Analysis of Nitrous Oxide and Helium in Air: Greenhouse Gas Inventory

Joseph Khoury, Alfred Sattler, Bryan Linsley, Keith Magers, and Jean Lee
Joint Water Pollution Control Plant
Water Quality Laboratory
Sanitation Districts of Los Angeles County
24501 Figueroa Street, Carson CA 90745
[email protected]

ABSTRACT
We have developed an analytical method using gas a chromatograph equipped with a pulsed
discharge detector (GC-PDD) that is capable of analyzing nitrous oxide (N2O) from 50 ppbv up
to 500 ppmv. Samples are collected in 4-layer metalized gas sampling bags. Total time of
analysis is 17 minutes, and no appreciable interference from water or CO2 has been observed.

To measure dissolved N2O, an equilibrium headspace method was developed. The headspace
N2O gas was analyzed by GC-PDD, and this result was converted into the corresponding
dissolved N2O concentration. The approximate calibration range of dissolved N2O is from 0.002
µg/mL to 22 µg/mL.

A method for analysis of helium in air was developed using a gas chromatograph equipped with
a microvolume thermal conductivity detector (TCD). The quantifiable range ran from 5 ppmv
(the ambient air level) up to 12,000 ppmv using two calibration curves.

KEYWORDS
Nitrous oxide, helium, dissolved nitrous oxide, greenhouse gas, wastewater, gas chromatography,
GC-PDD, analytical method

INTRODUCTION
Wastewater facilities have been identified as sources of greenhouse gases. The waste disposal
and treatment sector, including wastewater facilities, has been estimated to contribute 3.4% of
overall emissions, including 18.1% of overall methane emissions and 2.3% of overall nitrous
oxide emissions. These emission levels have caused government regulators to become interested
in establishing controls on the wastewater industry. [CEC 2006]

The Sanitation Districts of Los Angeles County (the Sanitation Districts) are interested in
quantifying the major greenhouse gas emissions from wastewater and related treatment
processes. The data will be used to estimate the impact wastewater treatment has on greenhouse
gas emissions and to establish a baseline on current emissions. Of the three main anthropogenic
greenhouse gases, the Sanitation Districts Laboratories already use established protocols for the
analysis of carbon dioxide (CO2) and of methane (CH4). However, it had no established method
for the analysis of N2O, which can be produced during wastewater treatment
nitrification/denitrification processes and during composting. The engineering staff uses a multi-
gas, photoacoustic field analyzer, the Innova 1412 (LumaSense Technologies, Inc.) that is able to
quantitatively measure N2O down to about 1 ppmv, somewhat above the ambient air level of 310
ppbv. A reliable laboratory method is necessary to corroborate field measurements and to be
able to quantitatively measure at least down to the ambient air levels, and preferably below.

Aqueous nitrous oxide has been analyzed using an electrochemical probe [Anderson 2001],
direct injection of aqueous samples into a GC-ECD system, and by equilibrium headspace
analysis [Elkins 1980]. The probe measurements can be done in the field; however, the detection
limit is about 13 µg/L (a factor of 6 times higher than the equilibrium headspace technique).
Direct injection of aqueous samples has a detection limit of about 5 µg/L. The primary
disadvantages for this direct injection procedure is contamination of the GC system by solids
present in the wastewater samples. These solids cannot be prefiltered from the sample without
loss of N2O. We chose the equilibrium headspace method since it has the highest sensitivity,
about 2 µg/L, and is the least likely to develop contamination problems.

In addition, engineers used helium as a tracer gas to conduct precise airflow measurements, so a
reliable method to analyze for helium in air was also needed. Because of helium’s inertness and
high ionization potential, the only practical means of measurement is to employ a gas
chromatograph equipped with a thermal conductivity detector (GC-TCD). The thermal
conductivity of helium differs from nitrogen by a factor of 6 and from argon by a factor of 8.7.
Since TCD sensitivity increases with thermoconductivity difference, nitrogen and argon are the
carrier gases of choice for helium analysis.

