A sweeping systematic review published in Discover Catalysis argues that nanocatalysts paired with Industry 5.0 digital technologies could reshape environmental remediation, allowing air, water and soil pollution to be tackled within a single, intelligent framework rather than as three separate problems. The review, led by Nima Alizadeh Raef, Siamak Pedrammehr and Fatemeh Hamedi and published on 1 October 2026, is open access and draws on a PRISMA 2020-compliant search of six scientific databases covering literature from January 2020 to March 2026.
The scale of the screening effort is considerable. The authors’ searches across Scopus, Web of Science Core Collection, ScienceDirect, Nature, IEEE Xplore and PubMed returned 1,928 citations. After duplicates and irrelevant records were removed, 378 papers passed title and abstract screening, and 196 studies met the inclusion criteria for qualitative synthesis. Of those, 138 papers contained enough quantitative data to feed a meta-analysis. To qualify, studies had to involve nanocatalysts used in experimental or pilot-scale processes incorporating some element of the Industry 5.0 approach, and to address pollutants in more than one environmental compartment or demonstrate multi-pollutant degradation.
What distinguishes this review from earlier, single-compartment surveys is its framing. The authors define
The authors define “multi-matrix remediation” as the recorded removal of pollutants in more than one environmental matrix, either simultaneously through a treatment train or in stages using the same catalyst platform. This is deliberately distinguished from “multi-pollutant removal”, which refers to removing different classes of pollutants simultaneously within a single matrix. The review also reserves the term “Industry 5.0-enabled nanocatalysis” for systems in which digital technology has been coupled, artificially or experimentally, to the functioning of the nanocatalysis, including monitoring or optimisation, and contrasts these with conceptually interoperable systems and those merely considered to have potential for artificial intelligence implementation.
On that point, the authors are candid about the state of the evidence. Most of the studies they examined concern distinct, consecutive or theoretical combinations of matrices rather than a demonstrated simultaneous treatment of air, water and soil in one container system. Quantitative catalytic performance has been relatively frequently documented, they note, while the integration of artificial intelligence and machine learning, continuous flow operation, multi-matrix testing and catalyst recovery schemes are not consistently reported across the literature. Their conclusions echo this caution: intelligent nanocatalysts such as single-atom, dual-atom and heterostructured systems have moved beyond tackling individual pollutants in a single medium, and a growing body of evidence points toward their use across air, water and soil applications, but largely evaluated separately rather than as demonstrated integrated multi-matrix systems.
The review identifies several design advances from 2022 to early 2026 as central to the field’s progress. Single atom catalysts and dual atom catalysts exhibit close to one hundred percent atom efficiency and allow fine tuning of electronic properties through coordination engineering. Iron-nitrogen four-coordinate sites and copper-iron dual active sites supported on nitrogen-doped carbon have been shown to be effective in activating molecular oxygen and peroxymonosulfate for decomposing volatile organic compounds and non-biodegradable organic compounds in aqueous solutions. Machine learning-based screening approaches have reportedly accelerated the search for suitable metal-carrier systems from years to months.
Hierarchical metal-organic framework materials represent another strand of progress. Through pyrolysis and post-synthesis modification, researchers have fabricated multi-layered materials combining microporous structures suited to selective adsorption with mesoporous and macroporous structures that enable rapid mass transport. These materials have performed successfully in gas, aqueous and soil phases. Bimetallic cobalt-nickel heterostructures formed from ZIF-67@ZIF-8 precursors have performed well in the simultaneous photocatalytic reduction of chromium(VI) ions and oxidation of organic dyes. In soil remediation specifically, MOF-derived cobalt-nickel-carbon catalysts showed relatively high efficiency in degrading polycyclic aromatic hydrocarbons in soils and in leachates containing heavy metals, with magnetic features that make recycling feasible.
Stimuli-responsive nanocatalysts form a third category. These incorporate photo-, electro-, magneto- and pH-responsive elements that allow on-demand and adaptively selective regulation of catalytic activity. One example highlighted is a core-shell structured titanium dioxide@MXene material with internalised magnetic nanoparticles, which provides both magnetic recycling and magnetic regulation of catalytic processes, making it suitable for use in digital twin-based process systems.
Computational methods feature prominently in the review’s account of how these materials are designed. Density functional theory calculations have been essential for elucidating the electronics of active sites and reaction mechanisms. For iron-copper dual metal atom catalysts, such calculations indicate a synergistic electronic interaction between neighbouring metal atoms that lowers the energy barrier for oxygen activation and pollutant adsorption by 0.8 to 1.2 electronvolts relative to single-atom analogues. Machine learning models built on calculated adsorption energies, adsorption strengths and other physical properties have predicted new material combinations with accuracy above eighty-five percent. The authors stress, however, that the predictive capability of these models is fundamentally constrained by the accuracy of the descriptors used, the representativeness of the data, model approximations and the need for experimental validation. They also caution that linear correlations reported from study to study, with coefficients above 0.91, are suitable only for the particular matched datasets for which they are reported and cannot be generalised across the broad set of heterogeneous catalysts covered by the review.
A distinctive contribution of the paper is a proposed quantitative indicator, the Synergistic Multi-Matrix Remediation Index, or SMMRI. The index combines normalised removal efficiencies in air, water and soil, a ratio of the average apparent rate constant across matrices to a reference rate constant for a benchmark catalyst, and a long-term stability factor defined as the number of cycles until twenty percent activity loss, each weighted by factors determined by expert judgment or an entropy method, with the weights summing to one. The authors argue that unlike removal-efficiency metrics applying to a single matrix, this composite measure would permit comparison of a catalyst’s efficiency across an array of matrices.
