Scalable laser micro- and nanostructuring of mould inserts for functional injection-moulded polymer surfaces

The paper demonstrates a scalable manufacturing route for functional polymer surfaces by combining high-speed laser structuring of steel mould inserts with injection moulding replication. The authors address the key limitation of industrial adoption—the low throughput of conventional single-beam laser texturing—by developing a selective process acceleration strategy that distinguishes between two structuring regimes and applies the most appropriate technique to each.

For deep microhole drilling, throughput was increased up to 20-fold by exploiting the high repetition rate (1 MHz) of an ultrashort-pulse fibre laser in single-beam mode. The paper states: "The productivity gain for deep microhole drilling follows directly from the repetition-rate increase. A reference process at 50 kHz requires 2.0 s to deliver 100 000 pulses, whereas the IPG source delivered the same pulse count in 0.10 s at 1 MHz. This corresponds to a nominal 20-fold reduction in irradiation time per hole without beam splitting."

For nanostructuring by laser-induced periodic surface structures (LIPSS), a line-beam shaping approach using a spatial light modulator enabled a 35-fold productivity gain, reaching processing speeds exceeding 100 cm2 min−1, while maintaining sub-micrometre fidelity. The paper reports: For nanostructuring, the Perla laser combined with line-beam shaping produced LIPSS at 105 cm2 min−1, approximately 35× faster than the 3 cm2 min−1 single-beam process, while maintaining sub-micrometre fidelity.

The structured mould inserts were used for injection moulding of polypropylene (PP), PA66 and ABS parts. Replication experiments confirmed successful transfer of micro- and nanostructures, with PP showing the highest fidelity. The paper notes: "For PP, replication of microholes resulted in round micropillars with a height of 25 µm, not reaching the full depth of the 100 µm hole... Overall, PP exhibited the best LIPSS replication quality, followed by PA66 and ABS."

Vacuum-assisted injection moulding further improved feature height by 56–283% across materials. Specifically: from 25 µm to 39 µm for PP (+56 %), from 13 µm to 23 µm for PA66 (+77 %), and from 3 µm to 11.5 µm for ABS (+283 %).

Functional evaluation demonstrated that all laser-textured PP surfaces increased the static water contact angle relative to the untreated reference (108°), with values up to 134° in the Wenzel regime. The paper states: "Square-mesh structures achieved the highest contact angle (133.8°). Pillar microstructures increased the contact angle by 14.5° to 122.5°. Nanostructures alone had a minor effect, increasing the contact angle by 4° on flat surfaces and 2° on top of microstructures. All droplets exhibited sliding angles >80°, indicating a Wenzel wetting state."

Structured PA66 surfaces reduced bacterial retention by up to 99.8% for E. coli and 90% for S. aureus compared to unstructured controls. The paper notes: "A reduction of up to 99.8% for E. coli and approximately 90% for S. aureus was measured for textures combining micropillars and LIPSS... The lower reduction observed for S. aureus (90%) compared to E. coli (99.8%) is consistent with the attachment point theory described by Lutey et al. However, the authors caution: Preliminary tests of bacterial adhesion were performed... The experiments were conducted on a limited number of samples and therefore the results should be regarded as indicative rather than conclusive."

Lap-shear testing showed that laser textures increased the joint shear strength of PP bonded with a non-optimised adhesive by up to 30-fold. The paper reports: "When an adhesive optimised for the given material was used (ABS bonded with an ABS-specific adhesive), the shear strength remained relatively stable across the various laser-treated geometries... In contrast, when no dedicated adhesive was available (polypropylene bonded with the same adhesive as used for ABS to simulate a non-optimal situation), the effect of the texture on joint strength was much more pronounced. In this case, shear strength improvements of up to 30-fold were recorded compared with the reference."

The paper concludes: "This study demonstrates that high-precision micro- and nanostructuring of steel mould inserts for injection moulding can be achieved with industrially relevant throughput by combining ultrashort-pulse laser processing with advanced beam-shaping strategies... By combining sub-micrometre precision with high-throughput processing and compatibility with standard injection moulding, this work bridges the gap between laboratory-scale laser structuring and industrial production of functional plastic components."

Improvements to AI Systems:

1. Multi-Physics Process Optimization Agent

• Improvement: Train a reinforcement learning model on the dual-regime laser structuring parameters (repetition rate, beam shaping, pulse count) to dynamically select optimal processing strategies for arbitrary target geometries.

