Abstract
The increased utilisation of lithium-ion batteries in the last years does not come without cost. Due to thermal runaway and exothermic degradation reactions, portable batteries pose enormous risks to waste management systems and infrastructure in their end-of-life phase. All over Europe, the number of waste fires caused by lithium-ion batteries are rising. The risk of a battery fire is mainly influenced by the probability and severity of a thermal runaway or exothermic degradation, which depends on the current state of charge (SOC) of the respective battery. In order to determine the distribution of the SOC which is one of the main influence factors to waste fires caused by lithium-ion batteries, 980 waste battery cells were representatively sampled, manually dismantled and analysed using a prototypic laboratory test stand. Approximately 24% of the analysed cells and batteries had a residual SOC of at least 25%, and approximately 12% had a residual SOC of at least 50%. Hence, approximately every fourth to eighth portable battery threatens to cause a waste fire when critically damaged. Furthermore, a distinct relationship between the actual cell voltage and the residual SOC was found for end-of-life portable batteries.
Introduction
Since lithium-ion batteries were first marketed in the early 1990s, their importance to our technology-driven lifestyle increases rapidly. Lithium-based battery types captured almost every market segment of portable electronic devices or household appliances. However, this development does not come without costs. Besides the predictable safety risks of lithium-ion batteries in their use phase, in the end-of-life phase they certainly put waste management facilities, personnel and infrastructure at enormous risks of fire (Nigl and Pomberger, 2018). Needless to say, that a rising number of large waste fires is contradicting environmental endeavours (Nigl et al., 2020) because a single open waste fire may have a greater environmental impact than the annual emissions of Sweden’s incinerator plants (Ibrahim et al., 2013).
Although waste fires can be traced back to lithium-ion batteries in many cases, before 2020 there was hardly any statistically valid data available (Autischer et al., 2020). In any case, the number of fires in waste management is increasing sharply (Mikalsen et al., 2020; Nigl et al., 2019).
The battery directive of the European Union (EU) (2006/66/European Commission (EC)) defines portable batteries as batteries and accumulators which are all sealed, could be carried by an average person without difficulty and are neither automotive or industrial batteries.
Separate collection schemes for portable batteries are available for more than two decades. However, recent studies from Austria show that vast amounts of these batteries are still discarded in other municipal solid waste streams (e.g. residual household waste, lightweight packaging waste, metal packaging waste and waste of electrical and electronic equipment). With about 718 kg per year, most of these wrongfully discarded batteries are entering the residual household waste collection systems (Nigl et al., 2020).
Considering the fire triangle, lithium-ion batteries act as a heat and/or ignition sources combined with waste (as fuel) and atmospheric oxygen (as an oxidising agent) – deliver the basic conditions for a self-sustaining exothermic chain reaction. A circumstance, that is very specific for waste management systems (Nigl and Pomberger, 2018).
In general, lithium-ion batteries’ thermal runaway or exothermic degradation reactions can be caused by over-charging, short-circuiting, deep discharging (over-discharging) and mechanical and thermal abuse (Lisbona and Snee, 2011). However, in waste management systems, the by far most problematic causes are the latter: mechanical and thermal abuse (Nigl and Pomberger, 2020).
The probability that a lithium-ion battery undergoes thermal runaway or exothermic degradation is mainly influenced by the current state of charge (SOC) of the respective battery. Hence, the SOC is one of the main influence factors to waste fires caused by lithium-ion batteries.
According to Mikolajczak et al. (2011), even severe crushing of cells below approximately 50% SOC will not lead to a severe reaction. However, Golubkov et al. (2015) showed that after thermal abuse, lithium-ion cells with Li x (Ni0.80Co0.15Al0.05)O2 cathodes (NCA) displayed an unmistakable thermal runaway when SOC was ⩾25%. In the same way, lithium-ion cells with Li x FePO4 cathode (LFP) showed mild exothermic reactions when SOC was ⩾25% and pronounced thermal runaway when SOC was ⩾50%. According to Larsson et al. (2014) and Baba et al. (2002) even when the SOC is below the mentioned thresholds, exothermic reactions can occur, albeit far less intense.
Lisbona and Snee (2011) stated that new and old batteries had been reported to show significant safety performance differences under test conditions. Accordingly, abuse tests on used batteries may be more suitable to characterise hazards and risks associated with used and waste batteries. Unfortunately, in scientific literature, abuse tests with lithium-ion batteries are usually done with new batteries. Larsson et al. (2014) showed in experiments (with ~2- to 3-year-old cells) that the exothermic reactions are far less intense when the cells are not fully loaded.
