RESEARCH REPORT 122 THE EFFECT OF EVA-MODIFIED BITUMENS ON ROLLED ASPHALTS CONTAINING DIFFERENT FINE AGGREGATES

Transcription

1 TRANSPORT AND ROAD RSARCH LABORATORY Department of Transport RSARCH RPORT 122 TH FFCT OF VA-MODIFID BITUMNS ON ROLLD ASPHALTS CONTAINING DIFFRNT FIN AGGRGATS by J Carswell The views expressed in this report are not necessarily those of the Department of Transport Pavement Materials and Construction Division Highways Group Transport and Road Research Laboratory Crowthorne, Berkshire, RG11 6AU 1987 ISSN

2 Ownership of the Transport Research Laboratory was transferred from the Department of Transport to a subsidiary of the Transport Research Foundation on I st April This report has been reproduced by permission of the Controller of HMSO. xtracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged.

3 CONTNTS Page Abstract 1 1. Introduction 1 2. Binders Penetration Softening point Viscosity (45 C) Viscosity ( C) 2 3. Discussion of rheological changes 2 4. Aggregates 3 5. Properties of asphalt mixes Marshall design Deformation resistance 9 6. Relationships between different properties of materials Conclusions Acknowledgements References 15 CROWN COPYRIGHT 1987 xtracts from the text may be reproduced, except for commercial purposes, provided the source is acknowledged

4 TH FFCT OF VA-MODIFID BITUMNS ON ROLLD ASPHALTS CONTAINING DIFFRNT FIN AGGRGATS ABSTRACT The changes in the rheological properties of petroleum bitumens after the addition of ethylenevinyl acetate co-polymer (VA) are described and the consequent effect on their use in rolled asphalt wearing course mixtures is reported. In particular, the changes in the Marshall stability and deformation resistance using four asphalt sands of different stability are quantified. The relationship between these properties and the effect of the VA component is discussed in detail, and equations are derived relating the Marshall stability of a conventional rolled asphalt wearing course to the Marshall stability of a rolled asphalt containing bitumen modified with 5 per cent VA for the same resistance to permanent deformation. 1 INTRODUCTION The use of bitumens modified with an ethylene-vinyl acetate co-polymer (VA) is now fairly widespread in the UK particularly in weather conditions likely to affect adversely the laying of conventional wearing courses or where more deformation-resistant surfacings are required. Denning and Carswell (1981) have detailed the effect of a particular grade of VA on the physical properties of the bitumen and on the rolled asphalt mixtures made with this binder. This work has led to a draft specification being introduced to cover the use of this particular grade (18_+2 per cent vinyl acetate, melt flow index 150_+40) in rolled asphalt wearing courses. The principal effect of adding 5 per cent of this VA to a petroleum bitumen is to increase the resistance to permanent deformation of a rolled asphalt wearing course material to a far greater degree than would be expected from the measured change in Marshall stability. This finding raises the question as to whether low stability fine aggregates, at present unsuitable for surfacing heavily-trafficked sites, may be used with an VA modified bitumen to achieve a performance similar to that of conventional asphalts made with higher stability fine aggregates. As a result of this research the option would be available to use either a locally-available, lower-stability sand with a more-expensive modified binder or a higher stability sand, with high haulage costs, with a conventional binder. This report quantifies the effect of two petroleum bitumens modified with ethylene-vinyl acetate copolymer on rolled asphalt wearing course mixtures containing four different fine aggregates. The fine aggregates were chosen to represent a wide range of stability characteristics. 2 BINDRS The following binders were used in this project: a 50 pen bitumen, a bitumen, the bitumen plus 2.5 per cent and 5.0 per cent VA (18 per cent vinyl acetate, 150 melt flow index) and the bitumen plus 5.0 per cent of the same VA. The bitumen/va binders were blended using a low shear mixer for a period of approximately 2 hours at a temperature of 160 C. Rheological tests were carried out on the 5 binders. 2.1 PNTRATION The penetration of each binder at 25 C was determined according to the Institute of Petroleum method 49/1984 (Institute of Petroleum 1984a). The results are given in Table 1 and show that the addition of VA to a bitumen reduces the penetration by about one grade. 2.2 SOFTNING POINT Ring and ball softening point measurements were carried out according to the Institute of Petroleum method 58/1984, (Institute of Petroleum (1984b)) and the results are tabulated in Table 1. The increase in softening point is marked, being 5 C for the 2.5 per cent addition, 12 C for the 5.0 per cent addition to the bitumen and 15 C for this level of addition to the bitumen. It appears from this limited information that the change in softening point with the addition of VA may be linear, particularly for a given bitumen and polymer. 2.3 VISCOSITY (45 C) The viscosity of each binder at 45 C was measured using the Sliding-Plate Rheometer (Denning and Carswell 1981). The values of viscosity at a shear rate of 0.05s -1 and at a strain of 1 are shown in Table 1. Previous work (Denning and Carswell 1981) has shown a good correlation between this measure of viscosity and the wheel-tracking rate of a rolled asphalt (with 30 per cent coarse aggregate) made with the same binder. Table 1 also shows the nominal elastic recovery of each binder measured with the same rheometer, determined after a period equal to the loading time; it can be seen that the VA confers a degree of elasticity to the binder, in proportion to the polymer content, indicating that the

