*1.6. Study Objectives*

The article presents a new method for risk evaluation and the assessment of portable batteries and. especially, lithium-based batteries in waste management systems strongly focused on the first term of Equations (1) and (2).

The following research questions are answered:


### **2. Materials and Methods**

Risk analysis is the systematic use of information in order to identify hazards and estimate risks to individuals, property, and the environment. There are three main steps in the analysis of risks: (1) hazard identification, (2) frequency analysis, and (3) consequences analysis [16].

Because the exact number of fires triggered by lithium-ion batteries is not even approximately known [1], no direct probability of occurrence can be achieved [13]. Therefore, an alternative approach is taken in this study to quantify the risks of lithium-ion batteries in waste management systems.

The probability of a battery-caused fire incident (*FIBC*) can be expressed as the product of two terms: (1) the probability of the presence of a hazardous battery and (2) the probability of critical damage done to the present battery (Equation (3)):

*F IBC* = *P*(*presence o f a hazardous battery*) × *P*(*critical damage to the present battery*) (3)

Figure 1 displays the overall approach and the boundaries of this study. After the basic step of literature research, two different workflows were followed. The first was covered by two previous publications and it is about the presence of a hazardous battery; the second, is about the critical damage done to a present battery (which, consequently, leads to a battery-caused fire) and it is covered by this study. Afterwards, the results are combined in the overall risk modelling and assessment.

**Figure 1.** Overall approach and boundaries of this study (grey box).

More specifically, the first part of Equation (3) depends on the following influence factors (upper part of Table 3): (1) battery content in the respective waste stream [items/tonne], (2) percentage of lithium batteries within these batteries (both [5]), and (3) the percentage of lithium batteries that contain a critical state of charge (SOC) [13]. The second part of the equation depends on the (1) number and intensity of potential damaging incidents, (2) the size and shape of the respective batteries, and (3) the energy content of the batteries and ambient conditions (lower part of Table 3).

**Table 3.** Overview of the relevant influence factors on the presence of a fire-hazardous battery (1.1–1.3) and on critical damage to the present battery (2.1–2.3).


The parameters (1) the degree of damage, (2) size, and the shape (construction type) of the portable batteries were investigated for this study. Portable batteries that were collected during waste sorting and the characterisation campaign of the input fractions residual waste, lightweight packaging waste, metal packaging waste (compared to previous publication of [5]), and commercial waste and specific output waste fractions (of treatment plants for residual waste and lightweight packaging).

### *2.1. Degree of Damage*

In Table 4, the description of the defined damage classes (DC) and respective degree of damage is given. DC 1 to 3 are considered as minorly damaged and DC 4 to 7 as majorly damaged. The classification was made visually during the manual documentation of the portable batteries in the laboratory. Sample pictures of the damage classes of portable batteries are given in Supplementary Materials.

**Table 4.** Overview of damage classes and description of the degree of damage.


Thereafter, the distribution of the degree of damage in the investigated waste fractions was used in order to model the potential damages, the portable batteries must endure between both (1) the intention to dispose of and the arrival at the waste treatment site and (2) the arrival at the treatment site and after running through the treatment process. Because of financial and time restraints, this was done for the fractions of residual household waste and lightweight packaging waste only.

As a background reference for the degree of damage that is occurring already during the use phase, the separated collection of portable batteries was also sampled and investigated accordingly.

### *2.2. Size and Shape of Batteries*

The size was measured as the maximal edge length (mm) of the portable batteries, and the shape (the construction type, e.g., round cell, button cell, prismatic cell, and pouch cell) of the portable batteries was documented. These parameters are considered to have an influence on the likelihood of critical damage (causing a waste fire). Hence, the distribution patterns of these parameters were investigated and documented.

### *2.3. Energy Content and Surrounding Conditions*

When considering the various possible scenarios, the exact thermodynamic modelling of the potential ignition of waste that is caused by a portable battery must be based on many parameters. Thus, many underlying assumptions would be necessary.

