Life cycle assessment of a new smart label for intelligent packaging

Within the past years, there has been a growing demand for sustainable, cost-efficient on-line sensing of chemical and physical properties and locations of products. Measuring of products’ physical properties, such as temperature and humidity, could improve product safety and efficiency of logistic operations. In the future measurement of temperature of food items could also aid in reducing food wastage. The aim of this study was to calculate the life cycle environment impacts of a temperature logger, hereafter called smart label, primarily targeted for the monitoring of the packed food products. According to the results, the largest normalised impacts of the smart label production are resource use (both use of fossil fuels and use of minerals and metals), eutrophication and particulate matter formation. The main materials causing these impacts were the printed electronics inks and adhesives. In addition, energy used in the production, and plastics used as substrates had large impacts on the results. It should be noted that the present calculations have mainly been made on a laboratory scale. The impacts are likely to get smaller on an industrial scale with more efficient production. In the future, the label could potentially bring environmental benefits through product savings when used in products with high environmental load.


Introduction
Within the past years, there has been a growing demand for sustainable, cost-efficient on-line sensing of chemical and physical properties and locations of products.Measuring of products' physical properties, such as temperature and humidity, could improve product safety and efficiency of logistic operations.(Hakola et al 2022).Intelligent sensing and monitoring technologies can also be applied for tracking and tracing in material and product supply chains (Musa et al 2014).Smart tags or labels are one solution for such technologies (figure 1).
According to Sohail et al (2018) intelligent packaging technologies provide means for controlling packed product quality, provide more convenience to consumers, market and brand the products, and control counterfeiting and theft.Intelligent packaging is a relatively new application domain, and it has specific needs, such as cost-efficient, recyclable, and lightweight electronics.According to estimates, in 2030 almost 21 billion packages will feature an electronic feature (Das 2019).To ensure that these technologies do not carry a large environmental footprint, attention needs to be paid on their environmental impacts.Potential solutions for achieving this include implementation of recyclable, material efficient and energy-autonomous monitoring technologies.Benefits from using the intelligent sensing technologies should always be critically evaluated against the potential environmental burden they carry.
For smart labels to have as low environmental burden as possible, and potentially reduce the overall life cycle impacts of the product contained in the package, they need to be designed considering environmental impacts over their whole life cycle.Yet only few studies have so far analysed the environmental impacts of producing smart labels or tags.Glogic et al (2021) analysed a multi-layer anti-counterfeit label printed on a paper.The journal article presented some comparison of the environmental advantages of printing the label on a paper instead of plastic, reporting that using paper resulted in a 10%-20% in almost all the impact categories studied.Shou and Domenech (2022) analyse the potential of applying blockchain technology through a smart label for data collection for LCA of fashion industries.However, the study does not assess the environmental impacts of the smart labels specifically.Some further studies have been published on the environmental impacts of other types of printed electronics, such as food industry.Yet, this question is not quite straight-forward and depends on the origin of the waste materials.Increase in bioenergy and biomaterial demand has resulted in high demand for many biowaste or bio-based side products (Liu et al 2014, Välimäki et al 2020, Sudheshwar et al 2023a, 2023b).Sudheshwar et al (2023a) present an analysis of the carbon footprint of a semi-quantitative drug detection and point-of-care device using paper-based printed electronics.
Several studies have pointed out the potential of smart labels for food packaging in reducing food waste (e.g.Barone and Aschemann-Witzel 2022).According to FAO, the global amount of food wastage is assumed to be 1.6 billion tonnes per year, with editable parts amounting to 1.3 billion tonnes (FAO 2013).According to FAO (2013), most of the carbon footprint of food waste is caused by cereals (34%), meat and vegetables (both causing 21% of the total).
Along with consumer behaviour, one of the main reasons for food loss and waste in high-or mediumincome countries is the lack of coordination between different actors in the food supply chain (Heising et al 2017).Presently, food trader or consumer is primarily informed about the quality of the product through fixed 'best before' or 'use by' dates given in the packages.They do not give information of the state of each package or the way the product has been treated during the production chain.Yet, according to Heising et al (2017), provided the conditions of a product during the supply chain have been optimal, its quality can still be acceptable after the expiry date on the package.
The aim of this study was to analyse the life cycle environmental impacts of a new smart label for sensing in temperature monitoring and data logging.The analysis extended over the whole life cycle of the product.Furthermore, the potential impacts resulting from reduced food wastage using the smart label were also analysed and compared to the impacts of producing the smart labels.

