Beyond the recycling bin: can Milan recover more of its residual waste?
· 8 min read
When we think about recycling, attention usually focuses on households and businesses: separating packaging, collecting food waste and ensuring recyclable materials reach the correct bins. But what happens to the material left behind?
Even cities with advanced separate-collection systems generate substantial quantities of mixed, ‘residual’ waste. Residual waste often still contains potentially recyclable resources, yet it is frequently directed towards energy recovery rather than recycling, due to the difficulty and cost associated with post-consumer sorting.
However, regulatory pressure is increasing. The European Union is increasingly prioritising reuse and recycling over energy recovery through incineration (Potting et al. 2017), and is accelerating the transition to a circular economy through policies such as the Waste Framework Directive, which sets ambitious recycling targets for 2025, 2030 and 2035 (Directive (EU) 2018/851).
As residual waste continues to represent a significant portion of collected waste, exploring methods to treat it is increasingly becoming not an option, but an imperative. Our research explored whether technological innovation could help recover additional materials from residual waste, using Milan as a case study.
Milan provides an interesting starting point precisely because its separate-waste-collection system is already well established. According to data provided by AMSA, the city’s waste-management operator, Milan collects approximately 638,000 tonnes of municipal waste annually, of which 63.8% is collected separately. This leaves approximately 231,000 tonnes, or 36.2%, in the residual waste stream, which is subsequently directed towards energy recovery.
The analysis of the residual stream reveals substantial quantities of potentially recoverable materials. The largest identifiable streams represent paper and cardboard (30.6%) followed by plastic bottles and packaging of residual waste (19%). Not all of this material is necessarily suitable for recycling, particularly when contaminated or mixed with other waste, but its presence still represents an opportunity.
Milan’s existing separation system already works quite well, and our aim was not to redesign it, but to explore what could be recovered from the 231,000 tonnes of waste that remain. This led us to the following question:
Could an additional sorting stage recover materials before they are incinerated, and could the additional recovery deliver environmental benefits while remaining economically viable?
We began by evaluating and modelling different sorting technologies, drawing on the European Commission’s Best Available Techniques Reference Document for Waste Treatment, technical literature, specifications provided by manufacturers, and operational input from AMSA and A2A, focusing on several different technologies:
Each technology performs a separate function: Mechanical systems open waste bags and separate materials by size. Magnets and eddy-current separators recover ferrous and non-ferrous metals. Optical sorting technologies use spectral characteristics to identify materials, while AI-assisted robotic systems recognise and physically pick individual objects. Our initial assessments showed promising results, but also revealed a fundamental limitation: no single technology could effectively address the diversity of Milan’s residual waste.
We therefore shifted from evaluating individual machines to modelling an integrated sorting line using a mass-balance model. The logic is straightforward: each technology prepares the material for the next. Opening bags exposes previously enclosed objects, screening creates more manageable waste streams, and material-specific technologies subsequently recover selected fractions. At each stage, some waste is successfully sorted into processable material, while remaining waste continues down the sorting line.
To quantify these interactions in practical terms, we developed a model that follows individual material categories through successive processing stages, and projects their separation rate at each stage, based on factors such as the material size, potential contamination and the observed efficiency of the technology. This approach ensures that recovered material can be traced to its removal stage, and allows for potential changes in one variable, such as the fraction of plastic waste, or the efficiency of a technology, to influence the entire system.
Under the assumed operating conditions, an integrated sorting system could recover approximately 64,342 tonnes of material annually from Milan’s 231,000 tonnes of residual waste, corresponding to a potential recovery rate of 27.9%.
Paper and cardboard, together with plastic bottles and packaging, account for approximately 84% of the material projected to be recovered. This is not surprising considering their large share of the input material. Smaller contributions come from metals, textiles, rigid plastics and glass.
The model also distinguishes between physically separating an object and recovering material suitable for recycling. For instance, only 50% of sorted paper and cardboard and 25% of sorted textiles are assumed to qualify as ‘recovered’, as set out by AMSA. More importantly, all paper towels and plastic bags, representing a significant portion of residual waste, were set at 0% recoverable. These assumptions reflect downstream material-quality and acceptance constraints. Nevertheless, they suggest that meaningful material recovery may remain possible even after extensive separate collection.
Recovering materials can also reduce the environmental impacts associated with producing goods from virgin resources. Using material-property data from Ansys Granta, we estimated the potential energy, greenhouse-gas and water-footprint implications of the recovered quantities.
Our model identifies two potential sources of avoided CO₂ emissions: reducing the amount of material sent for incineration and substituting recycled materials for virgin production. The model projected that 77,700 tonnes of CO₂ could potentially be avoided by rerouting materials, particularly plastics, away from waste-to-energy combustion, and further avoided emissions of approximately 42,000 tonnes of CO₂ annually with regards to the substituting of virgin-material production with recycling. Combined, the model estimates approximately 119,700 tonnes of CO₂ per year in potential avoided emissions.
Material recovery also introduces an important energy trade-off. Diverting combustible waste from incineration reduces electricity generation from waste-to-energy plants. Therefore, the model constructed an energy bridge, showing the gross energy savings associated with using recycled material instead of producing virgin material, and then subtracting from that the energy generation foregone by not routing material to the waste-to-energy plant. Our model estimates approximately 479 GWh of annual gross primary-energy savings from recycling. After subtracting approximately 136 GWh of foregone waste-to-energy electricity, the estimated energy saving remains around 343 GWh annually. The fact that routing waste away from energy production through burning can still produce energy savings, provided that the material is recycled, speaks to the energy intensity of material production processes, and the further importance of using recycled material.
We also calculated a gross avoided virgin-production water footprint of approximately 59.7 million cubic metres annually, although this estimate is more uncertain because it depends heavily on material proxies for paper and textiles. Crucially, it should not be interpreted as water physically saved in Milan, but as an indication of potential avoided water demand associated with virgin-material production.
Although our findings are promising, translating them into an operational sorting facility would require further validation and testing. The model relies partly on assumptions where detailed Milan-specific information was unavailable. For example, particle-size distributions were adapted from ADEME’s French residual-waste dataset, while waste characteristics were modelled based on proxy materials. Furthermore, materials must meet downstream recycling requirements. Successfully separating a material does not guarantee that it can ultimately be recycled.
A broader concern remains regarding the economic viability of such a solution. While the technologies can effectively recover the materials, sufficient downstream recycling capacity is still required to process them at scale. In addition, recycled materials must be able to compete economically with virgin alternatives. Achieving this will likely require supportive policy measures and sustained political commitment to strengthen demand for secondary raw materials and create the conditions for a viable circular economy.
Our research suggests that the circular-economy discussion should not end once waste has been classified as residual. In Milan, our model indicates that nearly 28% of the remaining mixed waste could potentially be recovered through an integrated sequence of sorting technologies. The broader lesson is that effective residual waste management is possible, but it depends on the performance of the entire system, not individual machines.
For cities seeking to improve resource efficiency, residual-waste sorting deserves closer consideration not as a substitute for waste prevention and separate collection, but as an additional opportunity to retain valuable materials within the circular economy.
illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.
The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.
Gabriele Coppa
Circularity · Recycling
Michael Wright
Circularity · Environmental Sustainability
Michael Wright
Biodiversity · AI
Deutsche Welle
AI · Circularity
Procurement Magazine
Corporate Sustainability · Circularity
Waste Management Review
Circularity · Public Governance