Application Note

Energy and Chemicals

Analysis of Trace Carbon Dioxide and

Permanent Gas Impurities in Fuel Cell
Hydrogen and High-Purity Hydrogen
by GC
Achieve low detection limits using an Agilent 8890
GC fitted with a pulsed discharge helium ionization
detector (PDHID)

Author Abstract
Li, Wan
An Agilent 8890 GC fitted with a plug and play pulsed discharge helium ionization
Agilent Technologies
detector (PDHID) was used to detect ppb to low-ppm concentration carbon dioxide,
(Shanghai) Co. Ltd.
carbon monoxide, methane, argon, oxygen, nitrogen, and other impurities in high-
purity hydrogen. The GC method provided low detection limits, good reproducibility,
and excellent linearity, meeting the requirements specified in GB/T 3634.2-2011,
GB/T 37244-2018, and ISO 14687-2019.
Introduction Traditional GC methods based on simultaneous qualitative and quantitative
flame ionization detection (FID) with detection of impurities in high purity
China has set a goal of becoming methanizer have a detection limit for hydrogen. Carbon dioxide (CO2), carbon
carbon-neutral by 2060. To help achieve carbon monoxide of 50 to 100 ppb, monoxide (CO), methane (CH4), argon
the target, the Chinese government is which is above the sensitivity needed (Ar), nitrogen (N2), and other impurities
supporting the development of new for some users of hydrogen. However, a were measured from a single injection
technologies such as hydrogen energy GC fitted with a pulsed discharge helium into the GC. The method was evaluated
and hydrogen fuel cells for vehicles, ionization detector (PDHID) can lower the against the requirements for sensitivity,
and hydrogen refueling stations. The carbon monoxide detection limit below reproducibility, and linearity specified in
importance of the hydrogen fuel cell 50 ppb, which is welcomed by many GB/T 3634.2-20112, GB/T 37244-20183,
vehicle industry to China is underlined users in China. and ISO 14687-20194.
by its inclusion in various strategic
The PDHID uses a pulsed DC discharge
plans. The plans include the "National
in helium as its ionization source. The Experimental
Innovation-Driven Development Strategy
column effluent, which flows counter to
Outline", "Made in China 2025", and Samples
the discharge helium flow, is ionized by
"Medium and Long-Term Development Details of the hydrogen calibration
photons from the helium discharge. The
Plan for the Automobile Industry". To standard (bought from Air Liquide) are
PDHID is universal except for neon (Ne)
date, many automobile companies given in Table 1. The standard contained
which has an ionization potential (IP) of
around the world have launched fuel approximately 10 ppm of CO2, CO, CH4,
21.56 eV. The IP is higher than the energy
cell passenger vehicles, and cities such Ar, O2, and N2. To test the dynamic range
of the He metastable (19.8 eV), leading
as Beijing and Shanghai have built a and detection limits of the method, an
to low ionization of Ne. The PDHID is a
network of hydrogen refueling stations. Agilent dynamic dilution system was
highly sensitive detector with minimal
As hydrogen is increasingly used to used to dilute the calibration sample.
detectable levels (MDLs) to fixed gases
power vehicles, strict requirements have Ultrahigh purity hydrogen (99.999%
in the low ppb range.1
been imposed to limit pollution from grade, Air Liquide) was used as
any impurities in the gas. Specifications In this study, an Agilent 8890 GC fitted
the diluent.
for the purity of hydrogen used in with a plug and play PDHID (Valco
other applications are also becoming Instruments Co., Inc.) was used for the
more stringent. Hydrogen is used in
the manufacturing and processing of Table 1. Hydrogen samples and the concentration of impurities in the hydrogen samples after each level
large-scale integrated circuits (ICs), of dilution.
smelting and processing of high-purity
Calibration Concentration of the Sample Following a Dilution (ppb)
metals, development and production Standard Diluent (H2)
of liquid hydrogen and its derivative Flow Flow Dilution
(mL/min) (mL/min) Factor CO2 CO CH4 Ar O2 N2
products, carrier gas used for precision
4.78 0 1 10,100 10,000 10,000 9,900 9,950 9,910
instruments (chromatographs, etc.),
4.78 4.92 2.03 4,977 4,928 4,928 4,879 4,903 4,883
and diluents used in standard gas
4.78 19.17 5.01 2,016 1,996 1,996 1,976 1,986 1,978
mixtures. Therefore, hydrogen-purity
testing, especially for high-purity 2.84 25.66 10.04 1,006 996 996 987 992 988

hydrogen and ultra-high-purity hydrogen, 2.84 53.96 20.00 502 500 497 492 498 493

is fast becoming a critical requirement 2.84 139.16 50.00 202 200 200 198 199 198

