The findings of this study are contingent upon the establishment of our proprietary battery gigafactory in Germany21. As part of the factory planning process, the material flows depicted here were determined, the filter technology was designed in accordance with these specifications, and the relevant planning permission was applied for. Concurrently, the methodologies and data were employed in the conceptualization of additional battery manufacturing facilities across Europe, with a particular focus on the development of the abatement system.
Glossary
NMC811
Li1Ni0.8Mn0.1Co0.1O2
Lithium nickel manganese cobalt oxide
LFP
Li1Fe1P1O4
Lithium iron phosphate
Carbon black
C
Carbon black
PVDF
(C2H2F2)n
Polyvinylidene fluoride
NMPC
5H9
NON-Methyl-2-pyrrolidone
Graphite
C
Graphite
SiO2
SiO2
Silicon dioxide
SBR
C8H8, C4H6
Styrene-butadiene rubber
CMC
CH2CO2
HCarboxymethyl cellulose
Al
Al
Aluminum
Cu
Cu
Copper
Al2O3
Al2O3
Aluminum oxide
HF
HF
Hydrofluoric acid
EMC
C4H8O3
Ethyl methyl carbonate
VC
C2H3
ClVinyl chloride
FEC
C3H3FO3
Fluoroethylene carbonate
CO
CO
Carbon monoxide
H2
H2
Hydrogen
C2H4
C2H4
Ethylene
CH4
CH4
Methane
C2H6
C2H6
Ethane
CO2
CO2
Carbon dioxide
VOC
–
Volatile organic compound
Cell chemistries and cell formats use
We are analyzing a battery gigafactory, once designed for the production of NMC811 battery cells and once designed for LFP battery cells. We use prismatic cell housings with the dimensions 100 x 288 x 30 mm for NMC811 battery cells and 100 x 288 x 42 mm for LFP battery cells. The LFP battery cells are thicker, due to the higher thickness of the LFP coating and thus of the electrode stack. The NMC811 cells have a storage capacity of 124 Ah, the LFP cells of 136 Ah. For better comparability, both cells each have the same nominal energy of 446 Wh. A production line can manufacture 34 cells per minute. In a gigafactory with 4 production lines, this corresponds to an annual production capacity of 32 GWh/a for NMC811 cells and 28 GWh/a for LFP cells.
Determination of process emissions before the abatement system
The following describes how the process emissions before the abatement system were determined. The data is based on information provided by the equipment manufacturers or on standard calculation methods.
Dosing and mixing
The generated powder particles in the air during dosing are sucked off directly for each powder separately. The dust generation is calculated via a “mine dump approach” based on amount of material, bulk density and discharge height according to VDI 3790 part 3:
$${q}_{{Ab}}={q}_{{norm},{korr}} \, {{\cdot }} \, {\rho }_{S} \, {{\cdot }} \, {k}_{U}$$
(1)
with
$${q}_{{norm},{korr}}={q}_{{norm}} \, {{\cdot }} \, {k}_{H} \, {{\cdot }} \, 0.5 \, {{\cdot }} \, {k}_{{equipment}}$$
(2)
and
$${k}_{H}={\left(\frac{{H}_{{free}}+{H}_{{tube}}+{k}_{{friction}}}{2}\right)}^{1.25}$$
(3)
The emission factor of dumping qAb is given in g/tMateria,. The empirical correction factor for the transshipment equipment is defined as kequipment = 1 because there is a continuous loading equipment. The friction factor is set to kfriction = 0 as there is a loading head, i.e. irrespective of the pitch. The local factor is kU = 0.1 due to the presence of a hopper with dust extraction. Because the slurry is stirred under vacuum, during mixing there is no exhaust air. Binder mixing, dry mixing and buffer tanks were not taken into account in the calculation. Regarding emissions to water, a fairly frequent cleaning interval of 15 minutes per day with 20 m³/h flow of deionized water is assumed to clean the mixer pipes and tanks. This cleaning process produces wastewater contaminated with slurry with a dilution of 1:20. The weight distribution of the substances in the wastewater can be derived using the bill of materials (BOM). The calculated data is confirmed by the machine equipment manufacturer and by own factory acceptance tests.
Coating and drying
Emission in coating and drying are wastewater from cleaning the nozzles, NMP remaining in the exhaust air in cathode coating and water vapor remaining in the exhaust air in anode coating. In coating and drying no particle emissions in the exhaust air occur.
The weight distribution of the substances in the wastewater is the same as in the mixing process. The solvent recovery of NMP is carried out by pre-cooling of the stream in heat exchanger with the recirculating air, followed by condensation with cooling and chilled water – reducing the NMP concentration in the air by 87%. Approximately 20% of the circulating air is diverted as exhaust air. Thus 3% of the initial amount of NMP is emitted, while 97% are recovered or kept in the close loop. For anode coating, where water is used as solvent, no solvent recovery takes place. The entire air stream is exhausted in the environment. The data is provided by the machine equipment manufacturer and obtained/confirmed by own factory acceptance tests.