Material and methods
The product analysed in this paper is a temperature logger, hereafter called smart label, primarily targeted for the monitoring of the packed food products.The label is energy autonomous, and it can be attached to primary or secondary packaging of expensive or affordable products, respectively, for monitoring the quality of the packed products during shipment, in storage or in the shops.Thus, the logger can be utilized to improve product safety and to avoid product spoilage (e.g.food waste) by identifying products with quality problems, and thereby overcoming the current practice of throwing away the whole batch when single items or items in a single secondary packaging are found to be damaged.Another advantage is automated identification of the compromised packages through the wireless connection instead of manual work.
In order to analyse the potential environmental aspects related to the smart label, a life cycle assessment (LCA) was conducted.LCA is a commonly used method for analysing the environmental impacts of product, service or system throughout its life cycle streams, and this should be considered when calculating the impacts of their use (Hellweg and Milà I Canals 2014).
The present approach (Rebitzer et al 2004, Guinée et al 2011) enables achievement of a comprehensive understanding of the environmental impacts of the product.In this study, main focus was on the raw materials used for the production of the smart label as the most concise information was available on them.As the described smart label is not yet manufactured commercially and not even on a pilot scale, energy use of its production was estimated based on data on potential production methods from the literature or expert estimations.The recovery and disposal of the final product were included in the analysis.As the label is relatively small, and the materials used in it are very small in quantity, it was assumed that it would end up either in waste incineration or landfills as part of the impurities originating from the recycled packages.
In the analysis, the system was divided in five different production phases.In a further analysis the impacts were divided to different types of components or resources, and waste management, namely: surface mount device (SMD) components, energy used in production, printed electronics inks and adhesives, poly(lactic acid) (PLA) and other plastics, waste and others.Bio-based film in Phases I and V was assumed to be PLA made of corn (Benavides et al 2019).
The life cycle impact assessment (LCIA), an essential part of LCA, was conducted by applying the impact assessment methods recommended in the EU product environmental footprint method (Zampori and Pant 2019).The life cycle assessment was conducted with the LCA calculation tool SULCA ('SULCA-Sustainability tool for Ecodesign, Footprints & LCA' 2024).Most of the life cycle inventory (LCI) data were taken from the Ecoinvent database (Wernet et al 2016, see also table 1).

Smart label
Aim of the smart label is the monitoring of the packed goods helping to decrease the environmental impacts of food waste.The label consists of an IC optimized sensor chip embedded with NFC interface for temperature monitoring and data logging.Furthermore, the smart label includes organic photovoltaics (OPVs) for light harvesting at indoors, supercapacitor (SC) for energy storing, printed and hybrid-integrated circuitry, and infographics in 120 mm × 98 mm size.The structure of the label is presented in figure 2(a) and the function in figure 2(b).Energy autonomous smart label was designed to provide 2-3.5 V operational voltage and 3 µA average current consumption.The 24 h energy autonomy was reached by harvesting indoor light for four to eight hours per day and by providing storage power for 16-20 h per day.The 0.65 J energy storage enabled to have 2-4 measurement per hour (Hakola et al 2022).
As the aim was to make the smart label energy autonomous, its energy source is rechargeable.This also increases the lifetime of the electronic features in comparison to primary batteries (Tuukkanen et al 2016).In addition, the label does not require external recharging during use phase.As the intelligent packages are primarily used in indoor environments, the energy harvesting is conducted from artificial light sources under low light conditions.The 24 h daily energy demand is satisfied by harvesting indoor light for 4 to 8 h per day, and by providing storage power for 16-20 h per day.The layouts of the demonstrator and the component structures were experimentally optimized to fulfil the specifications of the system.The following subsections present this development.

Process flow
It has been calculated that additive printing based electronic manufacturing uses even five times less energy than subtractive processes (Nassajfar et al 2021).In addition, less process chemicals are used, and less material waste created.Thereby, additive manufacturing methods were used in this work.First antenna and circuitry were printed on PLA base substrate using screen printing.After that OPV was gravure-printed and finished with evaporated top layers next to the antenna and, the SC was doctor-blade coated on top of OPV.SMDs were integrated on printed circuitry with adhesive bonding technology.The cover film with infographics was flexo-printed on regenerated cellulose film (NatureFlex™) and laminated on top of the label (Futamura Group 2024).The process flow is described in figure 3.