• Capability: The AI can automatically generate a manufacturing plan that balances throughput (e.g., 20× faster drilling vs. 35× faster LIPSS) and fidelity, given a user-specified micro/nanostructure design and material.

2. Predictive Replication Fidelity Model

• Improvement: Build a neural network that learns the mapping from mould insert topography (hole depth, LIPSS periodicity) and injection moulding conditions (material type, vacuum assistance) to final part feature height and shape fidelity, using the paper’s data (PP: 25→39 µm, PA66: 13→23 µm, ABS: 3→11.5 µm).

• Capability: The AI can predict replication quality for new material–geometry combinations and recommend process adjustments (e.g., vacuum pressure, mould temperature) to achieve target feature heights, reducing trial-and-error in manufacturing.

3. Wetting-State Classifier and Surface Designer

• Improvement: Use the contact angle data (108° baseline, 133.8° max, Wenzel regime) to train a generative model that proposes surface texture patterns (pillar, square-mesh, LIPSS) for a desired hydrophobicity or sliding angle.

• Capability: The AI can design a mould texture that achieves a specified contact angle (e.g., >130°) while predicting whether the wetting state will be Wenzel or Cassie-Baxter, enabling tailored anti-fouling or self-cleaning surfaces.

4. Antibacterial Surface Optimization via Multi-Objective AI

• Improvement: Develop a Bayesian optimization framework that combines the bacterial retention reduction data (99.8% for E. coli, 90% for S. aureus) with geometric parameters (micropillar spacing, LIPSS orientation) to maximize antibacterial efficacy while minimizing manufacturing time.

• Capability: The AI can output a Pareto front of texture designs balancing bacterial reduction, production speed, and mechanical durability, and flag confidence intervals based on the paper’s limited-sample caveat.

5. Adhesive Joint Strength Predictor

• Improvement: Create a transfer-learning model that uses the lap-shear data (30× improvement for non-optimized adhesive, stable for optimized adhesive) to predict joint strength as a function of texture geometry, adhesive type, and polymer compatibility.

• Capability: The AI can recommend the optimal laser texture for a given adhesive–polymer pair, or identify when texture is unnecessary (e.g., when a dedicated adhesive already provides high strength), saving processing time and cost.

6. Industrial-Scale Process Digital Twin

• Improvement: Integrate the throughput data (1 MHz drilling, 105 cm2/min LIPSS) into a digital twin simulator that models the entire pipeline—laser structuring, mould replication, and functional testing—to optimize production line scheduling and yield.

• Capability: The AI can simulate “what-if” scenarios (e.g., changing laser source, mould steel grade, or injection pressure) and predict final part performance (contact angle, bacterial adhesion, shear strength) in real time, enabling rapid prototyping of manufacturing cells.

7. Uncertainty-Aware Material Recommendation System

• Improvement: Implement a probabilistic model that accounts for the paper’s explicit caveats (e.g., limited bacterial adhesion samples, PP not reaching full hole depth) to output confidence intervals for any predicted surface property.

• Capability: The AI can flag low-confidence predictions (e.g., ABS replication fidelity) and suggest additional experiments or alternative materials, improving reliability in industrial deployment.

Functional polymer surfaces with tailored wettability, antibacterial and adhesion properties are increasingly required in medical, packaging and consumer applications. Laser structuring of steel mould inserts followed by injection moulding offers a scalable manufacturing route, but conventional single-beam texturing has limited throughput. Here, we present a selective acceleration strategy that applies different laser techniques to micro- and nanostructuring. Deep microhole drilling was accelerated up to 20-fold by operating an ultrashort-pulse fibre laser at a repetition rate of 1 MHz in single-beam mode. For laser-induced periodic surface structures (LIPSS), line-beam shaping with a spatial light modulator increased productivity by 35-fold, reaching processing speeds above 100 cm squared min-1 while retaining sub-micrometre fidelity. Replication experiments with polypropylene (PP), PA66 and ABS confirmed successful transfer of micro- and nanostructures, with PP showing the highest fidelity. Vacuum-assisted injection moulding increased replicated feature height by 56--283 percent. All laser-textured PP surfaces showed higher static water contact angles than the untreated reference, reaching approximately 134 in the Wenzel regime. Structured PA66 surfaces reduced bacterial retention by up to 99.8 percent for E. coli and approximately 90 percent for S. aureus. Laser texturing also increased the shear strength of PP joints bonded with a non-optimised adhesive by up to 30-fold. This approach provides a practical, coating-free route to functional polymer components and helps bridge the gap between laboratory laser texturing and industrial injection moulding.