Using a Varian 4900 Micro GC equipped with a TCD, this laboratory was able to detect nitrous
oxide at 10 ppmv, but not at 1 ppmv or below. Nitrous oxide in air has been quantified down to
10 ppmv using infrared spectroscopy [NIOSH 1994, Saier 1954]. An IR Gas Filter Correlation
N2O Analyzer (Teledyne Advanced Pollution Instrumentation, Inc Model 320EU) is described as
having a lower detectable limit of 10 ppbv. Nitrous oxide has also been detected down to
ambient atmospheric levels using GC-TCD, but a specialized series of traps was required to
concentrate 20 liters of air before injection onto a GC column [Hahn 1972]. In order to directly
detect nitrous oxide at ambient atmospheric concentrations, a more sensitive detector is required.
The development of the electron capture detector (ECD) has made such direct measurements
possible [Wentworth 1973, Rasmussen 1976], and GC-ECD has been used to measure N2O in air
[Czepiel 1995, Keller 1993]. The pulsed discharge helium ionization detector (PDD) has also
been shown to have this level of sensitivity, and has been used to detect nitrous oxide.
[Wentworth 1992] We chose the PDD since it would provide the necessary sensitivity, without
the additional concerns and expense of having a radioactive detector. It is also less affected by
contamination with oxygen, water, or hydrogen sulfide, and can be used for a wider range of
compounds (excluding helium and neon).

METHODOLOGY
Sample Collection
Air and other gaseous samples are collected in bags composed of layered materials (SKC Inc.
FlexFoil) that prevent permeation in and out of the bag, and are designed for sampling low
molecular weight compounds. Aqueous samples are collected with zero headspace in 40-mL
amber vials with Teflon coated silicone septum caps. Copper sulfate is added as a preservative.
Nitrous Oxide Analysis
A loop injection system is used to draw a sample of known volume from a bag and to inject it
onto a chromatographic column in a Hewlett Packard 5890 gas chromatograph equipped with a
VICI PDD (model D-2). In order to accommodate the stringent gas purity requirement of the
PDD, helium carrier gas is sent through a special heated purifier (VICI, Model HP2). The
purification substrate is a non-evaporable gettering alloy contained in a welded assembly, and is
activated by heating. The gas from the purifier flows to a tee, one side of which goes through a
restrictor (50 mL/min) and on to the discharge gas inlet on the detector. The other side of the tee
flows (2 mL/min) to the injector valve. A stainless steel transfer line supplies carrier gas from
the injector valve to the head of the capillary GC column.

The analytical column is a 30m x 0.53mm i.d. 3.00 µm Carbon PLOT (J&W Scientific) capillary
column. A particle trap (2.5 m x 0.53 mm i.d.) is attached to the back end of the column and
connected to the column inlet side of the PDD. The column oven is temperature programmed
starting at 30°C, held for 4 minutes, and then ramped at 15°C/min to 50°C. The detector
temperature is set at 100°C.

Dissolved Nitrous Oxide Analysis

To measure dissolved N2O, the entire liquid sample was quantitatively transferred to a 1 L
evacuated glass bulb equipped with a septum sidearm. After transfer, the bulb was brought to
atmospheric pressure with pure nitrogen. The bulb was then shaken for five minutes to speed
equilibration of the headspace and liquid. A syringe equipped with a shutoff valve was used to
withdraw approximately 200 mL of headspace gas from the glass bulb. The contents of the
syringe were then transferred to a FlexFoil sample bag, and then analyzed for N2O.

The N2O concentration in the original liquid sample is

calculated using the measured N2O concentration in the Figure 1 Valve configuration for
headspace, the volume of the liquid sample, the volume of sample loop injection system.
the preservative, the volume and temperature of the bulb, and Load
published N2O solubility data. However, for the 40 mL
liquid sample in a 1 L bulb, the headspace to liquid volume
ratio is 24:1. Under these conditions, a Henry’s Law
calculation reveals that the amount of N2O remaining in the
liquid is negligible (only about 2% of the total N2O). This
means that under these conditions, the calculation of the
dissolved N2O concentration in the original sample is not
strongly dependent on the N2O solubility equations.
Inject
Helium Analysis
Loop Injector
For helium analysis, a loop injection system is used to draw a
sample of known volume from a gas-sampling bag, and to
inject it onto a chromatographic column. We assembled a 6-
port, 2-position Valco valve onto a microelectronic actuator
and configured it in a standard loop-injection configuration
shown in Figure 1. Carrier gas is directed from the injector
to Position 4 of the valve using stainless steel tubing, 1/32” O.D., 0.45 mm I.D. (Entech
Instruments). Position 5 from the valve is connected to the column using 1/16”, 0.5 mm I.D.
stainless steel tubing. The sample line is connected to Position 1 of the valve using 1 mm ID
tubing equipped with an in-line filter (2 µm screen) and a needle attachment for sampling from
bags. A sample loop is connected across Positions 3 and 6 of the valve.

