by thoroughly washing the aggregate cases, raising five to optimum adhesion. with distilled water and also by priming it with phenol or ANILINE-FURFURAL(DIP). Aired Specimens. Priming with a phenol-furfural mixture. resulted in optimum adhesion in all nine cases; this rating Averaging the different bituminous materials and the was obtained only by two of the unprimed specimens. two methods of curing, priming treatments with anilineSoaked Specimens. Priming was beneficial in eight and furfural, phenol-furfural, and furfural alone rate highest; neutral in one out of nine cases, raising five to optimum no treatment shows lowest rating. The high rating of washadhesiveness. ing with distilled water is remarkable. But this washing FURFURAL. Aired Specimens. Priming was beneficial will cost more than the priming of the natural chat with in seven and neutral in two out of nine cases. Of the primed furfural or its resinous derivatives. specimens, eight out of nine showed optimum adhesiveness. The data reported indicate that in many cases the adSoaked Specimens. Priming was beneficial in all nine hesiveness of bitumen to siliceous minerals can be improved cases, raising the rating to 100 in eight out of nine cases. by precoating the aggregate with furfural and with its resinThe results obtained with these primers and especially ous derivatives. Since only small amounts of the priming with furfural and its resinous derivatives are quite remarkable. substances are necessary to produce a marked change in the The fact that only very thin (theoretically one-molecular) surface properties of the mineral, the cost of about 10 cents films are necessary to produce a remarkable change in adper pound of these materials should not be prohibitive. hesiveness puts such primers in the range of economic applicaBefore any money is spent for a primer in a specific case, the tion. The added cost in bituminous construction can be value of the primer should be tested by laboratory experimade a nominal one if the right types of coating methods are ments. The time of soaking desirable as a pretreatment used. It should be further remarked that the thinner the before tests are made in the washing machine depends upon film of priming substance, the better is its effect. To inthe climate and topographic conditions to which the pavevestigate the effectiveness of these primers further, tests ment will be exposed and will have to be determined experiwere made with samples of Joplin chat, a material which is mentally. Saville (I) showed that foF Missouri conditions, commercially available in large quantities. 40-hour soaking resulted in the best correlation of test results TESTSON JOPLIN CHAT. Table VII, which contains the with actual behavior on the road. data on Joplin chat, shows the following facts: In the present investigation tests were performed on samAired Specimens. Bitumen 10 showed optimum adples which had riot been soaked a t all and on samples soaked hesion i o the chat, which was not changed by any treatment. TABLE VII.

ADHESIVENESS TESTSON JOPLINCHAT^

7

TESTSON BITUMEN-TREATED SOILSPECIMENS TAKENFROM EXPERIMENTAL PROJECT TABLE VIII. RESULTSOF WATERIMMERSION ON ROUTE100, FRANKLIN COUNTY,Mo., AND REMOLDED IN LABORATORY

-

7 Activator % Water Intake VOl. -Station% of 1 day 8 days From To Bitumen" bitumen Material Ratio a b a 6 782 40 785 30 Slow-curing road oil (SC-3) 3.0 Aniline-furfural 3:2 8 15 24 25 785 30 795 f 0 3 4.5 Aniline-furfural 3:2 18 15 25 23 796 03 806 00 3.6 High-temp tar .. 31 27 31 30 806 00 815 60 3.0 High-temp: tar-furfural 3:2 1; 9 25 25 60 824 24 3.8 Low-temp. tar-furfural 3 :1 6 22 23 815 825 65 832 28 3 .O Aniline-furfural 3:2 8 15 24 25 832 2 8 841 -I- 94 1.5 Aniline-furfural 3:2 15 19 24 26 841 94 852 38 Medium-curing cutback (MC-4) Samples swell and soften; cannot be handled 852 38 858 60 Emulsion A Samples swell and soften. cannot be handled 60 866 35 Emulsion B Samples do not swell but'crumble when handled 858 Samples do not swell but crumble when handled 866 35 871 95 Emulsion C 871 95 886 90 Medium-curing tar (TM-2) Samples break when handled 90 891 22 Medium-curing tar (TM-2) Samplerr break when handled 886 891 22 903 35 Medium-curing cutback (MC-3) Samples break when handled: brown ooze 903 35 908 28 Slow-owing road ojl (SC-3) Samples break when handled 2 8 913 72 Slow-curing road 011 (SC-4) Samples soften and slake' brown ooze 908 913 72 921 22 Slow-curing road oil (SC-4) Samples soften and slake: brown ooze 22 927 4 5 Slow-anring road oil (SC-3) Samples break when handled: some swelling 921

+ + + + + + + + + + + + + + + + + +

5

+ + + + + + + + + + + + + + + +

8 per cent by weight on basis of material passing No. 40 sieve.

.

b

by Weight after: ,16 days 36 days a b a b 27 28 27 25 28 26 . . 34 37 28 27 28 29 26 25 26 27 27 2 8 27 28 2f 29

-

Penetration after 36 Daysb 130 82 33 35

...

106 80 103 163

...

201

76

82 114

100 grams, 10 seconds. I

1368

INDUSTRIAL AND ENGINEERING CHEMISTRY

for different periods of time. I n the case of the primed gravel, soaking for one week was considered to represent a more severe procedure than that of Saville. The shorter airing and soaking times allowed for the Joplin chat were chosen because the greatest danger of stripping exists for this aggregate during the limited time of mixing and compacting on the road. Again it deserves emphasis that the primary interest was in the functional relation of the data rather than in their absolute values. As stated before, different standards of adhesiveness have to be set up for different locations according to climatic and topographic conditions.

