Styrenic block copolymers as bitumen modifiers for improved roofing sheets.

Waterproofing is characterized by a continuous improvement in terms of performance and durability. Bitumen itself, largely used in manufacturing roofing sheets, will hardly match the novel application requirements if not properly modified. To this aim, polymeric materials are very effective.

The state of the art seems to be represented by the use of styrenic thermoplastic elastomers (SBS), as they show the best balance between technical and economical requirements.

Their effect is based on a peculiar mechanism: Although being the minor component, i.e. 10-15%, SBS copolymers are able to build a continuous phase. The polybutadiene mid-segments absorb a great portion of the maltene fraction of bitumen by swelling, while styrenic end blocks show poor compatibility against bitumen, and segregate in hard-phase domains acting as physical crosslinks for the soft polybutadiene segments (ref. 1).

The resulting morphology gives the composite a pseudorubbery plateau, resembling the behavior of a highly swollen rubber network. The elastic properties are enhanced in the range of temperatures relevant for the application (ref. 2), while the ease of processing is not affected, as in the temperature range involved in manufacturing SBS copolymers show their thermoplastic behavior (refs. 1 and 3).

The aim of this work is to investigate the complex interactions between bitumen and SBS copolymers, which result in dramatic alterations of the viscoelastic behavior. The use of dynamic-mechanical spectrometry as a powerful tool to obtain a more detailed and complete assessment of the material performances is examined, in comparison with the traditional testing methods commonly adopted by the waterproofing industry.

Experimental

Materials

* A star-shaped telechelic styrene-butadiene block copolymer (Europrene Sol T 161/B; Sty/Bu ratio 30/70: Mw=224 x 10, Mn=160 x 10) was used.

* A commercial grade of bitumen Solea (180/200 penetration), supplied by IP (Italiana Petroli) was used as received.

Preparation of the samples

A Brabender Plastograph L 651, equipped with a one liter sigma-blades mixing head was chosen to accomplish mixing of bitumen with thermoplastic rubber under controlled conditions. The adopted apparatus allows an easy and accurate control of two fundamental mixing parameters like rotor speed and temperature. A continuous recording of mixing torque and mixture temperature is also available.

The experimental procedure was the following: Temperature was set at 190 [degree] C and the warm mixer filled with the proper quantity of cold neat bitumen. After a preheating step, necessary to warm up the bitumen at the onset temperature, stirring was started, gradually increasing speed to 100 rpm. When temperature and mixing torque resulted in constant values, the thermoplastic rubber was added.

The final mixture was then directly poured in the molds, designed in order to obtain the proper test specimens for physical and mechanical characterization.

Fluorescence microscopy 

The morphology of the bitumen-rubber mixtures was examined by optical fluorescence microscopy, performed with a Leitz Laborlux 12 instrument, at a 100x magnification.

Traditional bitumen testing

Neat bitumen and bitumen-rubber mixtures were characterized by the following physical tests, commonly used in the waterproofing industry:

* Viscosity at various temperatures by a coaxial cylinders viscometer Brookfield DV II RVT.

* Penetration ASTMD 5

* Ring and ball softening point ASTM D 36

* Bending on mandrell UNI 8202/15.

Dynamic-mechanical measurements

Dynamic mechanical measurements on neat bitumen and bitumen-rubber mixtures have been carried out on a Rheometrics RDS II, operating in torsion and using a parallel plate geometry (8 mm diameter). The gap height was ranging between 5 and 6 mm and, in order to prevent slippage, test samples have been glued.

Results and discussion

Building the rubbery matrix

In order to investigate the interactions taking place between bitumen and polymer during the hot mixing, a blend based on 100 parts of bitumen and 13 parts of SBS copolymer was prepared at different mixing times, using the apparatus and the experimental procedure previously described. The morphology of the final mixtures was analyzed by fluorescence microscopy. A very effective phase contrast is obtainable as SBS copolymers emit a yellow-green fluorescence while the neat bitumen appears dark. Nevertheless, on a deeper examination by the same technique of every single bitumen fraction, as obtained according to ASTM 2007 method, only saturated and asphaltenes appear dark, while aromatics, resins and naphthenics are fluorescent (ref. 4).

In figure 1 the micrographs of the blend are reported, with reference to different mixing times. The building-up of a continuous elastomeric phase can be easily identified.

On the basis of the foregoing observations and considering that the SBS weight fraction in the blend is extremely low, it can be reasonably assumed that the continuous phase is actually composed by the thermoplastic elastomer swelled by the miscible bitumen fractions.

Although the mixing conditions used are less severe than the industrial ones, in which high shear mixers are employed, the SBS swelling and the consequent phase-inversion occur in a relatively short time leading then to a steady morphological arrangement of the system. This behavior is mainly due to the thermoplastic nature of SBS copolymers, which makes the mixing with bitumen extremely easy and effective.