A microvolume thermal conductivity detector (Valco Instruments Co., Inc. Model TCD2) had
been installed onto a Varian 3800 gas chromatograph. The design of this detector has minimized
cell volume to accommodate capillary columns without the use of make-up gas, thereby
optimizing the sensitivity of the detector at low carrier gas flow rates. The microvolume TCD is
operated in constant temperature mode in which changes in thermal conductivity are measured
by the change in current required to keep the filament at a constant temperature. To minimize
effects of column temperature, the signal is referenced to a flow of gas passing through a ~1
meter coil of 1/16” stainless steel tubing (1 mm I.D.) inside the GC oven, and directed onto the
reference side of the detector.

This analysis employed a molecular sieve column (30 m, 0.53 mm I.D., 50 µm film) with a 2.5
m particle trap installed between the column outlet and the sample side of the microvolume
TCD. We set the flow rate to 5.1 mL/min and adjusted the reference flow to the same value.
The TCD base temperature was set at 100°C with the filament held at 230°C.

Optimizing Sensitivity
A few preliminary experiments using 10 µL and 50 µL loops suggested that a larger loop would
be required in order to calibrate down to 5 ppmv. The helium peak-width was about 2.5 seconds
(15 ppmv helium sample) at a flow rate of 5 mL/min. Switching to a 100-µL loop increased the
peak height without significantly changing the peak-width. Fluctuations in the baseline, 5 mV
high (and similar in width and intensity to low level helium peaks) were preventing further
attempts to decrease the level of detection. Turning off the GC oven eliminated these
fluctuations, leaving only 300 µV, high-frequency noise. At this point, ambient levels of helium
in the air (5.2 ppmv) were detectable. Additionally, switching from nitrogen to argon carrier gas
increased our response by 20% without increasing the noise level. Attempts to further optimize
the detector response by increasing or decreasing the carrier gas flow rate did not succeed.

RESULTS and DISCUSSION Figure 2 Portion of the N2O low cal. curve.
Nitrous Oxide Analysis
0
Two calibration curves were constructed. The low
Detector Response

curve (Figure 2) spanned the range from 0.05 to 50 0

ppmv and was produced by preparing eight dilutions
of nitrous oxide in nitrogen. Dilutions at 0.05, 25 and 0
50 ppmv were prepared by injecting 200 µL, 40 and
80 mL of 200 ppmv N2O gas standard (Scott) into a 4- 0

layer foil Bag and diluting to 800 mL with nitrogen.

0
Dilutions at 0.3, .0.5, 1, 2, 5, and 10 ppmv were
0 1 2 3 4 5
prepared by injecting 1.2, 2, 4, 8, 20 and 40 mL of 200 N2O concentration
ppm v/v (ppmv)
ppmv N2O gas standard and 16 mL of a 91% CO2 gas
standard into a 4-layer foil bag and diluting to 800 mL with nitrogen. A linear fit of the curve
forced through the origin and weighted by 1/x had a correlation coefficient of 0.9999.

The high calibration curve spanned the range from 25 to 500 ppmv with seven points. Dilutions
at 25, 105, 187.5, 250, 350, 400, and 500 ppmv were prepared by injecting 0.10, 0.42, 0.75, 1.0,
1.4, 1.6, and 2.0 mL of 99.0 % N2O gas standard (Oxygen Services) into 5-L bags and diluting to
4.0 L with nitrogen. The curve was best fit with a quadratic forced through the origin and
weighted by 1/x; it had a correlation coefficient of 0.9999.

Method Detection Limit The method detection limit was estimated using a mixture of 0.05
ppmv of N2O in nitrogen. The mixture was prepared by injecting 200 µL of 200 ppmv N2O gas
standard (Scott Specialty Gases) into 800 mL of nitrogen in a 4-layer foil bag. The sample was
analyzed a total of seven times over the course of a day. The calculated MDL using in this
mixture was 0.01 ppmv. Average recovery was 103%, with 9 % RSD.

The global average ambient level of nitrous oxide in the air is approximately 0.310 ppmv. We
took advantage of this ambient level of N2O in air, using it as a secondary-source standard, to
evaluate the performance of this method. Three samples of ambient outdoor air were analyzed
on a single day. Using 0.310 ppmv as the expected value, calculated recoveries ranged from
99% to 105%.

Interferences Nitrous oxide elutes on Figure 3 Nitrous oxide, spiked to 0.25 ppm into reactor gas
the tail of the large CO2 peak. In sample, elutes at 3.3 min. on the tailing edge of CO2 the large
order to examine whether or not N2O peak at 2.8 min. The inset is an expanded view of the N2O peak.
recovery has any CO2 concentration
dependence, a series of dilutions of

Nitrous Oxide
N2O in CO2 were prepared and
analyzed. The N2O recovery
appeared to be independent of the
Detector Response

CO2 concentration. Furthermore, gas

from the CO2 vent of a secondary
treatment reactor was used to study
Nitrous Oxide

other possible matrix interferences.