Furfural in Bituminous Soil Stabilization Experiments in this laboratory have demonstrated that soils can be stabilized with furfural resins without added bituminous materials. But such a procedure is a t present uneconomical. It is economically possible to use small percentages of furfural and resinifying mixtures of furfural with other chemicals as activators in bituminous soil stabilization. The purpose of these activators is a double onenamely, to help in an easier mixing and dispersion of the bitumen in the soil material and to improve the water and weather resistance of the resulting soil-bitumen mixtures. After promising laboratory experimentation, test sections were built in 1937 on Route 100, Franklin County, Mo.; a slow-curing road oil (SC-3) was used as the basis in each section : 1. 33 gallons aniline, 21 gallons furfural, 1805 gallons oil 2. 165 gallons aniline, 109 gallons furfural, 6058 gallons oil 3. 74 gallons furfural, 112 gallons tar, 6197 gallons oil 4. 51 gallons furfural, 165 gallons of low-temperature tar, 5693 gallons oil 5. 79 gallons aniline, 53 gallons furfural, 4392 gallons oil 6 . 58 gallons aniline, 38 gallons furfural, 6414 gallons oil The activators were added to the supply tanks while they were being filled with the road oil. No additional mixing was necessary to secure the dispersion through the oil. While the road has not been in service long enough to show any great differences in permanent stabilization effects due to the activators, the use of the latter resulted in an easier mixing of the bitumen with the soil and in a much more thorough dispersion of the oil throughout the soil, than was obtained without the activators. Also, soil samples taken from the completed road and tested in the laboratory showed the distinct beneficial effect of the activators. The results of the tests on these stabilized soils are shown in Table VIII. The samples were taken out of the different sections and liberated by hand-picking of larger and smaller pieces of aggregate. The bitumen-treated soil was then dried, pulverized to pass the 40-mesh sieve, and, after addition of water, molded into circular specimens, 1 inch thick and 1.5 inches in diameter. The specimens were slowly dried in air in order to prevent crack formation until they had obtained constant weight. Then they were entirely immersed in distilled water. The absorption of water was determined, where possible, after 1, 8, 16, and 36 days, respectively. After 36-day immersion, the specimens were also tested for softening with the Humboldt penetration needle (100 grams, 10 seconds). The molded specimens prepared in this way from reclaimed stabilized soil possess a great inherent porosity. Therefore, the penetration data appear to possess more weight than the data on water intake.

Summary The influence of different pretreatments on the stripping resistance of bitumen-coated aggregate was determined with

VOL. 30, NO. 12

the use of a wash test. The different pretreatments included washing of the aggregate (before coating) with distilled water, aqueous hydrochloric acid solution, and aqua regia, respectively, and priming with aniline, phenol, furfural, aniline-furfural, phenol-furfural, and lead acetate, respectively. In certain cases the wash test was preceded by curing periods in air or in water. The data obtained brought out that: 1. The stripping resistance of bitumen-coated aggregate depends upon the properties of the bitumen, the surface properties of the mineral, and the time and condition of contact between bitumen and mineral surface. 2. Soaking followed by the wash test is a more time-consuming but also a more discriminating procedure. 3. The effect of washing the aggregate with hydrochloric acid and aqua regia before coating depends upon the character of the mineral surface and that of the bitumen; but washing with aqua regia tends t o improve the adhesiveness more than washing with hydrochloric acid. 4. With three gravels from Alabama the average influence of priming with furfural and its derivatives on stripping resistance can be seen from the following average rating: Treatment None Aniline Aniline-furfural vapor Phenol-furfural Aniline-furfural Furf ural

1 Week in Air 46 59.4 68.7 68.7

76 74

1 Week in Water 19 31 38.6 58.3 60.5 70

5. With Joplin chat the average effect of washing or priming on the stripping resistance is indicated in the following rating: Treatment None Washing with aqua regia Priming with henol Priming with g a d acetate Washing with distilled water Priming with aniline Priming with furfural Priming with aniline-furfural Priming with phenol-furfural

Av. Rating 73 79 79 81 90 90 92 94 94

Samples taken from a bituminous soil stabilization project, in some sections of which resin-forming primers were used, were tested in the laboratory. It was found that specimens containing aniline-furfural, high-temperature tar-furfural, and low-temperature tar-furfural admixtures showed better resistance to water than samples which did not contain these materials. The evidence appears to show that furfural and its resinous derivatives might be destined to fill an important place in bituminous road construction. While these materials cannot be expected to cure all the evils for which lack of adhesiveness is often blamed, they deserve careful consideration in cases where stripping is encountered. Laboratory testing of the effectiveness of furfural and other priming substances should always precede any application on the road.

Acknowledgment Credit is due the writer’s assistants, G. W. Eckert, E. G . Powell, and E. B. Shipley, for performing the laboratory tests; and F. V. Reagel and R. C. Schappler of the Missouri State Highway Department, and especially E. F. Kelley of the United States Bureau of Public Roads for kind criticism and permission to publish this investigation.

Literature Cited (1) Saville, V. B., and Axon, E. O., Proc. Asphalt Paving Tech., 9,86-101 (1937). ( 2 ) Winterkorn, H. F., Proc. Montana Natl. Biturnin. Conf., 3 , 190214 (1937). (3) Winterkorn, H.F.,and co-workers, Proc. Asphalt Paving Tech., 9,63-85 (1937).

RECEIVED June 13, 1938.