Fiber alignment alone does not tell the whole mechanical story in a two-dimensional network, according to a preprint. When the model compared different fiber laws at the same degree of alignment, strongly buckling networks were more than an order of magnitude softer. The result suggests that how fibers carry load matters alongside the pattern they form.
How the model works
The work develops a continuum framework, meaning it treats the network as a smoothly varying material rather than tracking every fiber. At each point, it assigns a probability distribution to fiber directions, assumes an affine deformation in which the fibers follow the local bulk motion, and determines the deformation map by imposing mechanical equilibrium. In plain terms, the model combines local fiber orientation with force balance to predict how the material changes shape.
It applies that framework to two planar settings: a network pulled in one direction and a single cell-like region that contracts equally in all directions. The first setting tests alignment and density changes under uniaxial stretch. The second follows how a local contraction changes the surrounding network.
For a check against discrete calculations, the study used finite-element fiber networks with a mean fiber length of 20 micrometres, a mean thickness of 0.2 micrometres, and mean connectivity set to either 8 or 12.
Buckling changes the early response
In the stretching problem, the calculations reduce several geometric quantities to one mechanically selected number, k, an anisotropy ratio set by the requirement that the sides remain traction-free. Once k is known, the model's orientation distribution, nematic order, density ratio and fraction of buckled fibers are all functions of it. Nematic order here is a measure of how strongly fibers favor a common direction.
The model predicts that the early density response depends on the buckling ratio. With a ratio of about 0.12 or lower, the network initially densifies. With weaker buckling, and in the material-linear reference case, it initially becomes less dense before densifying closer to the maximum stretch.
At larger stretches, however, the density and nematic-order curves nearly collapse onto a generally positive relationship. In other words, once the network is pulled far enough, more alignment tends to appear alongside greater density in the model, even though their early behavior can differ.
Stiffness has another story
The mechanical response is more dramatic for nonlinear fibers. The modeled axial modulus, a measure of resistance to stretching along the load, rises by several orders of magnitude as stretch increases and diverges at the limiting stretch. The material-linear reference stays nearly constant. In that reference, the small-strain Poisson ratio, which compares sideways and lengthwise deformation, is 1/3, and the normalized axial modulus is also 1/3.
The model also fixes a limiting relationship between anisotropy and lateral response. At the stiffening lock, anisotropy diverges and the Poisson ratio reaches the reciprocal of the single-fiber limiting strain, regardless of the buckling ratio. That result is a property of the model's limit rather than an estimate from observations.
A local contraction travels farther
The same contrast appears around a contracting cell-like region. Radial displacement falls off with distance as an inverse first power in the material-linear case, but as an inverse 0.28 power when the buckling ratio is 0.05. Because the latter decline is slower, the modeled displacement extends farther from the contracting region.
Far from the region, the model links nematic order and excess density through the same power-law exponent. In the material-linear limit, density excess falls as the fourth inverse power of distance. This gives the framework a structural prediction beyond the immediate neighborhood: the spatial reach of alignment and densification is tied to the same decay pattern.
At the reported cell boundary, the anisotropy ratio k was 1.66 for a limiting strain of 0.3 and a buckling ratio of 1, and 1.53 for the same limiting strain with a buckling ratio of 0.05. The material-linear reference reached 1.90. Radial stretch was 1.07 in the buckling network, compared with 1.33 in the material-linear network. In this comparison, the buckling model accommodated the imposed contraction with less radial extension.
Where the model starts to miss
The authors tested the affine picture against non-affine discrete finite-element simulations. At low k, the predicted and simulated orientation distributions agreed closely. At larger k, the affine calculation became too sharply peaked around the loading axis, although it continued to describe the tails well.
That mismatch marks a limit to the model's reach. The conclusions concern the two planar settings and the discrete comparisons described here. No statistical uncertainty interval or quantitative theory-simulation error metric is reported. Within those boundaries, the clearest message is narrower: alignment and densification can track one another at larger stretches, but alignment alone is not a sufficient guide to stiffness.
Publication status
The document is a preprint identified as arXiv:2608.25125v1 in the cond-mat.soft category and dated 25 August 2026. The work was partially supported by Grant No. 2022197 from the United States-Israel Binational Science Foundation.
Paper data and sources
Original title: Micromechanical statistical model links induced nematic order to mechanical response in fiber networks
Authors: Ehud Haimov, Yoni Koren, Ayelet Lesman et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-25
DOI: Not available
Original paper · Full text