5 binder is less Newtonian in character than a bitumen at this temperature. 2.4 VISCOSITY ( C) Measurements of viscosity in the range 100 to 200 C were made on each binder using a Haake viscobalance, which utilizes the principle of a falling sphere in a fluid. This temperature range is useful as it corresponds to the mixing, laying and rolling temperatures of rolled asphalt wearing course materials. The viscosity curves for each binder are shown in Figure 1 and equiviscous temperatures, relating to mixing and rolling temperatures, are O t~ v d BINDR o Bitumen o Bitumen per cent VA Bitumen + 5 per cent VA Bitumen Bitumen + 5 per cent VA I I I I I I I I I I I Temperature ( C) Fig. 1 Relationship between HAAK viscosity and temperature shown in Table 1. The temperature values shown for 300 poise are extrapolated from the experimental results and should be treated with caution. The effect of the VA co-polymer is to render the binder more viscous, but not markedly so, and is in keeping with the increased softening point values. However, previous work has shown that the VAmodified bitumens do not require higher mixing and laying temperatures in commercial practice, possibly because the VA confers a degree of shear susceptibility to the binder as it departs from Newtonian behaviour. 3 DISCUSSION OF RHOLOGICAL CHANGS For a 2.5 per cent addition of VA to a bitumen the high temperature viscosity characteristics remain relatively unchanged but, over the typical temperature range experienced on the road, the binder is generally stiffer and slightly more elastic. At 45 C it is nearly twice as viscous as the bitumen, indicating an increased resistance to permanent deformation when used as a binder in a wearing course material. The effects of adding 5 per cent VA to the and the bitumens are virtually the same; the penetration is reduced by about one grade and there is a substantial increase in the ring and ball softening point. At 45 C, the elasticity of the polymer modified binders were an order of magnitude greater than the respective base bitumens and their viscosities were also increased by a factor of 4 to 5. This stiffening effect was observed in the normal mixing and laying temperature range but not to the extent that modifications to existing commercial practice were required. TABL 1 Rheological properties of the binders Binder Penetration (0.1 mm) RSB Softening Pt. ( C) Apparent viscosity (45 ) Strain=l, Shear rate = 0.05s -1 (poise) lastic recovery (per cent) quiviscous mixing and rolling temperatures ( C) 2 poise 10 poise 50 poise 300 poise per cent VA +5 per cent VA + 5 per cent VA s s s * 84* 8* 75* 78* *extrapolated values

