—10 °C ranged from 0。2% to 0。4% [45]。 Combined above, it indicates that BEJ mate-

average temperatures。 In total, 1825 cycles were loaded to reflect the effect of   daily

temperature cycles in 5 years。 The obtained test results highlighted the importance of material relaxation to joint performance。 All of the samples exhibited significant fatigue cracking。 The materials with higher adhesion and lower tensile strength tended to fail in a tensile mode with a crack forming through the APJ material while stronger APJs showed bond cracking [9,11]。 The observed weak points of fatigue failure were in agreement with the zones of high stress concentrations indicated by  finite  element analysis。

Thermal stress restrained specimen tests can be used to evaluate cold weather performance of joint mixtures。 Test results indicate the temperature at which joint fracture happens even with zero movement as well as the glass transition temper- ature, Tg。 The fracture temperature or Tg can be used as an indicator as the mini- mum temperature for joint application。 Research done by Bramel and Dolan proposed a design guideline and a modified standard AASHTO test to qualify the APJ materials on the basis of relaxation and glass transition temperature。 Joint materials should not be installed where the lowest anticipated temperature is be- low the glass transition temperature of the material。 The material time constant, t75, for the load to relax to 75% of its initial load, was introduced to ensure that the stress relaxation of joint material is faster than the applied thermal displace- ment。 The values of t75 of the three materials  tested  varied  from  3 min  to 76,000 min。 It was found that the joint materials with t75 smaller than 15 min were able to relax most of the stress induced by joint movements [9,11]。

3。3。2。 High temperature property

The joint material should be stiff enough not to be deformed under wheel load yet flexible enough to deform and accommodate the bridge deck movement。 For the purpose of ranking joint materials, Heng and King used Marshall testing as the qual- ity control testing。 Mashall properties including stability and flow were used as the criteria for permanent deformation resistance and flexibility。 It indicated that mix- tures with higher values of Marshall stability and flow showed better performance [29]。 Xiao performed a study on the water stability of compound modified asphalt mixtures of bridge expansion joints。 The compound modified binder consisted of bitumen of Penetration 70 dmm, crumb rubber powder, resin modified agent and aromatic oil。 All of the six asphalt mixtures with various binder-aggregate ratios showed a retained Marshall stability larger than 90% and indirect tensile strength ratio higher than 85%, indicating good resistance to water damage [43]。

APJs and the adjacent asphalt pavements are subjected to the same wheel loads。 This indicates that APJs should have good rutting resistance and permanent defor- mation resistance comparable to normal hot mix asphalt。 Because the stone skele- ton of joint materials is not as strong as that of asphalt concrete, the rutting resistance of the former strongly relies on the high temperature stability of the used binder。 Rutting testing indicated that the dynamic stability (number of wheel pas- sages to induce 1 mm rut depth) of joint materials is ranging from 1000 passes/mm to 2000 passes/mm, which is comparable to asphalt mixtures prepared using based bitumen but is much lower compared to modified asphalt mixtures with a dynamic stability of approximately 5000 passes/mm [34,43]。 According to Chinese specifica- tion JTG F40-2004, Conventional asphalt mixtures should have an acceptable limit of dynamic stability at least larger than 800 passes/mm for highway asphalt pave- ments [45]。 It is thus suggested that for the mixture design purpose, APJ should establish a similar acceptable limit of dynamic  stability。