Application Note

Environmental

Determination of 14 Impurity
Elements in Lithium Carbonate
Using ICP-OES
Routine quality control of raw materials used to
produce cathode material for lithium ion batteries

Authors Introduction
Feng Wenkun As part of their efforts to combat climate change, many countries are reducing
Agilent Technologies the amount of energy that is generated from burning fossil fuels by switching to
(China) Co., Ltd. renewable energy sources. However, to meet the energy demands of the transport,
heating, or industrial sectors from renewables, efficient and cost-effective energy
storage technologies are needed. Lithium (Li)-ion batteries are widely used in
consumer electronic devices due to their large capacity, high voltage, light weight,
and lack of memory effects. These properties also make them attractive for use
in electric or hybrid vehicles. Among the components of a Li-ion battery (cathode,
anode, separator, and electrolyte), the materials used for the cathode account
for more than 40% of the total cost of the battery. Lithium carbonate (Li2CO3) is a
key raw material that is used to prepare cathode materials and the quality of the
material determines the price. Also, any impurities present in the raw materials
used to make the battery can affect its performance and sample dilutions or multiple readings of the same sample,
lifespan. Therefore, impurity analysis of battery-materials is increasing sample throughput. Also, avoiding extra sample
vital for suppliers of the raw materials, battery manufacturers, dilutions prevents introducing potential errors caused by
and emerging industries such as the electric vehicle power large dilutions or reducing the analyte signal below its
battery sector (1–3). detection limit. The 5110 features a cooled cone interface
The elemental analysis of battery-grade Li2CO3 is often based (CCI) that improves the reliability of the results by reducing
on the methods outlined in International Electrotechnical the formation of EIE interferences. Instrument operating
Commission (IEC) 62321 standard and in Chinese standard parameters are shown in Table 1.
GB/T 11064.16-2013 (4, 5). Both standards use an external
calibration ICP-OES method for the analysis. The presence Table 1. Operating parameters of the 5110 VDV ICP-OES.
of many lithium ions in the plasma from Li2CO3 can affect
Parameter Setting
the analysis of easily ionized elements (EIEs) such as Na
Viewing Mode Axial
and K by ICP-OES, leading to false-positive results. To
Read Time (s) 10
minimize or eliminate the EIE interferences, the plasma
of the ICP-OES can be viewed radially. But radial view Replicates 3

measurements typically lack the sensitivity required for Stabilization Time (s) 8

the analysis of trace elements in battery-grade Li2CO3. Fast Pump (rpm) 60

Also, the ultraviolet visible (UV-Vis) spectrophotometric Pump Speed (rpm) 12
method adopted for the measurement of silicon (Si) in GB/T Power (kW) 1.2
11064.16-2013 lacks accuracy and is complicated. Being Plasma Gas Flow (L/min) 12
able to determine all target elements, including Si, in a single Auxiliary Gas Flow (L/min) 1.0
analysis by ICP-OES would simplify the analysis, improve the Nebulizer Gas Flow (L/min) 0.7
quality of the data, and increase productivity.
In this study, the Agilent 5110 Vertical Dual View (VDV) ICP-OES
was used for the measurement of Al, Ca, Cd, Cu, Fe, K, Mg, Background correction
Mn, Na, Ni, Pb, S, Si, and Zn in Li2CO3 using the method of The ICP Expert instrument control software includes
standard addition (MSA). MSA was used to minimize the easy-to-use background correction techniques including
differences between the standards and samples. It also Fitted Background Correction (FBC). FBC accurately corrects
allows Na and K to be measured axially, ensuring that all background structures without user intervention or method
elements are measured in a single plasma view mode. development. An example of automatic background fitting
using FBC from this study is shown for Pb 220.353 in Figure 1.
Experimental FBC provides accurate correction of the background
Instrumentation structure that can arise from other emission lines in the
The Agilent 5110 VDV ICP-OES with a vertically orientated sample, allowing more accurate measurements (6).
torch was used in this study. The sample introduction
system consisted of a SeaSpray concentric glass nebulizer,
double-pass glass cyclonic spray chamber, and an Easy-fit
demountable VDV 1.8 mm inner diameter (id) injector torch.
The 5110 ICP-OES uses a solid-state radio frequency (SSRF)
generator operating at 27 MHz to deliver a plasma with the
stability and robustness necessary for the analysis of high
matrix samples such as Li2CO3. The fast processing speed of
the VistaChip III CCD detector, up to 1 MHz pixel processing
speed, reads the full wavelength range in a single exposure.
With such fast analysis speeds, the argon gas consumption
per sample is typically lower than other ICP-OES instruments.
The detector also provides a wide dynamic linear range for
effective analysis of trace and major elements in the same
Figure 1. Automatic FBC has accurately modeled the Nb 220.363 interfering
run. The wide analytical range eliminates the need for extra peak on Pb 220.353 nm.