Substrate
In the prior analyses comparing different bio-based films, it was found out that PLA was the most compatible biobased substrate for printed electronics processing at 80  regenerated cellulose film Natureflex™), it obtained the best performance and durability against heat, humidity, and solvents, providing a good optical transmission (Luoma et al 2022(Luoma et al , 2023)).For that reason, PLA films were selected for the smart label development.PLA film was chosen as a base substrate for the labels (commercial fossil PET was used as a reference).

Antenna and circuitry
Antenna and printed circuitry with bridges were fabricated by screen printing thermally curable silver traces and UV-curable acrylate-based dielectric.Silver traces were dried in the convection oven at 100 • C for 60 min and dielectric layers were UV-cured using a power level of 100% (DELO Delolux 202 lamp, wavelength 365 nm).The line and gap width of the antenna pattern was 500 µm, and the minimum line and gap widths of the circuitry were 200 µm and 150 µm, respectively.

Printed OPVs
The aim of the OPVs development was to fabricate low environmental impact OPVs that would work under specified lighting conditions and thereby meet the system requirements.In practise this implies using earth abundant, non-hazardous materials, and replacement of fossil materials with bio-based materials.Furthermore, the thin film device structures were fabricated using printing, which typically is more material and energy efficient than competing methods and treated thermally after printing at 100   current after floating the module for 1 h at 3 V was 5 µA.

Component assembly
In addition to energy harvesting and storage components, the smart label contains IC and power management circuitry components.These were assembled by dispensing isotropic conductive adhesive and pickand-placing the rigid components on printed circuitry.After thermal curing at 80 • C, UV-curable adhesive was dispensed for mechanical support.The lead design principle of the circuitry was minimizing the number of components and required energy resulting in desired performance indicators.As an example, two components (i.e.low leakage Schottky and Zener diode) solution was investigated to minimize the component count, but it did not reach the system performance.Therefore, more sophisticated power management integrated circuit with higher number of SMDs were implemented.

Data used in the LCI
Data for the LCI was mainly taken from the Ecoinvent LCA Database (Wernet et al 2016).Ecoinvent is a very extensive database, containing LCI and life cycle impact assessment (LCIA) for thousands of processes.For some processes also other data sources were used (tables 1 and 2).
For the LCA, all the material and energy inputs used in the production of the label were included in the analysis.For the inks all material data was not available as such.Therefore, they were constructed from different components (graphite ink), material safety data (paper separator), or only covered some impacts, or part of the life cycle (PVA, P3HT, PCBM) (table 1).Energy consumption for smart label processing was based on actual measurements and estimated throughput times from pilot-scale fabrication.However, Ecoinvent database was used for the thermal evaporation and for the printing of the graphics due to the lack of available in-house measurement data.

