n o isotopocule measurements using laser …...revised: 15 april 2020 – accepted: 21 april 2020...

Atmos Meas Tech 13 2797ndash2831 2020httpsdoiorg105194amt-13-2797-2020copy Author(s) 2020 This work is distributed underthe Creative Commons Attribution 40 License

N2O isotopocule measurements using laser spectroscopy analyzercharacterization and intercomparisonStephen J Harris12 Jesper Liisberg3 Longlong Xia4 Jing Wei5 Kerstin Zeyer5 Longfei Yu5 Matti Barthel6Benjamin Wolf4 Bryce F J Kelly1 Dioni I Cendoacuten2 Thomas Blunier3 Johan Six6 and Joachim Mohn5

1School of Biological Earth and Environmental Sciences UNSW Sydney NSW Australia2Australian Nuclear Science and Technology Organisation Lucas Heights NSW Australia3University of Copenhagen Niels Bohr Institute Copenhagen Denmark4Karlsruhe Institute of Technology IMK-IFU Garmisch-Partenkirchen Germany5Empa Laboratory for Air PollutionEnvironmental Technology Duumlbendorf Switzerland6ETH Zuumlrich Department of Environmental Systems Science Zuumlrich SwitzerlandThese authors contributed equally to this work

Correspondence Joachim Mohn (joachimmohnempach)

Received 25 November 2019 ndash Discussion started 5 December 2019Revised 15 April 2020 ndash Accepted 21 April 2020 ndash Published 28 May 2020

Abstract For the past two decades the measurement ofnitrous oxide (N2O) isotopocules ndash isotopically substitutedmolecules 14N15N16O 15N14N16O and 14N14N18O of themain isotopic species 14N14N16O ndash has been a promisingtechnique for understanding N2O production and consump-tion pathways The coupling of non-cryogenic and tuneablelight sources with different detection schemes such as directabsorption quantum cascade laser absorption spectroscopy(QCLAS) cavity ring-down spectroscopy (CRDS) and off-axis integrated cavity output spectroscopy (OA-ICOS) hasenabled the production of commercially available and field-deployable N2O isotopic analyzers In contrast to traditionalisotope-ratio mass spectrometry (IRMS) these instrumentsare inherently selective for position-specific 15N substitutionand provide real-time data with minimal or no sample pre-treatment which is highly attractive for process studies

Here we compared the performance of N2O isotopelaser spectrometers with the three most common detectionschemes OA-ICOS (N2OIA-30e-EP ABB ndash Los Gatos Re-search Inc) CRDS (G5131-i Picarro Inc) and QCLAS(dual QCLAS and preconcentration trace gas extractor(TREX)-mini QCLAS Aerodyne Research Inc) For eachinstrument the precision drift and repeatability of N2O molefraction [N2O] and isotope data were tested The analyz-ers were then characterized for their dependence on [N2O]gas matrix composition (O2 Ar) and spectral interferences

caused by H2O CO2 CH4 and CO to develop analyzer-specific correction functions Subsequently a simulated two-end-member mixing experiment was used to compare the ac-curacy and repeatability of corrected and calibrated isotopemeasurements that could be acquired using the different laserspectrometers

Our results show that N2O isotope laser spectrometer per-formance is governed by an interplay between instrumentalprecision drift matrix effects and spectral interferences Toretrieve compatible and accurate results it is necessary to in-clude appropriate reference materials following the identi-cal treatment (IT) principle during every measurement Re-maining differences between sample and reference gas com-positions have to be corrected by applying analyzer-specificcorrection algorithms These matrix and trace gas correctionequations vary considerably according to N2O mole fractioncomplicating the procedure further Thus researchers shouldstrive to minimize differences in composition between sam-ple and reference gases In closing we provide a calibrationworkflow to guide researchers in the operation of N2O iso-tope laser spectrometers in order to acquire accurate N2Oisotope analyses We anticipate that this workflow will as-sist in applications where matrix and trace gas compositionsvary considerably (eg laboratory incubations N2O liber-ated from wastewater or groundwater) as well as extend to

Published by Copernicus Publications on behalf of the European Geosciences Union

2798 S J Harris et al N2O isotopocule measurements using laser spectroscopy

future analyzer models and instruments focusing on isotopicspecies of other molecules

1 Introduction

Nitrous oxide (N2O) is a long-lived greenhouse gas with a100-year global warming potential nearly 300 times that ofcarbon dioxide (CO2 Forster et al 2007) and is the largestemission source of ozone-depleting nitrogen oxides in thestratosphere (Ravishankara et al 2009) In 2019 the glob-ally averaged [N2O] reached approximately 332 ppb com-pared to the pre-industrial level of 270 ppb (NOAAESRL2019) While this increase is known to be linked primarily toincreased fertilizer use in agriculture (Bouwman et al 2002Mosier et al 1998 Tian et al 2015) understanding the un-derlying microbial processes producing and consuming N2Ohas proved more challenging and individual source contri-butions from sectors such as agricultural soils wastewatermanagement and biomass burning to global bottom-up esti-mates of N2O emissions have large uncertainties (Denmanet al 2007) Stable isotopes are an effective tool for dis-tinguishing N2O sources and determining production path-ways which is critical for developing appropriate mitigationstrategies (Baggs 2008 Ostrom and Ostrom 2011 Toyodaet al 2017)

The N2O molecule has an asymmetric linear structure(NNO) with the following most abundant isotopocules14N15N16O (15NαminusN2O) 15N14N16O (15NβminusN2O)14N14N18O (18OminusN2O) and 14N14N16O (Yoshida andToyoda 2000) The terms 15Nα and 15Nβ refer to therespective central and terminal positions of nitrogen (N)atoms in the NNO molecule (Toyoda and Yoshida 1999)Isotopic abundances are reported in δ notation whereδ15N= R(15N14N)sampleR(

15N14N)referenceminus 1 denotesthe relative difference in per mil (permil) of the sample versusatmospheric N2 (AIR-N2) The isotope ratio R(15N14N)equals x(15N)x(14N) with x being the absolute abun-dance of 14N and 15N respectively Similarly ViennaStandard Mean Ocean Water (VSMOW) is the interna-tional isotope-ratio scale for δ18O In practice the isotopeδ value is calculated from measurement of isotopoculeratios of sample and reference gases with the latter beingdefined on the AIR-N2 and VSMOW scales By extensionδ15Nα denotes the corresponding relative difference ofisotope ratios for 14N15N16O14N14N16O and δ15Nβ for15N14N16O14N14N16O The site-specific intramoleculardistribution of 15N within the N2O molecule is termed 15N-site preference (SP) and is defined as SP= δ15Nα minus δ15Nβ

(Yoshida and Toyoda 2000) The term δ15Nbulk is usedto express the average δ15N value and is equivalent toδ15Nbulk

= (δ15Nα + δ15Nβ)2Extensive evidence has shown that SP δ15Nbulk and δ18O

can be used to differentiate N2O source processes and

biogeochemical cycling (Decock and Six 2013 Denk etal 2017 Heil et al 2014 Lewicka-Szczebak et al 20142015 Ostrom et al 2007 Sutka et al 2003 2006 Toy-oda et al 2005 2017 Wei et al 2017) Isotopocule abun-dances have been measured in a wide range of environ-ments including the troposphere (Harris et al 2014 Roumlck-mann and Levin 2005 Toyoda et al 2013) agriculturalsoils (Buchen et al 2018 Ibraim et al 2019 Koumlster etal 2011 Mohn et al 2012 Ostrom et al 2007 Park etal 2011 Peacuterez et al 2001 2006 Toyoda et al 2011 Ver-hoeven et al 2018 2019 Well et al 2008 Well and Flessa2009 Wolf et al 2015) mixed urbanndashagricultural environ-ments (Harris et al 2017) coal and waste combustion (Har-ris et al 2015b Ogawa and Yoshida 2005) fossil fuel com-bustion (Toyoda et al 2008) wastewater treatment (Harriset al 2015a b Wunderlin et al 2012 2013) groundwa-ter (Koba et al 2009 Minamikawa et al 2011 Nikolenkoet al 2019 Well et al 2005 2012) estuaries (Erler etal 2015) mangrove forests (Murray et al 2018) stratifiedwater impoundments (Yue et al 2018) and firn air and icecores (Bernard et al 2006 Ishijima et al 2007 Prokopiouet al 2017) While some applications like laboratory incu-bation experiments allow for analysis of the isotopic signa-ture of the pure source most studies require analysis of thesource diluted in ambient air This specifically applies to ter-restrial ecosystem research since N2O emitted from soils isimmediately mixed with background atmospheric N2O Tounderstand the importance of soil emissions for the globalN2O budget two-end-member mixing models commonly in-terpreted using Keeling or MillerndashTans plots are frequentlyused to back-calculate the isotopic composition of N2O emit-ted from soils (Keeling 1958 Miller and Tans 2003)

N2O isotopocules can be analyzed by isotope-ratio massspectrometry (IRMS) and laser spectroscopic techniqueswith currently available commercial spectrometers operatingin the mid-infrared (MIR) region to achieve highest sensitiv-ities IRMS analysis of the N2O intramolecular 15N distri-bution is based on quantification of the fragmented (NO+mz 30 and 31) and molecular (N2O+ mz 44 45 and46) ions to calculate isotope ratios for the entire molecule(15N14N and 18O16O) and the central (Nα) and terminal(Nβ ) N atom (Toyoda and Yoshida 1999) The analysis ofN2O SP by IRMS is complicated by the rearrangement of Nα

and Nβ in the ion source while analysis of δ15Nbulk (4544)involves correction for NN17O (mass 45) IRMS can achieverepeatability as good as 01 permil for δ15N δ18O δ15Nα andδ15Nβ (Potter et al 2013 Roumlckmann and Levin 2005) butan interlaboratory comparison study showed substantial de-viations in measurements of N2O isotopic composition inparticular for SP (up to 10 permil) (Mohn et al 2014)

The advancement of mid-infrared laser spectroscopic tech-niques was enabled by the invention and availability of non-cryogenic light sources which have been coupled with dif-ferent detection schemes such as direct absorption quantumcascade laser absorption spectroscopy (QCLAS Aerodyne

Atmos Meas Tech 13 2797ndash2831 2020 httpsdoiorg105194amt-13-2797-2020

S J Harris et al N2O isotopocule measurements using laser spectroscopy 2799

Research Inc (ARI) Waumlchter et al 2008) cavity ring-downspectroscopy (CRDS Picarro Inc Berden et al 2000) andoff-axis integrated cavity output spectroscopy (OA-ICOSABB ndash Los Gatos Research Inc Baer et al 2002) to realizecompact field-deployable analyzers In short the emissionwavelength of a laser light source is rapidly and repetitivelyscanned through a spectral region containing the spectrallines of the target N2O isotopocules The laser light is cou-pled into a multi-path cell filled with the sample gas and themixing ratios of individual isotopic species are determinedfrom the detected absorption using Beerrsquos law The wave-lengths of spectral lines of N2O isotopocules with distinct17O 18O or position-specific 15N substitution are uniquedue to the existence of characteristic rotationalndashvibrationalspectra (Rothman et al 2005) Thus unlike IRMS laserspectroscopy does not require mass-overlap correction How-ever the spectral lines may have varying degrees of overlapwith those of other gaseous species which if unaccountedfor may produce erroneous apparent absorption intensitiesOne advantage of laser spectroscopy is that instruments cananalyze the N2O isotopic composition in gaseous mixtures(eg ambient air) in a flow-through mode providing real-time data with minimal or no sample pretreatment which ishighly attractive to better resolve the temporal complexityof N2O production and consumption processes (Decock andSix 2013 Heil et al 2014 Koumlster et al 2013 Winther etal 2018)

Despite the described inherent benefits of laser spec-troscopy for N2O isotope analysis applications remain chal-lenging and are still scarce for four main reasons

1 Two pure N2O isotopocule reference materials(USGS51 USGS52) have only recently been madeavailable through the United States Geological Survey(USGS) with provisional values assigned by the TokyoInstitute of Technology (Ostrom et al 2018) The lackof N2O isotopocule reference materials was identifiedas a major reason limiting interlaboratory compatibility(Mohn et al 2014)

2 Laser spectrometers are subject to drift effects (eg dueto moving interference fringes) particularly under fluc-tuating laboratory temperatures which limits their per-formance (Werle et al 1993)

3 If apparent δ values retrieved from a spectrometer arecalculated from raw uncalibrated isotopocule mole frac-tions referred to here as a δ-calibration approach aninverse concentration dependence may be introducedThis can arise if the analyzer measurements of iso-topocule mole fractions are linear yet the relation-ship between measured and true mole fractions has anon-zero intercept (eg Griffith et al 2012 Griffith2018) such as due to baseline structures (eg interfer-ing fringes Tuzson et al 2008)

4 Laser spectroscopic results are affected by mole frac-tion changes of atmospheric background gases (N2 O2and Ar) herein called gas matrix effects due to thedifference of pressure-broadening coefficients (Nara etal 2012) and potentially by spectral interferences fromother atmospheric constituents (H2O CO2 CH4 COetc) herein called trace gas effects depending on theselected wavelength region The latter is particularlypronounced for N2O due to its low atmospheric abun-dance in comparison to other trace gases

Several studies have described some of the above effectsfor CO2 (Bowling et al 2003 2005 Griffis et al 2004Griffith et al 2012 Friedrichs et al 2010 Malowany etal 2015 Pataki et al 2006 Pang et al 2016 Rella etal 2013 Vogel et al 2013 Wen et al 2013) CH4 (Eyeret al 2016 Griffith et al 2012 Rella et al 2013) and re-cently N2O isotope laser spectrometers (Erler et al 2015Harris et al 2014 Ibraim et al 2018 Waumlchter et al 2008)However a comprehensive and comparative characterizationof the above effects for commercially available N2O isotopeanalyzers is lacking

Here we present an intercomparison study of com-mercially available N2O isotope laser spectrometers withthe three most common detection schemes (1) OA-ICOS(N2OIA-30e-EP ABB ndash Los Gatos Research Inc) (2) CRDS(G5131-i Picarro Inc) (3) QCLAS (dual QCLAS and tracegas extractor (TREX)-mini QCLAS ARI) Performancecharacteristics including precision repeatability drift and de-pendence of isotope measurements on [N2O] were deter-mined Instruments were tested for gas matrix effects (O2Ar) and spectral interferences from enhanced trace gas molefractions (CO2 CH4 CO H2O) at various [N2O] to de-velop analyzer-specific correction functions The accuracyof different spectrometer designs was then assessed duringa laboratory-controlled mixing experiment designed to sim-ulate two-end-member mixing in which results were com-pared to calculated expected values as well as to those ac-quired using IRMS (δ values) and gas chromatography (GCN2O concentration) In closing we provide a calibrationworkflow that will assist researchers in the operation of N2Oand other trace gas isotope laser spectrometers in order toacquire accurate isotope analyses

2 Materials and methods

21 Analytical techniques

Operational details of the laser spectrometers tested in thisstudy including wavenumber regions line positions and linestrengths of N2O are provided in Table 1 In Fig 1 selectedN2O rotational lines are shown in combination with the ab-sorption lines of the atmospheric most abundant IR-activetrace gases (H2O CO2 CH4 CO and O3) within the differ-ent wavenumber regions used by the analyzers Figure 1 can

httpsdoiorg105194amt-13-2797-2020 Atmos Meas Tech 13 2797ndash2831 2020


be used to rationalize possible spectral interferences withindifferent wavenumber regions

211 OA-ICOS (ABB ndash Los Gatos Research Inc)

The N2OIA-30e-EP (model 914-0027 serial number 15-830ABB ndash Los Gatos Research Inc USA) tested in this studywas provided by the University of New South Wales (UNSWSydney Australia) and is herein referred to as OA-ICOS I(Table 1) The instrument employs the OA-ICOS techniqueintegrated with a quantum cascade laser (QCL) (Baer etal 2002) In short the QCL beam is directed off axis intothe cavity cell with highly reflective mirrors providing anoptical path of several kilometers For further details on theOA-ICOS technique the reader is referred to the webpage ofABB ndash Los Gatos Research Inc (ABB ndash Los Gatos ResearchInc 2019) and Baer et al (2002)

