Cellulose is hydrophilic. Most engineering thermoplastics are hydrophobic. Put them together without a compatibilizer and you get a composite where the reinforcement barely talks to the matrix. The nanocrystals are physically present in the polymer, but the interface between them transfers load poorly. Flexural modulus improves because the stiff crystals resist deformation regardless of interface quality. But flexural strength and impact resistance — properties that depend on the interface transferring stress from matrix to reinforcement — remain disappointing.
The Interface Problem in Quantitative Terms
We measured the effect of interface quality by comparing untreated NC-200 with silane-treated NC-200, both at 22 wt% in PA6. Untreated NC-200/PA6 flexural strength: 116 MPa. Silane-treated NC-200/PA6 flexural strength: 142 MPa — a 22% improvement. Flexural modulus changed minimally: 5.7 GPa untreated versus 5.9 GPa treated. The modulus insensitivity confirms that modulus is a bulk property driven by crystal stiffness, while strength depends on how well the crystal-matrix interface transmits stress before failure.
The 22% strength gain is significant enough to move NC-200/PA6 from a "not quite good enough" candidate to a viable selection for semi-structural automotive parts. Several of our qualification programs started with untreated NC-200, failed the customer's flexural strength specification by 10-15%, and were re-evaluated with the silane-treated grade. The treated grade passed. Surface treatment literally determines whether the material qualifies or does not.
What Silane Coupling Agents Do
A silane coupling agent is a bifunctional molecule. One end — the alkoxy groups (typically methoxy or ethoxy) — reacts with hydroxyl groups on the cellulose surface through hydrolysis and condensation, forming covalent Si-O-C bonds. The other end — an organic functional group chosen to match the matrix polymer — provides chemical compatibility with the thermoplastic. For PA6, we use aminopropyltriethoxysilane (APTES), where the amine group reacts with PA6 carboxyl end groups during melt processing to form amide linkages.
The result is a molecular bridge: cellulose-O-Si-CH2CH2CH2-NH-CO-PA6. This bridge transfers mechanical load from the matrix to the reinforcement through covalent bonds rather than relying on weak van der Waals interactions at an uncompatibilized interface. The bond strength of a covalent Si-O-C linkage is approximately 450 kJ/mol, compared to roughly 5-10 kJ/mol for van der Waals attraction between a cellulose hydroxyl and a PA6 amide group. The covalent bridge is 50-100x stronger at the molecular level.
Our Treatment Protocol
We treat CNC in the aqueous paste stage, before compounding. The 25 wt% CNC paste is diluted to 5 wt% in deionized water. APTES is added at 3 wt% relative to dry CNC mass — for a 1 kg dry-CNC batch, that is 30 grams of APTES. The mixture is stirred at 60°C for 2 hours, which allows the APTES ethoxy groups to hydrolyze (forming silanol) and condense onto the cellulose surface hydroxyl groups.
After treatment, the suspension is washed once by centrifugation and re-dispersion in deionized water to remove unreacted APTES. The washing step is important — excess silane in the compound acts as a plasticizer and reduces modulus. We verify treatment success by FTIR spectroscopy, looking for the characteristic Si-O-C stretching peak at 1050 cm⁻¹ and the amine N-H bending peak at 1560 cm⁻¹ on the treated CNC surface.
The treated paste is then reconcentrated to 25 wt% and fed into the twin-screw extruder using the same compounding parameters as untreated CNC. The silane treatment does not change the compounding process — it only changes the surface chemistry of the CNC entering the process. Screw profile, barrel temperatures, feed rates, and throughput are identical between treated and untreated grades.
Why We Chose APTES Over Other Silanes
The silane coupling agent market offers hundreds of chemistries. We evaluated five candidates for CNC/PA6: APTES (aminopropyltriethoxysilane), GPTMS (glycidoxypropyltrimethoxysilane), MPTMS (methacryloxypropyltrimethoxysilane), VTMS (vinyltrimethoxysilane), and HDTMS (hexadecyltrimethoxysilane).
APTES produced the highest flexural strength improvement (22%) and the best retention of impact resistance. GPTMS was second at 18% flexural strength improvement, but the epoxy functional group showed some crosslinking with the PA6 matrix that increased brittleness. MPTMS and VTMS showed minimal improvement (5-8%) because their organic functional groups (methacrylate and vinyl) do not react with PA6 during melt processing. HDTMS made the CNC hydrophobic — improved dispersion in PP — but the long alkyl chain did not provide reactive compatibility with PA6.
The choice is matrix-dependent. For polypropylene matrices, we use a maleic anhydride grafted PP (MAPP) compatibilizer rather than a silane, because the anhydride group reacts with cellulose hydroxyl groups more effectively than any silane in a non-polar matrix. MAPP is added during compounding at 5 wt% relative to CNC loading. For polyester matrices (PET, PBT), GPTMS is the preferred silane because the epoxy functional group reacts with polyester hydroxyl end groups.
Cost Impact of Surface Treatment
APTES costs approximately $35-45 per kilogram in drum quantities. At 3 wt% treatment level on CNC, the APTES cost per kilogram of dry CNC is approximately $1.10-1.35. For NC-200/PA6 pellets at 22 wt% CNC loading, the silane treatment adds approximately $0.25-0.30 per kilogram of compound.
The dilution, treatment, washing, and reconcentration steps add approximately 8 hours of processing time per batch compared to the untreated route. At our current scale, this adds roughly $0.40 per kilogram of compound in labor and equipment time. Total cost increase for silane-treated NC-200/PA6: approximately $0.65-0.70 per kilogram, or roughly 8-10% above the untreated grade price.
Whether the cost is justified depends on the application. For parts that pass strength specifications with the untreated grade, the surface treatment is unnecessary cost. For parts that need the 22% strength improvement to qualify, the $0.65/kg premium is trivial compared to the alternative — switching to a more expensive reinforcement material or redesigning the part geometry.
Quality Control for Treated CNC
We verify surface treatment on every batch using three methods. First, FTIR spectroscopy as described above — the Si-O-C and N-H peaks must be present and above threshold intensity. Second, water contact angle on a pressed CNC film — untreated CNC has a contact angle of approximately 25° (very hydrophilic); APTES-treated CNC has a contact angle of 55-65° (moderately hydrophobic). Third, the flexural strength of a compound made from the treated CNC must fall within our specification range of 135-155 MPa at 22 wt% loading in PA6.
If any of the three checks fails, the batch is quarantined and re-treated. In practice, batch failures are rare — approximately 1 in 20 — and are almost always caused by insufficient treatment time due to temperature controller malfunction in the treatment vessel. The 60°C requirement is critical; at 40°C, APTES hydrolysis is too slow and the 2-hour treatment time is insufficient for complete surface coverage.
For customers interested in the silane-treated grade, request NC-200S (the "S" designates silane-treated) when ordering sample kits. Both treated and untreated grades are available for side-by-side evaluation. Contact info@soarceusa.org for pricing and availability.