Against that background, this study’s objective is to determine how many lithium-ion batteries are critically loaded in their end-of-life phase and hence how many pose a safety risk when they are mechanically damaged or exposed to heat. Specific knowledge about the residual SOC distribution is crucial, especially to profound risk analysis and assessment of the fire hazards of portable batteries in waste management systems. The test hypothesis for the distribution pattern of the residual SOC is shown in Figure 1. It is assumed that most of the batteries are almost empty, containing SOC values between 0% and 10–20%. Additionally, the article aims to investigate the influence of battery-related influence factors, such as the battery types or electrochemical systems.
Test hypothesis for the distribution pattern of residual SOC.
Materials and methods
Approach
For this study, a prototypical laboratory test stand for measuring the discharging parameters of portable batteries was planned and constructed. End-of-life batteries were sampled from the (mixed) portable battery collection, dismantled and the product specifications of the individual batteries or cells were researched. Afterwards, the batteries were discharged following the defined test procedure, and the discharge protocols were finally analysed and assessed (see Figure 2).
Approach of the study.
Prototypical laboratory test stand for portable batteries
The test stand was planned and constructed in collaboration of the Chair of Waste Processing Technology and Waste Management and the Chair of Electrical Engineering (both at Montanuniversität Leoben, Leoben, Austria). A controller was then programmed for automated measurement. The setup of the test stand enables the discharge of up to four cells or batteries with a maximum voltage of 14 V simultaneously under defined discharge parameters (e.g. current, voltage and time). Both voltage and current (0.1–5.0 A) were set manually. The corresponding residual SOC is then determined using the test protocol.
The test stand was calibrated before commissioning and the results were validated based on the measurement of new and fully charged cells.
Details about the test stand’s construction, components and the test procedure are provided in Supplemental Material.
Sampling of waste portable batteries
Two representative samples of used portable batteries from the mixed collection were taken from a national collection and sorting facility for portable batteries. On the one hand, samples were taken at two different times (dates) to ensure seasonal stratification. On the other hand, the sampling of batteries from various regional origins ensured representative sampling concerning effects of population density, social structure and other regional aspects. Each representative sample consisted of several incremental samples.
Sampling focused on lithium-ion batteries. The investigation of primary lithium batteries was dispensed with since exact product specifications for these batteries are generally extremely rare. Therefore, the exact remaining SOC of the batteries can only be determined in a few cases (with reasonable effort).
The representative samples included portable batteries from mobile phones, laptops, power tools, household appliances, toys, model making and other cordless electrical and electronic devices.
Until now, no documented incidents of fire in waste management systems were caused by conventional portable batteries, such as alkaline, zinc-carbon (both primary), nickel-cadmium or nickel-metal hydride (both secondary). Therefore, these battery types were not considered in the sampling campaign.
Additionally, portable batteries that had shown in pre-tests that successful SOC testing could not be conducted were not sampled. This primarily included batteries, so badly affected by the end-user that the preparation for measuring the remaining charge was associated with increased safety risks or unreasonable effort.
Dismantling
Since it is impossible to discharge portable batteries containing a battery management unit (BMU) beyond the end-of-discharge voltage, these BMUs had to be dismantled beforehand. Multi-cell batteries and battery packs, including batteries from laptops, tools or household appliances, also had to be dismantled beforehand.
Research of battery data
Initially, the determination of product specifications is required in order to discharging and subsequently calculating the residual SOC. Therefore, thorough literature research was done. The specifications include (1) nominal voltage, (2) nominal capacity, (3) nominal energy, (4) the battery type (e.g. form factor) and (5) the electrochemical system of the battery cells to be tested.
If the research did not produce any precise results, the capacity and nominal voltage were calculated from the information on the batteries as far as possible. For cells of the type 18,650 where no information was given or was not illegibly due to abrasion, a nominal capacity of 2.2 Ah and a nominal voltage of 3.7 V were assumed. This corresponded with the experience gained in the course of the tests. Other cell types than 18,650 where no information was given or found out had to be rejected.