The specific analyzer tested here was manufactured inJune 2014 and has had no hardware modifications since thenIt is also important to note that a more recent N2OIA-30e-EPmodel (model 914-0060) is available that in addition quanti-fies δ17O We are unaware of any study measuring N2O iso-topocules at natural abundance and ambient mole fractionswith the N2OIA-30e-EP The only studies published so farreporting N2O isotope data apply the N2OIA-30e-EP eitherat elevated [N2O] in a standardized gas matrix or using 15Nlabeling including Soto et al (2015) Li et al (2016) Konget al (2017) Brase et al (2017) Wassenaar et al (2018) andNikolenko et al (2019)

212 CRDS (Picarro Inc)

Two G5131-i analyzers (Picarro Inc USA) were used in thisstudy a 2015 model (referred to as CRDS I serial num-ber 5001-PVU-JDD-S5001 delivered September 2015) pro-vided by the Niels Bohr Institute University of CopenhagenDenmark and a 2018 model (referred to as CRDS II serialnumber 5070-DAS-JDD-S5079 delivered June 2018) pro-vided by Karlsruhe Institute of Technology Germany (Ta-ble 1) In CRDS the beam of a single-frequency continuouswave (cw) laser diode enters a three-mirror cavity with aneffective pathlength of several kilometers to support a con-tinuous traveling light wave A photodetector measures thedecay of light in the cavity after the cw laser diode is shut offto retrieve the mole fraction of N2O isotopocules For moredetails we refer the reader to the webpage of Picarro Inc(Picarro Inc 2019) and Berden et al (2000)

Importantly the manufacturer-installed flow restrictorswere replaced in both analyzer models as we noted reducedflow rates due to clogging during initial reconnaissance test-ing In CRDS I a capillary (inner diameter ID 150 micromlength 81 mm flow 252 cm3 minminus1) was installed whileCRDS II was equipped with a critical orifice (ID 75 micromflow 125 cm3 minminus1) Both restrictors were tested and con-firmed leakproof Both analyzers had manufacturer-installed

permeation driers located prior to the inlet of the cavitywhich were not altered for this study In December 2017CRDS I received a software and hardware update as per themanufacturerrsquos recommendations The CRDS II did not re-ceive any software or hardware upgrades as it was acquiredimmediately prior to testing

To the best of our knowledge the work presented in Lee etal (2017) and Ji and Grundle (2019) discusses the only pub-lished uses of G5131-i models A prior model (the G5101-i)which employs a different spectral region and does not offerthe capability for δ18O was used by Peng et al (2014) Er-ler et al (2015) Li et al (2015) Lebegue et al (2016) andWinther et al (2018)

213 QCLAS (Aerodyne Research Inc)

Three QCLAS instruments (ARI USA CW-QC-TILDAS-SC-D) were used in this study One instrument (QCLASI serial number 046) purchased in 2013 was provided byKarlsruhe Institute of Technology Germany and two instru-ments purchased in 2014 (QCLAS II serial number 065)and 2016 (QCLAS III serial number 077) were supplied byETH Zuumlrich Switzerland (Table 1) QCLAS I was used in allexperiments presented in this study while QCLAS II and IIIwere only used to assess the reproducibility of drift reportedin Sect 31

All instruments were dual-cw QCL spectrometersequipped with mirror optics guiding the two laser beamsthrough an optical anchor point to assure precise coinci-dence of the beams at the detector On the way to the de-tector the laser beams are coupled into an astigmatic multi-pass cell with a volume of approximately 2100 cm3 in whichthe beams interact with the sample air The multiple passesthrough the absorption cell result in an absorption path lengthof approximately 204 m The cell pressure can be selectedby the user and was set to 533 mbar as a trade-off betweenline separation and sensitivity This set point is automaticallymaintained by the TDL Wintel software (version 11489ARI MA USA) which compensates for variations in vac-uum pump speed by closing or opening a throttle valve at theoutlet of the absorption cell

QCLAS instruments offer great liberty to the user as thesystem can also be operated with different parameter set-tings such as the selection of spectral lines for quantifica-tion wavenumber calibration sample flow rate and pres-sure Thereby different applications can be realized fromhigh-flow eddy covariance studies or high-mole-fraction pro-cess studies to high-precision measurements coupled to acustomized inlet system In addition spectral interferencesand gas matrix effects can be taken into consideration bymulti-line analysis inclusion of the respective spectroscopicparameters in the spectral evaluation or adjustment of thepressure-broadening coefficients The spectrometers used inthis study (QCLAS IndashIII) were tested under standard set-tings but were not optimized for the respective experiments



Tabl

e1

Ove

rvie

wof

the

wav

elen

gth

regi

ons

line

posi

tions

and

line

stre

ngth

sof

N2O

isot

opoc

ules

and

key

oper

atin

gpa

ram

eter

sfo

reac

hla

sers

pect

rom

eter

test

edin

this

stud

yN

TP

stan

dsfo

rnor

mal

tem

pera

ture

and

pres

sure

Det

ectio

nsc

hem

eN

2OW

aven

umbe

rIs

otop

ocul

esL

ine

posi

tions

Flow

rate

Cel

lC

ell

Inte

rnal

Eff

ectiv

eM

easu

rem

ent

Ref

eren

ces

(mod

el

rang

ere

gion

(cmminus

1 )li

ne(c

m3

minminus

1 )te

mpe

ratu

repr

essu

repl

umbi

ngvo

lum

efr

eque

ncy

man

ufac

ture

r)(p

pb)

(cmminus

1 )st

reng

th(c

mminus

1 (

C)

(hPa

)vo

lum

eat

NT

P(s

)(m

olec

ule

cmminus

2 ))

(cm

3 )(c

m3 )

OA

-IC

OS

I30

0ndash10

000

021

921

ndash219

25

14N

14N

16O

21924

04

919times

10minus

2060

436

6193

060

50

100

Bae

reta

l(2

002)

(N

2OIA

-30e

-EP

21924

44

919times

10minus

20A

BB

ndashL

osG

atos

AB

Bndash

Los

Gat

os21

924

83

375times

10minus

19R

esea

rch

Inc

Res

earc

h14

N15

N16

O21

923

13

31times

10minus

21(2

019)

Inc

)15

N14

N16

O21

923

32

968times

10minus

21

14N

14N

18O

21921

31

113times

10minus

21

CR

DS

IampII

300ndash

1500

2195

7ndash2

196

314

N14

N16

O21

962

15

161times

10minus

2025

2(C

RD

SI)

402

100

404

223

41(C

RD

SI)

Pica

rro

(G51

31-i

21

962

45

161times

10minus

2012

5(C

RD

SII

)2

54(C

RD

SII

)In

c(2

019)

Pica

rro

14N

15N

16O

21957

62

734times

10minus

21

Inc

)15

N14

N16

O21

958

02

197times

10minus

21

14N

14N

18O

21959

51

431times

10minus

21

QC

LA

SI

IIamp

III

300ndash

9000

021

877

ndash218

815

14N

14N

16O

21880

42

601times

10minus

2113

0a20

a53

3a

2100

104a

100

aW

aumlcht

eret

al

(CW

-QC

-TIL

DA

S-SC

-D

and

14N

15N

16O

21879

43

294times

10minus

21(2

008)

Aer

odyn

e22

031

ndash220

34

15N

14N

16O

21878

53

274times

10minus

21H

arri

set

al

Res

earc

h14

N14

N18

O22

032

81

794times

10minus

21(2

014)

Inc

)

TR

EX

-QC

LA

SI

300ndash

1500

ab

2203

1ndash2

203

4a14

N14

N16

O22

031

02

710times

10minus

21ndash

b20

a35

6a

620

20a

100

aIb

raim

etal

(20

18)

(mod

ified

22031

11

435times

10minus

21

CW

-QC

-TIL

DA

S-76

-CS

14N

15N

16O

22033

69

798times

10minus

22

Aer

odyn

eR

esea

rch

15N

14N

16O

22032

07

016times

10minus

22

Inc

)14

N14

N18

O22

032

81

794times

10minus

21

aD

ualQ

C-T

ILD

AS

and

min

iQC

-TIL

DA

Sar

efle

xibl

esp

ectr

omet

erpl

atfo

rms

whi

chca

nbe

used

with

diff

eren

tpar

amet

erse

tting

sT

hein

dica

ted

num

bers

wer

ech

osen

fort

hede

scri

bed

expe

rim

ents

bT

hem

iniQ

C-T

ILD

AS

spec

trom

eter

isus

edin

com

bina

tion

with

apr

econ

cent

ratio

nde

vice

(Ibr

aim

etal

20

18)

the

indi

cate

dN

2Oco

ncen

trat

ion

rang

eis

prio

rto

prec

once

ntra

tion

cT

hepr

econ

cent

ratio

nndash

min

iQC

-TIL

DA

Ssy

stem

isus

edin

are

petit

ive

batc

hcy

cle

with

outa

cont

inuo

ussa

mpl

ega

sflo

w(I

brai

met

al

2018

201

9)



Figure 1 N2O isotopocule absorption line positions in the wavenumber regions selected for (a) OA-ICOS (b) CRDS and (c d) QCLAStechniques Regions of possible spectral overlap from interfering trace gases such as H2O CO2 CH4 and CO are shown The abundance-scaled line strengths of trace gases have been scaled with 10minus1 to 102 (as indicated) because they are mostly weaker than those of the N2Oisotopocules

QCLAS I was operated as a single laser instrument us-ing laser 1 to optimize spectral resolution of the frequencysweeps It is important to note the mixing ratios returned bythe instrument are solely based on fundamental spectroscopicconstants (Rothman et al 2005) so that corrections such asthe dependence of isotope ratios on [N2O] have to be imple-mented by the user in the postprocessing

To our knowledge QCLAS instruments have so far pre-dominately been used for determination of N2O isotopiccomposition in combination with preconcentration (see be-low) or at enhanced mole fractions (Harris et al 2015a bHeil et al 2014 Koumlster et al 2013) except for Yamamoto etal (2014) who had used a QCLAS (CW-QC-TILDAS-SC-S-N2OISO ARI USA) with one laser (2189 cmminus1) in com-bination with a closed chamber system To achieve the preci-sion and accuracy levels reported in their study Yamamoto etal (2014) corrected their measurements for mixing ratio de-

pendence and minimized instrumental drift by measuring N2gas every 1 h for background correction These authors alsoshowed that careful temperature control of their instrument inan air-conditioned cabinet was necessary for achieving opti-mal results

214 TREX-QCLAS

A compact mini QCLAS device (CW-QC-TILDAS-76-CSserial number 074 ARI USA) coupled with a preconcen-tration system (TREX) was provided by Empa SwitzerlandThe spectrometer comprises a continuous-wave mid-infraredquantum cascade laser source emitting at 2203 cmminus1 andan astigmatic multipass absorption cell with a path lengthof 76 m and a volume of approximately 620 cm3 (Ibraim etal 2018) (Table 1) The TREX unit was designed and man-ufactured at Empa and is used to separate the N2O from



the sample gas prior to QCLAS analysis Thereby the initial[N2O] is increased by a factor of 200ndash300 other trace gasesare removed and the gas matrix is set to standardized con-ditions Before entering the TREX device CO is oxidized toCO2 using a metal catalyst (Sofnocat 423 Molecular Prod-ucts Limited GB) Water and CO2 in sample gases wereremoved by a permeation dryer (PermaPure Inc USA) incombination with a sodium hydroxide (NaOH)magnesiumperchlorate (Mg(ClO4)2) trap (Ascarite 6 g 10ndash35 meshSigma Aldrich Switzerland bracketed by Mg(ClO4)2 2times15 g Alfa Aesar Germany) Thereafter N2O is adsorbedon a HayeSep D (Sigma Aldrich Switzerland) filled trapcooled down to 1251plusmn 01 K by attaching it to a copperbaseplate mounted on a high-power Stirling cryocooler (Cry-oTel GT Sunpower Inc USA) N2O adsorption requires5080plusmn 0011 L of gas to have passed through the adsorp-tion trap For N2O desorption the trap is decoupled from thecopper baseplate while slowly heating it to 275 K with a heatfoil (diameter 622 mm 100 W HK5549 Minco ProductsInc USA) Desorbed N2O is purged with 1ndash5 cm3 minminus1

of synthetic air into the QCLAS cell for analysis By con-trolling the flow rate and trapping time the [N2O] in theQCLAS cell can be adjusted to 60ndash80 ppm at a cell pressureof 356plusmn 004 mbar A custom-written LabVIEW program(Version 1801 National Instruments Corp USA) allows re-mote control and automatic operation of the TREX So farthe TREX-QCLAS system has been successfully applied todetermine N2O emission as well as N2O isotopic signaturesfrom various ecosystems (eg Mohn et al 2012 Harris etal 2014 Wolf et al 2015 Ibraim et al 2019)

215 GC-IRMS

IRMS analyses were conducted at ETH Zuumlrich using a gaspreparation unit (Trace Gas Elementar Manchester UK)coupled to an IsoPrime100 IRMS (Elementar ManchesterUK) [N2O] analysis using gas chromatography was alsoperformed at ETH Zuumlrich (456-GC Scion Instruments Liv-ingston UK) GC-IRMS analyses were conducted as part ofexperiments described further in Sect 248 Further analyti-cal details are provided in Sect S1 in the Supplement

22 Sample and reference gases

221 Matrix and interference test gases

Table 2 provides O2 Ar and trace gas mole fractions of ma-trix gases and interference test gases used during testing Thefour matrix gases comprised synthetic air (matrix a MesserSchweiz AG Switzerland) synthetic air with Ar (matrix bCarbagas AG Switzerland) synthetic air with Ar CO2 CH4and CO at near-ambient mole fractions (matrix c CarbagasAG Switzerland) and high-purity nitrogen gas (N2 MesserSchweiz AG Switzerland) Matrix gases were analyzed inthe World Meteorological Organization (WMO) Global At-

mosphere Watch (GAW) World Calibration Center at Empa(WCC Empa) for CO2 CH4 H2O (G1301 Picarro IncUSA) and N2O and CO (CW-QC-TILDAS-76-CS ARIUSA) against standards of the National Oceanic and At-mospheric AdministrationEarth System Research Labora-toryGlobal Monitoring Division (NOAAESRLGMD) The[N2O] in all matrix gases and N2 were below 03 ppb Thethree gas mixtures used for testing of spectral interferencescontained higher mole fractions of either CO2 CH4 or CO inmatrix gas b (Carbagas AG Switzerland) which preventedspectroscopic analysis of other trace substances

222 Reference gases (S1 S2) and pressurized air(PA1 PA2)

Preparation of pure and diluted reference gases

Two reference gases (S1 S2) with different N2O isotopiccomposition were used in this study Pure N2O referencegases were produced from high-purity N2O (Linde Ger-many) decanted into evacuated Luxfer aluminum cylinders(S1 P3333N S2 P3338N) with ROTAREX valves (MatarItaly) to a final pressure of maximum 45 bar to avoid con-densation Reference gas 1 (S1) was high-purity N2O onlyFor reference gas 2 (S2) high-purity N2O was supplementedwith defined amounts of isotopically pure (gt 98 ) 14N15NO(NLM-1045-PK) 15N14NO (NLM-1044-PK) (CambridgeIsotope Laboratories USA) and NN18O using a 10-port two-position valve (EH2C10WEPH with 20 mL sample loopValco Instruments Inc Switzerland) Since NN18O was notcommercially available it was synthesized using the follow-ing procedure (1) 18O exchange of HNO3 (18 mL SigmaAldrich) with 97 H18

2 O (5 mL Medical Isotopes Inc)under reflux for 24 h (2) condensation of NH3 and reac-tion controlled by LN2 and (3) thermal decomposition ofNH4NO3 in batches of 1 g in 150 mL glass bulbs with abreakseal (Glasblaumlserei Moumlller AG Switzerland) to produceNN18O The isotopic enrichment was analyzed after dilu-tion in N2 (999999 Messer Schweiz AG) with a Vi-sion 1000C quadrupole mass spectrometer (QMS) equippedwith a customized ambient pressure inlet (MKS InstrumentsUK) Triplicate analysis provided the following composition3625plusmn 010 of NN16O and 6375plusmn 076 of NN18O

