Resources & Publications

Vietnam discards hundreds of thousands of tonnes of durian rind every year. This study treats that waste as two products rather than one: a single integrated process recovers the flavonoids first, then converts the residual husk into carboxymethyl cellulose and casts it into a hydrogel — so nothing is thrown away between steps. Citric acid at 10% extracted 281.33 mg QE/g of flavonoids, and the remaining biomass yielded cellulose at 27.10% with 93.39% purity. The resulting hydrogel releases its payload steadily over 24 hours and is bioactive in its own right: antioxidant, anti-inflammatory and antibacterial against three pathogens. The paper closes with a costing, which is the part most valorisation studies leave out.

• One process, two products: flavonoids and cellulose are co-recovered from the same durian rind, with the residue from the first step feeding the second rather than becoming waste.
• Citric acid at 10% was the optimal extraction medium — 281.33 mg QE/g of total flavonoids, and cellulose recovered at 27.10% yield with 93.39% purity.
• The carboxymethyl cellulose hydrogel released 65.34% of its load over 24 hours, a controlled profile rather than a burst.
• Bioactive as made: antioxidant IC₅₀ of 55.75 µg/mL (DPPH), anti-inflammatory IC₅₀ of 610.70 µg/mL (BSA denaturation), and 10–14 mm inhibition zones against Staphylococcus aureus, Bacillus subtilis and Escherichia coli.
• Includes an economic evaluation of the integrated route — the step that decides whether a valorisation process ever leaves the laboratory.
• Dr Nguyễn Hoàng Chinh (Deakin University) is the senior (last) author of the nine-author team, which spans Ton Duc Thang University, Deakin and Texas A&M.
• Open access under CC BY-NC-ND — free to read.

A materials answer to the problem that has stalled metal–organic frameworks for years: MOFs capture microplastics superbly, but arrive as loose powders that cannot hold their structure inside a working filter. This study locks ZIF-8 into a double-network hydrogel through UV-initiated cross-linking, turning a laboratory powder into a membrane water can actually be run through. In continuous-flow column tests on laundry wastewater — among the largest sources of microplastic pollution — the optimised composite removed 97% of PET and 98% of PE, against the 32.6% and 76.5% reported by earlier MOF studies, and still removed 96% after six reuse cycles.

• 97% PET and 98% PE removed from real laundry wastewater; 96% retained after six reuse cycles, with stable operation for about four hours.
• ZIF-8.1@H6mg — ZIF-8 embedded in a PVA/PEGDA double-network hydrogel, photo-cross-linked under UV in ten minutes, then tested in a 30 mm continuous-flow filtration column.
• The gain over earlier MOF work (32.6–76.5%) comes from making the powder structurally deployable, not from a better sorbent — the removal mechanism is traced to the interpenetrating porous network, zeta potential and surface functional groups.
• Lưu Thị Trang is first author: she led conceptualisation, methodology and formal analysis, and wrote the original manuscript. AURORA 2025 Third Prize winner; PhD candidate at UTS and lecturer at Hanoi Architectural University.
• Open access — free to read and download from Springer Nature.

A dual-function breakthrough in cancer bioelectronics: this flexible, implantable silicon carbide (SiC) multi-electrode array both listens to cancer cells and eliminates them. Leveraging SiC's wide-bandgap stability and biocompatibility, the device delivers near-unity signal transmission from 0.1–100 Hz even after 2,000 bending cycles, records stable multichannel activity from breast cancer cells (MDA-MB-231), and triggers targeted irreversible electroporation to ablate them — without relying on heat. In vivo testing confirmed no inflammation, fibrosis or adverse tissue response, positioning flexible SiC electrode arrays as a platform for precision-targeted cancer monitoring and treatment in a single device.

A breakthrough in computational biology: the authors fuse attention-weighted multi-view graph convolutional networks with a variational graph autoencoder and a gated MLP to predict lncRNA–disease associations with striking accuracy (AUC 0.9584, AUPR 0.9590). By unifying lncRNA, disease and miRNA similarity networks into one heterogeneous graph, the model tackles data sparsity and class imbalance head-on — offering a powerful new tool for prioritising disease-linked lncRNA candidates and accelerating therapeutic discovery.

