Finite Chain Phantom Networks

• Finite chain phantom networks are cross-linked polymer models where chains interact solely through connectivity, finite extensibility, and idealized fluctuations.
• Cycle rank density (ξ) quantifies independent loops and serves as a master coordinate to collapse elasticity and rupture data across various network formulations.
• Rupture studies reveal that increased connectivity boosts strength while reducing extensibility, guiding design strategies for improved fracture toughness.

Searching arXiv for papers on finite chain phantom networks and closely related phantom network fracture/elasticity topics. Finite chain phantom networks are cross-linked polymer network models in which chains interact only through connectivity, junctions fluctuate non-affinely, and excluded-volume and entanglement constraints are neglected, while each strand nevertheless has a finite contour length, finite bond extensibility, or a finite rupture threshold. In recent work, this class has included end-linked star-polymer and linear-prepolymer networks represented by Gaussian or FENE-like bead–spring strands, as well as ideal freely jointed chain formulations with deformation-dependent junction fluctuations. Across these formulations, the central structural variable is the cycle rank density $\xi$, which counts independent loops per node and organizes both elasticity and rupture, while finite loops, strand-length statistics, and concentration determine the limits of that organization (Masubuchi et al., 6 Jul 2026, Masubuchi, 2024, Nanavati et al., 2024).

1. Conceptual definition and model assumptions

The phantom approximation treats a polymer network as a connectivity graph with fluctuating junctions. Strands are idealized as Gaussian springs, FENE-like springs, or ideal freely jointed chains; nonbonded interactions are absent; and chains can cross freely because there are no topological constraints from entanglements. In this setting, elasticity is generated by connectivity and entropic chain response rather than by excluded-volume or tube-like constraints (Sorichetti et al., 2021, Nanavati et al., 2024).

The “finite chain” qualifier has two closely related meanings in this literature. In simulation studies of fracture, each strand contains a finite number of beads or bonds, and rupture is introduced by deleting bonds whose extension exceeds a prescribed threshold; finite extensibility may be encoded either explicitly through a FENE-like spring or effectively through a bond-cut criterion applied to Gaussian strands (Masubuchi et al., 6 Jul 2026, Masubuchi, 2024). In analytical elasticity work, finite-chain behavior refers to non-Gaussian single-chain statistics with a bounded contour length, as in ideal freely jointed chains, for which the end-to-end distance satisfies $r \leq L_c = Nb$ and the entropic force diverges at lock-up (Nanavati et al., 2024).

Architecturally, the dominant model systems are end-linked networks of star prepolymers or stoichiometric linear-prepolymer/linker mixtures. Star networks usually parametrize the precursor by functionality $f$ and arm length $N_a$, with a mechanically effective strand length $N_s = 2N_a+1$ when two arms are linked end-to-end (Masubuchi, 30 Jan 2026, Masubuchi, 2024). End-linking chemistries based on binary A/B reactions are often used to suppress primary loops or odd-order loops, whereas random end-linking of identical reactive groups allows primary loops and thus introduces topological defects that remain important for both modulus and rupture (Masubuchi et al., 2023, Masubuchi, 2024).

2. Topological descriptors, cycle rank, and finite loops

The principal topological descriptor is the cycle rank density,

$\xi = \frac{E - V + 1}{V},$

for a connected percolating graph with $E$ edges and $V$ vertices. It measures the number of independent loops per node and increases with functionality and conversion because both add redundant load-bearing paths (Masubuchi et al., 6 Jul 2026). In star-polymer and end-linking simulations, $\xi$ is computed on the percolated cluster, with branch points counted as vertices and strands between them counted as edges; extenders with two reacted arms may lengthen strands without contributing new branch points (Masubuchi, 2023).

This graph-theoretic variable has two distinct roles. First, it acts as an effective mean-field coordinate for mechanics: data across different $(f,p)$ or $r \leq L_c = Nb$0 conditions collapse when plotted against $r \leq L_c = Nb$1 rather than against precursor chemistry alone (Masubuchi, 2024, Masubuchi et al., 2023). Second, it encodes the effect of loop defects in a coarse but not exhaustive way. In Michael Lang’s analysis of phantom modulus with finite loops, the classical ideal-tree result

$r \leq L_c = Nb$2

holds only for loop-free tree networks of monodisperse Gaussian strands and identical junction functionality (Lang, 2021).

Loop corrections require further distinction between pending and nonpending cycles. Under the resistor analogy and ideal loop gas approximation, only pending loops reduce the modulus, so the cycle rank after removing pending structures would suffice. Lang showed, however, that this analogy becomes only approximate once finite loops are present, because equilibrium strand conformations and force balance at loop junctions cannot generally be satisfied simultaneously (Lang, 2021). In the exact calculation for cyclic defects inserted into an otherwise perfect phantom network, each pending loop reduces the modulus by $r \leq L_c = Nb$3 independent of functionality, while nonpending loops produce additional, functionality-dependent reductions; for a two-chain loop,

$r \leq L_c = Nb$4

and for larger loops the correction approaches

$r \leq L_c = Nb$5

These results imply that finite loops do more than merely renormalize connectivity counts: they also alter junction fluctuations and time-averaged chain conformations (Lang, 2021).

