How to quickly characterize construction and demolition wastes? Traditional and advanced portable solutions in comparison

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE How to quickly characterize construction and demolition wastes? Traditional and advanced portable solutions in comparison Alessandra Mobili Dept. of Materials, Environmental Sciences and Urban Planning Università Politecnica delle Marche Ancona, Italy [email protected] Maria Teresa Calcagni Department of Industrial Engineering and Mathematical Sciences Università Politecnica delle Marche Ancona, Italy [email protected] Gian Marco Revel Department of Industrial Engineering and Mathematical Sciences Università Politecnica delle Marche Ancona, Italy [email protected] Jacopo Donnini Dept. of Materials, Environmental Sciences and Urban Planning Università Politecnica delle Marche Ancona, Italy [email protected] Gloria Cosoli Department of Industrial Engineering and Mathematical Sciences Università Politecnica delle Marche Ancona, Italy [email protected] Faculty of Engineering Università Telematica eCampus Novedrate, Italy [email protected] Simona Sabbatini Dept. of Materials, Environmental Sciences and Urban Planning Università Politecnica delle Marche Ancona, Italy [email protected] Francesca Tittarelli Dept. of Materials, Environmental Sciences and Urban Planning Università Politecnica delle Marche Ancona, Italy [email protected] Giovanni Salerno Department of Industrial Engineering and Mathematical Sciences Università Politecnica delle Marche Ancona, Italy [email protected] Elena Leoni Dept. of Agricultural, Food and Environmental Sciences Università Politecnica delle Marche Ancona, Italy [email protected] Valeria Corinaldesi Dept. of Materials, Environmental Sciences and Urban Planning Università Politecnica delle Marche Ancona, Italy [email protected] Abstract—Nowadays it is pivotal to develop innovative solutions fostering sustainability and circularity in the construction sector. This paper presents a comparison between a traditional method (e.g. FTIR) and two advanced portable solutions, namely MicroNIR and HSI, to characterize materials. The former is based on Near Infrared spectroscopy and is commonly used for materials characterization; conversely, the latter is based on image analysis. In the framework of the European project RECONSTRUCT, these techniques will be applied to construction and demolition waste (CDW) field with the aim of developing a classification database, and therefore a library, that can be used for the management of wastes in a fast, reliable, and cost-effective way in view of extending sustainability and circularity models also to the construction sector, with the final aim of CDW valorization. Keywords—CDW, FTIR, NIR, HSI, sustainability, building life cycle, waste valorization, measurements. I. INTRODUCTION Construction and demolition waste (CDW) accounts for most of the waste flux in Europe [1]. In 2018, it has been estimated that one third (36%) of the total wastes produced in the European Union by economic and households activities (approximately 2.317 million tons) came from the construction sector [2]. CDW is generated by construction and demolition of buildings and infrastructures as well as road maintenance works [3]. When excluding soils, it comprises a wide range of materials, including concrete, bricks, wood, glass, metals, and plastics [2]. According to the Waste Framework Directive (WFD) 2008/98/EC, EU member states were obliged to reach 70% by mass of CWD recovery by 2020 [4]. In time, EU countries increased their recovery rate and some of them exceeded this target already in 2016 [1], avoiding the disposal in terms of incineration and landfilling by reintroducing wastes into the economy. CWD disposal through landfilling and stacking can lead to several environmental issues such as depletion of space and noise, air, and water pollution [5]. Moreover, landfilling can cause tragic accidents and can affect the health of residents living nearby [6]. Although many countries established new markets for recovered CDW, the European construction sector remains anything but circular. Circular economy deviates from the socalled “take-make-consume-dispose” linear economic model which cannot be considered sustainable seeing as it assumes that natural resources are available, abundant, easy to source, and cheap to dispose of. Indeed, circular economy envisages a restorative process of maintaining the utility and the value of a certain material or component for as long as possible. In this way, new virgin materials and energy demand are minimized, while reducing extraction procedures, emissions generation, and waste management, lowering in turn the environmental impact [7]. By closely looking at data, it is evident that in some EU member states high rates of CDW recovery entail backfilling or low-grade applications [1]. Most of CDW is utilized for public works in constructions nonoccupied by humans such as roads, storage facilities, and nonurban applications, giving that current demolition