Lignin can break the strength-toughness trade-off in thermoplastic polyurethane elastomers (TPUEs) because its rigid, three-dimensional aromatic backbone works as a multifunctional chain extender rather than as a filler. Grafted into a polyurethane backbone at about 10 mol%, lignin polyols build a hierarchical hydrogen-bond network that dissipates energy through sacrificial bond rupture while the material stays melt-processable. The best formulation reaches a tensile strength of 127.7 MPa and a toughness of 1.43 GJ/m³, which is among the highest combined values reported for any polyurethane elastomer.
Polymer engineers have lived with an uncomfortable compromise for decades. Make an elastomer stronger and it usually turns brittle. Make it tougher and it tends to go soft. Thermoplastic polyurethane elastomers sit in the middle of that problem: they process beautifully and their microphase-separated structure is endlessly tunable, but the linear or weakly branched chains of a conventional formulation keep them inside the same bargain.
That bargain is now being renegotiated with a material most people associate with wood pulp and paper mill waste.
Why strength and toughness usually fight each other
Strength and toughness sound like the same thing. They are not.
Strength is how much stress a material carries before it fails, the peak of the stress-strain curve. Toughness is how much energy it absorbs before it fails, the area under that curve. A ceramic is strong and not tough. A rubber band is tough and not strong. Getting both at once needs a material that carries enormous load and dissipates enormous energy, which usually means two mechanisms working against each other inside one network.
Thermoplastic polyurethanes are built from soft segments (usually a polyether or polyester macrodiol) and hard segments formed when a diisocyanate reacts with a chain extender. The hard segments cluster into domains that act as physical crosslinks and reinforcing filler, and the soft segments supply elasticity. This microphase separation is why TPU is so versatile.
It also explains why the trade-off is so stubborn. Most attempts to escape it tighten one bolt and loosen another:
- Linear chain extenders with dense hydrogen bonding, acylhydrazides and similar motifs, push tensile strength past 75 MPa, but toughness stays near 390 MJ/m³.
- Dynamic covalent bonds add real toughness and even self-healing, while tensile strength rarely clears 40 MPa.
- Branched extenders add topological constraints that usually drag overall mechanical performance down.
- Bio-derived extenders such as castor oil and tannic acid improve the sustainability story and introduce dynamic interactions, but strength stays below 50 MPa and toughness below 120 MJ/m³.
None of these strategies builds a genuine three-dimensional network capable of multiscale energy dissipation. They modify interactions along a chain, and they leave the network architecture alone.
What makes lignin different from a conventional chain extender
Lignin is the most abundant aromatic biopolymer on Earth, and its structure is unusual in ways that turn out to matter for elastomers.
A typical chain extender is a small, flexible, linear molecule with two reactive ends. Lignin is nothing like that. It carries a rigid three-dimensional aromatic skeleton instead of a floppy aliphatic chain, a high density of hydroxyl groups (both aliphatic and phenolic), and π-conjugated aromatic motifs that stack and interact with each other.
In a TPU formulation that combination lets one lignin molecule do three jobs at once: multi-arm crosslinker, rigid reinforcing segment, and supramolecular interaction hub. The distance between a passive biofiller and that kind of network architect is the whole point of the approach.
Turning raw lignin into a reactive polyol
Technical lignin as isolated from biomass is not reactive enough, and not soluble enough, to make a good chain extender. The fix here is a one-step treatment with a recyclable ternary deep eutectic solvent (choline chloride, oxalic acid dihydrate, and glycerol), which extracts lignin from poplar wood and grafts glycerol onto it in the same operation.
The numbers tell the story:
| Property | Milled wood lignin | DES-extracted lignin polyol |
| Total hydroxyl content | 6.61 mmol/g | 10.92 mmol/g |
| Aliphatic OH | 5.17 mmol/g | 8.62 mmol/g |
| Phenolic OH | 1.44 mmol/g | 2.30 mmol/g |
| Number-average molecular weight (Mn) | n/a | 971 g/mol |
| Weight-average molecular weight (Mw) | n/a | 3,497 g/mol |
| Polydispersity (PDI) | n/a | 3.601 |
That moderate molecular weight matters as much as the hydroxyl count does. It keeps the polyol soluble and reactive enough to avoid phase heterogeneity during polymerization, which is where most lignin-polyurethane systems fall apart.
The base formulation uses PTMEG (Mn = 2,000 g/mol) as the soft segment, HMDI as the isocyanate, and a disulfide-containing dihydrazide chain extender. Lignin polyols then go in as an additional, spatially distributed dynamic extender.
Inside the network: what lignin actually changes
A denser, more dynamic hydrogen-bond environment
Infrared spectroscopy shows the hydrogen-bond environment is denser, and also reorganized.