6 D m m m Coarse aggregate (granite) TABL 2 Properties of aggregates and filler Fine aggregates Southport Bristol Beach Channel Whitegate AImington Filler limestone Sieve size per cent (by mass passing) 14 mm 10 mm 3.3 mm 3.35 mm 2.36 mm 300/~m ]00/~m ;)12/~m 75/~m 100 5O O O Specific gravity (Mg/m 3) Bulk density (Mg/m 3) The bitumen modified with 5 per cent VA showed a similar viscosity-temperature relationship over the mixing and laying range to that of a petroleum bitumen but had a higher softening point and an increased viscosity at 45 C. This suggests that in commercial production an asphalt containing this binder would provide a more deformation resistant surfacing without involving additional energy costs. Denning and Carswell (1981) have shown that asphalt surfacings containing VA modified bitumens may show better workability, with adequate compaction being achieved below the rolling temperature, equivalent to 50 poise, as measured by the Haake viscobalance. At present, the reasons for this are not clear and may be due either to the non- Newtonian behaviour of the binder at these temperatures or to physical-chemical reactions within the VA itself. This effect, if reproduced generally, should allow more time for compaction of an asphalt, thereby either extending the working season, or allowing more successful laying in adverse weather conditions. The problems associated with adverse weather working are the subject of a report by Daines (1985). Further work is needed to quantify the effect of the VA co-polymer on the compaction of asphalt in the temperature range C. 4 AGGRGATS All wearing course mixtures were made with granite coarse aggregate (Croft) and limestone filler (Flowers). The coarse aggregate and the filler gradings, together with their specific gravities are shown in Table 2. The variables in the mixtures were the type of binder and the type of fine aggregate. Four fine aggregates were used ranging from rounded to angular shaped particles, yielding a wide range of Marshall stability values. The fine aggregates were all sands and their gradings, specific gravities and bulk densities are recorded in Table 2. The Southport Beach and Bristol Channel fine aggregates are composed mainly of rounded particles, the difference between them being that the Southport material is more single-sized. The Whitegate fine aggregate was chosen to represent an average stability sand and the Almington fine aggregate for producing high Marshall stability asphalt wearing course mixes. 5 PROPRTIS OF ASPHALT MIXS 5.1 MARSHALL DSIGN The optimum binder content for each binder for each fine aggregate was determined using the procedure in British Standard BS 594 (1973) (British Standards Institution 1973) except that 30 per cent coarse aggregate was also included in the test specimens. The values of mix density, compacted aggregate density, Marshall stability and percentage voids are plotted against binder content for each fine aggregate and binder in Figures 2 to 5.

7 BINDR [] Bitumen O Bitumen per cent VA Bitumen + 5 per cent VA ix Bitumen Bitumen + 5 per cent VA ~-" ~ 2.24 g "o 2.23 x j/ " m == 2,22 o_ o I I I I I A z v 5.0 Q. o3 4.0 X. 5.0 o > I I I Fig. 2 Marshall Design on rolled asphalt (30 per cent stone content) with Southport Beach sand

8 BINDR D Bitumen O Bitumen per cent VA Bitumen + 5 per cent VA z~ Bitumen Bitumen + 5 per cent VA & ~ x \ o~ Q I I I I I I ' -- - A z +~ > U3 4.0 Q. x ~= O > I I I Fig. 3 Marshall Design on rolled asphalt (30 per cent stone content) with Bristol Channel sand

9 BINDR O Bitumen O Bitumen per cent VA Bitumen + 5 per cent VA z~ Bitumen Bitumen + 5 per cent VA ~" ~ (3 v.x j O ~J I I I 1.96 I I I m A B X _~ =.. > Fig I I I Marshall Design on rolled asphalt (30 per cent stone content) with Whitegate sand

10 BINDR n Bitumen O Bitumen per cent VA Bitumen + 5 per cent VA z~ Bitumen Bitumen + 5 per cent VA ~ ~ 2.oa m ~ 2.24 X (~ o I I I I I I I X 7.0.~ \ 5.0 I I I I I I Fig. 5 Marshall Design on rolled asphalt (30 per cent stone content) with Almington sand

11 TABL 3 Marshall properties at optimum binder content for each binder and for each fine aggregate MATRIAL PROPRTIS AT OPTIMUM Compacted Marshall Optimum binder aggregate Marshall content Mix density density Stability Flow Quotient (per cent by mass) (Mg/m 3) (Mg/m 3) (kn) (ram) (kn/mm) SOUTHPORT BACH FIN AGGRGAT Voids in mix (per cent) +2.5 per cent VA + 5 per cent VA + 5 per cent VA BRISTOL CHANNL FIN AGGRGAT +2.5 per cent VA +5 per cent VA +5 per cent VA WHITGAT FIN AGGRGAT +2.5 per cent VA + 5 per cent VA +5 per cent VA ALMINGTON FIN AGGRGAT per +5 per cent VA + 5 per cent VA Table 3 shows the properties of each mix at optimum binder contents; these were calculated in the normal way as the mean of the binder contents of the maxima found for mix density, for compacted aggregate density and for Marshall stability. The results showed that for each fine aggregate the type of binder made little difference to the optimum binder content, to the mix and compacted aggregate densities or to the void content of the mix. The only differences found were in the values of Marshall stabilities where the VA modified bitumens gave higher results. For the bitumen plus 2.5 per cent VA the increase in Marshall stability was just over 10 per cent for the Whitegate and Almington sands, and over 20 per cent for the Southport Beach and Bristol Channel sands. Marshall quotient values were also higher because the flow values remained relatively unchanged, except in the case of Whitegate sand where an unusually high flow value resulted in a slightly lower Marshall quotient. For the two grades of bitumen modified with 5 per cent VA the percentage increases in Marshall stability were of the order of per cent for the lower stability sands, but only about 10 per cent for the Almington sand, probably because the angularshaped particles of this sand have an over-riding effect. A lower than expected increase in Marshall stability was recorded for the bitumen plus 5.0 per cent VA when using the Southport Beach sand; the reason for this is not clear. One point of particular interest from the Marshall design studies was that the optimum binder contents were about 1 per cent lower for the lower stability sands than for the higher stability materials; when