High [N2O] reference gases (S1-a90 ppm S1-b90 ppm S1-c90 ppm S2-a90 ppm) with a target mole fraction of 90 ppmwere prepared in different matrix gases (a b c) using a two-step procedure First defined volumes of S1 and S2 weredosed into Luxfer aluminum cylinders (ROTAREX valveMatar Italy) filled with matrix gas (a b and c) to ambientpressure using N2O calibrated mass flow controllers (MFCs)(Voumlgtlin Instruments GmbH Switzerland) Second the N2Owas gravimetrically diluted (ICS429 Mettler Toledo GmbHSwitzerland) with matrix gas to the target mole fraction Am-bient [N2O] reference gases (S1-c330 ppb S2-c330 ppb) witha target mole fraction of 330 ppb were prepared by dosing



Table 2 O2 Ar content and trace gas concentrations for matrix and interference test gases Trace gas concentrations of matrix gases wereanalyzed by WMO GAW WCC Empa against standards of the NOAAESRLGMD For trace gas concentrations of interference test gasesmanufacturer specifications are given Reported O2 and Ar content is according to manufacturer specifications The given uncertainty is theuncertainty stated by the manufacturer or the standard deviation for analysis of n cylinders of the same specification

Gas Abbreviation O2a Ara CO2

b CH4b COb N2Ob n

() () (ppm) (ppb) (ppb) (ppb)

Matrix gases

Synthetic air Matrix a 205plusmn 05 ndash lt 1 lt 25 lt 200 lt 025 4Synthetic air+Ar Matrix b 2095plusmn 02 095plusmn 001 lt 05 lt 15 lt 150 lt 015 3Synthetic air+Ar+CO2+CH4+CO Matrix c 2095plusmn 04 095plusmn 002 397plusmn 3 2004plusmn 20 195plusmn 3 lt 015 9Nitrogen (60) N2 lt 000003 lt 00001 lt 02 lt 1 lt 1 lt 005 2

O2a Ara CO2

a CH4a COa N2Oa n

() () () (ppm) (ppm) (ppb)

Interference test gasesCO2 in synthetic air+Ar CO2 in matrix b 2106plusmn 02 094plusmn 001 402plusmn 004 na na na ndashCH4 in synthetic air+Ar CH4 in matrix b 2079plusmn 04 096plusmn 002 na 199plusmn 4 na na ndashCO in synthetic air+Ar CO in matrix b 2095plusmn 04 095plusmn 002 na na 206plusmn 04 na ndash

a Manufacturer specifications b Analyzed at WMO GAW WCC Empa ldquonardquo indicates data not analyzed due to very high concentration of one trace substance which affects spectroscopic analysisof other species

S1-c90 ppm or S2-c90 ppm into evacuated cylinders with a cali-brated MFC followed by gravimetric dilution with matrix c

Analysis of reference gases and pressurized air

Table 3 details the trace gas mole fractions and N2O isotopiccomposition of high and ambient [N2O] reference gases aswell as commercial pressurized air (PA1 and PA2) used dur-ing testing Trace gas mole fractions of high [N2O] referencegases were acquired from the trace gas levels in the respec-tive matrix gases (Table 2) while ambient [N2O] referencegases and target as well as background gases were analyzedby WCC Empa The isotopic composition of high [N2O] iso-tope reference gases in synthetic air (S1-a90 ppm S2-a90 ppm)was analyzed in relation to N2O isotope standards (Cal1ndashCal3) in an identical matrix gas (matrix a) using laser spec-troscopy (CW-QC-TILDAS-200 ARI Billerica USA) Thecomposition of Cal1ndashCal3 is outlined in Sect S2

For high-mole-fraction reference gases in matrix b and c(S1-b90 ppm S1-c90 ppm S2-c90 ppm) the δ values acquired forS1-a90 ppm and S2-a90 ppm were assigned since all S1 and S2reference gases (irrespective of gas matrix) were generatedfrom the same source of pure N2O gas Direct analysis of S1-b90 ppm S1-c90 ppm and S2-c90 ppm by QCLAS was not feasi-ble as no N2O isotope standards in matrix b and c were avail-able The absence of significant difference (lt 1 permil) in N2Oisotopic composition between S1-b90 ppm and S1-c90 ppm inrelation to S1-a90 ppm (and S2-c90 ppm to S2-a90 ppm) was as-sured by first statically diluting S1-b90 ppm S1-c90 ppm andS2-c90 ppm to ambient N2O mole fractions with synthetic airThis was followed by analysis using TREX-QCLAS (as de-scribed in Sect 214) against the same standards used forS1-a90 ppm S2-a90 ppm isotope analysis

Ambient mole fraction N2O isotope reference gases (S1-c330 ppb S2-c330 ppb) and PA1 and PA2 were analyzed byTREX-QCLAS (Sect 214) using N2O isotope standards(Cal1ndashCal5) as outlined in Sect S2

23 Laboratory setup measurement procedures anddata processing

231 Laboratory setup

All experiments were performed at the Laboratory for AirPollutionEnvironmental Technology Empa Switzerlandduring June 2018 and February 2019 The laboratory wasair conditioned to 295 K (plusmn1 K) with plusmn05 K diel variations(Saveris 2 Testo AG Switzerland) with the exception of ashort period (7 to 8 July 2018) where the air conditioningwas deactivated to test the temperature dependence of ana-lyzers Experiments were performed simultaneously for allanalyzers with the exception of the TREX-QCLAS whichrequires an extensive measurement protocol and additionaltime to trap and measure N2O (Ibraim et al 2018) and thuscould not be integrated concurrently with the other analyzers

Figure 2 shows a generalized experimental setup used forall experiments Additional information for specific experi-ments is given in Sect 24 and individual experimental se-tups are depicted in Sect S3 Gas flows were controlled us-ing a set of MFCs (model high-performance Voumlgtlin In-struments GmbH Switzerland) integrated into a MFC con-trol unit (Contrec AG Switzerland) All MFCs were cali-brated by the manufacturer for whole air which according toVoumlgtlin Instruments is valid for pure N2 and pure O2 as wellOperational ranges of applied MFCs ranged from 0ndash25 to 0ndash5000 cm3 minminus1 and had reported uncertainties of 03 oftheir maximum flow and 05 of actual flow To reduce the



uncertainty of the flow regulation the MFC with the small-est maximum flow range available was selected The sum ofdosed gas flows was always higher than the sum of gas con-sumption by analyzers with the overflow exhausted to roomair Gas lines between gas cylinders and MFCs as well asbetween MFCs and analyzers were 18primeprime stainless steel tub-ing (type 304 Supelco Sigma Aldrich Switzerland) Manualtwo-way (SS-1RS4 or SS-6H-MM Swagelok Switzerland)or three-way valves (SS-42GXS6MM Swagelok Switzer-land) were used to separate or combine gas flows

232 Measurement procedures data processing andcalibration

With the exception of Allan variance experiments performedin Sect 241 all gas mixtures analyzed during this studywere measured by the laser spectrometers for a period of15 min with the last 5 min used for data processing Cus-tomized R scripts (R Core Team 2017) were used to ex-tract the 5 min averaged data for each analyzer Whilstthe OA-ICOS and QCLAS instruments provide individual14N14N16O 14N15N16O 15N14N16O and 14N14N18O molefractions the default data output generated by the CRDS an-alyzers are δ values with underlying calculation schemes in-accessible to the user Therefore to remain consistent acrossanalyzers uncalibrated δ values were calculated for OA-ICOS and QCLAS instruments first using literature valuesfor the 15N14N (00036782) and 18O16O (00020052) iso-tope ratios of AIR-N2 and VSMOW (Werner and Brand2001)

Each experiment was performed over the course of 1 dand consisted of three phases (1) an initial calibration phase(2) an experimental phase and (3) a final calibration phaseDuring phases (1) and (3) reference gases S1-c330 ppb andS2-c330 ppb were analyzed On each occasion (ie twice aday) this was followed by the analysis of PA1 which wasused to determine the long-term (day-to-day) repeatability ofthe analyzers Phase (2) experiments are outlined in Sect 24Throughout all three phases all measurements were system-atically alternated with an Anchor gas measurement the pur-pose of which was two-fold (1) to enable drift correctionand (2) as a means of quantifying deviations of the measured[N2O] and δ values caused by increasing [N2O] (Sect 244)the removal of matrix gases (O2 and Ar in Sect 245) or ad-dition of trace gases (CO2 CH4 and CO in Sect 246) Ac-cordingly the composition of the Anchor gas varied acrossexperiments (see Sect 24) but remained consistent through-out each experiment A drift correction was applied to thedata if a linear or non-linear model fitted to the Anchor gasmeasurement over the course of an experiment was statisti-cally significant at p lt 005 Otherwise no drift correctionwas applied

In Sect 243 (repeatability experiments) and 248 (two-end-member mixing experiments) trace gas effects werecorrected according to Eqs (1) (2) and (3) using derived

analyzer-specific correction functions because the CO2 CH4and CO composition of PA1 in Sect 243 and the gas mix-tures in Sect 248 varied from those of the calibration gasesS1-c330 ppb (S1) and S2-c330 ppb (S2)

[N2O]tcG = [N2O]measG

minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)

δtcG = δmeasGminussum

x

(1δ(1[x]GδmeasG

))(2)

and

1[x]G = [x]Gminus[x]S1+ [x]S2

2 (3)

where [N2O]tcG and δtcG refer to the trace-gas-corrected[N2O] and δ values (δ15Nα δ15Nβ or δ18O) of sample gasG respectively [N2O]measG and δmeasG are the raw uncor-rected [N2O] and δ values measured by the analyzer for sam-ple gas G respectively 1[N2O] and 1δ refer to the offseton the [N2O] or δ values respectively resulting from the dif-ference in trace gas mole fraction between sample gas G andreference gases denoted 1[x]G [x]G is the mole fraction oftrace gas x (CO2 CH4 or CO) in sample gas G and [x]S1and [x]S2 are the mole fractions of trace gas x in referencegases S1-c330 ppb and S2-c330 ppb It is important to note thatthe differences in CO2 and CH4 mole fractions in S1-c330 ppband S2-c330 ppb are 2 orders of magnitude smaller than thedifferences with PA1

Thereafter δ values of trace-gas-corrected mole-fraction-corrected (Sect 248 only) and drift-corrected measure-ments from the analyzers were normalized to δ values onthe international isotope-ratio scales using a two-point lin-ear calibration procedure derived from values of S1-c330 ppb(S1) and S2-c330 ppb (S2) calculated using Eq (4) (Groumlning2018)

δCalG =δrefS1minus δrefS2

δcorrS1minus δcorrS2middot(δcorrGminus δcorrS1

)+δrefS1 (4)

where δCalG is the calibrated δ value for sample gas Gnormalized to international isotope-ratio scales δrefS1 andδrefS2 are the respective δ values assigned to referencegases S1-c330 ppb and S2-c330 ppb δcorrS1 and δcorrS2 are therespective δ values measured for the reference gases S1-c330 ppb and S2-c330 ppb which if required were drift cor-rected and δcorrG is the trace-gas-corrected mole-fraction-corrected (Sect 248 only) and drift-corrected (if required)δ value measured for the sample gas G

24 Testing of instruments

An overview of all experiments performed in this study in-cluding applied corrections and instruments tested is pro-vided in Table 4



Table 3 Trace gas concentrations and N2O isotopic composition of high and ambient N2O concentration reference gases and pressurized airTrace gas concentrations of high concentration reference gases were retrieved from the composition of matrix gases used for their production(see Table 2) trace gas concentrations in ambient concentration reference gases and pressurized air were analyzed by WMO GAW WCCEmpa against standards of the NOAAESRLGMD The N2O isotopic composition was quantified by laser spectroscopy (QCLAS) andpreconcentration ndash laser spectroscopy (TREX-QCLAS) against reference gases previously analyzed by the Tokyo Institute of Technology

Gas CO2 CH4 CO N2O δ15Nα vs δ15Nβ vs δ18O vs(ppm) (ppb) (ppb) (ppb) AIR-N2 AIR-N2 VSMOW

(permil) (permil) (permil)

High N2O concentration reference gases

S1-a90 ppm lt 1 lt 25 lt 200 sim 90000 054plusmn 017 115plusmn 006 3946plusmn 001S1-b90 ppm lt 05 lt 15 lt 150 sim 90000 054plusmn 017 115plusmn 006 3946plusmn 001S1-c90 ppm 397plusmn 3 2004plusmn 20 195plusmn 3 sim 90000 054plusmn 017 115plusmn 006 3946plusmn 001S2-a90 ppm lt 1 lt 25 lt 200 sim 90000 5143plusmn 006 5514plusmn 009 10009plusmn 003S2-c90 ppm 397plusmn 3 2004plusmn 20 195plusmn 3 sim 90000 5143plusmn 006 5514plusmn 009 10009plusmn 003

Ambient N2O concentration reference gases

S1-c330 ppb 39978plusmn 004 2022plusmn 02 195plusmn 03 32745plusmn 006 092plusmn 039 144plusmn 025 3912plusmn 018S2-c330 ppb 39862plusmn 004 2020plusmn 02 193plusmn 03 32397plusmn 006 5238plusmn 010 5561plusmn 012 9959plusmn 003

High N2O concentration source gas (SG) for two-end-member mixing experiments (Sect 248)

SG1-a90 ppm lt 1 lt 25 lt 200 sim 90000 minus2435plusmn 032 minus2294plusmn 033 3179plusmn 012SG2-a90 ppm lt 1 lt 25 lt 200 sim 90000 5143plusmn 006 5514plusmn 009 10009plusmn 003

Pressurized air

Pressurized air (PA1) 20055plusmn 007 2582plusmn 02 187plusmn 02 32651plusmn 006 1583plusmn 003 minus339plusmn 014 4466plusmn 002Pressurized air (PA2) 43799plusmn 036 2957plusmn 03 275plusmn 04 33350plusmn 009 1581plusmn 007 minus331plusmn 0004 4472plusmn 004

Figure 2 The generalized experimental setup used for all experiments conducted in this study The gases introduced via MFC flows A Band C were changed according to the experiments outlined in Sect 24 Tables 2 and 3 provide the composition of the matrix gases (MFC B)interference test gases (MFC C) and high [N2O] concentration reference gases (MFC A) Laboratory setups for each individual experimentare provided in Sect S3



Table 4 Overview of the experiments performed in this study

Experiment Sections Aims Corrections applied Instruments tested Comments

Instrumental 241 Short-term None OA-ICOS I Conducted at N2Oprecision 31 precision optimal CRDS I amp II concentrations(Allan integration QCLAS I sim 326deviation) timemaximum QCLAS II amp III 1000 10 000 ppb

precision (ambient only)and drift TREX-QCLAS I

Temperature 242 Temperature effects None OA-ICOS Ieffects 32 on [N2O] and CRDS I amp II

isotope deltas QCLAS I

Repeatability 243 Repeatability Drift OA-ICOS I Conducted at N2O(short term 33 CRDS I amp II concentrations sim 326sim 2 h) QCLAS I 1000 10 000 ppb

TREX-QCLAS I

Repeatability 243 Repeatability Drift delta OA-ICOS I Conducted at sim 326 ppb(long term 33 calibration CRDS I amp II N2O usingsim 2 weeks) trace gas QCLAS I PA1

effecta TREX-QCLAS I

N2O mole 244 [N2O] effects Drift OA-ICOS I CRDS 300 to 1500 ppb N2Ofraction 34 on isotope deltas CRDS I amp II OA-ICOS QCLAS 300 toeffects and derive correction QCLAS I 90 000 ppb

functions

Gas matrix 245 Gas matrix effects Drift OA-ICOS I Conducted at N2Oeffects (N2 35 on [N2O] and isotope QCLAS I concentrationsO2 and deltas and derive TREX-QCLAS I sim 330 660Ar) correction functions 990 ppb

Trace gas 246 Trace gas effects Drift OA-ICOS I Conducted at N2Oeffects 36 on [N2O] and CRDS I amp II concentrations(H2O CO2 isotope deltas and QCLAS I sim 330 660CH4 CO) derive correction TREX-QCLAS I 990 ppb

functions (except H2O)