2
Reagents, standard solutions, and samples Method detection limits
High purity nitric acid was bought from Suzhou Jingrui The MDLs shown in Table 2 are based on three sigma
Chemical Company (China). Calibration standards were measurements of 11 samples of the matrix solution that
prepared using Agilent standards, including a 10 mg/L were measured continuously. All MDLs were well below
multi-element standard solution 2A and a 10 mg/L 1.0 mg/kg, allowing impurities to be detected and monitored
multi-element standard solution-4. A commercial Li2CO3 at low levels in battery-grade Li2CO3.
sample was bought for the study.
Table 2. Method detection limits for 14 elements.
Sample preparation
To prepare the Li2CO3 sample for analysis, 0.3 g of solid Element and MDL in Solution Element and MDL in Solution
Wavelength (nm) (μg /L) Wavelength (nm) (μg /L)
sample was weighed into a PTFE digestion tube. 2–3 mL of
concentrated HNO3 was added slowly. (SiO2 is removed from Al 396.152 1.1 Mn 257.610 0.03
Li2CO3 during the extraction process, so no HF was added Ca 422.673 0.080 Na 589.592 0.08
during the digestion of Li2CO3). The tube was covered and Cd 226.502 Ni 216.555
0.16 1.9
heated on a hotplate for 120 minutes at 120 °C until
Cu 327.395 0.29 Pb 220.353 8.2
the solution was clear and transparent. After cooling,
the solution was made up to 30 mL with ultrapure water Fe 238.204 0.42 S 181.972 6.0
(Millipore Milli-Q). For the long-term stability test, the sample K 766.491 0.24 Si 251.611 5.6
was spiked at 50 μg /L. Mg 279.553 0.030 Zn 213.857 0.65
At the same time as preparing the sample, another aliquot
of the Li2CO3 sample was prepared to be used as a matrix Sample analysis and spike recovery
solution for the MSA calibration. A digestion blank was also
The digest of the commercial Li2CO3 sample was analyzed
prepared by digesting 2–3 mL HNO3 and making up to 30 mL.
using the 5110 ICP-OES method and the results are given in
The sample, spiked sample, and digestion blank were Table 3. To test the accuracy of the measurements, a spike
prepared in duplicate. recovery test was carried out at a concentration of 50 μg/L.
Preparation of MSA calibration standards The recoveries of all elements were between 95 and 102%
To prepare the multi-element calibration standards, 100 μL of (Table 3), confirming the suitability of the method for the
the Agilent multi-element standard solution-2A was diluted to determination of impurities in Li2CO3 at trace levels.
5 mL with the matrix solution to a concentration of 0.2 mg/L.
50, 100, 200, 500, and 1000 μL of this intermediate standard Table 3. Sample and spike recovery results for elemental impurities in
battery-grade Li2CO3.
solution were then diluted to 2 mL with the matrix solution.
This method produced a series of standards solutions at Element and Measured Sample + Spike Recovery
Wavelength (nm) Concentration in (%)
0.005, 0.01, 0.02, 0.05, and 0.1 mg/L. at 50 (μg /L)
Sample (µg/L)
A series of standard solutions of S and Si were prepared Al 396.152 1.3 52 100
at 0.01, 0.02, 0.05, and 0.1 mg/L. An intermediate standard
Ca 422.673 19 68 99
solution for S and Si was prepared by diluting 100 μL of
Cd 226.502 50.5 101
multi-element standard solution-4 to 5 mL with the matrix <MDL

solution. 100, 200, 500, and 1000 μL of the 0.2 mg/L Cu 327.395 0.50 52 102
intermediate standard were diluted to 2 mL with the Fe 238.204 1.8 51 99
matrix solution.
K 766.491 13 63 100