12 TABL 4 Wheel-tracing rates (45 C) at optimum binder content for 30 per cent wearing course mixtures containing different fine aggregates Binder Optimum binder content (per cent) Density (Mg/m 3) Wheeltracking rate (ram/h) Optimum binder content (per cent) Density (Mg/m 3) Wheeltracking rate (mm/h) SOUTHPORT BACH BRI, ( ~TOL CHANNL per cent VA + 5 per cent VA + 5 per cent VA WH ITGAT ALMINGTON +2.5 per cent VA + 5 per cent VA + 5 per cent VA translated into commercial practice, this would reduce bitumen costs by about 12 per cent and, to some extent, compensate for the increased cost of the polymer modified binder. But, an asphalt surfacing with this reduced binder content may be less durable and the results of these Marshall designs must be interpreted cautiously. However, the results of an experiment on the Winchester by-pass (Jacobs 1983) have shown that low binder content rolled asphalts can perform well and some rolled asphalts with 1.5 per cent below target binder content (derived from the procedure given in Section 3 of BS 594: 1973) performed adequately for 7 years. 5.2 DFORMATION RSISTANC Laboratory samples were manufactured at the optimum binder contents and at 1 per cent of optimum for each binder and fine aggregate; they were subjected to the TRRL wheel-tracking test at 45 C described by Jacobs (1981). The variations in wheel-tracking rate with binder content for each material are shown in Figures 6-9. Density values and wheel-tracking rates at the optimum binder contents are recorded in Table 4. For the 50 and bitumens the wheel-tracking rates showed a similar pattern, with the more stable sands giving the lower tracking rates. Only the asphalts made with bitumen and with the Whitegate and the Almington sands had wheeltracking rates less than 2mm/h, indicating suitability for use on very heavily trafficked roads; see Jacobs (1981). For the materials made with bitumen plus 2.5 per cent VA the wheel-tracking rates for three of the sands were reduced by a factor of 1.8, with all rates being less than 2mm/h. For the Almington sand the effect was less marked (reduction factor 1.35), probably because the high particle interlock of this sand was more significant than the contribution made by the VA polymer. The wheel-tracking rates for the asphalts containing the bitumens modified at the 5 per cent VA level were substantially reduced for all the fine aggregates, with all values below 1.5 mm/h, a reduction factor of 3 to 4. This indicates that asphalts containing low stability sands, when mixed with bitumen plus 5 per cent VA, should perform satisfactorily under the severest traffic loading conditions, even though the Marshall stability values for the Bristol Channel and Southport Beach sands suggest otherwise. The plots of variation of wheel-tracking rate with binder content in Figures 6 to 9, show that the addition of ethylene-vinyl acetate to a bitumen improves the tolerance of the rolled asphalt to variations in binder content. In practice, this would enable a higher-thanoptimum binder content to be specified without adversely affecting deformation resistance and this could be of benefit in certain applications. While the optimum binder content, determined by the Marshall design method, gives low wheel-tracking rates for each material, Figures 6 to 9 also show that, for a minimum deformation resistance, a lower binder content could be chosen, especially for the Whitegate and Almington fine aggregates. This confirms earlier work by Jacobs (1983). However, this trend is not apparent for all the fine aggregates in this study.