CO2 and 247 Effects of removal Drift OA-ICOS I Conducted at N2OCO removal 36 of CO2 (Ascarite) CRDS I amp II concentrations

and CO (Sofnocat) on QCLAS I sim 330 ppb[N2O] andisotope deltas

Two- 248 Test the ability Drift three-point OA-ICOS I (Exps 1ndash6) The workflowend-member 37 of the instruments concentration CRDS I amp II (Exps 1ndash4) provided inmixing to extrapolate a N2O dependence QCLAS I (Exps 1ndash6) Sect 43

source using a Keeling δ calibration TREX-QCLAS I (Exps 1ndash2) was appliedplot approach trace gas effecta GC [N2O]

and scaled with N2Ob IRMS [δ] (Exps 1ndash6)

a Derived from trace gas effect determined in Sect 36 b Derived from scaling effects described in Sect 362

241 Allan precision

The precision of the laser spectrometers was determinedusing the Allan variance technique (Allan 1966 Werle etal 1993) Experiments were conducted at different [N2O]ambient 1000 and 10 000 ppb For the Allan variance test-ing conducted at ambient [N2O] a continuous flow of PA1was measured continuously for 30 h For testing conducted

at 1000 and 10 000 ppb [N2O] S1-c90 ppm was dynamicallydiluted to 1000 or 10 000 ppb [N2O] with matrix gas c for10 h CRDS I and II were disconnected for the 10 000 ppbmeasurement because [N2O] exceeded the specified mea-surement range Daily drifts were estimated using the slopeof the linear regression over the measurement period normal-ized to 24 h (ie ppb dminus1 and permil dminus1)



242 Temperature effects

To investigate instrumental sensitivities to variations in am-bient temperature PA1 was simultaneously and continuouslymeasured by all analyzers in flow-through mode for a pe-riod of 24 h while the air conditioning of the laboratory wasturned off for over 10 h This led to a rise in temperature from21 to 30 C equating to an increase in temperature of approx-imately 09 C hminus1 The increase in laboratory room temper-ature was detectable shortly after the air conditioning wasturned off due to considerable heat being released from sev-eral other instruments located in the laboratory Thereafterthe air conditioning was restarted and the laboratory temper-ature returned to 21 C over the course of 16 h equating toa decrease of roughly 06 C hminus1 with most pronounced ef-fects observable shortly after restart of air conditioning whentemperature changes were highest

243 Repeatability

Measurements of PA1 were taken twice daily oversim 2 weeksprior to and following the experimental measurement periodto test the long-term repeatability of the analyzers Measure-ments were sequentially corrected for differences in trace gasconcentrations (Eqs 1ndash3) drift (if required) and then δ cal-ibrated (Eq 4) No matrix gas corrections were applied be-cause the N2 O2 and Ar composition of PA1 was identicalto that of S1-c330 ppb and S2-c330 ppb TREX-QCLAS I mea-surements for long-term repeatability were collected sepa-rately from other instruments over a period of 6 months Re-peatability over shorter time periods (25 h) was also testedfor each analyzer by acquiring 10 repeated 15 min measure-ments at different N2O mole fractions ambient (PA1) 1000and 10 000 ppb N2O

244 N2O mole fraction dependence

To determine the effect of changing [N2O] on the measuredδ values S1-c90 ppm was dynamically diluted with matrix cto various [N2O] spanning the operational ranges of the in-struments For both CRDS analyzers mole fractions between300 to 1500 ppb were tested while for the OA-ICOS I andQCLAS I mixing ratios ranged from 300 to 90 000 ppb Be-tween each [N2O] step change the dilution ratio was sys-tematically set to 330 ppb N2O to perform an Anchor gasmeasurement For each instrument the effect of increasing[N2O] on δ values was quantified by comparing the mea-sured δ values at each step change to the mean measuredδ values of the Anchor gas and was denoted 1δ such that1δ = δmeasuredminus δAnchor and 1δAnchor = 0 The experimentwas repeated on three consecutive days to test day-to-dayvariability

245 Gas matrix effects (O2 and Ar)

Gas matrix effects were investigated by determining the de-pendence of [N2O] and isotope δ values on the O2 or Ar mix-ing ratio of a gas mixture For O2 testing Gases 1 2 and 3(N2) were mixed to incrementally change mixing ratios ofO2 (0 ndash205 O2) while maintaining a consistent [N2O]of 330 ppb (Table 5) As an Anchor gas Gas 1 (S1-a90 ppm)was dynamically diluted with Gas 2 (matrix a) to produce330 ppb N2O in matrix a O2 mole fractions in the variousgas mixtures were analyzed with a paramagnetic O2 analyzer(Servomex UK) and agreed with expected values to within03 (relative) For Ar testing Gas 1 (S1-b90 ppm) was dy-namically diluted with Gas 2 (matrix b) to produce an An-chor gas with sim 330 ppb N2O in matrix b Gases 1 2 and 3(N2+O2) were then mixed to incrementally change mixingratios of Ar (0003 ndash095 Ar) while a consistent [N2O]of 330 ppb was maintained Ar compositional differenceswere estimated based on gas cylinder manufacturer specifi-cations and selected gas flows The effects of decreasing O2and Ar on [N2O] and δ values were quantified by compar-ing the measured [N2O] and δ values at each step change tothe mean measured [N2O] and δ values of the Anchor gasand were denoted 1[N2O] and 1δ similar to Sect 244Deviations in O2 and Ar mixing ratios were quantified bycomparing the [O2] and [Ar] at each step change to the mean[O2] and [Ar] of the Anchor gas and were denoted 1O2 and1Ar such that for example 1O2=O2 measuredminusO2 Anchor and1O2 Anchor=0 Both O2 and Ar experiments were triplicated

In addition O2 and Ar effects were derived for N2O molefractions of sim 660 and sim 990 ppb These experiments wereundertaken in a way similar to those described above exceptAnchor gas measurements were conducted once (not tripli-cated)

246 Trace gas effects (CO2 CH4 CO and H2O)

The sensitivity of [N2O] and δ values on changing tracegas concentrations was tested in a similar way to those de-scribed in Sect 245 In short Gas 1 (S1-b90 ppm) was dy-namically diluted with Gas 2 (matrix b) to create an An-chor gas with 330 ppb N2O in matrix b Gases 1 2 and 3(either CO2 CH4 or CO in matrix gas b) were mixed toincrementally change the mixing ratios of the target sub-stances (17ndash2030 ppm CO2 001ndash1025 ppm CH4 and 014ndash214 ppm CO) while maintaining a consistent gas matrix and[N2O] of 330 ppb (Table 5) Trace gas mole fractions in theproduced gas mixtures were analyzed with a Picarro G2401(Picarro Inc USA) and agreed with predictions within betterthan 2 ndash3 (relative) Similar to Sect 244 the effects ofincreasing CO2 CH4 and CO on [N2O] and δ values werequantified by comparing the measured [N2O] and δ valuesat each step change to the mean measured [N2O] and δ val-ues of the Anchor gas and were denoted 1[N2O] and 1δSimilar to Sect 245 deviations in CO2 CH4 and CO mix-



Table 5 Gas mixtures used to test effects of gas matrix (O2 Ar) or trace gases (CO2 CH4 and CO) on [N2O] and isotope deltas Gas 1was dynamically diluted with Gas 2 to make up an Anchor gas with [N2O] of sim 330 ppb which was systematically measured throughout theexperiments to (1) enable drift correction and (2) quantify deviations of the measured [N2O] and δ values caused by the removal of matrixgases (O2 and Ar in Sect 245) or addition of trace gases (CO2 CH4 and CO in Sect 246) Gases 1 2 and 3 were combined in differentfractions to make up sample gas with identical [N2O] but varying mixing ratio of the target compound

Target compound Gas 1 Gas 2 Gas 3 Mixing range

O2 S1-a90 ppm (N2O+N2+O2) N2+O2a N2 0 ndash205 O2

Ar S1-b90 ppm (N2O+N2+O2+Ar) N2+O2+Arb N2+O2a 0003 ndash095 Ar

CO2 S1-b90 ppm (N2O+N2+O2+Ar) N2+O2+Arb CO2 in N2+O2+Arb 172ndash2030 ppm CO2CH4 S1-b90 ppm (N2O+N2+O2+Ar) N2+O2+Arb CH4 in N2+O2+Arb 0014ndash1025 ppm CH4CO S1-b90 ppm (N2O+N2+O2+Ar) N2+O2+Arb CO in N2+O2+Arb 014ndash213 ppm CO

a Matrix a 205 O2 in N2 b Matrix b 2095 O2 095 Ar in N2

ing ratios were quantified by comparing the measured [CO2][CH4] and [CO] at each step change to the mean measured[CO2] [CH4] and [CO] of the Anchor gas Each experimentwas triplicated The interference effects were also tested atsim 660 ppb and sim 990 ppb N2O

The sensitivity of the analyzers to water vapor was testedby firstly diluting Gas 1 (S1-c90 ppm) with Gas 2 (matrix c)to produce an Anchor gas with 330 ppb N2O This mix-ture was then combined with Gas 3 (also matrix c) whichhad been passed through a humidifier (customized setup byGlasblaumlserei Moumlller Switzerland) set to 15 C (F20 JulaboGmbH Germany) dew point By varying the flows of Gases 2and 3 different mixing ratios of water vapor ranging from0 to 13 800 ppm were produced and measured using a dew-point meter (model 973 MBW Switzerland) H2O effectswere quantified as described above but [N2O] results wereadditionally corrected for dilution effects caused by the ad-dition of water vapor into the gas stream Water vapor de-pendence testing was not performed on the TREX-QCLASI as the instrument is equipped with a permeation dryer atthe inlet

247 CO2 and CO removal using NaOH (Ascarite) andSofnocat

The efficiency of NaOH and Sofnocat for removing spec-tral effects caused by CO2 and CO was assessed by repeat-ing CO2 and CO interference tests (Sect 246) but with therespective traps connected in line These experiments weretriplicated but only undertaken at sim 330 ppb N2O NaOHtraps were prepared using stainless steel tubing (OD 254 cmlength 20 cm) filled with 14 g Ascarite (0ndash30 mesh SigmaAldrich Switzerland) bracketed by 3 g Mg(ClO4)2 (Alfa Ae-sar Germany) each separated by glass wool The Sofno-cat trap was prepared similarly using stainless steel tubing(OD 254 cm length 20 cm) filled with 50 g Sofnocat (Sofno-cat 423 Molecular Products Limited GB) and capped oneach side with glass wool

248 Two-end-member mixing

The ability of the instruments to accurately extrapolate N2Osource compositions was tested using a simulated two-end-member mixing scenario in which a gas with high N2O con-centration considered to be a N2O source gas (SG) was dy-namically diluted into a gas with ambient N2O concentration(PA2) considered to be background air N2O mole fractionswere raised above ambient levels (denoted as1N2O) in threedifferent scenarios ranging (1) 0ndash30 ppb (2) 0ndash700 ppb and(3) 0ndash10 000 ppb In each scenario two isotopically differ-ent source gases with high N2O concentration were usedone source gas (SG1-a90 ppm) was 15N depleted comparedto PA2 and a second source gas (SG2-a90 ppm) was 15N en-riched compared to PA2 (Table 3) The three different mixingscenarios and two different source gases resulted in a total ofsix mixing scenarios (referred to as Exps 1ndash6) During eachexperiment PA2 was alternated with PA2+SG in four dif-ferent mixing ratios to give a span of N2O concentrationsand isotopic compositions required for Keeling plot analysisEach experiment was triplicated OA-ICOS I and QCLAS Iwere used in all experiments (Exps 1ndash6) CRDS was usedfor 1N2O 0ndash30 ppb and 0ndash700 ppb (Exps 1ndash4) and TREX-QCLAS was only used for 1N2O 0ndash30 ppb (Exps 1ndash2)

To test the robustness of trace gas correction equations de-rived for each analyzer in Sect 36 NaOH and Sofnocat trapswere placed in line between the PA2+SG mixtures and theanalyzers such that we could ensure a difference in CO2 andCO mole fractions between the measured gas mixture andreference gases (S1-c330 ppb S2-c330 ppb) The experimentswere also bracketed by two calibration phases (S1-c330 ppbS2-c330 ppb) to allow for δ calibration followed by two phaseswhere the N2O concentration dependence was determined

Gas samples for GC-IRMS analysis were taken in thesame phase (last 5 min of 15 min interval) used during theminute prior to the final 5 min used for averaging by thelaser-based analyzers The gas was collected at the commonoverflow port of the laser spectrometers using a 60 mL sy-ringe connected via a Luer lock three-way valve to the nee-



dle and port The 200 mL samples were taken at each con-centration step A 180 mL gas sample was stored in pre-evacuated 110 mL serum crimp vials for isotopic analysisusing IRMS IRMS analyses were conducted at ETH Zuumlrichusing a gas preparation unit (Trace Gas Elementar Manch-ester UK) coupled to an IsoPrime100 IRMS (ElementarManchester UK) The remaining 20 mL were injected in apre-evacuated 12 mL Labco exetainer for [N2O] analysis us-ing gas chromatography equipped with an electron capturedetector (ECD) performed at ETH Zuumlrich (Bruker 456-GCScion Instruments Livingston UK) After injection sam-ples were separated on HayeSep D packed columns with a5 CH4 in Ar mixture (P5) as carrier and make-up gasThe GC was calibrated using a suite of calibration gases atN2O concentrations of 0393 (Carbagas AG Switzerland)102 (PanGas AG Switzerland) and 317 ppm (Carbagas AGSwitzerland) For further analytical details see Verhoeven etal (2019) and Sect S1

For the laser-based analyzers data were processed as de-scribed in Sect 232 using the following sequential or-der (1) analyzer-specific correction functions determined inSect 36 were applied to correct for differences in trace gasconcentrations (CO2 CO) between sample gas and calibra-tion gases (2) the effect of [N2O] changes was corrected us-ing a three-point correction (3) a drift correction based onrepeated measurements of PA2 was applied if necessary and(4) δ values standardized to international scales (Eq 4) usingS1-c330 ppb and S2-c330 ppb

3 Results

Note that due to the large number of results acquired in thissection only selected results are shown in Figs 3 to 14 Thecomplete datasets (including [N2O] δ15Nα δ15Nβ and δ18Oacquired by all instruments tested) are provided in Sect S4

31 Allan precision

Allan deviations (square root of Allan variance) for 5 and10 min averaging times often reported in manufacturer spec-ifications at sim 327 1000 and 10 000 ppb [N2O] are shownin Table 6

At near-atmospheric N2O mole fractions of sim 3265 ppbboth CRDS analyzers showed the best precision and stabilityfor the measurement of δ15Nα δ15Nβ and δ18O (032 permilndash041 permil 041 permilndash045 permil 041 permilndash046 permil at 300 s averag-ing time respectively) while for the precision of [N2O]the OA-ICOS I and the CRDS II showed best performance(17times 10minus2 ppb at 300 s averaging time) (Figs 3 and S4-1Table 6) The Allan precision of CRDS and OA-ICOS ana-lyzers further improved with increasing averaging times andoptimal averaging times typically exceeded 15ndash3 h The pre-cision and daily drift of the OA-ICOS I and both CRDS an-alyzers were in agreement with manufacturer specifications

(ABB ndash Los Gatos Research Inc 2019 Picarro Inc 2019)The CRDS II outperformed the CRDS I for precision pre-sumably due to manufacturer upgradesimprovements in thenewer model The QCLAS spectrometers exhibited signifi-cant differences between instruments which might be due todifferences in the instrument hardwaredesign as instrumentswere manufactured between 2012 and 2016 or in the param-eter setting (such as cell pressure and tuning parameters) ofdifferent analyzers

Generally short-term (approximately up to 100 s) pre-cision of QCLAS instruments was compatible or superiorto CRDS or OA-ICOS but data quality was decreased forlonger averaging times due to drift effects Nonetheless theperformance of QCLAS I II and III generally agrees withAllan precision measurements executed by Yamamoto etal (2014) who reported 19 permilndash26 permil precision for δ valuesat ambient N2O mole fractions and 04 permilndash07 permil at 1000 ppbN2O QCLAS I which was tested further in Sect 32ndash37displayed the poorest performance of all QCLAS analyzersin particular for δ15Nβ The primary cause of the observedexcess drift in QCLAS I was fluctuating spectral baselinestructure (ARI personal communication 2019) which canbe significantly reduced by applying an automatic spectralcorrection method developed by ARI This methodology iscurrently in a trial phase and thus not yet implemented inthe software that controls the QCLAS instruments A briefoverview of the methodology is provided in Sect S5 andcorrected results for QCLAS I are provided in Table 6 Thismethodology is not discussed in detail here as it is beyondthe scope of this work Nonetheless QCLAS I achieved Al-lan deviations of sim 04 permil at 300 s averaging time for δ15Nα

and δ15Nβ at ambient N2O mole fractions when this correc-tion method was applied by ARI