practices produce streams lacking high-purity levels. Therefore, it is of utmost importance that the construction sector takes a step forward in CDW recycling, overcoming downcycling procedures and moving to more sustainable upcycling processes. To do that, innovative recycling methodologies should be pursued in order to reintroduce into the loop highquality and high-purity wastes as secondary raw materials that can give an added value to the final product. Actually, automated CWD sorting systems are based on mechanical techniques such as crushing, vibratory screening, and magnetic separation [8]. However, at the end of the process manual sorting is still required [9,10]. It is of extreme importance to sort waste materials also with a view of subsequent cataloguing according to the current legislation. In this regard, with the Decision 2000/532/EC, the Commission of the European Communities established to classify waste materials in unified European Waste Catalogue (EWC) codes. These codes are six-digit numbers formatted in three pairs (labelled as chapters, sub-chapters, and individual entry), categorizing wastes according to their production, transportation, handling, and treatment. EWC codes explicitly report also if a waste is hazardous (marked with * symbol) or not, since the primary purpose of the above-mentioned decision is to prevent harm both to people and to the environment. Indeed, wastes free of dangerous materials can be disposed or recycled whereas if a possible contamination has occurred, they cannot be reintroduced into the loop, but should be made inert and disposed through specific treatments by specialized companies, which inevitably increases management costs. Many studies focused on the advancement of CDW materials recognition, based on machine learning [11], artificial intelligence, and robotics [12]. However, these techniques only use visible images and therefore the ability to characterize materials is limited. Usually, they are techniques exploited in laboratory (controlled) conditions. In addition, artificial neural networks require a huge amount of data, images, and time to train the models in recognizing one material from another. Sensing and digital technologies can provide a relevant contribution in this context and different equipment can be exploited. Fourier transformed infrared (FTIR) spectroscopy is a traditional method that provides qualitative (through fingerprinting), semi-quantitative, and quantitative information on physical and chemical characteristics of materials. It can analyze solid, liquid, or gas samples. FTIR gives information regarding the twisting, rotating, bending, and vibration of the chemical bonding and can work in transmittance, reflectance, and absorbance. The spectrum obtained with this technique is a unique characteristic of the functional groups of the investigated material. Compared to other methods, FTIR has a high sensitivity, providing a fast, reliable, and robust analysis. Conversely, its main drawbacks lie in the difficulty in identifying complex samples with overlapping spectra and in the sensitivity to water; therefore, samples should be dried before analysis. Although portable FTIR instruments for noninvasive use in situ have been available for some years, they are still quite expensive and unwieldy. Near infrared (NIR) spectroscopy is a non-destructive and rapid technique commonly used to identify materials. However, one of its drawbacks lies in the difficult identification of black materials, since in the NIR spectral region their reflectance is very low [13]. One suggestion could be the use of midwave infrared (MWIR) spectroscopy [13,14] or its coupling with NIR spectroscopy, but a huge increase in complexity and cost of the system must be considered [15]. In a very recent paper, Reichert and Linß [16] proposed a preliminary study to estimate the water content of CDW concrete, bricks, tiles, and gypsum by NIR spectroscopy for their future use as recycled aggregates. They reported that this methodology gives accurate results especially when the material type is known a priori and when other properties are available (e.g. density, volume, porosity, surface area, etc.), as confirmed also by other authors [17]. They suggested also combining NIR with a high-resolution camera for defining CDW materials properties. Concerning this point, a recent research line consists in using hyperspectral imaging (HSI) for materials identification and characterization. HSI is a nondestructive technique that combines digital imaging with spectroscopy. Based on the camera used, it detects the spectrum (in terms of reflectance or intensity) of each pixel of the acquired image in different wavelength regions (visible (VIS) from 380 to 750 nm, NIR from 750 to 1000 nm, shortwave infrared (SWIR) from 1000 to 2500 nm, MWIR up to 5000 nm, etc.) [18]. Kotthaus and coworkers [19] proposed to use HSI analysis to set up a spectral library of urban materials through emittance and reflectance spectroscopy. They analyzed 74 sample materials and published a