A related consequence is that loop formation induces chain stretch. In loop-rich phantom networks, all loops tend to contract simultaneously to optimize conformations, and larger loops can therefore contain increasingly stretched chains. This makes a purely connectivity-based modulus estimate incomplete unless the fluctuation distribution of imperfect junctions is also modeled (Lang, 2021, Lang, 2021).

3. Elasticity from Gaussian, polydisperse, and finite-FJC viewpoints

At small strain, the baseline phantom-network modulus is reduced from the affine value by junction fluctuations. For Gaussian chains in a defect-free tree network,

$r \leq L_c = Nb$6

which recurs throughout the recent fracture literature as the constitutive reference for phantom elasticity (Masubuchi et al., 6 Jul 2026, Masubuchi et al., 2023). In uniaxial incompressible deformation, the corresponding nominal or true-stress forms are used as reference relations, although rupture simulations generally extract stress directly from minimized bead–spring configurations rather than imposing a closed-form constitutive law (Masubuchi et al., 6 Jul 2026, Masubuchi et al., 2024).

Finite-chain corrections enter by two routes. One route is non-Gaussian single-chain elasticity. “Phantom Networks of Finite Chains” develops an exact compact representation of ideal freely jointed chains and embeds it in an 8-chain geometry, then derives a phantom-network relaxation factor

$r \leq L_c = Nb$7

where $r \leq L_c = Nb$8. This factor reduces to the classical phantom prefactor $r \leq L_c = Nb$9 at small strain and tends to the affine limit as $f$0, because junction fluctuations are progressively suppressed near lock-up (Nanavati et al., 2024). In that formulation, the finite-chain phantom network is not merely a Gaussian network with a rupture cutoff; it is a non-Gaussian network whose fluctuation spectrum itself depends on deformation.

The second route is strand-length statistics. In randomly crosslinked phantom networks with exponential chain-length distributions, short strands are abundant, the exact end-to-end distribution becomes essential, and the modulus is not determined solely by the density of elastically active strands. Instead, the front factor

$f$1

depends on chain-length distribution and preparation protocol, and the shear modulus can even become a non-monotonic function of strand density for purely entropic reasons (Sorichetti et al., 2021). This directly qualifies the common simplification that phantom modulus is always a universal multiple of $f$2.

Concentration adds a further distinction. In star-polymer simulations with various strand densities, $f$3 collapses against $f$4 at fixed concentration, but the master curve depends on $f$5, especially at low concentration where mechanically ineffective fractions depress the modulus. Fracture metrics plotted against $f$6 also remain concentration-dependent, showing that small-strain stiffness does not uniquely encode the relevant topological and inhomogeneity information (Masubuchi et al., 2024).

4. Rupture mechanics and topology-controlled master curves

The most developed application of finite chain phantom networks is quasi-static rupture under uniaxial extension. The standard protocol is Brownian dynamics gelation in a periodic box, followed by repeated affine stretch increments, full energy minimization, and bond deletion whenever a bond length exceeds a critical threshold. Fracture is then defined by the loss of percolation or by the peak of the stress–strain curve, and the recorded quantities are the stretch or strain at break, the peak stress, and the work to rupture,

$f$7

depending on the strain measure used (Masubuchi, 2024, Masubuchi et al., 6 Jul 2026).

A central result is the emergence of $f$8-master curves. In end-linking networks with primary loops, $f$9, $N_a$0, and $N_a$1 collapse onto the same curves as loop-free star-network analogs when plotted against $N_a$2, implying that the mechanical effect of primary loops is embedded in the reduced effective connectivity (Masubuchi, 2024). In star-polymer networks without secondary loops, the quantities $N_a$3, $N_a$4, and $N_a$5 decrease monotonically with $N_a$6, showing that stress and work delivered per broken bond decline as network redundancy increases (Masubuchi et al., 2023). These results are not contradictory: per branch-point strength grows with $N_a$7, whereas per broken-bond efficiency decreases with $N_a$8.

For the stretch at break, a mechanical derivation has been proposed in which rupture is controlled by a soft, highly stretched strand in series with stiffer surrounding bundles. Identifying the number of parallel load-sharing strands with connectivity yields

$N_a$9

which reproduces the universal $N_s = 2N_a+1$0 collapse for both Gaussian and FENE spring networks over most of the studied range (Masubuchi, 30 Jan 2026). A related mechanistic picture in the larger 2026 statistical study is the bottleneck–parallel-bundle model: increasing $N_s = 2N_a+1$1 adds parallel load paths around the softest strand, raising stiffness and peak stress but reducing extensibility before the critical strand fails (Masubuchi et al., 6 Jul 2026).

Arm molecular weight introduces a second scaling variable. In star networks with $N_s = 2N_a+1$2, the simulations give

$N_s = 2N_a+1$3

with $N_s = 2N_a+1$4 (Masubuchi, 2024). This places strand length and connectivity on distinct but coupled scaling axes.