At [N2O] of 1000 ppb the precision of δ values measuredby all analyzers except CRDS I significantly improved dueto greater signal-to-noise ratios Whilst the performance ofOA-ICOS I was similar to that of CRDS II for δ15Nα andδ15Nβ (024 permil and 024 permil for CRDS II 028 permil and 037 permilfor OA-ICOS I at 300 s averaging time) CRDS II displayedthe best precision for δ18O (021 permil at 300 s averaging time)Also notable was the improved performance of the 2018model (CRDS II) compared to the 2015 model (CRDS I)QCLAS analyzers showed the best 1 s precision for δ valuesbut beyond 100 s δ measurements were still heavily affectedby instrumental drift resulting in lower precision especiallyfor QCLAS I When the spectral correction method describedin Sect S5 was applied QCLAS I achieved Allan deviationsof sim 02 permil at 300 s averaging time for δ15Nα and δ15Nβ at1000 ppb N2O

At [N2O] of 10 000 ppb all analyzers showed excellentprecision with QCLAS I II and III outperforming OA-ICOSI for precision of δ15Nα and δ15Nβ (collectively better than010 permil at 300 s averaging time for both δ15Nα and δ15Nβ )QCLAS II had the best precision for [N2O] (12 ppb at 300 saveraging time) OA-ICOS I and QCLAS III were the only



analyzers tested in this study that could be used to measureδ18O at 10 000 ppb N2O OA-ICOS I attained a precision of017 permil while QCLAS III attained a precision of 048 permilboth with 300 s averaging time QCLAS I achieved Allandeviations of sim 002 permilndash003 permil at 300 s averaging time forδ15Nα and δ15Nβ at 10 000 ppb N2O when the spectral cor-rection method (Sect S5) was applied

The precision of instruments on [N2O] measurements at1000 and 10 000 ppb N2O might not be representative be-cause of small fluctuations in the final gas mixture producedby the MFCs which were likely amplified due to the smalldilution ratios Moreover the different relative increases inAllan deviation compared to measurements at 3265 ppbmight have been caused by the different internal plumbingvolumes flow rates and spectral fits used for the analyzerswhich could scale or add to the increased Allan deviationintroduced via the MFCs Therefore the indicated [N2O]precisions should be considered as a pessimistic estimateNonetheless the observed decline in [N2O] precision for allanalyzers was around 1 order of magnitude when changingfrom atmospheric N2O mole fractions to 1000 ppb N2O andfrom 1000 ppb to 10 000 ppb N2O


All instruments tested showed significant effects albeit tovarying degrees on their measurements due to the changein laboratory temperature (Figs 4 and S4-2) The OA-ICOSI displayed no clear temperature effects for [N2O] δ15Nα

and δ15Nβ but displayed a moderate temperature depen-dence for δ18O measurements (up to 14 permil deviation fromthe mean) with measurement drift closely paralleling thelaboratory temperature (r2

= 078) Both CRDS instrumentsdisplayed smaller shifts in [N2O] (up to 014 ppb deviationfrom the mean) δ15Nα δ15Nβ and δ18O that occurred partic-ularly when the laboratory temperature had an acute changeQCLAS I showed a strong temperature dependence on δ15Nα

(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability

The best long-term repeatability for δ values was achievedby TREX-QCLAS I with 060 permil for δ15Nα 037 permil forδ15Nβ and 046 permil for δ18O even though measurements weretaken over a 6-month period (Table 7) The best repeatabilitywithout preconcentration was achieved by CRDS analyzerswith 052 permilndash075 permil for CRDS II and 079 permilndash083 permil forCRDS I for all δ values OA-ICOS I achieved repeatabilitybetween 1 permilndash2 permil (147 permil 119 permil and 217 permil for δ15Nα δ15Nβ and δ18O respectively) QCLAS I isotopic measure-ments attained repeatability of 54 permil and 86 permil for δ15Nα

and δ15Nβ respectively Short-term repeatability results for10 repeated 15 min measurements periods over 25 h are pro-vided in Sect S6

34 Dependence of isotopic measurements on N2O molefraction

There was an offset in measured δ values resulting from thechange in [N2O] introduced to the analyzers (Figs 5 and S4-3) A linear relationship between 1δ15Nα β and 1δ18O val-ues with [1N2O] was observed across all analyzers How-ever examination of the residuals from the linear regressionrevealed varying degrees of residual curvature highlightingthat further non-linear terms would be required to adequatelydescribe and correct for this mole fraction dependence Re-peated analysis of [N2O] dependencies on consecutive daysshowed similar trends indicating that the structure of non-linear effects might be stable over short periods of time Nev-ertheless there were small variabilities in δ values at a givenN2O mole fraction which could be due to the inherent uncer-tainty of the measurement andor day-to-day variations in themole fraction dependence The standard deviation of individ-ual 5 min averages of δ15Nα β and δ18O also varied accord-ing to the [N2O]measured by each analyzer due to variationsin the signal-to-noise ratio (Sect S7)


351 Gas matrix effects at ambient N2O mole fractions

With the exception of TREX-QCLAS I all instruments dis-played strong O2 dependencies for [N2O] and δ values(Figs 6 and S4-4) For these instruments linear regressionsbest described the offset of measured [N2O] and δ valuesresulting from the change in O2 composition of the matrixgas Importantly CRDS I and II displayed different degreesof O2 interference on [N2O] and δ values suggesting thatthese dependencies were either analyzer-specific or differ-ences were due to hardwaresoftware modifications betweendifferent production years Preconcentration prior to analy-sis as performed in TREXndashQCLAS I eliminated O2 depen-dencies as the gas matrix was normalized to synthetic air(205 O2)

The change in Ar composition of the matrix gas causedminor yet measurable interferences on [N2O] and δ mea-surements (Fig S4-5) The range investigated was betweenapproximately 0 and 095 Ar as anticipated for N2O insynthetic air (no Ar) reference gas versus a whole air (withAr) sample gas The effects observed for a 095 changein [Ar] were significantly smaller than that observed for O2but might extend to a similar range for sample and referencegases with higher differences in [Ar] The interference ef-fects were found to be best described by second-order poly-nomial functions though we expect that a linear fit wouldserve equally well if a larger change in [Ar] was investi-gated Although most functions to describe the dependenceon Ar across all instruments were statistically significant(p lt 005) maximum effects did not transgress the repeata-



Table 6 Key parameters for instrument stability retrieved from Allan variance experiments for [N2O] δ15Nα δ15Nβ and δ18O precision(1σ ) at 300 and 600 s averaging times and daily drift at various N2O concentrations The 1σ data refer to Allan deviation (square root ofAllan variance)

Instrument 1σ 1σ N2O 1σ 1σ δ15Nα 1σ 1σ δ15Nβ 1σ 1σ δ18ON2O N2O drift δ15Nα δ15Nα drift δ15Nβ δ15Nβ δ18O δ18O δ18O drift(ppb) (ppb) (ppb dminus1) (permil) (permil) (permil dminus1) (permil) (permil) (permil dminus1) (permil) (permil) (permil dminus1)

(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O

CRDS I 30times 10minus2 23times 10minus2 20times 10minus2 041 027 022 045 038 018 046 034 003CRDS II 17times 10minus2 12times 10minus2 32times 10minus2 032 023 35times 10minus3 041 031 014 041 028 010OA-ICOS I 17times 10minus2 13times 10minus2 15times 10minus2 108 082 007 079 052 050 169 114 234QCLAS I 63times 10minus2 46times 10minus2 37times 10minus2 124 141 680 345 422 1581 nd nd ndQCLAS Ilowast 21times 10minus2 24times 10minus2 012 039 037 071 042 055 483 nd nd ndQCLAS II 95times 10minus3 11times 10minus2 100 108 144 020 060 072 005 nd nd ndQCLAS III 25times 10minus2 36times 10minus2 075 081 123 009 078 122 004 097 151 013

sim 1000 ppb N2O

CRDS I 77times 10minus1 60times 10minus1 029 088 067 067 089 073 139 081 067 032CRDS II 21times 10minus1 13times 10minus1 054 024 020 036 024 018 023 021 013 086OA-ICOS I 17times 10minus1 12times 10minus1 102 028 023 093 037 025 054 067 044 015QCLAS I 33times 10minus1 24times 10minus1 103 047 061 725 083 111 801 nd nd ndQCLAS Ilowast 14times 10minus1 10times 10minus1 12times 10minus3 019 023 061 020 022 011 nd nd ndQCLAS II 20times 10minus1 24times 10minus1 411 052 049 004 022 019 43times 10minus3 nd nd ndQCLAS III 10times 10minus1 16times 100 161 081 137 006 072 118 003 038 054 005

sim 10000 ppb N2O

OA-ICOS I 17times 100 13times 100 130 012 010 012 015 009 054 017 012 035QCLAS I 33times 100 23times 100 374 006 007 109 009 011 082 nd nd ndQCLAS Ilowast 46times 10minus1 38times 10minus1 010 003 003 002 002 003 010 nd nd ndQCLAS II 12times 100 99times 10minus1 351 009 007 29times 10minus3 009 008 70times 10minus3 nd nd ndQCLAS III 13times 100 16times 100 661 010 017 34times 10minus3 010 013 58times 10minus3 048 065 28times 10minus3

lowast Data were reprocessed by Aerodyne Research Inc technicians using an automatic spectral correction method This method corrects data that were influenced by changing baseline structureFurther information on this method is provided in Sect S5 ldquondrdquo indicates not determined

Figure 3 Allan deviation (square root of Allan variance) plots for the OA-ICOS I (blue) CRDS I (red) CRDS II (black) QCLAS I (green)QCLAS II (purple) and QCLAS III (brown) at different N2O mole fractions (sim 327 1000 and 10 000 ppb) The dashed lines represent a slopeof minus05 (logndashlog scale) and indicate the expected behavior for Gaussian white noise in each analyzer The Allan deviations of all analyzerstested were reproducible on three separate occasions prior to the test results presented here Allan deviation plots for δ15Nβ and δ18O areprovided in Sect S4 (Fig S4-1 in the Supplement)



Figure 4 Examples of the dependency of different measurements on laboratory temperature (C) for OA-ICOS I (blue) CRDS I (red)CRDS II (black) and QCLAS I (green) The complete dataset is provided in Sect S4 (Fig S4-2) The laboratory temperature is indicated bya solid orange line and was allowed to vary over time Cell temperatures for each instrument are also plotted for comparison The analyzersbegan acquiring measurements at 0000 CEST on 8 July 2018 capturing the end of the rising limb of the laboratory temperature Results areplotted as the deviation from the mean without any anchoring to reference gases

Table 7 Summary of the measured [N2O] δ15Nα δ15Nβ and δ18O and associated 1σ at 300 s averaging times based on repeated measure-ments of PA1

Instrument n N2O 1σ N2O δ15Nα 1σ δ15Nα δ15Nβ 1σ δ15Nβ δ18O 1σ δ18O(ppb) (ppb) (permil) (permil) (permil) (permil) (permil) (permil)

CRDS I 22 32666 030 1586 079 minus230 083 4448 081CRDS II 22 32672 026 1571 052 minus286 064 4440 075OA-ICOS I 22 32649 007 1529 147 minus211 119 4401 217QCLAS I 22 32682 016 1392 535 minus297 857 ndash ndashTREX-QCLAS I 28 32670 129 1572 060 minus282 037 4431 046Empa-assigned values 3 32651 006 1581 007 minus331 0004 4472 004

bility (1σ ) of the Anchor gas measurements TREX-QCLASI measurements were not impaired by gas matrix effects

352 Continuity of gas matrix corrections at higherN2O mole fractions

When mole fractions of 660 and 990 ppb N2O were mea-sured by the laser spectrometers O2 interference effects on

[N2O] and δ values were well described using linear re-gression albeit with different slopes to those obtained for330 ppb N2O (Figs 6 and S4-4 Sect S8)

We could not adequately predict the nature in which theslopes of the interference effects scaled with N2O mole frac-tions Overall this suggests that interference effects wereanalyzer-specific and varied according to instrument-specific



Figure 5 Deviations of the measured δ15Nα δ15Nβ and δ18O values according to 1[N2O] for the OA-ICOS I (blue) CRDS I (red) CRDSII (black) and QCLAS I (green) Measurements span the manufacturer-specified operational ranges of the analyzers The experiment wasrepeated on three separate days A linear regression is indicated by the solid line and a residual plot is provided above each plot Individuallinear equations coefficients of determination (r2) and p values are indicated above each plot The remaining plots for δ15Nβ and δ18O areprovided in Sect S4 (Fig S4-3)

parameters rather than due to bona fide scaling of thepressure-broadening effect Therefore to account for com-bined effects of [O2] and [N2O] changes on measurementsa user would be required to perform a series of laboratorytests across the range of expected [O2] and [N2O] In an ex-emplary approach we applied a series of empirical equations(Eqs 5ndash6) to predict the offset of measured [N2O] and δ val-ues caused by changes in [O2] as a function of [N2O] intro-duced to the analyzers in this study

1[N2O]measmix(1[O2]A [N2O]expmix

)=

(A middot [N2O]2

expmix+B middot [N2O]expmix

)middot1[O2]A (5)

1δmeasmix(1[O2]A [N2O]expmix

)=

(a middot [N2O]2

expmix+ b middot [N2O]expmix+ c)middot1[O2]A (6)

where1[N2O]measmix and1δmeasmix are the measured off-sets on [N2O] and δ values for the gas mixtures introduced tothe analyzers as reported in Sect 351 respectively1[O2]A

is the difference in O2 mole fraction between the gas mixtureand Anchor gas as reported in Sect 351 [N2O]expmix is theexpected [N2O] of gas mixtures introduced to the analyzercalculated based on gas flows and cylinder compositions ofGases 1 2 and 3 as reported in Sect 245 A and B and ab and c are analyzer-specific constants

Using Eqs (5) and (6) to fit values for the constants A andB for1[N2O]measmix and a b and c for1δmeasmix resultedin a total of 11 analyzer-specific values (Sect S8) With gas-specific constants established interferences on [N2O] and δmeasurements for a sample gas G for a given analyzer can becorrected using Eqs (7)ndash(8)

[N2O]mcG

=

minus(1+B middot1[O2]G)+

radic(1+B middot1[O2]G)

2+ 4 middotA middot1[O2]G middot [N2O]measG

2 middotA middot1[O2]G(7)

δmcG = δmeasG

minus

(a middot [N2O]mcG

2+ b middot [N2O]mcG+ c

)middot1[O2]G (8)



Figure 6 Deviations of the measured [N2O] δ15Nα δ15Nβ and δ18O values according to 1O2 () at different N2O mole fractions (330660 and 990 ppb) for the OA-ICOS I (blue) CRDS I (red) CRDS II (black) QCLAS I (green) and TREX-QCLAS I (brown) The remainingplots for [N2O] δ15Nα and δ18O are provided in Sect S4 (Fig S4-4) The standard deviation of the Anchor gas (plusmn1σ ) is indicated by dashedlines Data points represent the mean and standard deviation (1σ ) of triplicate measurements Dependencies are best described using linearregressions which are indicated by a solid line Individual equations coefficients of determination (r2) and p values are indicated aboveeach plot for the 330 ppb N2O data only

where [N2O]mcG and δmcG are the matrix-corrected [N2O]and δ values of sample gas G respectively1[O2]G is the dif-ference in O2 mole fraction between sample gas G and refer-ence gases Correction using Eqs (7)ndash(8) removes the O2 ef-fect to a degree that corrected measurements from Sect 351are typically within the uncertainty bounds of the anchor(Sect S8)

Although Ar effects seemingly scaled with increased N2Omole fractions we did not derive scaling coefficients for Arbecause the derived Ar correction equations at 330 660 and990 ppb N2O were typically not statistically significant atp lt 005 These interferences also did not always exceedthe repeatability of Anchor gas measurements Although wecould have tested for effects for [Ar] changes greater than095 we limited our experiments to [Ar] expected in tro-pospheric samples