library covering both the VIS–SWIR and the long-wave infrared (LWIR) spectral regions. Xiao et al. [20] proposed an online classification of CDW materials (brick, concrete, wood, foam, and plastic) based on the coupling of two cameras. They used a 2D industrial camera to select a region of interest (ROI) of 5x5 pixels and a NIR HS camera with a wavelength range of 900-1700 nm to obtain spectral information about objects in the ROI. This method was developed to reduce the amount of data and the time required by a sole HS camera. After training the system with 50 pieces for each material, they found that around 1200 and 1400 nm the five types of materials can be distinguished clearly thanks to their characteristic absorption peaks. In this regard, HSI has been proposed to assess the quality of waste with the aim of consequent valorization. Serranti et al. [21] employed HSI based sensing devices working in the spectral range of 1000-1700 nm. They found that it is possible to discriminate between recycled aggregates and different contaminants such as bricks, gypsum, plastic, and wood, to name a few, ensuring the quality control of the recycled flow stream. This paper presents a preliminary study aimed at proposing a database for CDW classification, namely those having a EWC code with chapter no.17, based on techniques working in different spectral ranges, hence providing very comprehensive information. In details, the FTIR traditional laboratory method will be compared to advanced technologies based on NIR spectroscopy and HSI analysis in order to have a correlation between laboratory and portable solutions. II. MATERIALS AND METHODS Different types of natural stones and ceramic materials were studied in different spectral ranges. In particular, the latter were concrete and a red clay brick, belonging to EWC classes 17.01.01 and 17.01.02, respectively, whereas the former were two stones belonging to 17.05.04 EWC class. At first, to characterize univocally all materials, X-ray diffraction (XRD) analysis was performed by using a Bruker AXS D8 Advance diffractometer/reflectometer. Resulting XRD spectra cover a diffraction angle (2θ) between 5° and 85°. X- ray diffractograms (not reported for the sake of brevity) show that the studied materials were indeed concrete, clay brick, quartz, and clinochlore (a phyllosilicate mineral). In Fig. 1 all the investigated materials before being tested are reported. The traditional spectral method used for materials characterization was performed with a Spectrum GX 1 FTIR spectrometer equipped with a Perkin-Elmer® Autoimage microscope, having a photoconductive HgCdTe, MCT, array detector, which operates at liquid N2 temperature. The spectrometer covers the entire IR spectral range from 4000 to 400 cm−1 (i.e., 2500-25000 nm). Samples were analysed in transmission mode using the KBr pellet technique. Before materials investigation, background spectra were acquired on clean KBr pellet. Absorption spectra were obtained by 32 scans. Baseline (polynomial line fit) and smoothing were also performed. The first advanced spectral method was performed with a portable NIR spectrometer. MicroNIR OnSite is produced by Viavi Solution® and is enabled by linear variable filter (LVF) technology. It works in a wavelength range between 950 and 1650 nm (i.e., from 10526 to 6061 cm-1) and at an operating temperature comprised between -20 and 50 °C. It is equipped with a 128-pixel InGaAs photodiode array detector and each measurement requires from 0.25 to 0.50 s to be acquired. The second advanced spectral method used to characterize materials consisted of a commercial hyperspectral camera. The camera was a Hinalea 4250 model, working in the VISNIR spectral range between 400 and 1000 nm (i.e., from 25000 to 10000 cm-1) and having sensor spatial resolution of 2.3 MP and a spectral resolution of 4 nm. It allows to detect the spectrum of a pixel or a selected ROI, based on the black and white calibrations. The measurement configuration is completed by one (or more) halogen lamp, which provides a more homogenous lighting in the spectral range. The total intensity of the lamps must be constant during calibration and subsequent acquisitions. III. RESULTS AND DISCUSSION Results obtained from the three different instruments working in different spectral ranges are reported in Fig. 2. It is important to highlight that data close to lowest and highest wavelengths of spectral ranges of each instrumentation should be discarded from the analysis due to the high noise level. FTIR spectroscopy performed between 4000 to 400 cm-1 (i.e., 2500-25000 nm) on the four samples provided the following results: the concrete sample (blue line in Fig.2) exhibits characteristic bands of calcium carbonate (1798, 1427, 874, and 760 cm-1) and silica (1080 cm-1). The bands at 1640 and 3450 cm-1 are typical of water, which, in this case, is chemically bound to the material itself. Generally, water presents a broad band around 3500 cm-1, which shifts to lower wavenumbers when the water molecule is retained within the structure