Dielectric modeling of asphalt mixtures and relationship with density

Chieh Min Chang, Jian-Shiuh Chen, Tz Bin Wu

Research output: Contribution to journalArticlepeer-review

33 Citations (Scopus)


The paper is aimed at investigating the dielectric response of the dense-graded hot-mix asphalt mixture (HMA) using an open-ended coaxial probe technique over a wide frequency range. The effects of asphalt content, air voids and density on dielectric response were examined. The results indicated that the dielectric constant was frequency- independent in the range of 3-7 GHz. Beyond the 3-7 GHz range, the dielectric constant decreased with increasing frequency due to polarization phenomenon. A linear relationship between the dielectric constant and density of the compacted asphalt mixture was identified. A higher density of HMA induced a reduction in air voids, thus increasing the dielectric constant. The asphalt content showed limited effect on the dielectric constant. The dielectric mixing model was proposed to obtain the dielectric constant of the heterogeneous asphalt mixtures according to the volume fraction and the dielectric constant of the constituents in asphalt mixtures. In addition, the density prediction model was developed to estimate density of the dense-graded asphalt mixture using dielectric constant, air voids and asphalt content. Comparisons of predicted density and measured values showed a good agreement with R2=0.94. Both the dielectric mixing and the density prediction models offer a great potential for applying the electromagnetic technique to improve the nondestructive testing of asphalt pavements.

Original languageEnglish
Pages (from-to)104-111
Number of pages8
JournalJournal of Transportation Engineering
Issue number2
Publication statusPublished - 2010 Jul 19

All Science Journal Classification (ASJC) codes

  • Civil and Structural Engineering
  • Transportation


Dive into the research topics of 'Dielectric modeling of asphalt mixtures and relationship with density'. Together they form a unique fingerprint.

Cite this