13 o BINDR Bitumen o Bitumen per cent VA Bitumen + 5 per cent VA Bitumen Bitumen + 5 per cent VA BINDR Bitumen o Bitumen per cent VA Bitumen + 5 per cent VA Bitumen Bitumen + 5 per cent VA!0.0! 10.0 D --L == z= D # 0.0 I I I I! Fig. 6 Wheel-tracking rates of rolled asphalt (30 per cent stone content) with various binders using Bristol Channel fine aggregate Fig. 8 Wheel-tracking rates of rolled asphalt (30 per cent stone content) with various binders using Whitegate fine aggregate BINDR o Bitumen o Bitumen per cent VA Bitumen + 5 per cent VA Bitumen Bitumen + 5 per cent VA BINDR o Bitumen o Bitumen per cent VA Bitumen + 5 per cent VA Bitumen Bitumen + 5 per cent VA 10.0 / A. 5.0 v L == Y ~1 O ~ 0.0 I I I I I I Fig. 7 Wheel-tracking rates of rolled asphalt (30 per cent stone content) with various binders using Southport Beach fine aggregate Fig. 9 Wheel-tracking rates of rolled asphalt (30 per cent stone content) with various binders using Almington fine aggregate 10

14 6 RLATIONSHIPS BTWN DIFFRNT PROPRTIS OF MATRIALS The changes in the rheological properties of modified binders have already been discussed and provide a consistent pattern in this limited study. Relationships have been found between the softening point of a binder and the wheel-tracking rate by Jacobs (1981) who used Redhill sand and this information together with the data from four fine aggregates from this study are shown in Figure 10. Similarly, the effect of the apparent viscosity of a binder (45 C) on the wheel-tracking rate has been reported by Denning and Carswell (1981) for Redhill fine aggregate and regression lines for each fine aggregate are constructed in Figure 11. The relationships shown in both Figures 10 and 11 are distinct for each fine aggregate; it will be noted that the slopes are generally the same and the values are tabulated in Table 5. These different relationships probably reflect or are a function of material differences; In order to examine the role of the ethylene-vinyl acetate component in the Marshall stability values and wheel-tracking rates, these are considered with and without the addition of polymer. The effect of the addition of VA to bitumen on the Marshall stabilities for each fine aggregate is shown in Table 6 and on the wheel-tracking rates in Table 7. For the 2.5 per cent level of addition to the bitumen, the increase in Marshall stability values averaged 17 per cent, and the wheel-tracking rates are reduced by a factor of 0.6 (range 0.5 to 0.7). At the 5.0 per cent level of addition to both grades of bitumen the resultant Marshall stabilities were 35 per cent higher, with the increase for the Almington sand being somewhat smaller for the reasons discussed earlier. The wheel-tracking rates were reduced by a factor of 0.3 (range 0.25 to 0.35) which was found to be independent of type of fine aggregate. Currently, bitumen modified with 5 per cent VA is used in some asphalts therefore the results for this material will be examined in more detail. Figure 12 shows the relationship between Marshall stability and wheel-tracking rate on a logarithmic scale for both the unmodified and VA modified asphalts. A J= 0.5 v Bristol Channel ~-" ~ Southport \ A.C v 0.5 ~.='q~:--~ Southport ~.=="~,~ -_ -- -~.-,r- ~ Bristol Channel ~ -0.5 White~ Almington Redhill o o eo. Almington ~ ~~~'-""- ~ Whitegate ~'~ -3 I I I I Ring and ball softening point ( C).J -0.9 I I Log apparent viscosity (poise) Fig. 10 Relationship between log wheel-tracking rate and ring and ball softening point for each fine aggregate Fig. 11 Relationship between log wheel-tracking rate and log apparent viscosity for each fine aggregate TABL 5 Regression slopes and correlation coefficients for fine aggregates Softening point/wheel tracking Apparent viscosity (45 C)/Wheel-tracking rate regression rate regression Fine aggregate Slope Correlation coefficient Slope Correlation coefficient Southport Beach Bristol Channel Whitegate Almington Redhill t * t basis: LR 1003 * basis: LR