For example, the ignition temperature of different plastic types ranges from 350 ◦C (polyethylene) to 580 ◦C (polytetrafluoroethylene) [17]. Furthermore, [18] presented significantly lower values for the initial combustion temperature for plastics (254–354 ◦C) and paper materials (240–260 ◦C). Golubkov et al. [11] showed that these temperatures are easily met when lithium-ion batteries (of type 18650) with an SOC >25% (for NCA) or SOC >50% (for LFP) undergo thermal runaway. However, it is known that a lower heating rate provides a lower ignition temperature, and vice versa [19]. A significant discrepancy in that regard can be seen, as [16] used a constant heating rate of 20 K/min. for defining the ignition temperature. Jhu [20] showed, under adiabatic conditions in a closed test can, that the peak temperature of the thermal runaway reaction of a fully charged 18650 lithium-ion cells is able reach its maximum temperature of 654.3 ◦C in 0.49 s starting from 125.2 ◦C. In these experiments, four cells from different worldwide battery producers have been investigated.

Regarding the released energy from the runaway reaction, [21] measured 102 to 218 kJ per Ah, depending on cell chemistry and respective SOC for LFP and NMC cells. Consequently, a typical lithium-ion cell (type 18650, 2.6 Ah) would release thermal energy in the range of 265 and 567 kJ. This energy can easily ignite various waste materials; still, the duration of the thermal runaway and ambient parameters will influence the ignition's success. The two main types of parameters can be distinguished between battery- and ambient-based.

The battery-based parameters depend on the type and SOC of the battery and they affect the following:


Ambient-based parameters are:


### *2.4. Risk Modelling and Assessment*

Regarding risk modelling, the following assumptions were made:


The probability of occurrence followed typical grading schemes for risk graphs/risk matrices (e.g., as low as reasonably practicable, ALARP) e.g., [13,24]:

The probability of occurrence:


Property damage/losses:


### **3. Results**

Table 5 shows the identification and risk assessment of the possible hazards and threats of portable batteries for different facility areas or processes along the value chain of a residual household waste system.

### *3.1. Damage Degree*

Table 6 presents the distribution of investigated portable batteries' maximal edge length. The main findings are:


• Larger portable batteries (maximal edge length >30 mm) show a relatively high share of the DC 4 to 7.

**Table 5.** Qualitative risk assessment of possible hazards and threats of portable batteries (waste stream: residual household waste).


**Table 6.** Distribution of maximal edge length of the investigated portable batteries.


In Figure 2, the distribution of the portable batteries' damage degree according to the waste stream in which they were found.

Only two per cent of the spent batteries show a relevant degree of damage, according to the background reference sample (separated portable battery collection, SBC) (DC 4 to 7). The vast majority of over 97% of the batteries are undamaged (DC 1). No portable battery was found, which was heavily bloated or in a post-thermal runaway state (DC 7).

The input fractions results show that, on the one hand, portable batteries can already be subject to varying degrees of mechanical damage during collection, e.g., due to the force that is applied in press containers of collection vehicles. That is particularly remarkable when compared to the separate collection of portable batteries, as batteries are or will be damaged only to a minor extent in this collection system.

On the other hand, there is a significant degree of damage to portable batteries in processing plants, for example, during the conditioning of the waste stream by a preshredder (e.g., in residual waste treatment) or a bag opener (e.g., in lightweight packaging waste sorting), but also during the manipulation of the waste stream (while using a wheel loader or gripper). Other impacts, such as (1) shock loads, which may occur due to different drop heights in treatment processes (e.g., by falling into an output box or from one conveyor belt to another), or (2) vibration of vibrating conveyors can also damage portable batteries.

Moreover, when comparing the residual household waste before (RHW (in)) and after (RHW (out)) treatment in a respective facility, it is noticeable that the percentage of

the lower degrees of damage (DC 1 to 3) are reduced, while the percentage of the higher damage classes (DC 4 to 6) are increased (from approximately 6% to over 40% in total).

**Figure 2.** Distribution of damage degree according to the waste stream in which the batteries were found (Legend: RHW = residual household waste, LPW = lightweight packaging waste, MPW = metal packaging waste, CW = commercial waste, SBC = separate portable battery collection; values < 2% not displayed).

When comparing lightweight packaging waste before (LPW (in)) and after (LPW (out)) treatment (sorting plant), the effect is similar, but yet not that severe, which results in an increase of DC 4 to 6 from almost 6% to 14% in total. The share of DC 4 to 6 in commercial waste (CW (in)) is very similar to the input fractions of residual household waste and lightweight packaging waste, and it is notably much higher in metal packaging waste (MPW (in)).