Testing procedure
The actual current for discharging the battery cells was determined depending on each individual cell’s nominal capacity. The discharge started at a maximum discharge current (corresponding to C = 0.5), which remained constant as long as the cell voltage remained unchanged. When the voltage dropped, the discharge current was reduced by 0.1 A. The measurement was continued until a minimum discharge current (corresponding to C = 0.2) or a remaining voltage of 0.1 V was reached for at least 3 seconds. The maximum and minimum discharge currents were determined using the following equations:
where Imax is the maximum and Imin is the minimum discharge current (in A), Cmax and Cmin are the corresponding dimensionless C-factors of 0.5 or 0.2 and CN correspond to the nominal capacity of the battery (in Ah).
A SOC of 0% was assumed for battery cells whose voltage was so low that discharge according to the test procedure was not possible. However, that only happened when the cell voltage was ⩽0.1 V beforehand.
In addition to the parameters relevant for the testing procedure, the electrochemical system of the batteries (e.g. LCO, NMC, NCA and LFP) and their design (e.g. round cell, prismatic cell and pouch cell) were documented in the course of carrying out the experiment (e.g. research, possible dismantling, discharge measurements and data preparation).
Analysis
MS Excel was used for the manipulation and assessment of the determined raw data and for the calculation of SOC. R was used for statistical tests. The test report contains the measured voltage and the current at different intervals. From these three parameters, the energy was calculated and added up for each time interval.
The course of current, voltage and energy was then shown in a diagram for each measured cell. From the voltage curve, first conclusions could be drawn on the cell’s stability and the final evaluation compared with the results for the SOC (%). On the other hand, this made it possible to subsequently check the discharge parameters.
All the practical work (e.g. dismantling, data research, testing and analysis) of the batteries took place in the project laboratories of the Chair of Waste Processing Technology and Waste Management.
Results and discussion
A total of 1026 battery cells were tested in this study, 444 from the first and 582 from the second representative sampling campaign (see Table 1). Considering the different sizes of the samples, arithmetic mean as well as a weighted average for the values SOC thresholds 25% and 50% are given.
Overview of tested battery cells.
Not all sampled battery cells were included in the analysis, since in individual cases, an excessive deviation of the SOC occurred. That is expected due to errors in the discharge process. Such errors included:
manual operating errors in the test setup (e.g. wrong or invalid voltage settings) or
loose contacts between one of the poles of the cell/battery and the testing infrastructure (during the testing procedure).
Hence, a total of 980 cells were analysed. The dismantling of the batteries was often very time-consuming. Reasons for complex and challenging dismantling processes were:
firmly glued battery modules or components (e.g. pouch cells),
battery constructions that counteract non-destructive disassembly; for example inner housing screw connections or plug connections that are inaccessible due to solder lugs (mainly for tool batteries)
enclosures and casings whose construction was not designed for disassembly at all (e.g. laptop batteries).
In regard to the eco-design directive of the EU (Directive 2009/125/EC), these aspects are very important because the reparability and recyclability of electrical and electronic equipment (EEE) products are often already compromised in the design phase when safety aspects of the products’ use phase are considered but requirements of the end-of-life phase are not.
State of charge
The statistical distribution of the residual SOC is shown in Figure 3. Overall, a total of 23.6% of the batteries analysed had a residual SOC of at least 25% and 12.0% a residual SOC of at least 50%. In any case, the latter can be assumed to show a thermal runaway’s clear reactions in the event of damage.
Distribution of the residual SOC (%) of the cells/batteries analysed.
However, the investigation of the residues of a waste battery fire in December 2018 (Nigl and Kern, 2019) shows that the distribution of residual SOC (sample C; n = 100) of a certain batch of end-of-life portable batteries can deviate significantly from the overall distribution shown in Figure 3. The distribution of sample C compared to the two representative samples A and B are displayed in Figure 4. The main two reasons for the difference are:
special circumstances due to the collection area or origin of end-of-life portable batteries; for example waste batteries from commercial, industry or other points of the collection (such as electronics specialty stores) at which an average composition of end-of-life portable batteries cannot be assumed
and statistical deviation due to a sample size, that is too small or not representative.
Comparison of the two representative samples A and B (of this study) and the non-representative sample C (Nigl and Kern, 2019).
That has to be considered, especially in the risk analysis and assessment for individual portable battery collection systems in waste management. Safety and fire prevention measures (e.g. size and material of containers and bin used for battery collection) have to be adapted regarding the different circumstances mentioned above.
Battery types and electrochemical systems
Neither the design (e.g. type or form factor) nor the electrochemical system of the investigated batteries affected the distribution of the residual SOC significantly. However, the result of the literature research for the electrochemical systems was incomplete and sometimes contradictory and therefore had to be questioned. The reasons therefore are:
often no information regarding the electrochemical system was available
and sometimes, different sources of information regarding a certain battery contradicted each other.