Permanent gas analysis run on the
3.2 3.25 3.3 3.35 3
sample indicated that there was a
high concentration of CO2 (2.5%).
One aliquot of the sample was spiked
to 0.25 ppmv (Figure 3) and another
to 2.5 ppmv of N2O; 4 aliquots of 2.6 2.8 3 3.2 3.4 3.6 3.8 4
another sample (taken at a later date) Retention Time (min)
were spiked to 0.25 ppmv and four
aliquots were spiked to 3.33 ppmv of N2O. In all, 10 analyses were performed, and recoveries
ranged from 98% to 103%, averaging 99% with a 2% RSD.

Gaseous Nitrous Oxide Analysis Results Twenty-five air samples were taken from aeration tanks
at a wastewater treatment plant and analyzed for N2O using this method. Values ranged from 0.2
ppmv to 340 ppmv. Three gas samples from the CO2 vent of a secondary treatment reactor from
another wastewater treatment plant were also analyzed. Values ranged from 1.0 ppmv to 1.2
ppmv.

Dissolved Nitrous Oxide Analysis Results

A total of sixteen aqueous samples were collected from the aeration tanks at a wastewater
treatment plant at various times over the course of two days. The samples were preserved in the
field and stored at 5°C overnight and analyzed the following day. Dissolved nitrous oxide
concentrations in the liquid ranged from 0.008 µg/L to 0.79 µg/L.

Helium Analysis
Two calibration curves were produced. The Figure 4 Helium high calibration curve spanning 500
low curve spanning the range from 5 to 1,000 to 12,000 ppmv
ppmv was constructed by preparing seven
dilutions of helium in zero air. A 5-ppmv
dilution was prepared by injecting 50 µL of
helium (carrier grade, 99.9995% purity) into
10.0 L of zero air in a Tedlar bag. Six Detector Response
dilutions at 25, 50, 100, 250, 500, and 1,000
ppmv were prepared by injecting 25, 50, 100,
250, 500, and 1,000 µL of helium into 1 L of
air in a foil bag. A linear fit of the curve had a
correlation coefficient of 0.9998.

The high curve (Figure 4) spanning the range

from 500 to 12,000 ppmv (0.05% to 1.2%)
was constructed by preparing eight dilutions
of helium in zero air. Dilutions at 500 1,000 0 2,000 4,000 6,000 8,000 10,000 12,000
2,000 3,000 5,000 8,000 10,000 and 12,000 ppmv
ppm
He concentration (ppmv)
ppmv were prepared by injecting 1, 2, 4, 6, 10,
16, 20, and 24 mL of helium into a foil bag and diluting to 2 L with using zero air. A linear fit of
the curve had a correlation coefficient of 0.9995

Method Detection Limit The method detection limit was estimated by spiking zero air to 5 ppmv
of helium (5 µL of helium was injected into 1 L of zero air in a foil bag). A total of 7 samples
were prepared and analyzed over the course of two days. The calculated MDL using spiked zero
air was 1 ppmv. Average recovery was 95%, with 6 % RSD.

Natural outdoor air has an ambient helium level of 5.2 ppmv. We took advantage of the stable
and well-documented level of helium in ambient air, using it as a secondary-source standard, to
evaluate the performance of this method. Additionally, ambient air contains neon at 18.2 ppmv,
presenting a challenge in terms of the chromatographic separation of helium and neon. A sample
of ambient outdoor air as well as humidified outdoor air were analyzed 7 times on a single day.
Using 5.2 ppmv as the expected value, the average recovery was 105%, with 7 % RSD. For
humidified air, the average recovery (calculated using 5.1 ppmv for helium in humidified air)
was 96%, with a 7 % RSD.
Interferences The expected interferences are neon Figure 5 Outdoor Air, with ambient levels of
and hydrogen, since no other known compounds elute He (5.2 ppmv) and Ne (18 ppmv), and spiked
as early as helium. Hydrogen is fully resolved from to 5 ppmv with hydrogen.
helium over the entire range of the calibration curve.

Hydrogen
At low concentrations of helium, neon in air (18

Neon
ppmv) is also fully resolved (Figure 5). Even up to

Helium
250 ppmv of helium, the neon peak is still resolvable

Detector Response
from the helium peak; however, at 500 ppmv and
above, the neon peak is no longer resolvable. The
detector response to 18 ppmv of neon corresponds to
an area equivalent to approximately 5 ppmv of
helium, so for outdoor air samples that have been
spiked with helium above 250 ppmv, the actual level
of helium in the sample may be approximately 5
ppmv less than the reported level due to the neon co-
elution. As helium concentrations increase, the 2 2.1 2.2 2.3 2.4
relative contribution of this error becomes less Retention Time (min)
significant or even negligible.