They are equally explicit that the index remains conceptual. It has not been proven experimentally against alternative benchmark catalysts, the uncertainty and sensitivity of the relative weightings have not been investigated, and the values presented in the review’s tables and figures simply illustrate what the index could be for particular catalysts rather than constituting independently proved values. No comparison to external catalyst data has been made, no sensitivity analysis of the weights has been performed, and no experiments using the index have been replicated between laboratories. The authors present it as a framework to guide future research, not as a metric tested against data, and call for validation studies specifically on the SMMRI.
The review devotes substantial attention to obstacles that could prevent the technology from reaching practical maturity. Chief among these is the trade-off between activity and stability: many single and dual atom catalysts offering high catalytic efficiency are thermodynamically unstable because of their unsaturated coordination state, leaving them susceptible to leaching, reconfiguration and poisoning by organic or inorganic species in soil and wastewater matrices. Scalability poses a further barrier, since most modern nanocatalysts are synthesised using elaborate laboratory approaches such as atomic layer deposition, controlled pyrolysis or solvothermal reactions that are difficult to apply economically at large scale, where preserving atomistic dispersion and structural homogeneity is challenging. High precursor costs and complicated syntheses keep many innovative materials unaffordable for industrial use.
Ecological and human safety concerns receive particular emphasis. The authors note that reported catalyst removal may refer to the removal of free metal ions and molecular fragments still attached to the support rather than the detachment of whole nanoparticles, whose movement, bioavailability and toxicity can differ from those of single particles, and that most of the studies cited provide no separate estimates for these two mechanisms. Understanding of the long-term fate of the nanomaterials, including agglomeration, oxidation, sulfidation and possible retention in soil and aquatic organisms, remains limited, as does knowledge of what happens after recovery. The authors argue that the total environmental benefit of large-scale nanocatalyst production and use should be judged through life-cycle assessment covering raw-material processing, manufacturing, use, recovery and disposal, not merely by contaminant removal effectiveness.
Digital integration introduces complications of its own. Digital twins and AI-enabled controllers depend on efficient and precise sensors and accurate modelling, yet some current nanocatalysts lack suitable sensors, cannot operate continuously, or are not designed for efficient functioning within edge-computing systems. The “black box” character of AI-optimised catalysts makes them difficult to validate. Standardisation is another gap: single-pollutant threshold approaches are ill-suited to representing the long-term behaviour of smart nanocatalysts, whose performance across different pollutants is coupled. The review also flags the poorly researched socio-technical dimension of Industry 5.0, noting that researchers tend to focus on technical aspects while neglecting operator training, human-machine interaction, socio-economic implications and equitable distribution of the technologies in developing countries.
Looking forward, the authors sketch a research roadmap ordered by feasibility. Within immediate reach, they place better catalyst design for single-atom, dual-atom and MOF-based systems and a systematic comparison protocol for MOF catalysts across multiple matrices. Autonomous, self-evolving catalyst designs and quantum-computing-aided discovery are positioned as achievable only after further advances across the intervening areas. They are particularly guarded about quantum computing, describing the combination of machine learning with quantum computing as a much longer-term and very speculative prospect, with no current proof of quantum advantage in practice for this type of materials simulation. Machine learning models describing pathway dynamics on gas-liquid-solid interfaces might, they suggest, make laboratory experiments less hit-or-miss, but realising the fuller vision would require quantum chemistry, robotics and active learning to mature and be linked successfully, which they judge unlikely within the next few years.
The review closes on a measured note. It concludes that complex nano-architectures deliver significant benefits in catalytic degradation efficiency and can tolerate some variability within individual matrices, but that demonstration of an integrated process removing pollutants from air, water and soil simultaneously in a single operating regime remains experimental. Overcoming the intertwined challenges of scalability, long-term stability, nanotoxicity and regulatory compliance, the authors argue, will require enduring collaboration across materials science, chemical engineering, environmental toxicology, data science and regulatory policy, with the next development step concerning not only catalytic action but also ease of synthesis, temporal stability, material recovery and sustainable lifecycle criteria. The dataset supporting the study is openly available in Zenodo.
Subject of Research: Industry 5.0-enabled nanocatalysts for multi-matrix environmental pollution remediation
Article Title: A critical review of industry 5.0-enabled nano-catalysts for synergistic air, water, and soil pollution abatement toward environmental sustainability
Article References: Raef, N. A., Pedrammehr, S., & Hamedi, F. (2026). A critical review of industry 5.0-enabled nano-catalysts for synergistic air, water, and soil pollution abatement toward environmental sustainability. Discover Catalysis, 3(1), Article 19. https://doi.org/10.1007/s44344-026-00050-x
Image Credits: AI Generated
DOI: 10.1007/s44344-026-00050-x
Keywords: nanocatalysts, Industry 5.0, environmental remediation, single atom catalysts, metal-organic frameworks, digital twins, artificial intelligence, air pollution, water treatment, soil remediation, systematic review, SMMRI
News Source: Bethany Barker. (October 28, 2026). Smart Nanocatalysts Promise Cleaner Air, Water and Soil, Review Finds. Scienmag.