An overview of the identified electrochemical systems is given in Table 2.
Overview of the identified electrochemical systems of the tested batteries/cells.
Cell voltage versus residual SOC
The relationship between measured residual cell/battery voltage and residual SOC is shown in Figure 5. Three aspects of the plot are remarkable: Firstly, cells/batteries with a measured cell voltage <3.0 V have a residual SOC close to zero. Secondly, cells or batteries with residual voltage <3.5 V are unlikely to undergo thermal runaway when exposed to heat or mechanical damage. In the management of waste batteries that fact could theoretically be the basis to check whether a waste cell/battery which looks critically damaged really poses a risk of fire.
Plot of the tested cell/battery voltage and the residual state of charge.
Third, the plot also shows that due to the very high range of variation above 3.5 V, only rough estimates of residual SOC can be made using the measured cell voltage. According to the plot, a cell with a residual voltage of 3.95 V could have a residual SOC between 30% and 85%.
The determined residual SOC referred to the nominal capacity of the batteries, which was taken from the product specifications. Due to the charge/discharge cycles which the cells have experienced in their life cycle, there is usually a reduction in the current capacity in the batteries’ end-of-life. This was not taken into account in the calculation since the determination of the actual capacity would involve considerable additional effort, which would have had a negative effect on the overall sample that could be examined in the project.
Conclusions
In this article, in-depth research of the residual SOC of waste lithium-ion batteries was conducted. A total of 980 cells/batteries was analysed using a prototypic laboratory test stand. The main findings of the work are:
Firstly, 23.6% of the cells and batteries analysed had a residual SOC of at least 25% and 12.1% had a residual SOC of at least 50%. That means that almost very fourth tested cell or battery is likely to show mild exothermic reactions, while almost every eighth cell or battery is very likely to show pronounced thermal runaway including severe exothermic reactions.
Secondly, significant deviations in the residual SOC distribution can occur, especially for small batches of end-of-life batteries when samples are not representatively taken. This is highly relevant for waste batteries collected from commercial origins (e.g. manufacturers or retailers of electrical and electronic equipment).
Thirdly, there is a distinct relationship between the actual cell voltage and the residual SOC of the respective cell: Below 3 V the residual SOC is close to zero, above approximately 3.2 V SOC is rising steeply on a linear basis. Results indicate that batteries with a cell voltage >3.5 V pose a serious risk of fire.
No statistical relationship regarding possible differences between the different types of batteries (e.g. electrochemical systems or types of construction) could be shown. However, it is concluded that changes in the composition of electrochemical systems that come onto the market are unlikely to shift the residual SOC’s average distribution. Whether that parameter has any influence on the future development of the distribution of residual SOC has to be clarified with the battery manufacturers’ assistance.
The authors conclude that the distribution of residual SOC is rather dependent on other factors, such as
consumer behaviour (e.g. hoarding of batteries and EEE in households) and
the ‘vital status’ of batteries in their end-of-life (e.g. their state-of-health and whether the batteries are still properly functioning, and how their self-discharge rates are).
Nevertheless, the determination of the average distribution of the residual SOC of waste portable lithium-ion batteries provides a better understanding and crucial data in order to assess the risks and hazards posed to waste management systems by portable batteries.
The SOC is one of the most important influence factors why lithium-ion batteries put waste management systems at increased risk of fire. Risk hot-spots (such as sample C of this study) have to be identified and safety and prevention measures adapted accordingly.
Further research would be indispensable to verify whether the distribution pattern is stable over more extended time periods.
Supplemental Material
sj-docx-1-wmr-10.1177_0734242X211010640 – Supplemental material for The fire risk of portable batteries in their end-of-life: Investigation of the state of charge of waste lithium-ion batteries in Austria
Supplemental material, sj-docx-1-wmr-10.1177_0734242X211010640 for The fire risk of portable batteries in their end-of-life: Investigation of the state of charge of waste lithium-ion batteries in Austria by Thomas Nigl, Tanja Bäck, Stefan Stuhlpfarrer and Roland Pomberger in Waste Management & Research
Footnotes
Acknowledgements
The authors would like to thank the project partners (Elektro Recycling Austria (ERA) GmbH, Green Tech Cluster GmbH and Saubermacher Dienstleistungs AG) for their valuable support.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the Austrian Research Promotion Agency (Österreichische Forschungsfoerderungsgesellschaft; FFG grant number 850753).
Supplemental material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