of various hydrogen producers and 2.84 281.66 100.18 101 100 100 99 99 99

industrial users. 1.50 298.50 200.00 51 50 50 50 50 50

2
Instrumentation Results and discussion sample with CO2, CO, CH4, Ar, O2, N2 at a
The 8890 GC PDHID was configured concentration of approximately 50 ppb.
as shown in Table 2 and instrument Chromatogram of hydrogen Figure 1B shows the chromatogram of
operating conditions are given in Table 3. A chromatogram of the hydrogen the diluted standard, and insets for CO2,
standard containing impurities at the Ar, CH4, and CO from 10 consecutive
Table 2. Agilent 8890 GC-PDHID system
10 ppm level obtained using the 8890 injections. Due to the adsorption and
configuration.
GC-PDHID is shown in Figure 1A. The temperature programming of the
Valve 1: 10-port valve
figure shows good separation of CO2, chromatographic column and tubing,
Valves 2 to 4: 6-port valves
Valve 5: 6-port sample switching valve CO, CH4, Ar, O2, N2, and other impurities it was difficult to determine O2 at the
Valve
System Loop volume: 0.25 mL in hydrogen. Heart-cut and secondary 50 ppb level, but it was possible to
Purge chamber uses helium to protect
separation methods were used to detect 50 ppb CO. Using the helium
analytical valves 1 to 4.
prevent CO2 and Ar from being affected purge chamber to protect the valve
Four capillary chromatographic columns
Analytical by the unpurged H2 matrix, while enabling body allowed the PDHID baseline to be
Columns 1 and 2: PLOT-Q
Column
Columns 3 and 4: Molsieve the separation of O2 and Ar in the reduced to less than 1,000 pA. This low
Pulsed discharge helium ionization same run. detector baseline level is helpful for the
Detector
detector (PDHID) analysis of ppb level impurities in high
Dynamic
The 10 ppm hydrogen standard was purity hydrogen.
Dilution
Pneumatic Control Module (PCM) diluted by a factor of 200 to obtain a
channel 1 is used for diluent control.
System

Table 3. Agilent 8890 GC-PDHID operating pA A Impurities at the

10 ppm level 1. CO2
conditions. 5,500 2. Ar
3. O2
Helium purge flow rate: 2 mL/min 4. N2
Valves
Ambient temperature 4,500 5. CH2
6. CO
Column 1: Constant pressure,
35 psi (approximately 10 mL/min)
3,500
Column 2: Constant pressure,
30 psi (approximately 10 mL/min)
Column Flow Column 3: 30 psi held for 5 min, 2,500
increased to 40 psi at the rate of
2 PSI/min and held for 16 min

Column 4: Constant pressure, 1,500

15 psi (approximately 10 mL/min)
Initial temperature at 50 °C,
maintained for 5 min, 500
Column then raised to 120 °C 0 2 4 6 8 10 12 14 16
at the rate of 20 °C/min, Minutes
Temperature
then maintained for 7.5 min. pA
B Impurities at the
Duration: 16 min
5,500 50 ppb level
150 °C
PDHID
Helium flow rate: 32 mL/min

Data Collection 4,500

5 Hz
Rate

3,500

2,500

1,500

500
0 2 4 6 8 10 12 14 16
Minutes

Figure 1. Typical chromatogram obtained from the analysis of impurities at the 10 ppm level (A) and
50 ppb level (B) in high-purity hydrogen using the Agilent 8890 GC-PDHID.

3
Reproducibility Linearity Good linearity was obtained in the
Table 4 shows the reproducibility results Each calibration standard (Table 1) 50 ppb to 10 ppm concentration range
for CO2, CO, CH4, Ar, and N2 obtained by was injected six times, and the average for all impurities, with R2 correlation
six consecutive injections of samples peak height of the impurities was coefficients above 0.995. There was also
with a concentration of around 1 ppm plotted against its concentration. Linear excellent linearity for CO, CH4, CO2, and Ar
and 50 ppb. See Table 1 for the specific regression analysis was performed at the 50 ppb to 1 ppm level. The offset
concentration of each gas. For the automatically by the instrument at the origin of the N2 calibration curve
1 ppm samples, the peak height relative software. The calibration curves suggests a background interference.
standard deviation (RSD) was below 1% and linear regression equations are The interference is likely due to trace
and for the 50 ppb samples, the peak shown in Figure 2. The insets show levels of the gas in the diluent gas, and
height RSD was below 5%. the concentration range from 50 ppb a small amount of air infiltration from
to 1 ppm for CO, CH4, CO2, and Ar, and the sample line. Excluding the results
Detection limits
100 ppb to 1 ppm for N2. from the 50 ppb standard, good linearity
The method detection limits (MDLs) was obtained for N2 in the 100 ppb to
in Table 4 were calculated using the ppm range.
50 ppb sample analysis results, based on
signal-to-noise ratio (S/N)=3 criteria. The
MDLs of the five impurities were all lower
than 20 ppb, meeting the requirements
specified in GB/T 3634.2-2011, GB/T
37244-2018, and ISO 14687-2019.