through hydrogen bonds. Similarly, the red clay brick (orange line in Fig. 2) displays characteristic bands of calcium carbonate but has a broad band in the range of 1200900 cm-1, where the bands of silica and gypsum (CaSO4) overlap. The presence of gypsum is confirmed by bands at 690 and 580 cm-1. In the quartz mineral (purple line in Fig. 2), the two predominant bands fall at 1070, 780, and 690 cm-1, indicating the presence of silica and TiO2, responsible for the white coloration of the sample. The fourth sample, a phyllosilicate mineral, namely a mica (green line in Fig. 2), is composed of silica, water, and metallic complexes (whose bands fall at 754 and 530 cm-1). The region containing water bands is, however, split into two bands corresponding to water bound to the molecular structure (3422 cm-1) and free water not engaged in hydrogen bonds (3625 cm-1). Spectra obtained with the portable MicroNIR are evaluated between 950 and 1650 nm. The four analysed materials show always a peak around 1410 and 1500 nm (e.g. 7000 cm-1), which was less evident for the red clay brick. This peak is attributed to the first O-H stretching overtone of water. The presence of water in all the four samples was confirmed also by NIR spectroscopy, which works in the range between 1000 and 2500 nm (e.g. from 10000 to 4000 cm-1). NIR spectroscopy showed an additional peak around 2000 nm (5000 cm-1) related to the O-H combination band of water (NIR spectra were not reported for the sake of brevity). While the presence of water is problematic for FTIR spectroscopy, it is necessary when dealing with quality control of materials, are they virgin or recycled. For example, for biomass materials like woodchips it is extremely important the evaluation of their moisture content, since it influences energetic, management, and economic aspects of their utilization [22]. Moreover, for recycled concrete aggregates the assessment of water content is fundamental, in view of a subsequent use in new concretes. Indeed, the estimation of water content of recycled aggregates is necessary, e.g., when calculating the amount of each component in the concrete mix-design. Spectra obtained with HSI were evaluated between 450 and 900 nm. The slopes of the curves are very different among the four investigated materials as well as the peaks appearing along the analyzed range. Each material is therefore characterized by a typical spectral fingerprint. As each sample consists of a surface with irregular shape and a peculiar morphology, the impact of the light source on the surface itself can result in a unique spectrum. Fig. 1. Visual aspect of investigated materials: a. concrete, b. red clay brick, c. quartz, and d. clinochlore. Fig. 2. Materials spectra investigated by FTIR spectroscopy, HSI analysis, and portable MicroNIR: concrete (blue lines), red clay brick (orange lines), quartz (purple line), and clinochlore (green line). IV. CONCLUSION AND FUTURE PERSPECTIVES This paper reports a preliminary study of CDW recognition and sorting in the perspective of creating a standardized database for materials management, in view of future recycling processes. With more details, several C&D wastes have been characterized by FTIR traditional technique and two portable solutions, e.g. NIR spectroscopy and a novel image-based tool based on HSI. Obtained results highlight that: • FTIR spectroscopy permits to univocally characterize materials, representing the most reliable solution to identify molecules of several compounds; however, it is quite expensive and sensitive to the presence of water; • The portable MicroNIR has always highlighted a unique broad peak around 1429 nm, which correspond to the first O-H stretching overtone of water. This shows that the portable MicroNIR is not a sufficient tool for characterizing the investigated materials (ceramics and stones), but it should be coupled to other techniques; • The novel solution based on HSI permits to work in situ, only if controlled ambient conditions are ensured, but has shown that each of the investigated materials has a unique spectral fingerprint. To validate the proposed solution, based on relatively lowcost and portable equipments, it is extremely important to have a representative sample of CDW materials for each investigated category with the aim of having a broad spectrum of categorization, properly considering the intrinsic variability of materials. In this way, it would be much easier and less subjected to errors to classify unknown wastes and to proceed with high-quality recycling with a special focus on upcycling, thanks to a thorough consideration of the uncertainty of the measurement chain. In the future, the authors plan to also characterize other types of CDW, like wood, paper, metals, and plastics. It should be stressed that this methodology can be extended also to other types of wastes different than those belonging to the chapter no. 17 of EWC code in view of validating the proposed solution and to create a more exhaustive database that can be used by as many stakeholders as