Results and discussion Mg 279.553 0.60 51 100

Calibration and linearity Mn 257.610 0.10 50 100

Using MSA, linear calibration curves were obtained for Na 589.592 11 60 98

all 14 elements. The calibration correlation coefficients Ni 216.555 <MDL 52.1 104
were greater than 0.9995 in all cases. Figure 2 shows Pb 220.353 51.2
<MDL 102
the calibration curves for seven representative elements
S 181.972 30 78 96
including Na, K, Ca, Cd, Fe, Mg, and Si.
Si 251.611 7.1 56 98

Zn 213.857 1.4 52 100

3
Figure 2. MSA calibration curves for Na, K, Ca, Cd, Fe, Mg, and Si.

4
Long-term stability References
The long-term stability of the 5110 VDV ICP-OES was tested
1. Determination of Elemental Impurities in
by continuously analyzing a sample solution of Li2CO3 spiked
Graphite-based Anodes using the Agilent 5110 ICP-OES,
at 50 μg/L for 2.5 h. The relative standard deviation (RSD)
Agilent publication, 5991-9508EN
of all elements was less than 2%, as shown in Table 4. The
excellent precision demonstrates the stability, reliability, and 2. Determination of Elements in Ternary Material
robustness of the 5110 ICP-OES over the duration of the Nickel-Cobalt-Manganese Hydride, Agilent publication,
analytical run. 5991-9506EN
3. Rapid Analysis of Elemental Impurities in Battery
Table 4. Long-term stability results for 14 elements measured in Li2CO3 Electrolyte by ICP-OES, Agilent publication, 5994-1937EN
over 2.5 h.
4. IEC 62321:2008, Electrotechnical products -
Element and RSD Element and RSD Determination of levels of six regulated substances
Wavelength (nm) (%) Wavelength (nm) (%) (lead, mercury, cadmium, hexavalent chromium,
Al 396.152 1.2 Mn 257.610 1.1 polybrominated biphenyls, polybrominated diphenyl
Ca 422.673 0.7 Na 589.592 0.8 ethers), accessed October 2020, since withdrawn
Cd 226.502 1.2 Ni 216.555 1.9 5. GB/T 11064.16-2013 Methods for chemical analysis of
Cu 327.395 0.6 Pb 220.353 1.7 lithium carbonate, lithium hydroxide monohydrate, and
Fe 238.204 1.2 S 181.972 1.9 lithium chloride. Part 16: Determination of the amounts
K 766.491 0.6 Si 251.611 1.8 of calcium, magnesium, copper, lead, zinc, nickel,
Mg 279.553 1.3 Zn 213.857 1.5 manganese, cadmium, and aluminum
Inductively-coupled plasmas Bulk atomic emission
spectrometry, accessed October 2020, https://www.
chinesestandard.net/PDF.aspx/GBT11064.16-2013
Conclusion
6. Fitted Background Correction (FBC)—fast, accurate
The Agilent 5110 VDV ICP-OES was used to quantify 14
and fully automated background correction, Agilent
impurity elements in battery-grade lithium carbonate—a raw
publication, 5991-4836EN
material that is used in to produce the cathode of a Li-ion
battery. Silicon was included in the ICP-OES method, avoiding
the need to use a separate measurement technique such
as UV-Vis. The fast processing speed of the Vista Chip III
detector enabled all data to be collected in 60 seconds,
boosting productivity, and saving on argon gas consumption
and energy costs.
Recoveries of all 14 elements spiked at 50 µg/L into the
Li2CO3 sample digest were within ±10% of the spiked
concentrations. And the RSD of the 2.5 h stability test results
was less than 2%. These results demonstrate the accuracy,
stability, and reliability of the method and its suitability for
the large-scale, routine QC of raw materials used in the
production of Li-ion batteries.

www.agilent.com/chem

DE44158.12

This information is subject to change without notice.

© Agilent Technologies, Inc. 2020

Printed in the USA, December 9, 2020
5991-9507EN