Calendering
Exhaust air is generated by suction at the cleaning brushes, which remove dust from the coating after the calendaring process. The mass of dust removed and emitted to a air flow is based on the equipment manufacturer’s specifications. The dust itself consists of the materials of the cathode and anode coating according to the BOM. The data is provided by the machine equipment manufacturer and obtained/confirmed by own factory acceptance tests.
Slitting
Exhaust air is created by the suction of the dust produced on the cutting blades. Dust amount is calculated via cutting speed of 60 m/min, cutting thickness of 25 µm and coating thickness. The cross section of the blade respectively the cutting edge, which cuts the material, is rectangular. The assumption is that this theoretical rectangle cutting cross section will turn into dust. The dust also consists of the materials of the coating according to the BOM and in addition of the aluminum or copper foils. The data is provided by the machine equipment manufacturer and obtained/confirmed by own factory acceptance tests.
Vacuum drying
The vaporized NMP or water vapor end up in the cold trap of the pump producing wastewater with 25% NMP concentration. The amount of wastewater in the cold trap for all vacuum ovens is specified by the manufacturer as 2 l/h. The data is provided by the machine equipment manufacturer and obtained/confirmed by own factory acceptance tests.
Separating and Z-Folding
Each separator and Z-folding machine achieve a throughput of 4 ppm. Therefore, parallelization is carried out here to achieve a throughput of 34 parts per minute, which is specified by the coater. The separator punches out 6 sheets/s and when folding, a sheet is inserted in 0.5 s. Dust amount is again calculated via cutting thickness of 25 µm and coating thickness. The data is provided by the machine equipment manufacturer and obtained/confirmed by own factory acceptance tests.
Packing and laser welding
In order to achieve the throughput of 34 ppm here as well, four laser welding points are required to connect the cap with the stacks to the can and thus to close the cell. The throughput is limited here by the speed of the laser robot of 70 mm/s. The emission is aluminum oxide fume with a quantity of 2 mg/s. In this way, the total emissions can be calculated over the length of the weld seam. The data is provided by the machine equipment manufacturer and obtained/confirmed by own factory acceptance tests.
Electrolyte filling
The emissions consist of the evaporating electrolyte components that are sucked out of the cell during vacuuming. Gas amounts are calculated via evaporation rates, the area of the filling port (diameter of 6 mm) and the number of open cells at the same time (eight to six). Among the components that vaporize are EMC, VC, FEC, TMP but also HF due to very small unavoidable amounts of water in the electrolyte. The data is provided by the machine equipment manufacturer.
Finalization
During degassing, formation gas is released as an emission, which contains carbon monoxide, hydrogen, ethene, methane, ethane, carbon dioxide, electrolyte carbonates and other decomposition products of carbonates. The gas amount is 7 ml/Ah. The possible proportions of the gases are between 1% and 100%, can only be determined experimentally and differ greatly from cell chemistry, cell design, environmental conditions etc. That’s why the large ranges are given here. For this study the mean value is used. After the second electrolyte filling, the filling opening is finally sealed using laser welding. The same emissions occur as in the assembly with evaporating electrolyte components and aluminum oxide fume. Therefore, the same calculation approach is used but with different boundary conditions (half as many cells open at the same time and shorter weld seam of the filling port). The data is provided by the machine equipment manufacturer.
The particle sizes are always given as a size distribution. There are three different parameters that are used to describe them. These are D10, D50 and D90 and indicate how many particles of the substance (10%, 50% or 90%) are smaller than the specified value. The particle sizes used in this study are based on a literature and information provided by the equipment manufacturers.
Determination of the legal emission limits in Germany
The basis for the emission limits in Germany is the “Bundes-Immissionsschutzgesetz” (Federal Immission Control Act)24 and the “Verordnung zur Durchführung des Bundes-Immissionsschutzgesetzes“(Ordinance on the Implementation of the Federal Immission Control Act)25 derived from it. The specific technical instructions for implementing these are the administrative regulation “Technische Anleitung zur Reinhaltung der Luft“ (Technical Instructions on Air Quality Control26) for emissions into the air. This lists limit values as mass flow or concentration for the analysed substances. Appendix 2 specifies how the atmospheric dispersion must be calculated. Appendix 6 specifies the maximum soil concentration that may be reached (“S value”) and according to which the stack height must be calculated. The documents are available online in German language. An automated translation using translation tools is possible to understand the specifications and to reproduce the calculations.
Determination of the filter performance of the abatement system
When calculating the filter performance, the mass flow is reduced by the filter-specific separation efficiency. For filtering dust, we use an F8 pre-filter according to EN 779:201228 and an H14 fine dust filter according to EN 1822-1:201929. For the F8 pre-filter, we use separation efficiencies of 95% for PM10, 85% for PM2.5 and 77.5% for PM1 dusts31. For the H14 filter, we use separation efficiencies of 99.995 %29. For the cold trap we use separation efficiencies of 99.9% and for the thermal oxidizer 99.5%. The separation efficiency of the NMP recovery system is 1 mg/m³. These numbers are provided by the abatement system manufacturers. Supplementary table 2 summarizes the used abatement systems as well as their achievable retention.