15 TABL 6 ffect of VA content on Marshall stability Factor increase with binder modifications Fine aggregate per cent VA per cent VA per cent VA Bristol Channel M 1.50 Southport Beach m 1.30 Whitegate D m 1.45 Almington m AVRAG FACTOR INCRAS TABL 7 ffect of VA content on wheel-tracking rate (45 C) Factor reduction Fine aggregate per cent 5.0 per cent VA VA per cent VA Bristol Channel Southport Beach Whitegate Almington AVRAG RDUCTION FACTOR W = M (unmodified -- all work) (W = 21/M(1.34) A J~ W = M I (unmodified -- present work only) -. / (W = 13.3/M (1-04) ~ / W= M "~ ~.~'~ (5 per cent VA modified) " ~ W = 5/M(0"93)) i.o Log Marshall Stability (30 per cent stone content) of mixes, M (kn) O..c: 8,-,, = 6 i-- ~ uj ~o. ~_a~ ~ _ 5 ~)~< 4 o. "' o'j co ~ 3 >cg ~-8~ 2 Nx 1 ~ o / Design curves constructed at / ~,. 85 per cent confidence level / / / j~/ -- /. J - ~ M=0"21M1"44 i i i I I I I I Marshall Stability (30 per cent stone content) of mixes containing unmodified bitumens, M (kn) - Fig. 12 Relationship between wheel-tracking rate and Marshall Stability for asphalts containing unmodified and VA modified bitumen Fig. 13 Marshall Stability curves for equivalent wheel-tracking rates for materials with modified and unmodified bitumens 12

16 The equation derived for the unmodified asphalt is 13.3 W-M~.04 and for the VA asphalt 5 W = M 0'93... (1) where W=wheel tracking rate (mm/h) and M = Marshall stability (30 per cent stone content) (kn)..... (2) where W= wheel tracking rate of VA asphalt (ram/h) M = Marshall stability of VA asphalt (30 per cent stone content) (kn). For the case of equivalent wheel-tracking rates, W=W, we have M~.04 M0.93 or M = 0.35M 1'12... (3) This equation is shown in Figure 13 and suggests that if the Marshall stability of a conventional mix is known, then, for the same performance in the wheeltracking test, a much lower Marshall stability is required with a binder containing 5 per cent VA. This finding would permit sands, currently considered unsuitable, to be used for heavily-trafficked roads when used in a mix containing VA. The standard deviation of the data used in the derivation of equation (3) has been used to develop the design curve in Figure 13 corresponding to the 85 per cent confidence level, which can be simplified for general use (in the range 4-12 kn for conventional materials) to the following approximation. M = 0.5M (4) This is given in Figure 14. The work described in this report covers only four fine aggregates and two unmodified bitumens. However earlier work by Jacobs (1981) and (1983), provides a greater range of both these variables for unmodified binders. Assuming that Marshall values for stone filled mixtures (30 per cent) are 1.3 times those for mortar mixtures (Jacobs 1983), all the reported work, together with the results of this project can be combined to form a general equation relating these two properties for conventional asphalts. The equation is W= 21 M (5) which differs to some extent from equation (1) above, and from earlier work (Jacobs 1981). The relationship in equation (5) is shown in Figure 12. If it is assumed that the relation between wheel tracking and Marshall stability of mixes containing VA that could have been made with the sands of Jacobs earlier work is also that obtained from the present work, then equation (5) can be combined with equation (2) to derive a relationship between X "6.,. o,,, O O u'.c: 8~ ~8 Fig M = 2J3M (all work) I I I I I I I I Marshall Stability (30 per cent stone content) of mixes containing unmodified bitumens, M (kn) Marshall Stabilities of modified and unmodified materials with equivalent wheel-tracking rates Marshall stability for VA modified and for unmodified asphalts at equivalent tracking rates (W = W): 21 5 M1.34 M0.93 or M = 0.21 M TM... (6) This equation is plotted in Figure 13. A similar design curve at the 85 per cent confidence level for Marshall stability values of mixes containing 5 per cent VAmodified bitumen is also shown. For all work the approximate design equation relating Marshall stability of unmodified asphalts to that of modified- VA asphalts for equivalent wheel-tracking rates is M = 2M... (7) This relationship is shown in Figure 14. It is equally plausible to assume that the proportional shift in relations between wheel tracking and stability obtained from the present work and from all results combined (shown in Figure 12) is the same as would be obtained for these two groupings of materials with VA added: in this case equation (3) would apply. However Figure 14 shows that at stabilities of greater than 4 kn equation (7) is more conservative and is therefore recommended for use. The recently revised British Standard 594 (1985) (British Standards Insititution 1985) requires a minimum Marshall stability of 6 kn for roads carrying over 6000 commercial vehicles per day. From equation (7), for the same deformation resistance, an VA asphalt would require a Marshall stability of 4 kn. Alternatively, if the Marshall stability of the VA asphalt was 6 kn, it would correspond to using a conventional asphalt of 9 kn stability with the equivalent deformation resistance. J 13