In the carbonyl stretching region around 1,640 cm⁻¹, the lignin-containing elastomer shows a higher fraction of disordered hydrogen bonds than the control. That reads like a downgrade, and it is the opposite: the bulky, rigid lignin skeleton pries segments apart and breaks up overly crystalline hydrogen-bond domains, which is what you need if bonds are going to rupture and reform under load.
Meanwhile the hydroxyl stretching region near 3,300 cm⁻¹ intensifies substantially. Hydrogen-bond density and strength both go up.
Variable-temperature FTIR and two-dimensional correlation spectroscopy (2D-COS) map the sequence as the material heats. What emerges is a dissociation-then-reorganization mechanism: weaker disordered bonds let go first and absorb energy, while the stronger, more organized array holds the network together and later reforms. Sacrificial bonds of this kind are central to how the material handles load.
Network constriction and sharper microphase separation
The molecular redesign propagates up to the mesoscale in three measurable ways.
The hard segments stop moving. DSC puts the hard-segment glass transition at 59.8 °C in the control elastomer and 79.4 °C in the lignin version, a 20 °C jump. DMA is more dramatic: the soft-segment transition stays near −57 °C, which means lignin partitions selectively into the hard phase and leaves soft-segment mobility alone, while the hard-segment transition climbs from 22.1 °C to 96.4 °C.
Effective crosslink density rises by an order of magnitude. Calculated from the rubbery plateau storage modulus, it goes from 4.79 × 10² mol/m³ to 4.82 × 10³ mol/m³. This is an effective network density rather than a purely chemical crosslink count, since it includes covalent urethane linkages plus reversible physical associations such as hydrogen bonding, π-π stacking, and hard-domain aggregation.
The morphology gets bigger and cleaner without crystallizing. Wide-angle X-ray scattering shows only broad halos for both materials, so crystallization is ruled out as the cause of stiffening. Small-angle X-ray scattering shows the primary peak shifting from q = 0.6 to 0.5 nm⁻¹, which by Bragg’s law expands the interdomain spacing from about 10.5 nm to 12.6 nm. Steric bulk from the lignin skeleton pushes the domains apart while the hydroxyls reorganize local hydrogen bonding so the separation stays well defined.
Microscopy confirms it: SEM shows island-like hard domains dispersed in the soft matrix that are absent from the control, AFM shows surface roughness rising from Ra = 1.08 nm to 1.29 nm with crisp hard/soft alternation, and TEM resolves multi-layered hard-phase nanodomains.
The material stays a thermoplastic through all of this. Solubility tests put the soluble fraction at 99.93% in DMF and 99.90% in THF, so no continuous covalent network formed. The stiffening comes from lignin-mediated covalent incorporation plus reversible physical association, which is why the material can still be melted and reprocessed.
What the molecular simulations reveal
All-atom molecular dynamics and DFT calculations on representative hard-segment fragments quantify what the spectroscopy implies.
| Metric | Control TPUE | Lignin TPUE |
| Total potential energy | −7,350 kcal/mol | −7,665 kcal/mol |
| Intermolecular H-bonds (100-frame run) | 250 to 280 | 300 to 310 |
| Peak H-bond length | 2.0 to 2.1 Å | 1.8 to 1.9 Å |
| Cohesive energy density | about 410 J/cm³ | above 520 J/cm³ |
| Total cohesive energy | about 2,000 kJ/mol | about 2,500 kJ/mol |
DFT binding energies between hard-segment fragments make the anchoring effect explicit: a conventional extender-extender pair sits at −7.20 kcal/mol, an extender-lignin pair at −8.74 kcal/mol, and a lignin-lignin pair at −12.28 kcal/mol.
Two findings deserve emphasis. Shorter hydrogen bonds are stronger hydrogen bonds, so lignin raises both the number of noncovalent crosslinks and their individual strength. Second, and this runs against intuition, hydroxyls left unreacted after incomplete conversion are not inert defects. Simulations across several model configurations show they act as secondary physical crosslinking nodes and add to cohesive energy and toughness.
The numbers: 127.7 MPa and 1.43 GJ/m³
At 10 mol% lignin polyol loading, the elastomer reaches:
- Tensile strength: 127.7 ± 9.52 MPa
- Toughness: 1.43 ± 0.11 GJ/m³
- Elongation at break: roughly 3,000%
- True rupture stress: 3.72 GPa
- Tear energy: 790.85 ± 37.75 kJ/m²
- Puncture force: 63.862 N
For scale, a 0.1 g filament of this material hoists a 5 kg mass, 50,000 times its own weight.