A further proof of the extent of bitumen/SBS interactions is given by the variation of bitumen viscoelastic behavior, analyzed by dynamic-mechanical spectrometry. The experimental procedure adopted was the following: Temperature sweeps have been carried out in the range -80 - 120 [degree] C, by fixed steps of 2 [degree] C and a conditioning time of two minutes. The test frequency was 1 Hz and the strain properly increased to a maximum of 20% at high temperature within the limits of linear viscoelasticity.

In figure 2 the curves of the storage modulus G' vs. temperature are reported both for the neat bitumen and the bitumen/SBS blend mixed 50 minutes. Modification with SBS copolymers results in a remarkable improvement in flexibility in the low temperature range; on the contrary, at higher temperatures the neat bitumen begins to flow as a viscous liquid while the blend still retains an appreciable elastic component. This proves the primary role of the elastomeric matrix, which resembles the behavior of a highly swollen rubber network.

In figure 3 the curves' storage modulus vs. mixing time of the bitumen/SBS mixture, measured respectively at -20 and 40 [degree] C are reported. It can be clearly observed that at both temperatures the modulus values vary only in the former step of the mixing process, while mixing times longer than 15-20 minutes are not effective.

This behavior is in good agreement with the variation in the morphological arrangement previously described. The main effect of the modification with SBS copolymers results in a dramatic reduction of bitumen thermal sensitivity, which represents the main technological drawback of neat bitumen.

Time-temperature superposition 

The complete assessment of materials' viscoelastic behavior should usually require a number of dynamic-mechanical experiments performed in a wide range of temperatures and frequencies. Unfortunately, even the up-to-date experimental apparatus are not able to cover the whole range of interest; both very low and very high frequencies are outside the range of practically available experimental conditions.

In this situation time-temperature superposition principle proves to be an attractive and valuable technique particularly when polymeric materials are examined. When it applies, there is an important practical consequence: Results obtained in limited intervals of temperature and frequency can easily be extrapolated on wider ranges allowing the performance prediction even when experimental capabilities fail.

The validity of time-temperature superposition for the blends of bitumen and SBS thermoplastic rubbers was tested by the following experimental approach.

Dynamic measurements were carried out on a sample of 100/13 bitumen/SBS blend, mixed for 50 minutes. A series of isotherms was performed in the temperature range -50/+120 [degree] C, with a fixed step of 10 [degrees] C; the frequency range investigated was [10.sup.-2]/[10.sup.2[ Hz, and four experimental points each frequency decade were taken. In order to keep torque values in the right range for the instrument, strain was properly increased, always within the limits of linear viscoelasticity, owing to the progressive sample softening taking place when test temperature is raised.

Plots of storage modulus (G') and loss modulus (G") vs. frequency, in the experimental temperature range, are shown in figures 4 and 5 respectively. According to the WLF time-temperature superposition principle (ref. 5), G' and G" experimental data were reduced to the reference temperature To=20 [degrees] C using the following equations: where [unkeyable] is the frequency, [unkeyable] is the density, To is the reference temperature and aT is the horizontal shift factor defined as:

Shifting each isotherm over the adjacent curve allows the empirical prediction of the aT values. As shown in figure 6 aT empirical values properly fit the well known WLF equation (ref.5): the constants co1 and co2 being distinctive for the material and depending on the choice of reference temperature. Empirical aT values were used in constructing both the master curves for G' and G" vs. frequency at the reference temperature of 20 degrees C (see figure 7).

The master curve for the loss angle, as reproduced in figure 8, was obtained by plotting, as a function of the frequency, the G"/G' ratio computed on the basis of the master curves in figure 7.

Therefore as regards the validity of time-temperature superposition, it can be pointed out that the following validation creteria (ref.5) were fulfilled in this work:

* a substantial overlapping of shifted curves with reasonable fit was found;

* the same values of aT were used in deriving both master curves (G' and G");

* the values of aT fitted very well the WLF equation.

As regards the viscoelastic behavior the trend of storage modulus (G'), showing a detectable change in slope in the same frequency region where the loss angle reaches the minimum value, can advise the presence of a pseudo-rubbery plateau (refs. 4 and 5). For crosslinked polymers the elastic plateau extends to extremely low frequencies, on the contrary in the present case flow occurs due to the time-depending nature of physicial crosslinks, settled up by the styrenic domains of the thermoplastic elastomer.

This statement is in close agreement with the foregoing discussion concerning the building up of a continuous rubbery matrix during the mixing of bitumen /SBS blends.

Primary performance assessment

Prefabricated bituminous membrances are composite materials with a multilayer structure consisting of non-woven glass or synthetic material impregnated and held together by a suitable bitumen-based binder. In this respect, binder plays a leading role giving to the overall-composite structure waterproofing and physical-mechanical characteristics necessary to fulfill performance and durability requirements.

In this work, attention was focused on the prediction of membrane performance examining the viscoelastic behavior of the basic binder component, namely the bitumen.