A complementary 2026 formulation separates fracture into two universal master curves: macroscopic softening governed by absolute stretch, with $N_s = 2N_a+1$5, and microscopic scission governed by relative stretch, with

$N_s = 2N_a+1$6

Within that framework, concentration enters through $N_s = 2N_a+1$7 (Masubuchi, 27 Feb 2026).

5. Statistical structure of rupture and extreme-value interpretation

The broadest statistical dataset to date comprises 30,000 phantom Gaussian star networks generated from 1,000 realizations for each of 30 conditions with $N_s = 2N_a+1$8–$N_s = 2N_a+1$9 and $\xi = \frac{E - V + 1}{V},$0–$\xi = \frac{E - V + 1}{V},$1. The earlier mean master curves survive unchanged at this scale, but the individual rupture events are intrinsically random (Masubuchi et al., 6 Jul 2026).

At fixed $\xi = \frac{E - V + 1}{V},$2, the fluctuation of $\xi = \frac{E - V + 1}{V},$3 is small—less than $\xi = \frac{E - V + 1}{V},$4 in absolute magnitude and with $\xi = \frac{E - V + 1}{V},$5 in almost all cases—whereas $\xi = \frac{E - V + 1}{V},$6, $\xi = \frac{E - V + 1}{V},$7, and $\xi = \frac{E - V + 1}{V},$8 scatter much more strongly, with relative standard deviations of $\xi = \frac{E - V + 1}{V},$9–$E$0, $E$1–$E$2, and $E$3–$E$4, respectively (Masubuchi et al., 6 Jul 2026). The within-condition variance of the breaking properties is therefore not explained by residual fluctuations in $E$5 alone. This is an important correction to any overly strong reading of the master-curve concept: $E$6 predicts conditional means, not the exact outcome of a single realization.

The distributions are also non-Gaussian in a structured way. At small $E$7, the $E$8 distribution is right-skewed; at large $E$9, the $V$0 distribution becomes left-skewed with a sharp upper edge, and a minimum-value Weibull form often provides the best empirical fit by AIC. The Pearson correlation $V$1 rises from approximately zero at small $V$2 to approximately $V$3 at large $V$4, indicating that stronger realizations also become more extensible only in highly connected networks (Masubuchi et al., 6 Jul 2026).

Because $V$5, $V$6, and $V$7 are extreme values of the network response, extreme-value theory is the natural interpretive framework. The 2026 study explicitly compares Gaussian, log-normal, and minimum-value Weibull forms, but also emphasizes that 1,000 realizations per condition are insufficient to stabilize tail-shape estimates under bootstrap resampling (Masubuchi et al., 6 Jul 2026). The observed trends are opposite to those of the random fuse model: in phantom chain networks, increasing connectivity narrows fluctuations and increases strength, whereas in brittle lattice models with externally imposed threshold disorder, weakest-link behavior is associated with strength reduction as system size grows.

6. Scope, limitations, and design implications

The scope of these results is sharply defined by the phantom assumption. The models omit excluded volume, entanglements, osmotic interactions, viscoelastic dissipation, and strain-induced crystallization; most rupture studies are also effectively athermal during loading because Brownian motion is turned off after gelation (Masubuchi et al., 6 Jul 2026, Masubuchi et al., 2024). Consequently, agreement with experiments is strongest for topology-dominated model gels and weakest when real materials derive toughness from solvent, entanglements, or crystallization.

This limitation is visible in comparisons with experiment. Mixed-functionality star simulations reproduce the qualitative observation of Fujiyabu et al. that introducing 3-arm stars into 4-arm networks can improve toughness at high conversion, but the simulated gains in $V$8 and $V$9 are smaller, plausibly because stretch-induced crystallization is absent from the phantom model (Masubuchi, 2023). Likewise, the concentration dependence of experimental tetra-PEG fracture is stronger than in phantom simulations because the latter lack osmotic swelling and related homogenization effects (Masubuchi et al., 2024).

Within the phantom class, however, the design rules are unusually clear. To increase breaking stress and rupture energy per branch point, one increases $\xi$0 by raising functionality and/or conversion; to maximize extensibility, one lowers $\xi$1, accepting lower strength and larger sample-to-sample scatter (Masubuchi et al., 6 Jul 2026). At high conversion, lower-functionality networks can outperform higher-functionality ones because they achieve smaller $\xi$2 and longer effective strands, whereas at low conversion the trend reverses because higher functionality is needed to achieve a mature percolated network (Masubuchi et al., 2023). Concentration provides an independent control axis: increasing $\xi$3 raises $\xi$4, $\xi$5, and $\xi$6, but fracture remains not solely describable by modulus because the statistics of load paths change with concentration (Masubuchi et al., 2024).

A common misconception is therefore that a single scalar—either modulus or cycle rank—fully determines finite chain phantom network mechanics. The recent literature supports a more precise statement: $\xi$7 is the master coordinate for conditional mean behavior within a given phantom class, but loop topology, concentration, strand-length statistics, and finite-chain constitutive choice still determine the modulus, the fluctuation structure, and the accuracy of extrapolation beyond that class (Masubuchi et al., 6 Jul 2026, Sorichetti et al., 2021, Nanavati et al., 2024).