A powerful step forward for precision oncology: this framework fuses clinical data, whole-slide histopathology images and gene expression profiles through a transformer-based multimodal architecture to predict breast cancer receptor status (ER, PR, HER2) — the key markers that guide treatment choice. Trained on 937 TCGA-BRCA cases and externally validated on CPTAC-BRCA, the model achieves outstanding AUCs of up to 0.96 while remaining biologically interpretable, offering a faster, more consistent complement to conventional IHC/ISH testing.

A landmark contribution to climate economics: this study is among the first to prove that not all heatwaves hit inflation the same way. Using a structural VAR to disentangle short-lived weather shocks from persistent climate shocks, then tracing their effects through smooth local projections, the authors reveal sharply heterogeneous inflation impacts across Australian states, seasons and spending categories — a finding with direct implications for central bank forecasting and climate-adjusted monetary policy.

An analysis of Australia's semiconductor ecosystem — strengths, investment environment, research institutions and strategic partnerships — with actionable recommendations for how Vietnam can leverage the bilateral relationship to accelerate its own semiconductor capability and supply chain integration.

A policy-oriented study examining Australia's green transition experience across energy, finance, regulation and industry — drawing practical lessons for Vietnam's own ESG agenda, green finance frameworks and sustainable development strategy in the context of accelerating global climate commitments.

A cutting-edge fusion of AI and finance: the authors deploy GPT-4o-mini and Gemini-1.5-flash to score sentiment across 18,253 economic news articles, then feed the results into an XGBoost model that predicts VN-Index movements with an outstanding R² of 0.9879. Explainable-AI tools (SHAP and LIME) pinpoint exactly which forces move Vietnam's stock market hardest — earnings and dividend news, foreign investment policy, and political stability — offering investors and policymakers an unusually precise, transparent read on market drivers.

A powerful behavioural-finance breakthrough: mining over 80,000 Vietnamese news articles with large language models, the authors uncover a striking inverted U-shaped relationship between investor negativity and firm distress — risk climbs with bad news, then paradoxically falls beyond a tipping point. Grounded in prospect theory and regret theory, and validated across 80 distressed listed firms (2010–2022), the model sharpens early-warning prediction well beyond traditional financial ratios alone.

The only publication of its kind: three years of Vietnam–Australia smart agriculture knowledge (2021–2023), compiled into one authoritative yearbook under the patronage of Vietnam's Ministry of Agriculture and Rural Development — with support from the Ministry of Planning and Investment, the Ministry of Science and Technology, the Vietnamese Embassy in Australia and the State Committee for Overseas Vietnamese Affairs.

• 15 in-depth articles by 20+ leading Vietnamese and international experts — from Texas A&M University, Microsoft AI research, and the former Director of Vietnam's Southern Horticultural Research Institute
• Gene technology in crop breeding, AI · IoT · Big Data farming, and 9 Agriculture 5.0 technology trends
• The complete smart agri-food value chain — from farm to table: processing, branding, export quality and cross-border digital transformation (2023–2033 outlook)
• Green and sustainable finance solutions for Vietnamese agriculture from HDBank

A landmark gas-sensing innovation registered at Sungkyunkwan University: by fusing 2D reduced graphene oxide with 3D zinc oxide nanorods on an AlGaN/GaN heterojunction, this patented Schottky diode sensor detects NO₂, SO₂ and formaldehyde at room temperature down to just 120 parts per billion — with response times as fast as 2 minutes. Ultra-low power and highly sensitive, the design is built for the next generation of IoT air-quality monitors, industrial safety systems and environmental sensing networks, backed by Korea's Nanotechnology and Materials Technology Development programme.

A rare double achievement in materials engineering: this registered Korean patent, developed at Sungkyunkwan University, introduces a single field-effect-transistor sensor whose gate insulator blends piezoelectric materials (P(VDF-TrFE), BaTiO₃, PbTiO₃) with ferromagnetic materials (CoFe₂O₄, Fe₃O₄, NiFe₂O₄, BaFe₁₂O₁₉) — allowing one compact device to measure pressure and magnetic field simultaneously and extract a magnetoelectric coupling coefficient in real time. Backed by Korea's Global Frontier R&D Program, the innovation points toward smaller, smarter multi-sensing chips for next-generation electronics and biomedical devices.