Helium Analysis Results Twenty-five air samples were taken from aeration tanks at a wastewater
treatment plant while helium was being used as a tracer gas for airflow measurements. Helium
concentrations ranged from 0.05% to 0.32%. Future studies may require analysis of lower
helium concentrations.

Evaluation of N2O Field Probe

To evaluate the performance of a photoacoustic Figure 6 Comparison of laboratory and field-
field analyzer, results from this instrument were probe data. Probe results (as % of the lab value)
compared with the GC-PDD results. Twenty-five are plotted vs. the lab value. On average, the
air samples were taken from aeration tanks at a probe returned 108% of the lab result.
wastewater treatment plant and were analyzed in the 180%
Probe Value (% of Laboratory Value)

laboratory for N2O using the GC-PDD method. 160%

During the bag sampling, field measurements were 140%
taken simultaneously using a photoacoustic multi- 120%
gas analyzer. Values ranged from 0.2 ppmv to 340 100%
ppmv. A comparison of the laboratory and field 80%
measurements is summarized in Figure 6. 60%
Anywhere from 1 to 5 field measurements were 40%
taken during the course of filling a bag sample.
20%
These points were averaged and the ratio of this
0%
average to the laboratory result is plotted vs. the 0 100 200 300
laboratory result. The average of the ratio of the N2O Concentration (ppmv)
field value to laboratory value is 108% (7% RDS).
For each data point, this ratio is within two standard
deviations (dotted lines) of the average (108%) except for points below 3.5 ppmv, which is
approaching the limit of detection of the photoacoustic probe device. Points below 3.5 ppmv
were not used to calculate the average and standard deviation.

CONCLUSIONS
We have developed a simple and reliable laboratory method to quantitatively measure nitrous
oxide in air down to 0.05 ppmv using GC-PDD with loop injection. In conjunction, we have
developed a headspace method to analyze dissolved N2O in wastewater down to 0.002 µg/L. We
have also developed a method to analyze for helium in air down to the ambient air level of 5
ppmv using GC-TCD analysis of bag grab samples.

REFERENCES
Andersen, K.; Kjaer, T.; Revsbech, N. P. (2001) An Oxygen Insensitive Microsensor for Nitrous
Oxide. Sensors and Actuators B-Chemical, 81 (1), 42-48.
California Energy Commission (CEC),(2006) Inventory Of California Greenhouse Gas
Emissions And Sinks: 1990 TO 2004
Czepiel, P.; Crill, P.; Harriss, R. (1995) Nitrous Oxide Emissions from Municipal Wastewater
Treatment. Environmental Science & Technology, 29 (9), 2352-2356.
Elkins, J.W. (1980) Determination of Dissolved Nitrous Oxide in Aquatic Systems by Gas
Chromatography Using Electron-Capture Detection and Multiple Phase Equilibration.
Analytical Chemistry, 52 (2), 263-267.
Hahn, J. (1972) Improved Gas Chromatographic Method for Field Measurement of Nitrous
Oxide in Air and Water Using a 5A Molecular Sieve Trap. Analytical Chemistry, 44 (11),
1889-92.
Keller, M.; Veldkamp, E.; Weitz, A.M.; Reiners, W.A. (1993) Effect of Pasture Age on Soil
Trace-Gas Emissions from a Deforested Area of Costa Rica, Nature 365 (6443), 244-246.
National Institute for Occupational Safety and Health (NIOSH), NIOSH Manual of Analytical
Methods, Fourth Edition, (1994), Method 6600, Nitrous Oxide.
Rasmussen, R.A.; Krasnec, J.; Pierotti, D. (1976) N2O Analysis in the Atmosphere via Electron
Capture-Gas Chromatography. Geophys. Res. Lett., 3, 615-18.
Saier, E.L.; Pozefsky, A. (1954) Quantitative Determination of Nitric Oxide and Nitrous Oxide
by Infrared Absorption. Analytical Chemistry, 26 (6),1079-1080.
Wentworth, W.E.; Freeman, R.R. (1973) Measurement of Atmospheric Nitrous Oxide Using an
Electron Capture Detector in Conjunction With Gas Chromatography. J Chromatogr. 79,
322-4.
Wentworth, W.E.; Vasnin, S.V.; Sterns, S.D.; Meyer, C.J. (1992) Pulsed Discharge Helium
Ionization Detector, Chromatographia, 34, 219-225.