17 TABL 8 Marshall stabilities required for wheel-tracking equivalence Wheel-tracking rate (mm/h) Marshall stability required for conventional asphalts M* (kn) Marshall stability required for VA-modified asphalts Mt (kn) * from W = 21 M [ design equation M=~M Table 8 sets out the Marshall stabilities of conventional materials calculated for a range of wheel-tracking rates from the general equation (5). The Marshall stabilities of the VA-modified materials for the same wheel-tracking rates are also calculated using the design equation (7). In Table 9 the values calculated for the Marshall stability of the asphalt made with the VA-modified bitumens using equations (4) and (7) are recorded for a range of Marshall stabilities of conventional materials for wheel-tracking equivalence. The two equations provide similar values over a wide range of Marshall stabilities. 7 CONCLUSIONS The addition of ethylene-vinyl acetate (VA) copolymer to a petroleum bitumen produces the following rheological changes: 1. A decrease in the penetration of the binder by just under one grade at the 2.5 per cent level of addition and by about one grade at the 5.0 per cent level. 2. An increase in the ring and ball softening point, with the increase being dependent on the concentration of VA. TABL 9 Prediction of Marshall stability for VA asphalts (5 per cent VA in bitumen) Marshall stability (M) of unmodified asphalt (kn) Predicted Marshall stability for VA asphalts (M) (using design equations) for equivalent wheel-tracking rate 1 M = ~M + 3/4* (kn) M=2M + (kn) * Basis: work done in this project + Basis: * plus work by Jacobs (1981) and (1983)

18 3. An increase in the elasticity of the binder at higher road temperatures while at the same time making the binder more viscous. 4. An increase in the viscosity at typical mixing and laying temperatures but not to the extent that the increase in the softening point value would imply. When VA-modified bitumens are used in rolled asphalt wearing course the following changes are produced: 1. Marshall stabilities are increased by about 20 per cent for the 2.5 per cent VA level, irrespective of fine aggregate and by about 35 per cent at the 5 per cent VA level, though the effect is less marked for the high stability sands. 2. The wheel-tracking rate is reduced by a factor of 0.6 for the 2.5 per cent level of VA addition and by a factor of 0.3 for the bitumen containing 5 per cent VA, with the reductions being generally independent of type of fine aggregate. Relationships between the Marshall stability of the conventional asphalt (M) and the Marshall stability of the VA modified asphalt (M) with the same resistance to permanent deformation was M=~M for the range of Marshall stabilities normally encountered. Thus VA-modified binders can be used to enhance the properties of fine aggregates which produce low Marshall stability rolled asphalts with conventional binders. 8 ACKNOWLDGMNTS The work described in this report forms part of the research programme of the Pavement Materials and Construction Division (Division Head: G F Salt) of the Highways Group of the TRRL. The research team included J H Denning, J Carswell, D M Colwill and Miss F Tetley. 9 RFRNCS BRITISH STANDARDS INSTITUTION (1973). Rolled asphalt (hot process) for roads and other paved areas. British Standard BS 594 British Standards Institution, London. BRITISH STANDARDS INSTITUTION (1985). Hot rolled asphalt for roads and other paved areas. Part 1 Specification for constituent materials and asphalt mixtures. British Standards BS 594 British Standards Institution, London. DAINS, M (1985). Cooling of bituminous layers and time available for their compaction. Department of Transport TRRL Report RR4: Transport and Road Research Laboratory, Crowthorne. DNNING, J H and CARSWLL, J (1981). Improvements in rolled asphalt surfacings by the addition of organic polymers. Department of the nvironment, Department of Transport, TRRL Report LR 989: Transport and Road Research Laboratory, Crowthorne. INSTITUT OF PTROLUM (1984a). IP Standards for Petroleum and its Products. IP4a/84 Penetration of bituminous materials. London, Institute of Petroleum. INSTITUT OF PTROLUM (1984b). IP Standards for Petroleum and its Products. IP58/84 Softening point of bitumen, ring and ball. London, Institute of Petroleum. JACOBS, F A (1981). Hot roled asphalt: effect of binder properties on resistance to deformation. Department of the nvironment, Department of Transport. TRRL Report LR 1003: Transport and Road Research Laboratory, Crowthorne. JACOBS, F A (1983). A33 Winchester by-pass: the performance of rolled-asphalts designed by the Marshall Test. Department of the nvironment, Department of Transport. TRRL Report LR 1082: Transport and Road Research Laboratory, Crowthorne. Printed in the United Kingdom for Her Majesty's Stationery Office (2774/87) Dd /87 C10 G