36 Trace gas effects (H2O CO2 CH4 and CO)

361 Trace gas effects at ambient N2O mole fractions

The apparent offset of [N2O] and δ values resulting from thechange in CO2 composition of the matrix gas was best de-scribed by linear functions (Figs 7 and S4-6) OA-ICOS Iexhibited discrete and well-defined linear interference effectsof CO2 on [N2O] δ15Nα δ15Nβ and δ18O (all r2 gt 095)likely due to crosstalk between CO2 absorption lines situatednear 219246 and 219233 cmminus1 Both CRDS instrumentsshowed CO2 interference effects of smaller magnitude for[N2O] δ15Nα and δ18O presumably due to CO2 absorption

lines at 219621 219572 and 219602 cmminus1 QCLAS I dis-played less well-defined CO2 interference effects for δ15Nβ which was possibly due to several overlapping absorptionlines of CO2 located near 218785 cmminus1 All linear functionsderived for the TREX-QCLAS I were not statistically signif-icant at p lt 005 As shown in Figs 7 and S4-6 the NaOHtrap was effective in removing the CO2 effect (if present)across the mole fraction ranges tested for all instruments

Similarly CH4 effects on apparent [N2O] and δ valueswere well described by linear functions (Figs 8 and S4-7)The largest effects were for CRDS I and II which both dis-played strong CH4 dependencies for δ15Nα and δ18O of sim-ilar magnitude This might be due to crosstalk of 14N15N16Oand 14N14N18O absorption lines with the respective CH4lines located at 219576 and 219595 cmminus1 For OA-ICOS Iminor CH4 effects were observed for δ15Nβ due to absorp-tion line overlap at 219233 cmminus1 QCLAS I did not displayany CH4 interference effect over the tested [CH4] range Lin-ear functions derived for the TREX-QCLAS I were not sta-tistically significant at p lt 005 The similarity between the[N2O] dependencies on CH4 mole fractions for OA-ICOS ICRDS I II and QCLAS I suggests that the apparent effectsmay be due to small fluctuations in the gas mixtures producedby the MFCs rather than a discrete spectral interference ef-fect

The CRDS analyzers showed minor interference effectsfor δ15Nα and δ15Nβ on [CO] (014ndash214 ppm) (Fig S4-8)likely due to crosstalk with CO absorption lines located at219569 and 219583 cmminus1 The magnitude of these effects



was similar for both models QCLAS I displayed interferenceeffects for δ15Nα and δ15Nβ caused by a CO absorption linelocated near 21879 cmminus1 although this effect did not exceedthe repeatability of the Anchor gas (containing no CO) overthe measurement range The effects of [CO] on δ values ac-quired using OA-ICOS I and TREX-QCLAS I were not sta-tistically significant at p lt 005 Similar to CH4 the resem-blance of [CO] effects to [N2O]measurements for OA-ICOSI CRDS I II and QCLAS I suggests that the apparent effectsmay be due to inaccuracies in the dynamic dilution processrather than a discrete spectral interference effect The Sofno-cat trap was effective in removing CO (if present) across themole fraction ranges tested for all instruments

OA-ICOS I exhibited large effects of [H2O] (0ndash13 800 ppm) on δ15Nβ (up to minus10 permil) and δ18O (up tominus15 permil) and minor dependencies for δ15Nα (up to 4 permil) and[N2O] (up to 1 ppb) across the range tested (Fig S4-9) ForQCLAS I the H2O effect was largest for δ15Nα (up to 20 permil)whilst minor effects for [N2O] (up to 2 ppb) were observed inrelation to the Anchor gas (no H2O) In contrast both CRDSinstruments showed no significant effects across the rangetested which is attributable to the installation of permeationdryers inside the analyzers by the manufacturer

362 Continuity of trace gas corrections at higher N2Omole fractions

Interference effects from CO2 CH4 and CO on apparent δvalues where significant inversely scaled with increasing[N2O] (Figs 7 8 S4-8 and Sect S8) The scaling of tracegas effects can be explained by simple spectral overlap of the14N15N16O 15N14N16O and 14N14N18O lines with those ofthe trace gas which results in the interference effects beinginversely proportional to the mixing ratio of N2O Howeverthere may be additional spectral overlap between the tracegas and the 14N14N16O peak resulting in an offset for themeasured [N2O] which introduces a further shift in the δ val-ues (as shown in Sect 34) The effect on the apparent [N2O]was less clear and was possibly confounded by inaccuraciesduring dynamic gas mixing In this study the scaling of in-terference effects from trace gases as a function of the [N2O]introduced to the analyzers could be described using Eqs (9)and (10)

1[N2O]measmix(1[x]A [N2O]expmix

)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)

1δmeasmix(1[x]A [N2O]expmix

)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)

where1[N2O]measmix and1δmeasmix are the measured off-sets on [N2O] and δ values for the gas mixtures introduced tothe analyzers as reported in Sect 361 respectively 1[x]Ais the difference in trace gas mole fraction between the gas

mixture and Anchor gas as reported in Sect 361 and Ax Bx ax and bx are constants that are trace gas and instrumentspecific The constant bx only occurs when there is spectraloverlap from the trace gas and 14N14N16O absorption lines

For a sample gas G the effect can then be corrected byusing Eqs (11) and (12)


minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)

In Eqs (11)ndash(12) the sum of the effect of all interfering gaseswith overlapping absorption lines is taken into account Sim-ilar to Sect 352 correction using Eqs (11)ndash(12) removesthe trace gas interference effects to the extent that correctedmeasurements from Sect 361 are within the repeatabilitybounds of the Anchor gas (Sect S8) Similar inverse rela-tionships have been described by Malowany et al (2015) forH2S interferences on δ13CminusCO2


Results for the two-end-member mixing experiment wereevaluated in two different ways First results for individualgas mixtures acquired by laser spectroscopy and GC-IRMSwere compared to expected [N2O] and δ values calculatedfrom N2O mole fractions and isotopic composition of end-members and mixing fractions Second source values wereextrapolated using a weighted total least squares regressionanalysis known as Keeling plot analysis (Keeling 1958)and compared to assigned δ values of the source gas usedin each experiment

371 Comparison with expected [N2O] and δ values

Triplicate measurements (meanplusmn1σ ) obtained using thelaser spectrometers and GC-IRMS were plotted against ex-pected [N2O] and δ values calculated using MFC flow ratesN2O mole fractions and isotopic composition of backgroundand source gases (Figs 12ndash15) Comparisons between indi-vidual laser spectrometer measurements and GC-IRMS areplotted in Sect S9 and are discussed only briefly below

OA-ICOS I

Generally there was good agreement of [N2O] between theOA-ICOS I and expected values although the analyzer over-estimated mole fractions at higher 1N2O during Exps 5and 6 There was excellent agreement between the OA-ICOSI and calculated expected δ values (all r2 gt 095 Figs 9and S4-10) Measurements for δ15Nα were mostly withinplusmn24 permil of expected values while δ15Nβ δ15Nbulk and SP



Figure 7 Deviations of the measured [N2O] δ15Nα δ15Nβ and δ18O values according to 1CO2 (ppm) at different N2O mole fractions(330 660 and 990 ppb) for the OA-ICOS I (blue) CRDS I (red) CRDS II (black) QCLAS I (green) and TREX-QCLAS I (brown) Theremaining plots for [N2O] δ15Nα and δ18O are provided in Sect S4 (Fig S4-6) The standard deviation of the Anchor gas (plusmn1σ ) is indicatedby dashed lines Data points represent the mean and standard deviation (1σ ) of triplicate measurements Dependencies are best described bylinear fits which are indicated by solid lines Individual equations coefficients of determination (r2) and p values are indicated above eachplot for the 330 ppb N2O data only

Figure 8 Deviations of the measured [N2O] δ15Nα δ15Nβ and δ18O values according to 1CH4 (ppm) at different N2O mole fractions(330 660 and 990 ppb) for the OA-ICOS I (blue) CRDS I (red) CRDS II (black) QCLAS I (green) and TREX-QCLAS I (brown) Theremaining plots for [N2O] δ15Nβ and δ18O are provided in Sect S4 (Fig S4-7) Data points represent the mean and standard deviation(1σ ) of triplicate measurements Dependencies are best described by linear fits which are indicated by solid lines Individual equationscoefficients of determination (r2) and p values are indicated above each plot for the 330 ppb N2O data only



were all within plusmn2 permil of expected values δ18O measure-ments were the poorest performing but were typically withinplusmn36 permil of expected values Similarly there was excellentagreement between OA-ICOS I and IRMS isotope values (allr2 gt 095) which agreed within 17 permilndash24 permil (Fig S9-2)The standard deviations of triplicate isotope measurementsdecreased dramatically with increasing 1N2O improvingfrom 1 permil to 2 permil during Exps 1 and 2 to typically betterthan 01 permil during Exps 5 and 6 Conversely the standarddeviations of triplicate sample measurements for [N2O] in-creased with increasing1N2O rising fromlt 01 ppb duringExps 1ndash4 to gt 1 ppb during Exps 5 and 6 Nonetheless allOA-ICOS I [N2O] measurements had better 1σ repeatabilitythan those acquired using GC The repeatability of the trip-licate isotope measurements with OA-ICOS I was typicallybetter than IRMS exclusively at higher 1N2O (gt 700 ppb)

CRDS I

[N2O] acquired by CRDS I were in good agreement with ex-pected values although the analyzer slightly underestimatedmole fractions at higher 1N2O during Exps 3 and 4 Therewas excellent agreement between the CRDS I and calcu-lated expected isotope values (all r2 gt 095 Figs 10 andS4-11) Measurements for δ15Nα and δ15Nβ were mostlybetter than plusmn11 permil of expected values while δ15Nbulk waswithin plusmn05 permil of expected values SP and δ18O measure-ments were typically within plusmn15 permil of expected valuesThere was excellent agreement between CRDS I and IRMSisotope values (all r2 gt 093) which agreed to within 05 permilndash19 permil (Fig S9-3) In general the standard deviations oftriplicate isotope measurements increased as a function of1N2O with the lowest deviations of 01 permilndash1 permil occurringwhen 1N2Olt 100 ppb However two triplicated measure-ments for δ15Nbulk had standard deviations better than 01 permilThe standard deviations of triplicate measurements for [N2O]also increased with increasing 1N2O mole fractions ris-ing from 003ndash007 ppb when 1N2O=sim 0 ppb (ie ambi-ent conditions) to sim 1 ppb when 1N2O=sim 700 ppb Withthe exception of one triplicate measurement all CRDS I[N2O] measurements had better 1σ repeatability than thoseacquired using GC Overall IRMS had slightly better re-peatability (most ranging from 01 permil to 1 permil) than CRDSI (most ranging from 01 permil to 2 permil) for isotopic measure-ments

CRDS II

Similar to results for CRDS I [N2O] acquired by CRDS IIwere in good agreement with expected values but slightly un-derestimated mole fractions at higher 1N2O during Exps 3and 4 There was excellent agreement between the CRDSII and calculated expected isotope values (all r2 gt 099Figs 11 and S4-12) Measurements for δ15Nα and SP weretypically better than plusmn08 permil of expected values while

δ15Nβ δ15Nbulk measurements were all within plusmn04 permil ofexpected values δ18O measurements were within plusmn10 permilof expected values There was excellent agreement betweenCRDS II and IRMS isotope values (all r2 gt 098) whichagreed within plusmn06 permilndash14 permil (Fig S9-4) The standard de-viations of triplicate isotope measurements typically de-creased as a function of1N2O with the lowest deviations oflt 01 permilndash03 permil occurring when 1N2O=sim 700 ppb Con-versely the standard deviations of triplicate sample mea-surements for [N2O] increased with increasing 1N2O ris-ing from 004ndash009 ppb when 1N2O=sim 0 ppb (ie ambi-ent conditions) to sim 04 ppb when 1N2O=sim 700 ppb AllCRDS II [N2O] measurements had better 1σ repeatabilitythan those acquired using GC There was no clear distinc-tion between CRDS II and IRMS triplicate repeatability withboth achieving triplicate repeatability ranging from 01 permil to1 permil for most isotopic measurements However the repeata-bility of SP CRDS II measurements was mostly better thanIRMS achieving triplicate repeatability between 01 permil and06 permil compared to 02 permilndash1 permil for IRMS

QCLAS I

There was good agreement of [N2O] between QCLAS I andexpected values however the analyzer underestimated molefractions at higher 1N2O during Exps 5 and 6 Unfortu-nately it is clear from the large spread of isotope valuesdepicted in Fig 12 that the standardized calibration schemeselected for the two-end-member mixing tests was insuffi-cient for acquiring accurate and precise isotopic measure-ments using QCLAS I For this reason we urge researchersnot to overinterpret such results as the implementation of aQCLAS-specific calibration procedure (in line with resultsfrom Sect 31 and 33) would improve results dramaticallyNonetheless QCLAS I obtained accurate results at higherN2O mole fractions (indicated in red in Figs 12 and S4-13)such that when 1N2Olt 700 ppb measurements were ex-cluded δ15Nα δ15Nβ δ15Nbulk and SP were withinplusmn30 permil14 permil 14 permil and 38 permil of calculated expected values re-spectively Similarly QCLAS I showed good agreement withIRMS only at higher 1N2O (gt 700 ppb Fig S9-5) Similarto OA-ICOS I the standard deviations of QCLAS I triplicateisotope measurements decreased dramatically with increas-ing 1N2O improving from sim 10 permil during Exps 1 and 2to typically between 01 permil and 1 permil during Exps 5 and 6Conversely the standard deviations of triplicate sample mea-surements for [N2O] increased with increasing1N2O risingfrom lt 01 ppb during Exps 1ndash4 to gt 1 ppb during Exps 5and 6 All QCLAS I [N2O] measurements had better 1σ re-peatability than those acquired using GC QCLAS I had trip-licate isotope measurement standard deviations comparableto IRMS only at higher 1N2O (gt 700 ppb)



Figure 9 Correlation diagrams for δ15Nbulk and SP measurements at various 1N2O mole fractions analyzed by OA-ICOS I plotted againstexpected values The remaining plots for [N2O] δ15Nα δ15Nβ and δ18O are provided in Sect S4 (Fig S4-10) The solid black line denotesthe 1 1 line while the dotted line indicates plusmn1σ of the residuals from the 1 1 line The dashed blue line represents a linear fit to the dataIndividual equations coefficients of determination (r2) and p values are indicated above each plot Each data point represents the meanand standard deviation (1σ ) of triplicate measurements The inset plots indicate the standard deviation (1σ ) of the triplicate measurementsachieved at different 1N2O mole fractions and the 1 1 line is similarly a solid line

Figure 10 Correlation diagrams for δ15Nbulk and SP measurements at various 1N2O mole fractions analyzed by CRDS I plotted againstexpected values The remaining plots for [N2O] δ15Nα δ15Nβ and δ18O are provided in Sect S4 (Fig S4-11) The solid black line denotesthe 1 1 line while the dotted line indicates plusmn1σ of the residuals from the 1 1 line The dashed red line represents a linear fit to the dataIndividual equations coefficients of determination (r2) and p values are indicated above each plot Each data point represents the meanand standard deviation (1σ ) of triplicate measurements The inset plots indicate the standard deviation (1σ ) of the triplicate measurementsachieved at different 1N2O mole fractions and the 1 1 line is similarly a solid line

TREX-QCLAS I

There was good agreement of N2O mixing ratios betweenthe TREX-QCLAS I and expected values Similarly therewas excellent agreement between the TREX-QCLAS I andcalculated expected isotope values (all r2 gt 097 Figs 13and S4-14) Measurements for δ15Nα δ15Nβ δ15Nbulk andSP were within plusmn029 permil 032 permil 023 permil and 041 permil ofexpected values respectively δ18O measurements were typ-ically within plusmn024 permil of expected values Generally thestandard deviations of triplicate isotope measurements de-

creased with increasing 1N2O improving from typically02 permilndash03 permil at low1N2O mole fractions (ambient) to closeto or better than 01 permil when 1N2O reached 30 ppb Con-versely the standard deviations of triplicate sample mea-surements for [N2O] increased with increasing 1N2O ris-ing from lt 03 to sim 1 ppb No comparison could be madebetween TREX-QCLAS I and IRMS measurements becauseTREX-QCLAS measurements were undertaken separatelyfrom the other instruments