On an Ashby chart of reported polyurethanes, this combination sits above petro-based linear extenders, petro-based branched extenders, plant-oil-based polyols, plant-based polyphenols, and commercial PU. Earlier lignin-based polyurethanes topped out below 100 MPa and below 1.0 GJ/m³.
Why 10 mol% is the sweet spot
Loading matters, and the curve is not monotonic.
Below 10 mol%, too few lignin junctions exist to form a continuous reinforcing network. At 10 mol% the network is continuous and hierarchical, and performance peaks. At 15 mol% performance falls: AFM shows surface roughness jumping to Ra = 11.8 nm and TEM shows extensive irregular hard-segment aggregates. Excess lignin-driven hydrogen bonding passes the dispersion limit, and those heterogeneous domains become stress concentrators that trigger early failure.
Control experiments are worth noting. Varying the conventional chain extender dosage up or down while holding the NCO/OH ratio constant made things worse in both directions, which rules out “more hard segment” as the explanation.
A second control is more telling. Erythritol and sorbitol, both purely aliphatic polyols, carry theoretical hydroxyl densities of 32.80 and 32.95 mmol/g, roughly three times the lignin polyol’s 10.92 mmol/g. If hydroxyl count were the whole story they should beat lignin. They reach only 37.35 MPa and 74.48 MPa.
The rigid three-dimensional aromatic architecture does the work, and hydroxyl density alone does not explain it. The aromatic skeleton disperses macroscopic stress spatially while the surrounding dynamic hydrogen bonds handle sacrificial rupture and reorganization.
Elastic recovery, damage tolerance, and notch insensitivity
High toughness usually costs elasticity. Here it largely does not.
Cyclic tensile testing shows moderate hysteresis on the first loading-unloading cycle, from disruption of weak disordered hydrogen bonds, then near-negligible hysteresis from the second cycle onward. The network settles into a more stable configuration and stays there.
| Strain level | Stress retention after 5 cycles | Elastic recovery |
| 500% | 84.5% | 97.4% |
| 1000% | strong recovery maintained | n/a |
| 1500% | strong recovery maintained | n/a |
| 2200% | 38.1% | 61.0% |
The strain-dependent behavior at extreme deformation is a textbook Mullins effect, and it works in the material’s favor. Stretching to 22 times original length without rupture requires a large population of noncovalent interactions to dissociate and soak up energy, which protects the covalent backbone from catastrophic scission. Even fractured and rejoined pieces stretched near breaking elongation show only about 20% residual strain.
The material is also markedly notch-insensitive. Growing cracks get blunted by the dynamic network, so a notched sample behaves far more like an unnotched one than a conventional high-strength elastomer would.
Still processable: casting, melt spinning, and 3D printing
High-performance networks usually cannot be reprocessed. This one can, because thermoplasticity here comes from the absence of a densely crosslinked permanent covalent network rather than from the absence of covalent bonds.
The material can be solution-cast, melt-processed, and melt-spun into uniform fibers, and it prints into complex load-bearing geometries. Temperature-dependent stress relaxation explains why. At the melting zone temperature of 120 °C, stress relaxes rapidly, so the dynamic network rearranges quickly and the melt flows well. Even at 95 °C, the cooler spinneret exit temperature, roughly 80% of internal stress dissipates within 1 minute, with full relaxation in 5 minutes.
That fast relaxation bleeds off residual shear stress accumulated during extrusion, which prevents melt fracture while leaving enough melt strength for continuous fiber drawing.
From elastomer to composite: carbon fiber and closed-loop recycling
A strong, tough, processable matrix is what fiber-reinforced composites need. Carbon fiber fabric laminates made by solution casting show:
| Property | Pristine CF fabric | Control TPUE/CF | Lignin TPUE/CF |
| Tensile strength | about 305 MPa | about 510 MPa | 913.7 MPa |
| Young’s modulus | n/a | n/a | 3.68 GPa |
| Max tear load | n/a | n/a | 130.2 N |
| Puncture resistance | n/a | n/a | 147.3 N |
That is roughly triple the pristine fabric and about 400 MPa above the control composite. SEM confirms complete impregnation and a smooth interface, so the stiff-but-tough matrix transfers load into the fibers instead of debonding.
The effect is not limited to carbon fiber:
- Glass fiber: 108 MPa to 604 MPa
- Kevlar fiber: 301 MPa to 1,249 MPa
Closed-loop recyclability
Because the network is supramolecular rather than permanently covalent, both the fiber and the matrix can be recovered. The elastomer dissolves in DMF at 80 °C over 4 hours, then the solution is cast into a PTFE mold and vacuum-dried at 75 °C for 12 hours.
After several cycles the reclaimed elastomer keeps more than 90% of its initial mechanical properties, and infrared spectroscopy shows the characteristic absorption peaks essentially unchanged. The dissolution-regeneration route does not attack the polymer backbone.
