Nanocellulose-Reinforced Thermoplastics for Melt Material-Extrusion Additive Manufacturing: A Bibliometric Map and Filament-to-Part Evidence Audit

Preprint (version 1, not peer reviewed) prepared for submission to the Journal of Thermoplastic Composite Materials, with the data and code that support it. Many studies propose nanocellulose-reinforced thermoplastics as bio-based feedstocks for melt-extrusion additive manufacturing, yet no systematic study has examined whether the reinforcement benefit measured on the filament is retained during printing. We searched eight databases on 25 October 2026 and retrieved 13,483 records (6,712 unique). After dual independent screening (Cohen's κ = 0.76), full-text assessment of uncertain records, and an audit of the rule-excluded remainder, we retained 130 documents published from 2016 to 2026, 74 of which were published since 2023. Poly(lactic acid) is the matrix in 57 documents; cellulose nanofibrils (66) and cellulose nanocrystals (54) appear in nearly equal numbers. The VOSviewer keyword co-occurrence map (52 terms, five clusters) separates a melt-compounding core from clusters on nanocrystal dispersion and shape memory, degradable biomedical matrices, mechanical performance and filament characterization, and interlayer adhesion, with interlayer adhesion as a peripheral term. Only 22 abstracts mention both the filament and a printed object of the same formulation, and 12 signal interlayer or build-direction behavior. Reported tensile-strength gains, from +33% for filament to +84% for printed parts, refer to non-equivalent objects. We propose a minimum extraction set and a property-retention ratio to enable researchers to compare future studies across the feedstock-to-part chain. Files: the preprint (PDF); Supplementary_Material_JTCM.xlsx (Tables S1–S5, Data S1); and the data and code folder (search strings, harmonised record identifiers with screening status, both reviewers’ decisions and justifications, full-text assessment and audit records, the 130 included documents, the VOSviewer thesaurus, map and network files, and the Python analysis scripts with reproduce_from_deposit.py). Abstracts and database keyword fields exported from Scopus, Web of Science and the other platforms are not included because of the platforms’ licence terms.