Table 4. Agilent 8890 GC-PDHID reproducibility data and detection limits.

10 ppm Dilute to 1 ppm Sample 10 ppm Dilute to 50 ppb Sample

RT RT Height Height RT RT Height Height Concentration MDL
Components (min) RSD (pA) RSD (min) RSD (pA) RSD S/N (ppb) (S/N=3)
CO2 5.2697 0.019% 347.22 0.96% 5.2867 0.025% 6.88 1.29% 196.9 54.1 0.8

Ar 7.9233 0.015% 175.81 0.17% 7.9372 0.015% 11.35 0.83% 324.7 53.0 0.5

N2 10.2213 0.009% 648.82 0.71% 10.2320 0.007% 237.84 0.49% 6808.2 53.1 0.02

CH4 12.1629 0.008% 294.62 0.22% 12.1709 0.007% 12.43 0.39% 355.7 54.6 0.5

CO 15.7888 0.005% 76.44 0.30% 15.8882 0.016% 2.73 1.93% 78.2 54.6 2.1

4
 CO CH4

80 y = 0.0779x – 2.1188 y = 0.0818x – 3.5159

3,000 300 y = 0.2969x – 3.2919 y = 0.2673x + 16.987
900 R² = 0.9995 R² = 0.9999 R² = 1 R² = 0.9988
60
200
40
20 100
2,000
600 0

Height (pA)
0 500 1,000 0
Height (pA)

0 500 1,000

300 1,000

0 0
0 2,000 4,000 6,000 8,000 10,000 0 2,000 4,000 6,000 8,000 10,000
Concentration (ppb) Concentration (ppb)

CO2 Ar
400 y = 0.3535x – 16.894 200
4,000 y = 0.3533x – 9.624 2,000 y = 0.1753x + 1.5703 y = 0.1704x + 6.0807
300 R² = 0.9967 R² = 0.9991 150 R² = 0.9998 R² = 0.9997
200 100
3,000 1,500
100 50
0 0
Height (pA)

Height (pA)

0 500 1,000 0 200 400 600 800 1,000

2,000 1,000

1,000 500

0 0
0 2,000 4,000 6,000 8,000 10,000 0 2,000 4,000 6,000 8,000 10,000
Concentration (ppb) Concentration (ppb)

N2
1,200 y = 0.7403x + 290.12
4,000 y = 0.3435x + 277.05
R² = 0.9991
800 R² = 0.9954

400
3,000
0
0 200 400 600 800 1,000
Height (pA)

2,000

1,000

0
0 2,000 4,000 6,000 8,000 10,000

Concentration (ppb)

Figure 2. Linearity of each impurity in the range of 50 ppb to 10 ppm.

5
Conclusion References
The Agilent 8890 GC fitted with a pulsed 1. Pulsed Discharge Detector Model
discharge helium ionization detector D-3-I-8890 Instruction Manual, Valco
(PDHID) was used successfully for the Instruments Co. Inc. VICI publication,
analysis of trace impurities in high-purity accessed December 2021, https://
hydrogen. Qualitative and quantitative www.vici.com/support/manuals/
detection of CO2, CO, CH4, Ar, O2, and d3_hp_8890.pdf
N2 were be achieved in the same run 2. GB/T 3634.2-2011 Hydrogen. Part 2:
from one injection. Using heart-cutting, Pure hydrogen, high pure hydrogen
the matrix effects of hydrogen were and ultrapure hydrogen
reduced, enabling baseline separation of
Ar, O2 and CO2. Separation of O2 and Ar 3. GB/T 37244-2018 Fuel specification
was also achieved without the need for for proton exchange membrane fuel
cryogenic conditions. cell vehicles—Hydrogen

The detection limit of each impurity 4. ISO 14687-2019 Hydrogen fuel

was less than 20 ppb, meeting the quality–Product specification
requirements of GB/T 3634.2-2011,
GB/T 37244-2018, and ISO 14687-
2019 standards. Also, good linearity
(R2 >0.995) was achieved for CO, CH4,
CO2, and Ar in the concentration range
of 50 ppb to 10 ppm and for N2 from
100 ppb to 10 ppm.
The helium purge chamber prevented
air leaking into the GC system from
the valve connection points, ensuring
the baseline level of the PDHID to be
kept under 1,000 pA. Maintaining a low
baseline is helpful for the analysis of real
samples containing impurities below the
ppb level.
The GC method enables producers and
users of high purity hydrogen to test
for a range of impurities, potentially
reducing pollution from fuel-hydrogen,
or improving yields and performances of
ICs or gas mixes.

www.agilent.com/chem

DE67222349

This information is subject to change without notice.

© Agilent Technologies, Inc. 2021

Printed in the USA, December 23, 2021
5994-4415EN