The design principle, generalized
The transferable lesson is about balancing two kinds of connectivity inside one network, and it is bigger than “add lignin.”
Covalent anchoring comes first. Lignin hydroxyls react with isocyanate to form urethane linkages, integrating lignin into the backbone and improving interfacial compatibility, so there is no phase-separated filler problem.
Reversible physical association comes second. The rigid aromatic structure restricts chain mobility, raises effective hard-segment content, and drives microphase separation. Hydrogen bonding between lignin hydroxyls and urethane groups, plus aromatic π-π interactions between lignin units, creates associations that dissipate energy under load and partially dissociate and reform at elevated temperature.
Strength comes from the dense dynamic crosslinks. Toughness comes from hierarchical hydrogen bonding and enhanced microphase separation, which provide large reversible energy dissipation. Processability comes from the physical network letting go when you heat it, rather than from the covalent backbone.
Different structural parameters govern those three contributions, so they can be tuned independently. That is what decoupling the strength-toughness trade-off means in practice.
The same logic should extend to epoxies, polyesters, and other polyurethane platforms through control of functional group reactivity, network density, and supramolecular interaction strength. It also points toward high-value use of an abundant aromatic biomass stream that industry currently burns for process heat.
Frequently asked questions
What is the strength-toughness trade-off in thermoplastic polyurethane elastomers? It is the observation that strategies which raise tensile strength usually reduce toughness, and vice versa. Conventional TPUEs have linear or weakly branched architectures, so tightening the network to gain strength removes the molecular mobility needed to dissipate energy, which makes the material brittle.
How does lignin improve strength and toughness at the same time? Lignin’s rigid aromatic backbone and dense hydroxyl groups let it act as a three-dimensional multifunctional chain extender. It anchors covalently into the polyurethane backbone while building a hierarchical, dynamic hydrogen-bond network. Under load, weaker bonds rupture sacrificially to absorb energy while the covalent network and stronger associations carry the stress.
What is the optimal lignin loading in a TPUE? About 10 mol%. Lower loadings fail to form a continuous reinforcing network. At 15 mol%, excessive hydrogen bonding causes irregular hard-segment aggregation (AFM roughness jumps to Ra = 11.8 nm), and those aggregates act as stress concentrators that degrade performance.
Is a lignin-modified thermoplastic polyurethane still recyclable? Yes. Solubility tests show a 99.93% soluble fraction in DMF and 99.90% in THF, confirming that no continuous covalent network formed. The material can be solution-cast, melt-processed, melt-spun, and 3D printed. In carbon fiber composites the matrix dissolves in DMF at 80 °C over 4 hours and can be re-cast, retaining over 90% of its mechanical properties across multiple cycles.
Does the improvement come only from having more hydroxyl groups? No. Erythritol and sorbitol have roughly three times the theoretical hydroxyl density of the lignin polyol (32.80 and 32.95 mmol/g versus 10.92 mmol/g) but reach only 37.35 MPa and 74.48 MPa. The rigid three-dimensional aromatic architecture is the decisive factor.
Are unreacted hydroxyl groups a defect in lignin-based polyurethanes? Molecular dynamics simulations across multiple model configurations show that exposed hydroxyl groups act as secondary physical crosslinking nodes. They contribute to heterogeneous physical entrapment, which promotes synergistic stress dissipation and adds to the cohesive energy.
Can lignin-based TPUE be used as a composite matrix? Yes. Carbon fiber laminates reach 913.7 MPa tensile strength and 3.68 GPa Young’s modulus, about triple the pristine fabric, with a tear load of 130.2 N and puncture resistance of 147.3 N. Glass fiber composites rise from 108 MPa to 604 MPa, and Kevlar composites from 301 MPa to 1,249 MPa.
Key takeaways
- Lignin works here as a network architect rather than a filler: multi-arm crosslinker, rigid segment, and supramolecular node in one molecule.
- A recyclable ternary deep eutectic solvent extracts and graft-modifies lignin in one step, raising total hydroxyl content from 6.61 to 10.92 mmol/g while keeping molecular weight low enough for homogeneous polymerization.
- The material reaches 127.7 MPa strength and 1.43 GJ/m³ toughness at 10 mol% loading, with about 3,000% elongation at break.
- Performance comes from balanced covalent anchoring and reversible physical association, not from hydroxyl density alone.
- It stays thermoplastic, 3D-printable, and closed-loop recyclable, and it lifts carbon fiber composite tensile strength to 913.7 MPa while allowing recovery of both fiber and matrix.
The same network that lets a 0.1 g filament hoist 5 kg also survives being dissolved and re-cast with more than 90% of its properties intact, a combination petro-based chain extenders have not managed.

