A Comparative Study of Cohesive Zone Models for Predicting Delamination Behaviors of Arterial Wall

Arterial tissue delamination, manifested as the failure between arterial layers, is a critical process in the rupture of atherosclerotic plaque, leading to potential life-threatening clinical consequences. Numerous models have been used to characterize the arterial tissue delamination. Few has investigated the effect of Cohesive Zone Model (CZM) shapes on predicting the delamination behavior of the arterial wall. In this study, four types of cohesive zone models (triangular, trapezoidal, linear-exponential and exponential-linear) were investigated to compare their predictability of the arterial wall failure. The Holzapfel-Gasser-Ogden (HGO) model was adopted for modelling the mechanical behavior of the aortic bulk material. The simulation results using CZM on the aortic media delamination were also compared with the results on mouse plaque delamination and human fibrous cap delamination. The results show that: 1) the simulation results based on the four shapes of CZMs match well with the experimental results, 2) the triangular and exponential-linear CZMs are in good agreement with the experimental force-displacement curves of mouse plaque delamination, 3) considering the viscoelastic effect of the arterial tissue, the triangular and exponential-linear CZMs match well with the experimental force-displacement curves of human fibrous cap delamination. Thus, triangular and exponential-linear CZMs can capture the arterial tissue failure response well.

💡 Research Summary

The paper investigates how the shape of cohesive zone models (CZMs) influences the prediction of arterial wall delamination, focusing on the medial layer where most dissection events occur. Four CZM traction‑separation laws are examined: triangular, trapezoidal, linear‑exponential, and exponential‑linear. The bulk arterial tissue is modeled with the anisotropic hyper‑elastic Holzapfel‑Gasser‑Ogden (HGO) formulation, which captures the matrix response and the two families of collagen fibers through separate strain‑energy terms, fiber orientation, dispersion, and stiffness parameters.

Finite‑element simulations are performed in ABAQUS using a user‑defined element (UEL) to implement the CZMs. The arterial strip geometry (4 mm × 1.2 mm × 0.05 mm) is meshed with 1920 eight‑node brick elements for the bulk and 72 zero‑thickness eight‑node cohesive elements along the predefined peeling path. Mesh size (0.05 mm) is chosen after a convergence study. Material parameters for HGO and CZM (interfacial stiffness K_w, critical energy release rate G_c, peak traction σ_c) are taken from prior experimental works.

Three sets of validation data are used: (1) aortic media peeling experiments by Sommer et al., (2) mouse atherosclerotic plaque peeling, and (3) human fibrous‑cap peeling, the latter incorporating viscoelastic behavior of arterial tissue. Simulated load‑per‑width versus displacement curves fall within the experimental range (25.5–28.5 mN/mm compared with 23–35 mN/mm). All four CZM shapes reproduce the overall trend, but distinct differences emerge in the softening stage and peak force.

Triangular and exponential‑linear CZMs match the experimental force‑displacement curves most closely for both mouse plaque and human fibrous‑cap tests. The triangular law provides a simple linear‑elastic rise to a peak traction followed by a rapid softening, which aligns well with the observed abrupt loss of load‑bearing capacity. The exponential‑linear law, which features an initial exponential increase in traction before a linear decline, captures the gradual stiffening due to collagen fiber recruitment and the subsequent softening observed in viscoelastic human tissue.

Linear‑exponential CZM yields the lowest peak forces (≈10 % lower than experiments) because its softening stage begins earlier and reduces traction more quickly. Trapezoidal CZM maintains the peak traction over a longer displacement, leading to higher initial load transmission but over‑predicting energy dissipation. Damage variable evolution analysis shows that linear‑exponential CZM accumulates damage fastest, followed by triangular, exponential‑linear, and trapezoidal, although this ordering does not directly translate to peak load values.

The study concludes that for arterial medial delamination, the triangular and exponential‑linear CZMs provide the most reliable predictions, especially when viscoelastic effects are considered. Triangular CZM offers a computationally efficient option with acceptable accuracy, while exponential‑linear CZM better reflects the underlying fiber‑recruitment mechanics. The findings underscore that the choice of CZM shape should be guided by the specific mechanical features of interest (e.g., initial stiffness, energy absorption, damage evolution) and the availability of material data. Future work is suggested to extend the analysis to patient‑specific geometries, combined pressure‑shear loading, and to integrate imaging‑based parameter identification for personalized risk assessment of arterial dissection.