Figure 11 Correlation diagrams for δ15Nbulk and SP measurements at various 1N2O mole fractions analyzed by CRDS II plotted againstexpected values The remaining plots for [N2O] δ15Nα δ15Nβ and δ18O are provided in Sect S4 (Fig S4-12) The solid black line denotesthe 1 1 line while the dotted line indicates plusmn1σ of the residuals from the 1 1 line The dashed black line represents a linear fit to thedata Individual equations coefficients of determination (r2) and p values are indicated above each plot Each data point represents the meanand standard deviation (1σ ) of triplicate measurements The inset plots indicate the standard deviation (1σ ) of the triplicate measurementsachieved at different 1N2O mole fractions and the 1 1 line is similarly a solid line

Figure 12 Correlation diagrams for δ15Nbulk and SP measurements at various 1N2O mole fractions analyzed by QCLAS I plotted againstexpected values The remaining plots for [N2O] δ15Nα δ15Nβ and δ18O are provided in Sect S4 (Fig S4-13) The solid black line denotesthe 1 1 line while the dotted line indicates plusmn1σ of the residuals from the 1 1 line The dashed green line represents a linear fit to thedata Individual equations coefficients of determination (r2) and p values are indicated above each plot Each data point represents the meanand standard deviation (1σ ) of triplicate measurements The inset plots indicate the standard deviation (1σ ) of the triplicate measurementsachieved at different 1N2O mole fractions and the 1 1 line is similarly a solid line Results for Exps 5ndash6 are highlighted in red with thedashed red line indicating a linear fit to this data

372 Source identification using Keeling analysis

Despite the excellent agreement between expected and mea-sured values across all experiments for OA-ICOS I CRDSI and II and TREX-QCLAS I the extrapolated source in-tercept values acquired using Keeling analysis showed largestandard errors especially for Exps 1 and 2 (1N2O=30 ppb) (Figs 14 and S4-15 Sect S10) This was mostlydue to the small mole fraction range (ie large inverse molefraction range) over which the regression line was extrapo-

lated in order to acquire the intercept value The cause ofthe erroneous intercepts values was likely two-fold (1) theextrapolated source was highly susceptible to measurementsacquired at background levels and due to the inherent greateruncertainty associated with measurements acquired at ambi-ent N2O mole fractions intercepts can be skewed accord-ingly and (2) any further non-linearity that could not betaken into account in the three-point concentration depen-dence correction applied Overall this implies that N2O iso-tope source studies using laser spectroscopy focusing on



Figure 13 Correlation diagrams for δ15Nbulk and SP measurements at various1N2O mole fractions analyzed by OA-ICOS I plotted againstexpected values The remaining plots for [N2O] δ15Nα δ15Nβ and δ18O are provided in Sect S4 (Fig S4-14) The solid black line denotesthe 1 1 line while the dotted line indicates plusmn1σ of the residuals from the 1 1 line The dashed green line represents a linear fit to thedata Individual equations coefficients of determination (r2) and p values are indicated above each plot Each data point represents the meanand standard deviation (1σ ) of triplicate measurements The inset plots indicate the standard deviation (1σ ) of the triplicate measurementsachieved at different 1N2O mole fractions and the 1 1 line is similarly a solid line

near-ambient N2O variations remain a challenging undertak-ing and one should expect large uncertainty in source esti-mates over small mole fraction changes

For Exps 3ndash6 however the accuracy of the source in-tercept and its standard error improved dramatically for allanalyzers on account of the decreasing uncertainty in mea-surement OA-ICOS I and both CRDS analyzers typicallyachieved withinplusmn2 permilndash5 permil of the assigned values for δ15Nα δ15Nβ δ15Nbulk SP and δ18O and had performance com-parable to or better than the GC-IRMS approach (Figs 14and S4-15) Similarly the standard error of all intercepts de-creased dramatically for Exps 3ndash6 and all analyzers typi-cally achieved better than plusmn1 permil standard error on derivedintercepts in Exps 5 and 6

4 Discussion

41 Factors affecting the precision and accuracy of N2Oisotopocule measurements using laser spectroscopy

A summary of results is presented in Table 8 Our resultshighlight that the precision at which laser-based analyzersacquire N2O isotopocule measurements is a function of N2Omole fraction the selected measuring and averaging timesand calibration frequency according to measurement stabil-ity The degree of accuracy obtained using different laserspectrometers is ultimately a function of the robustness ofcorrections aimed at removing matrix and trace gas effectsand the selected calibration procedure aimed at standardizingthe data to international scales

All spectrometers tested displayed temperature effects onisotope measurements which can be attributed to differencesin the lower state energies of the probed N2O isotopocule

lines (Sect S11) (eg Waumlchter et al 2008) The temper-ature sensitivities of all analyzers tested necessitates thatespecially when deployed in the field they be operated un-der temperature-controlled conditions (such as in maintainedfield stations)

The experiments performed in this study were undertakenusing a standardized protocol Calibration was performed onisotope δ values derived from raw uncalibrated isotopoculeamount fractions thus requiring [N2O] dependence cor-rections Alternative approaches aimed at calibrating iso-topocule amount fractions prior to deriving δ values werenot included in our study but have the potential to removethe need for this correction (eg Wen et al 2013 Flores etal 2017 Griffith 2018) if appropriate reference materialsbecome available Isotopocule calibration approaches wouldrequire a set of N2O standard gases with high-accuracy molefractions in addition to assigned δ values

All analyzers tested in this study showed significant effectsfrom changing O2 composition of the gas matrix Althoughthe magnitude of this effect ultimately varied across the an-alyzers and was dependent on N2O mixing ratios the effectof a change in O2 composition of 205 was typically onthe order of 10 permil to 30 permil for δ values Similar O2 depen-dencies have been reported by Erler et al (2015) for CRDSN2O isotope laser spectrometers as well as for CRDS H2Oisotope analyzers (Johnson and Rella 2017) The underly-ing reason for this effect is differences in N2 versus O2 (andAr) broadening parameters of the probed N2O isotopoculelines In short the N2 O2 (and Ar) broadening parametersdepend on rotational quantum numbers of the respective N2Olines (Henry et al 1985 Sect S11) Thus differences in therotational quantum numbers for a pair of isotopocules (eg14N15N16O14N14N16O) relate to a difference in their N2 O2



Figure 14 1δ15Nbulk and 1SP(EstimatedSourceminusTrueSource) values derived from the OA-ICOS I (blue) CRDS I (red) CRDS II (black)QCLAS I (green) and IRMS (purple) via Keeling analysis of the two-end-member mixing scenario The remaining plots for δ15Nα δ15Nβ

and δ18O are provided in Sect S4 (Fig S4-15) EstimatedSource = TrueSource is indicated by a solid black line at y = 0 and the dottedlines indicated plusmn2 permil deviation from y = 0 The change in concentration exceeding that of the background gas is indicated for Exps 1ndash2(1N2O=sim 30 ppb) 3ndash4 (1N2O=sim 700 ppb) and 5ndash6 (1N2O=sim 10000 ppb) Note that the QCLAS I results for Exps 1 and 2 are notdepicted to maintain clarity as they exceed the selected y-axis scale

Table 8 Summary of main findings presented in this study

Detection scheme (model OA-ICOS I CRDS I amp II QCLAS I TREX-QCLAS Imanufacturer) (N2OIA-30e-EP) (G5131-i) (CW-QC-TILDAS-SC-D) (CW-QC-TILDAS-76-CS)

Allan precision (300 s)δ15Nα δ15Nβ δ18O [permil]

3265 ppb N2O 079ndash169 032ndash046 039ndash345a ndsim 1000 ppb N2O 028ndash067 021ndash089 019ndash083a ndsim 10000 ppb N2O 012ndash017 nd 002ndash048a nd

Repeatability (3265 ppb N2O)N2O [ppb] 007 026ndash030 016 129δ15Nα δ15Nβ δ18O [permil] 119ndash217 052ndash083 535ndash857 037ndash060

Temperature effect (3265 ppb N2O)N2O [ppb Kminus1] 001 002 010 ndδ15Nα δ15Nβ δ18O [permil Kminus1] 036ndash260 025ndash065 3129ndash3732 nd

N2O mole fraction dependenceδ15Nα δ15Nβ δ18O [permil ppb (11N2O)] minus8296 to 2544 minus458 to 1353 minus66386 to 15 833 nd

O2 matrix effect (330 ppb N2O)N2O [ppb minus1 (1O2)] minus0044 024ndash0305 0351 nsδ15Nα δ15Nβ δ18O [permil minus1 (1O2)] 0874ndash1270 minus0279 to (minus1364) minus1111 ns

CO2 trace gas effects (330 ppb N2O)N2O [ppb ppmminus1 (1CO2)] 00011 00005 ns nsδ15Nα δ15Nβ δ18O [permil ppmminus1 (1CO2)] minus0009 to 0026 ns to (minus00019) ns to 00154 ns

CH4 trace gas effects (330 ppb N2O)N2O [ppb ppmminus1 (1CH4)] nsb minus0039 to (minus0056) nsb nsb

δ15Nα δ15Nβ δ18O [permil ppmminus1 (1CH4)] 0173 0085ndash250 ns ns

CO trace gas effects (330 ppb N2O)N2O [ppb ppmminus1 (1CO)] minus029 minus015 to (minus024) minus019 nsδ15Nα δ15Nβ δ18O [permil ppmminus1 (1CO)] ns minus053 to (minus241) ns to (minus404) ns

a Includes QCLAS I II and III b Likely due to inaccuracies during dynamic dilution (see text for details) nd not determined ns not statistically significant at p lt 005 andor r2 lt 05



and Ar broadening parameters Consequently differences inthe O2 or Ar content of the sample gas matrix and that ofthe reference gas affect measured isotope ratios and lead tochanges in apparent δ values Nonetheless the magnitude ofeffects reported for the CRDS analyzers in this study variedacross CRDS I (a 2015 model) and CRDS II (a 2018 model)as well as from those reported by Erler et al (2015) There-fore we recommended that in applications where O2 concen-trations vary such as groundwater estuaries stratified wa-terbodies and incubation studies researchers test individualanalyzers for their specific dependencies to allow for correc-tion This is especially important given that N2O productionand reduction processes in such environments are stronglycontrolled by O2 availability Although the Ar effects charac-terized in this study were not large it is nonetheless recom-mended as a precautionary measure that researchers ensurewhere possible the standard calibration gas Ar compositionis similar to that of the sample gas

The CO2 effects for OA-ICOS and CH4 effects for CRDSanalyzers must be considered for applications of these ana-lyzers where CO2 and CH4 may also co-vary such as duringdiel atmospheric monitoring in soil-flux chamber measure-ments incubation studies and even waterbodies (eg Erleret al 2015) These effects need to be either characterizedand corrected for by the user or the interfering gas quan-titatively removed To our knowledge there is currently nocommercially available technique to remove CH4 from a gasstream without affecting N2O and therefore independent co-analysis of CH4 is ultimately required to correct for these ef-fects post-measurement Similarly while water vapor effectscan in theory be characterized and corrected for all instru-ments we recommend that researchers remove water vaporfrom the gas stream prior to analysis Although not testedhere other studies have highlighted possible spectral inter-ference effects associated with elevated H2S and volatile or-ganic compounds (Erler et al 2015 Ostrom and Ostrom2017) but these may also be removed from gas streams us-ing chemical traps (eg Cu and activated carbon traps re-spectively)

The scaling of gas matrix and trace gas effects with [N2O]has important implications for any measurement setup thatrelies on post-measurement correction equations An equa-tion developed to correct for CH4 effects that was derivedusing a [N2O] of 330 ppb should not be implemented for asample gas containing 990 ppb N2O For example as shownin Fig 8 the measured interference effect on δ15Nα measure-ments acquired using CRDS II for 10 ppm [CH4] at 330 ppbN2O was 249 permil while at 990 ppb N2O it was 81 permil re-sulting in a 168 permil difference The scaling of interferenceeffects from trace gases has been reported previously forCO2CH4 laser spectrometers (Assan et al 2017 Malowanyet al 2015) This underlines the usefulness of removingH2O CO2 and CO with scrubbers prior to measurement asthis removes the need for correction equations to begin withand the scaling of corrections that can ensue We are unaware

of any studies that have shown that O2 interferences causedby pressure-broadening linewidth effects change as a func-tion of N2O mole fraction While we were unable to describethe scaling of the O2 effect sufficiently using correction func-tions based on theoretical deductions empirical equationsbased on experimental testing such as those developed inSect 352 and 362 could be implemented by researcherswhen covariation in both O2 and N2O in the sample gas isexpected Alternatively as shown in this study matrix andortrace gas effects can be removed by automated N2O precon-centration devices such as TREX (Ibraim et al 2018 Mohnet al 2010) similar to IRMS However such devices are notcommercially available complex to build and operate andrestrict sample frequency

42 Pre-measurement considerations

Our study clearly shows that knowledgeestimation of thematrix and trace gas composition of both reference standardsand sample gases and the differences between them are crit-ical for accurate N2O isotopocule analysis using laser spec-troscopy We acknowledge however that this may be diffi-cult to predict in certain applications without prior testing ofthe sample gas and therefore researchers should err on theside of caution

As a prerequisite to acquiring measurements using N2Oisotope laser spectrometers researchers will be required toconsider the accessory gas mixtures required to characterizetheir instrument For applications with significant variationsin matrix (O2 Ar) or trace gas (CO2 CH4 CO) composi-tions researchers will require gas mixtures containing thegas of interest in order to characterize the associated inter-ference effects for their laser spectrometer This also necessi-tates that appropriate interference detectors are implementedespecially O2 and CH4 analyzers given that these effects can-not be mitigated using chemical traps

In this study interference effects and the associated scal-ing of these effects according to the co-measured N2O molefraction were derived via dynamic dilution with various gasmixtures using MFCs This allowed for the introduction ofa wide range of gas mixtures to the analyzers for interfer-ence testing and consequently only a small amount of gasmixtures were required for all of the experiments outlined inSect 24 In comparison a much larger number of individualgas mixtures would have been required had they been pre-pared using static dilution techniques (see Erler et al 2015)The scaling of interference effects was sufficiently distin-guished by undertaking testing at three different N2O molefractions (N2O= 330 660 and 990 ppb) and we thereforerecommend this as a minimum criterion for researchers wish-ing to characterize this effect

Researchers should also consider the sample gas volumerequired for a given measurement application using a specificlaser spectrometer In our experience ensuring that five lasercavity cell volumes have been flushed prior to measurement



is best practice to negate any memory effects when these in-struments are operated using continuous flow-through con-figurations (as opposed to discrete sample measurements ina closed laser cavity) By following this procedure and usingthe operating parameters selected in this study (Table 1) thesample gas volume required for a single 300 s measurementis approximately 80 mL for CRDS II 150 mL for CRDS I600 mL for OA-ICOS I and 1200 mL for QCLAS I By com-parison TREX-QCLAS I requires approximately 5 L of sam-ple gas to allow for N2O preconcentration These samplegas volumes represent typical numbers for atmospheric ap-plications however instrument parameter settings such asflow rate and cell pressure which ultimately change the re-quired sample volume can be optimized depending on themeasurement application This is particularly the case forQCLAS instruments which can be operated with differentuser-adjustable settings For applications requiring discretesample analysis (eg the headspace analysis of δ15N andδ18O in N2O derived from dissolved NOminus3 ) high N2O con-centration gas samples with lower volumes can be introducedto these instruments using injection ports and dilution gases(eg Soto et al 2015 Wassenaar et al 2018) however wedid not test these capabilities in our study