In table 1, the properties obtained according to the traditional tests generally used in the waterproofing industry are reported, both for the neat bitumen and the 100/13 bitumen/ SBS blend.

In spite of the substantial modification of bitumen properties, due to the addition of SBS copolymer, it is difficult to draw, on the basis of these data, objective predictions about the performance of the final product.

Traditional tests are in fact conceived in order to establish a set of specifications for the bitumen and hardly match the actual conditions in service.

Most prefabricated membranes end up on buildings roofs which are naturally exposed to great climatic effects. In this respect thermal cycling, taking place according to the natural daily, seasonal or yearly frequencies, plays a fundamental role.

On the practical point of view, thermal cycling turns in mechanical cycling because of the roof structure which consists of different materials with different thermal expansion coefficients.

Therefore waterproofing membrane, and in particular the bitumen-based binder, must be able to withstand stresses induced by the alternate stretching and shrinking of the overall roof structure. According to the season, extreme environment conditions can produce winter embrittlement, leading to crack formation, or in turn, summer softening which impairs dimensional stability.

Primary performance of bitumen, when used as membrane binder, was analyzed on the basis of the above considerations using the dynamic data reported in the former section.

As previously mentioned the building of a continuous rubbery matrix enhances to a great extent bitumen elasticity. For this reason storage modulus was chosen as an effective indicator of viscoelastic behavior.

Using the experimental data reproduced in figure 4 a series of master curves, storage modulus vs. frequency, was obtained at different reference temperatures in the range 50/ 120 degreess C with a steo of 10 degree C.

Treatment of experimental data was performed on the basis of time-temperature superposition principle. Coefficients co1 and co2 in equation (4) were corrected, in order to provide coefficients C11 and cl2 corresponding to the actual reference temperature (T1), using the following equations:

Master curves, obtained in this way, provided the data used in constructing the plot of storage modulus vs. temperature as reported in figure 9. The range of temperature considered covers most of the in service conditions, while reference frequencies were selected according to the previously mentioned natural cycling of the roof. The highest frequency (1 Hz) can be representative for the thermal shock induced by a sudden summer storm or for the stresses that the roof undergoes during the construction and the maintenance.

Addition of SBS copolymer improves bitumen performance over the entire temperature range. In comparison with neat bitumen behavior, as represented by the dashed line, modified bitumen exhibits, at the same time, higher moduli at high temperatures and lower moduli at low temperatures. Moreover it is important to note that the performance of modified bitumen seems to be satisfactory even at the boundary conditions. The combination of low temperatures and high frequencies may cause embrittlement and failure, on the contrary high temperatures and low frequencies promote creep and flow. Neat bitumen just behaves in this way, while modified bitumen displays in turn flexibility or consistency enough to withstand both conditions.

Primary performance assessment, as outlined above trying to match final in service conditions, proves the effectiveness of dynamic-mechanical measurements when used in analyzing polymer-bitumen systems.

As regards the other waterproofing requirements, i.e. durability, dynamic measurements, combined with suitable aging procedures, could represent a promising technique.

Conclusions

Bitumen viscoelastic behavior can be modified to a great extent and carefully tuned, in order to match practical requirements, by the addition of styrenic thermoplastic elastomers.

Hot mixing of bitumen and SBS copolymers leads to the formation of a rubber-like continuous phase. This can be directly detected by fluorescence microscopy or rendered in evidence by dynamic-mechanical analysis.

Time-temperature superposition principle was proven valid, for bitumen/SBS blends, in the temperature range-50/120 degrees C. Empirical shift factors (aT) values fit well the WLF equation.

Dynamic-mechanical analysis has found to be a powerful technique for performance assessment, allowing the prediction of viscoelastic behavior in the final service conditions; in this respect storage modulus was chosen as an effective indicator of viscoelastic behavior. Modified bitumen performs much better than neat bitumen in all the experimental range investigated. This is particularly evident at boundary conditions, i.e. at the extreme temperature and

frequency values.

References

(1). Dianik, E., Gargani, L., "Bitumen modification with SBS thermoplastic elastomers," 4th Eurobitume Symposium, 276, (Spain 1989). (2.) Piazza, S., Arcozzi, A. and Verga, C., "Modified bitumens containing thermoplastic polymers," Rubber Chem. Tech., 53, 994 (1980). (3.) Kraus, G., "Modification of asphalt by blck polymers of butadiene and styrene," Rubber Chem. Tech., 55, 1389 (1982). (4.) Bouldin, M., Collins, J.M. and Berker, A., "Rheology and microstructure of polymer/ asphalt blends," presented at a meeting of the Rubber Division, American Chemical Society, Las Vegas, Nevada May 29-June 1, 1990. (5.) Ferry, J.D., "Viscoelastic properties of polymers," John Wiley, N.Y., (3rd edition), (1980).