Results
Below, life cycle impacts of the label production are presented (figures 5-7).In normalisation, the impacts are divided by a reference value (in this case the total global impacts per person) and are thus comparable to one other under the assumption that all the impacts were valued equally important.Overall, the largest normalised impacts of the smart label are resource use (both use of fossil fuels and use of minerals and metals), eutrophication and particulate matter emissions (figure 5).Of the total impacts, by far the highest normalised impact was the use of minerals and metals, with the normalised value of 3,4 × 10 −03 .As it was so much higher than the other impacts, it was left out of figure 5.The high normalised impacts of abiotic resource use, in particular minerals and metals highlights the large contribution of electronics to resource consumption.Of the other impact categories, the largest normalised impacts were the freshwater eutrophication, fossil fuel use and particulate matter impacts (figure 5).
When studying the contribution of the different production phases of the total impacts, Phase I (Fabrication of the antenna) had the largest contribution to the impacts in most of the impact categories, in particular freshwater eutrophication and land use (figure 6).Phase I (Fabrication of the antenna and circuitry) impacts largely result from the  production of silver paste5 , with its contribution ranging from over 90% to 50%.Its large contribution to land use impacts results from mining while the contribution to eutrophication impacts comes from phosphate production.Silver production results in large phosphorus emissions, which again are the main contributor to the eutrophication impacts to freshwaters.Also phases III (Fabrication of supercapacitor) and IV (SMD assembly) were important contributors to most of the impacts with the other phases having smaller roles.The impacts in phases III (Fabrication of supercapacitor) and IV (SMD assembly) largely originate from energy production, which can also be seen in that their contribution is larger in impact categories where emissions from energy generation, such as CO 2 , NO X and particulates, typically have a large role, e.g.acidification and particulate matter impacts.Also impacts on human toxicity mainly result from heavy metal emissions of heat production in phases II (Fabrication of OPV), III (Fabrication of supercapacitor) and IV (SMD assembly).
In addition to the production phases, the impacts were also analysed through different types of components (figure 7).It is evident that the main contributors in almost all the impact categories are the inks and adhesives.However, within this, most of the impacts result from silver paste production, its share ranging from about 74% (use of fossil fuels) to over 99% (use of minerals, metals, and other resources).The use of energy contributes about 20%-40% to total impacts in most of the impact categories.Based on the results, it is evident that replacing silver with a material with lower environmental burden, e.g.copper or carbon, would reduce the impacts considerably.However, it is important to note that copper oxidises rapidly, and therefore it might not be suitable to replace silver in all the cases.In addition, carbon has poorer conductivity as compared to metals, thus limiting its usage in printed electronics applications.
A further scenario was constructed where PLA was replaced with PET.Use of PET resulted in small reduction in all of the impact categories but these were reductions were only 1%-3%, except for ecotoxicity and water use.The higher impacts for PLA in these impact categories stemmed from maize grain production that was assumed to be the raw material of the PLA.These impacts would diminish considerably if the PLA was made of waste materials such as biowaste from food industry.Yet, this question is not quite straight-forward and depends on the origin of the waste materials.Increase in bioenergy and biomaterial demand has resulted in high demand for many biowaste or bio-based side streams, and this should be considered when calculating the impacts of their use.

Discussion and conclusions
There are potential ways of reducing the environmental impacts of the smart labels.First, the production in this case took place on a pilot scale, which that the impacts per one label would most likely be lowered when processing is optimized and implemented in a larger scale production, which typically is more efficient.In addition, if process heat recovery could be utilised for the process steps performed at elevated temperatures, the total efficiency of the production process would improve as well.Based on literature, it was estimated that the potential energy saving could be at least 20% through process heat recovery (Jouhara et al 2018).Process optimisation would further reduce the energy consumption.Furthermore, energy production was now assumed to originate from average European electricity and heat production, which has a relatively high environmental footprint.As energy consumption is responsible for 20%-30% of impacts in most of the impact categories, changing to renewable energy would notably decrease the impacts.However, it should be noted that when more labels are produced in future, the end-of-life phase needs development, e.g.development of methodologies to collect the labels among packaging waste (Hakola et al 2022).
Other potential and important ways of reducing the system's impacts can be obtained by replacing the silver paste used in the fabrication of the antenna, and by minimizing the size of the smart label for improved process efficiency.More labels per m 2 would lead to reduced consumption of materials and energy per one label.Silver was a dominant contributor to many of the impacts, and by replacing it when possible with a substance with less environmental impacts, such as copper, could reduce the impacts even by dozens of percentages (Nassajfar et al 2021).Replacement of PLA with paper would also reduce emissions as paper has lower environmental footprint in most impact categories (see also e.g.Liu et al 2014).Printability issues might arise from use of porous and rough paper surfaces, but sometimes optimized paper substrates with some roughness can improve conductivity of printed traces due to improved ink adhesion and better ink-substrate interactions (Jansson et al 2022).
In the LCA, the activated carbon used in phase III was assumed to be fossil carbon, even though in the smart label production it was actually biogenic carbon.However, the contribution of activated carbon to the impacts was less than 5% in all the impact categories.Thus, changing it would not have an important impact on the results.Through identification of spoiled and unspoiled food items, e.g. by recognizing items exposed to too high temperatures, the smart labels could reduce carbon footprints in other systems, and thus increase their own climate handprint (for carbon handprint, see e.g.Grönman et al 2019).
Yet, before wide-scale adoption of intelligent sensors in packages is possible, standardized and comprehensive tests on their reliability need to be adopted (Dodero et al 2021).Presently, the impacts of producing the labels are relatively high in comparison to e.g.food items, and they could only provide benefits in case the number of labels needed to save one item was fairly low, or products with high environmental footprint (such as meat or other animal products) were saved.Environmental load of the labels in relation to the products could also be reduced if the number of labels per products was reduced (e.g. by applying one label per a box containing 10 or 50 products).Further reductions in the environment impacts of the labels could be achieved by increasing the efficiency of the production through process optimization, reducing the material consumption by miniaturizing the label size and lowering the number of components through optimization.Also by replacing the raw materials or energy sources with processes that had lower environmental burden (such as silver with copper as mentioned earlier), the labels could potentially provide benefits even when used items with lower emissions.
In the analysis, the smart label is not recycled or otherwise recovered.The recycling of such a label is challenging, as it is very thin and contains several different substances, which are difficult to separate from one another.Also, in reality, the intelligent packaging has proven to be very difficult to recycle, which is a major challenge to its widespread use e.g.(Ahmad and Lim 2022).Thus, this is an important area of development in the future.Still, despite the problems in the packaging treatment, the potential benefits from reduced food waste may overcome them.For example, the amount of separately collected municipal biodegradable waste produced in Finland in 2020 was ca.494 000 tonnes while the amount of plastic waste was about 93 000 tonnes (Statistics Finland 2022).Therefore, even though packaging carries an environmental load, and producing intelligent packaging even increases that, it could potentially be compensated through the environmental savings from reduced food wastage.