possible to pursue the goal of circularity not only for the construction industry but also for the waste management sector. ACKNOWLEDGMENT This research activity was carried out within the RECONSTRUCT (A Territorial Construction System for a Circular Low-Carbon Built Environment) project, funded by the European Union’s Horizon Europe research and innovation programme under grant agreement no. 101082265. REFERENCES [1] European Environment Agency, Construction and demolition waste: challenges and opportunities in a circular economy, 2020. https://www.eea.europa.eu/publications/construction-and-demolitionwaste-challenges. [2] J. Moschen-schimek, T. Kasper, M. Huber-humer, Critical review of the recovery rates of construction and demolition waste in the European Union – An analysis of influencing factors in selected EU countries, Waste Manag. 167 (2023) 150–164. doi:10.1016/j.wasman.2023.05.020. [3] European Innovation Partnership on Raw Materials. Raw Materials Scoreboard, 2020. doi:10.2873/680176. [4] DIRECTIVE 2008/98/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 19 November 2008 on waste and repealing certain Directives, (2008). [5] A. Nawaz, J. Chen, X. Su, Exploring the trends in construction and demolition waste (C&DW) research: A scientometric analysis approach, Sustain. Energy Technol. Assessments. 55 (2023) 102953. doi:10.1016/j.seta.2022.102953. [6] O. Adedeji, Z. Wang, Intelligent waste classification system using deep learning convolutional neural network, in: Procedia Manuf. 2nd Int. Conf. Sustain. Mater. Process. Manuf. (SMPM 2019), 2019: pp. 607– 612. doi:10.1016/j.promfg.2019.05.086. [7] European Environment Agency. Circular economy in Europe. Developing the knowledge base, Copenhagen, Denmark, 2016. [8] Y. Ku, J. Yang, H. Fang, W. Xiao, J. Zhuang, Deep learning of grasping detection for a robot used in sorting construction and demolition waste, J. Mater. Cycles Waste Manag. 23 (2021) 84–95. doi:10.1007/s10163020-01098-z. [9] M.E. Ozdemir, Z. Ali, B. Subeshan, E. Asmatulu, Applying machine learning approach in recycling, J. Mater. Cycles Waste Manag. 23 (2021) 855–871. doi:10.1007/s10163-021-01182-y. [10] L. Weisheng, Y. Hongping, Off-site sorting of construction waste: What can we learn from Hong Kong?, Resour. Conserv. Recycl. 69 (2012) 100–108. doi:10.1016/j.resconrec.2012.09.007. [11] P.-Y. Wu, C. Sandels, K. Mjörnell, M. Mangold, T. Johansson, Predicting the presence of hazardous materials in buildings using machine learning, Build. Environ. 213 (2022) 108894. doi:10.1016/j.buildenv.2022.108894. [12] S. Dodampegama, L. Hou, E. Asadi, G. Zhang, S. Setunge, Revolutionizing construction and demolition waste sorting: Insights from artificial intelligence and robotic applications, Resour. Conserv. Recycl. 202 (2024) 107375. doi:10.1016/j.resconrec.2023.107375. [13] O. Rozenstein, E. Puckrin, J. Adamowski, Development of a new approach based on midwave infrared spectroscopy for post-consumer black plastic waste sorting in the recycling industry, Waste Manag. 68 (2017) 38–44. doi:10.1016/j.wasman.2017.07.023. [14] W. Becker, K. Sachsenheimer, M. Klemenz, Detection of Black Plastics in the Middle Infrared Spectrum (MIR) Using Photon UpConversion Technique for Polymer Recycling Purposes, Polymers (Basel). 9 (2017) 435. doi:10.3390/polym9090435. [15] W. Xiao, J. Yang, H. Fang, J. Zhuang, Y. Ku, A robust classification algorithm for separation of construction waste using NIR hyperspectral system, Waste Manag. 90 (2019) 1–9. doi:10.1016/j.wasman.2019.04.036. [16] I. Reichert, E. Linß, Water content estimation of recycled building materials based on near-infrared spectroscopy, Constr. Build. Mater. 412 (2024) 134827. doi:10.1016/j.conbuildmat.2023.134827. [17] S.R. Shukla, S.K. Sharma, Estimation of density, moisture content and strength properties of Tectona grandis wood using Near Infrared Spectroscopy, 2021 (2021) 1–12. doi:10.4067/s0718221x2021000100418. [18] G. Bonifazi, G. Capobianco, R. Palmieri, S. Serranti, Hyperspectral imaging applied to the waste recycling sector, Spectrosc. Eur. 31 (2019) 8–11. [19] S. Kotthaus, T.E.L. Smith, M.J. Wooster, C.S.B. Grimmond, Derivation of an urban materials spectral library through emittance and reflectance spectroscopy, ISPRS J. Photogramm. Remote Sens. 94 (2014) 194–212. doi:10.1016/j.isprsjprs.2014.05.005. [20] W. Xiao, J. Yang, H. Fang, J. Zhuang, Y. Ku, Development of online classification system for construction waste based on industrial camera and hyperspectral camera, PLoS One. 14 (2019) 1–16. doi:10.1371/journal.pone.0208706. [21] S. Serranti, R. Palmieri, G. Bonifazi, Hyperspectral imaging applied to demolition waste recycling: innovative approach for product quality control, J. Electron. Imaging. 24 (2015) 043003. doi:10.1117/1.JEI.24.4.043003. [22] E. Leoni, M. Mancini, G. Picchi, G. Toscano, Performance evaluation of NIR spectrophotometer simulating in-line acquisition for moisture content prediction of woodchips and comparison with hand-held NIR spectrophotometer, Fuel. 357 (2024) 130015. doi:10.1016/j.fuel.2023.130015.