Calculation of atmospheric dispersion
We use a Gaussian plume model according to VDI 3782 part 1 to calculate the atmospheric dispersion32. This is valid because the dispersion situation is stationary, the terrain is flat, the dispersion is undisturbed and wind speeds > 1 m/s are present32. According to VDI 3782 part 1, the concentration of substances in the air is calculated as follows:
$${{{\rm{c}}}}\left({{{\rm{x}}}},{{{\rm{y}}}},{{{\rm{z}}}}\right)= \frac{{{{{\rm{Q}}}}}_{0}}{2\cdot \pi \cdot {{{\rm{u}}}}\left({{{\rm{x}}}}\right)\cdot {{{{\rm{\sigma }}}}}_{{{{\rm{y}}}}}\left({{{\rm{x}}}}\right)\cdot {{{{\rm{\sigma }}}}}_{{{{\rm{z}}}}}\left({{{\rm{x}}}}\right)}\cdot \exp \left[-\frac{{{{{\rm{y}}}}}^{2}}{2\cdot {{{{\rm{\sigma }}}}}_{{{{\rm{y}}}}}^{2}\left({{{\rm{x}}}}\right)}\right]\cdot \\ \left\{{\sum }_{{{{\rm{n}}}}=-\infty }^{+\infty }\exp \left[-\frac{{\left[z-{h}_{n}^{+}\left(x\right)\right]}^{2}}{2\cdot {\sigma }_{z}^{2}\left(x\right)}\right]+{{{\rm{a}}}}\left({{{\rm{x}}}}\right)\cdot {\sum }_{{{{\rm{n}}}}=-\infty }^{+\infty }\exp \left[-\frac{{\left[z-{h}_{n}^{-}\left(x\right)\right]}^{2}}{2\cdot {\sigma }_{z}^{2}\left(x\right)}\right]\right\}$$
(4)
It is defined: the mass flow Q0, the wind speed u, the standard deviation of the Gaussian plume σy(x) and σz(x), the effective source height hn+ (reflected on the ground hn-) and a factor a(x) which takes into account the mass transfer from the air to the ground. As we are only considering the immediate surroundings of the factory, we use n = 0.
σy(x) and σz(x) are calculated as followed:
$${\sigma }_{y}(x)=F\cdot {x}_{r}{\left(\frac{x}{{x}_{r}}\right)}^{f}$$
(5)
$${\sigma }_{z}(x)=G\cdot {x}_{r}{\left(\frac{x}{{x}_{r}}\right)}^{g}$$
(6)
We assume a dispersion class III/1 (indifferent/neutral) according to Klug/Manier, see VDI 3782 part 633. For this following values are given: F = 0.640, f = 0.784, G = 0.215, g = 0.88532 (see VDI 3782 part 6, table 5a). The parameters in this standard were obtained by empirical experiments.
The chimney height is calculated on the basis of the dispersion calculation. The effective source height must be specified so that the “S-values” of the pollutants emitted are not exceeded. The S-value is defined as the “Concentration of the air pollutant that must not be exceeded” (TA-Luft, Chapter 5.5.2.2)26. The maximum permitted limit value for nickel and its compounds is defined as 0.00052 mg/m³ or 0.52 µg/m³ (TA-Luft, Annex 6, Chapter 5.5)26. However, the target value for nickel dust in the air in the EU is 0.02 µg/m³ 22.
Calculation of emissions to soil
The calculation formula for the emission in soil is specified according to VDI 3782 part 534 with:
$${F}_{d}={v}_{d}* c$$
(7)
The deposition velocity vd is specified as (see “Technische Anleitung zur Reinhaltung der Luft”, Annex 2, Chapter 4)26: vd = 0.01 m/s for particles with an aerodynamic diameter da of 2.5 to 10 µm. For an aerodynamic diameter da of 10 to 50 µm, vd = 0.05 m/s is specified. The limit value for emissions to soil for nickel and its inorganic compounds is also specified in “Technische Anleitung zur Reinhaltung der Luft”, at 15 µg/(m²d) (annual mean value), in Chapter 4.5.1, Table 626.
Disclaimer
We can neither confirm nor deny whether gigafactories with the emission values mentioned in scenario A (law-minimum) are or will actually be built in Germany. The final environmental assessment in accordance with the Federal Immission Control Act (BimSchG) and the associated planning permission are in the hands of the responsible authorities. In theory, they would have to prohibit a building permit as soon as the limit value for emissions in soil, averaged over the year, exceeds 15 µg/(m²d). As we used and applied for the filter technology from scenario B (achievable) in our own gigafactory project, we were unable to verify the effectiveness of the official assessment ourselves.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.