Conclusions
In conclusion, main impacts of the smart label originated from screen printing of the antenna and circuitry (materials and energy used for those processes), followed by fabrication of OPV, SC and component assembly.Within these, the main processes causing these impacts were the printed electronics inks and adhesives.Also, energy used in the production, and plastics used as substrates had a large impact on the results.Here the energy used in the heat production was assumed to represent present average European production, and electricity German production.Their environment load is expected to considerably decrease during the next decades, which will naturally reduce the product's environmental impacts as well.Furthermore, the energy consumption of the lab-scale fabricated smart label was based on measured pilot-scale fabrication conditions, even though some assumptions on evaporation, printing of graphics and production-phase process heat recovery have been made on how it would develop on an industrial scale.Production in larger scale would be more efficient and thus reduce the environmental impacts of the label.
The labels could potentially bring environmental benefits through product savings when used in products with high environmental loads.When production of the label is extended to an industrial level and thereby its environmental load reduced, savings could be achieved with lower number of labels, and for products with lower environmental load.

Figure 1 .
Figure 1.The potential applications for the case study of this paper (Image: Jonna Palojärvi and Kaisla Tiitinen, LAB University of Applied Sciences).

Figure 2 .
Figure 2. (a) Structure of the fabricated smart label (image: Sakari Ritvasalo, LAB University of Applied Sciences), (b) function of the smart label.
• C (Hengevoss et al 2016, Välimäki et al 2020).Material screening was based on literature and the experiments were executed with OPV device thin-film structure presented in figure 4(a).Structure comprised PLA substrate/poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) processed by gravure printing (G)/PEDOT:PSS (G) /electron donor (NF3000-P):acceptor (NF3000-N) blend (G)/lithium fluoride through thermal evaporation (E)/aluminium (E) OPV module for the smart label comprised eight serially connected cells to ensure the full energy-autonomy for the system under indoor light energy harvesting.These modules gave at maximum power point 143 µW power, 64 µA current and 2.19 V voltage under 1000 lux fluorescent light illumination.3.5.Coated SCIn the SC development, the target was to use environmentally friendly materials without weakening electrical and mechanical properties (figure4(b)).For the aqueous electrolyte, three cells connected in series provided 3.6 V required by the smart label.Current collectors (graphite ink) and electrodes (activated carbon ink) were doctor blade coated through stencils and thermally treated at 90 • C, and 60 • C, respectively, and a separator paper and NaCl electrolyte were applied.Furthermore, they were sealed by faceto-face assembling the stack with pressure sensitive adhesive tape (Fu et al 2022).The SC modules with three series connected SCs had capacitance of 150 mF and equivalent series resistance of 20 Ω.The leakage

Figure 4 .
Figure 4. (a) Schematic cross-section of (a) the organic photovoltaics (OPVs) device layer structure (b) schematic cross-section of the supercapacitor module consisting of three cells.

Figure 5 .
Figure 5. Normalised impacts producing the smart label.

Figure 6 .
Figure 6.Impacts divided to the different production phases.

Table 1 .
Bill of materials and data sources used in the LCI.
a Isotropic conductive adhesive (ICA) was left out of the analysis due to the very small amount used.

Table 2 .
Processes used for heat and electricity production, and waste treatment.