43 Measurement workflow

In line with our results we propose a step-by-step work-flow that can be followed by researchers to acquire N2Oisotopocule measurements (Fig 15) This workflow seeks tocover all sources of potential error tested in our study Notall steps will be applicable because interference effects varyacross analyzers For QCLAS analyzers which offer highversatility interference effects can also be approached bymulti-line analysis inclusion of interfering spectral lines oradaption of pressure-broadening parameters in the spectralfitting algorithm For specific applications such as incuba-tion experiments with He accessory injection units and se-tups using TREX related actions have to be taken While wetested several mono-variant (eg changes in [CH4] at con-stant [N2O]) and some bi-variant (eg changes in [CH4] and[N2O]) systems in our study more complex systems (egchanges in [CH4] [O2] and [N2O]) were not tested and de-viations from additive behavior are to be expected Depend-ing on the desired precision users may vary the measurementand averaging times and calibration frequency

44 What degree of accuracy can be achieved using thisworkflow

The simulated two-end-member mixing experiments con-ducted in this study show that when the workflow proposedabove is applied accuracy within plusmn05 permil can be achievedfor TREX-QCLAS plusmn04 permilndash16 permil for CRDS analyzers andplusmn16 permilndash36 permil for OA-ICOS analyzers Likewise the com-parison between the laser spectrometers and IRMS highlights

that cross-technique compatibility within plusmn1 permilndash25 permil canbe achieved for most N2O isotopocule measurements How-ever it is clear that the balanced (ie non-analyzer-specific)approach applied for the purpose of this comparative studydid not cater to QCLAS I Therefore a more specific calibra-tion protocol for the QCLAS I will likely yield better perfor-mance as shown in Table 6 It is worth noting that althoughthe results of our study are representative of the performanceof the instruments tested the magnitude of reported effectsand the performance are likely to vary within the same ana-lyzer models

Whilst the laboratory-simulated mixing experiment is notfully representative of naturally occurring two-end-membermixing per se the results are useful in comparing inter-cept accuracy and uncertainty amongst analyzers and againstIRMS Our results show that large uncertainties exist for N2Osource apportionment using Keeling analysis performed atnear-ambient N2O mole fractions Given the amount of cor-rections that are required in the experiment we have not de-tailed individual analyzer uncertainty budgets to quantify in-dividual sources of error on the intercept as it is beyond thescope of this study Nonetheless the reduction of uncertaintywith increasing 1N2O shown in Exps 1ndash6 in Sect 37 hasalso been shown in previous studies (eg Wolf et al 2015)Therefore by extension it is reasonable to assume that thecurrent largest source of uncertainty for ambient N2O mea-surements using laser spectroscopy is the inherent signal-to-noise ratio of the measurement

5 Conclusions

In this study we characterized and compared N2O isotopelaser-based analyzers with the three most common detec-tion schemes including OA-ICOS CRDS and QCLAS Ourresults show a number of factors that need to be carefullyconsidered to ensure precise and accurate measurements ofN2O isotopocules using laser spectroscopy The performanceof N2O isotope laser spectrometers depends on a complexinterplay between instrumental precision drift matrix gascomposition and associated spectral interferences that ulti-mately vary as a function of N2O mole fraction On this basiswe echo recommendations from Ostrom and Ostrom (2017)who cautioned not to underestimate the need for the carefulconsideration of analyzer-specific corrections These analyz-ers clearly do not represent ldquoplug and playrdquo devices ndash insteadone needs to carefully consider the desired application preci-sion and accuracy and develop appropriate calibration strate-gies to achieve these outcomes

Consequently we recommend calibration schemes thathave (1) a calibration frequency that is adequate for con-straining instrument drift over experimental periodlong-termmeasurements (2) temperature stability during measure-ment or the temperature effect adequately characterized andcorrected (3) a three-point or higher [N2O] effect correc-



Figure 15 Proposed measurement workflow for the operation of N2O isotope laser spectrometers Relevant sections of this study are shownnext to each step

tion that spans the range of expected [N2O] (if calibrationrelies on raw δ values derived from uncalibrated isotopoculeamount fractions ie a δ-calibration approach) and (4) ac-counted for the differences in matrix and trace gas composi-tion between the sample gas and reference gases wherebyeither analyzer-specific interference corrections have been

carefully characterized and applied or where possible in-terfering substances (CO2 CO H2O) removed using chemi-cal traps Correcting for interference effects becomes signifi-cantly more complicated once [N2O] exceeds ambient levelsrequiring a multitude of analyzer- and gas-specific constantsthat inevitably increase the number of gas mixtures required



by the user as well as the uncertainty of the measurementResearchers should therefore strive to implement measure-ment setups that require as few corrections as possible andthis will inherently decrease the combined uncertainty in themeasurement

Data availability All data are available from the corresponding au-thor upon request

Supplement The supplement related to this article is available on-line at httpsdoiorg105194amt-13-2797-2020-supplement

Author contributions SJH and JL guided by JM designed thestudy established the setup and coordinated the experiments Re-sponsibilities of the analyzers were as follows SJH BFJK and DICwere responsible for OA-ICOS JL and TB were responsible forCRDS I LX and BW were responsible for QCLAS I and CRDS IIJW and KZ were responsible for TREX-QCLAS I MB and JS wereresponsible for QCLAS II QCLAS III and GC-IRMS JM preparedthe non-commercial gas mixtures used in this study LY referencedthe δ values of applied gas mixtures to international scales SJH andJL performed the main data analysis and together with JM devel-oped correction algorithms SJH and JL wrote the manuscript withinput from all co-authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements We extend our thanks to Doug Baer RobProvencal and Frederick Despagne from ABB ndash Los Gatos Re-search Inc Renato Winkler Magdalena Hofmann and PaulosGetachew from Picarro Inc and Dave Nelson and Barry McManusfrom Aerodyne Research Inc for their helpful discussions and as-sistance concerning the analyzers We thank Christoph Zellwegerfor his assistance in analyzing matrix gas cylinders and providingthe Picarro G2401 for matrix interference testing and the NABELlaboratory at Empa for providing the dew-point generator for watervapor testing We want to thank Kristyacutena Kantnerovaacute from Empaand Daniel Zindel from ETH for the synthesis of 18O-labeled N2O

Financial support This research has been supported by theSwiss National Science Foundation (grant nos CRSII5_170876200021_163075) the Australian Endeavour Leadership Program(grant no ERF_RDDH_6743_2018) the European Metrology Pro-gramme for Innovation and Research (EMPIR) co-funded by theEuropean Unionrsquos Horizon 2020 research and innovation programand the EMPIR participating states (grant no 16ENV06 Metrol-ogy for Stable Isotope Reference Standards (SIRS)) the Cot-ton Research and Development Corporation (CRDC) (grant noANSTO1801) and the H2020 Marie Skłodowska-Curie Actions(EMPAPOSTDOCS-II (grant no 754364))

Stephen Harris is supported by PhD scholarships from the Aus-tralian Government the Australian Institute of Nuclear Scienceand Engineering (AINSE (scholarship no RG172557)) AustralianNuclear Science and Technology Organisation (ANSTO) (schol-arship no RG171945) and CRDC (scholarship no RG1626691)The UNSW Sydney N2OIA-30e-EP was purchased using fundsfrom the Australian National Collaborative Research InfrastructureStrategy (NCRIS) Groundwater Infrastructure Project (grant noRG134076) Jing Wei and Longfei Yu were supported by the SwissNational Science Foundation (SNSF) within grants CRSII5_170876and 200021_163075 as well as the EMPAPOSTDOCS-II pro-gram which received funding from the European Unionrsquos H2020Marie Skłodowska-Curie Actions (EMPAPOSTDOCS-II (grant no754364))

Review statement This paper was edited by Huilin Chen and re-viewed by three anonymous referees

References

ABB-Los Gatos Research Inc Off-Axis Integrated Cavity OutputSpectroscopy (OA-ICOS) available at httpwwwlgrinccomadvantages last access 15 August 2019

Allan D W Statistics of atomic frequency standards P IEEE 54221ndash230 1966

Assan S Baudic A Guemri A Ciais P Gros V and Vo-gel F R Characterization of interferences to in situ observa-tions of δ13CH4 and C2H6 when using a cavity ring-down spec-trometer at industrial sites Atmos Meas Tech 10 2077ndash2091httpsdoiorg105194amt-10-2077-2017 2017

Baer D S Paul J B Gupta M and OrsquoKeefe A Sensitive ab-sorption measurements in the near-infrared region using off-axisintegrated-cavity-output spectroscopy Appl Phys B-Lasers O75 261ndash265 httpsdoiorg101007s00340-002-0971-z 2002

Baggs E M A review of stable isotope techniques for N2O sourcepartitioning in soils recent progress remaining challenges andfuture considerations Rapid Commun Mass Sp 22 1664ndash1672 httpsdoiorg101002rcm3456 2008

Berden G Peeters R and Meijer G Cavity ring-down spectroscopy Experimental schemes and ap-plications Int Rev Phys Chem 19 565ndash607httpsdoiorg101080014423500750040627 2000

Bernard S Roumlckmann T Kaiser J Barnola J-M FischerH Blunier T and Chappellaz J Constraints on N2O bud-get changes since pre-industrial time from new firn air and icecore isotope measurements Atmos Chem Phys 6 493ndash503httpsdoiorg105194acp-6-493-2006 2006

Bouwman A F Boumans L J M and Batjes N HModeling global annual N2O and NO emissions fromfertilized fields Global Biogeochem Cy 16 1080httpsdoiorg1010292001GB001812 2002

Bowling D R Sargent S D Tanner B D and EhleringerJ R Tunable diode laser absorption spectroscopy for sta-ble isotope studies of ecosystem-atmosphere CO2 exchangeAgr Forest Meteorol 118 1ndash19 httpsdoiorg101016s0168-1923(03)00074-1 2003



Bowling D R Burns S P Conway T J MonsonR K and White J W C Extensive observations ofCO2 carbon isotope content in and above a high-elevationsubalpine forest Global Biogeochem Cy 19 GB3023httpsdoiorg1010292004gb002394 2005

Brase L Bange H W Lendt R Sanders T and DaumlhnkeK High Resolution Measurements of Nitrous Oxide(N2O) in the Elbe Estuary Front Mar Sci 4 162httpsdoiorg103389fmars201700162 2017

Buchen C Lewicka-Szczebak D Flessa H and WellR Estimating N2O processes during grassland renewaland grassland conversion to maize cropping using N2Oisotopocules Rapid Commun Mass Sp 32 1053ndash1067httpsdoiorg101002rcm8132 2018

Decock C and Six J How reliable is the intramolecu-lar distribution of 15N in N2O to source partition N2Oemitted from soil Soil Biol Biochem 65 114ndash127httpsdoiorg101016jsoilbio201305012 2013

Denk T R Mohn J Decock C Lewicka-Szczebak DHarris E Butterbach-Bahl K Kiese R and Wolf BThe nitrogen cycle A review of isotope effects and iso-tope modeling approaches Soil Biol Biochem 105 121ndash137httpsdoiorg101016jsoilbio201611015 2017

Denman K L Brasseur G Chidthaisong A Ciais P Cox PM Dickinson R E Hauglustaine D Heinze C Holland EJacob D Lohmann U Ramachandran S da Silva Dias P LWofsky S C and Zhang X Couplings between changes in theclimate system and biogeochemistry in Climate change 2007the physical science basis contribution of working group I tothe Fourth Assessment Report of the Intergovernmental Panel onClimate Change edited by Solomon S Qin D Manning MChen Z Marquis M Averyt K B Tignor M and MillerH L Cambridge University Press Cambridge United Kingdomand New York NY USA 499ndash587 2007

Erler D V Duncan T M Murray R Maher D T SantosI R Gatland J R Mangion P and Eyre B D Apply-ing cavity ring-down spectroscopy for the measurement of dis-solved nitrous oxide mole fractions and bulk nitrogen isotopiccomposition in aquatic systems Correcting for interferencesand field application Limnol Oceanogr-Meth 13 391ndash401httpsdoiorg101002lom310032 2015

Eyer S Tuzson B Popa M E van der Veen C RoumlckmannT Rothe M Brand W A Fisher R Lowry D Nisbet EG Brennwald M S Harris E Zellweger C EmmeneggerL Fischer H and Mohn J Real-time analysis of δ13C- andδDminusCH4 in ambient air with laser spectroscopy method devel-opment and first intercomparison results Atmos Meas Tech 9263ndash280 httpsdoiorg105194amt-9-263-2016 2016

Flores E Viallon J Moussay P Griffith D W and Wiel-gosz R I Calibration strategies for FT-IR and other isotoperatio infrared spectrometer instruments for accurate δ13C andδ18O measurements of CO2 in air Anal Chem 89 3648ndash3655httpsdoiorg101021acsanalchem6b05063 2017

Forster P Ramaswamy V Artaxo P Berntsen T Betts R Fa-hey D W Haywood J Lean J Lowe D C Myhre GNganga J Prinn R Raga G Schulz M and Van Dorland RClimate Change 2007 Changes in Atmospheric Constituents andin Radiative Forcing The Physical Science Basis Contributionof Working Group I to the Fourth Assessment Report of the In-

tergovernmental Panel on Climate Change edited by SolomonS Qin D Manning M Chen Z Marquis M Averyt K BTignor M and Miller H S Cambridge University Press Cam-bridge UK New York NY USA 2007

Friedrichs G Bock J Temps F Fietzek P KoumlrtzingerA and Wallace D W Toward continuous monitor-ing of seawater 13CO2

12CO2 isotope ratio and pCO2Performance of cavity ringdown spectroscopy and gasmatrix effects Limnol Oceanogr-Meth 8 539ndash551httpsdoiorg104319lom20108539 2010

Griffis T J Baker J M Sargent S D Tanner B D andZhang J Measuring field-scale isotopic CO2 fluxes withtunable diode laser absorption spectroscopy and microme-teorological techniques Agr Forest Meteorol 124 15ndash29httpsdoiorg101016jagrformet200401009 2004

Griffith D W T Calibration of isotopologue-specific optical tracegas analysers a practical guide Atmos Meas Tech 11 6189ndash6201 httpsdoiorg105194amt-11-6189-2018 2018

Griffith D W T Deutscher N M Caldow C Kettlewell GRiggenbach M and Hammer S A Fourier transform infraredtrace gas and isotope analyser for atmospheric applications At-mos Meas Tech 5 2481ndash2498 httpsdoiorg105194amt-5-2481-2012 2012

Groumlning M TEL Technical Note No 01 SICalib User Manual(Stable Isotope Calibration for routine δ-scale measurements)Ver 216j 2018

Harris E Nelson D D Olszewski W Zahniser M Potter KE McManus B J Whitehill A Prinn R G and Ono SDevelopment of a spectroscopic technique for continuous onlinemonitoring of oxygen and site-specific nitrogen isotopic com-position of atmospheric nitrous oxide Anal Chem 86 1726ndash1734 httpsdoiorg101021ac403606u 2014

Harris E Joss A Emmenegger L Kipf M Wolf B MohnJ and Wunderlin P Isotopic evidence for nitrous oxide pro-duction pathways in a partial nitritation-anammox reactor WaterRes 83 258ndash270 httpsdoiorg101016jwatres2015060402015a

Harris E Zeyer K Kegel R Muumlller B Emmeneg-ger L and Mohn J Nitrous oxide and methane emis-sions and nitrous oxide isotopic composition from wasteincineration in Switzerland Waste Manage 35 135ndash140httpsdoiorg101016jwasman201410016 2015b

Harris E Henne S Huumlglin C Zellweger C Tuzson B IbraimE Emmenegger L and Mohn J Tracking nitrous oxide emis-sion processes at a suburban site with semicontinuous in situmeasurements of isotopic composition J Geophys Res-Atmos122 1850ndash1870 httpsdoiorg1010022016JD025906 2017

Heil J Wolf B Bruumlggemann N Emmenegger L Tuzson BVereecken H and Mohn J Site-specific 15N isotopic signa-tures of abiotically produced N2O Geochim Cosmochim Ac139 72ndash82 httpsdoiorg101016jgca201404037 2014

Henry A Margottin-Maclou M and Lacome N N2- andO2-broadening parameters in the ν3 band of 14N16

2 O JMol Spectrosc 111 291ndash300 httpsdoiorg1010160022-2852(85)90006-2 1985

Ibraim E Harris E Eyer S Tuzson B Emmenegger L SixJ and Mohn J Development of a field-deployable methodfor simultaneous real-time measurements of the four most



abundant N2O isotopocules Isot Environ Health S 54 1ndash5httpsdoiorg1010801025601620171345902 2018

Ibraim E Wolf B Harris E Gasche R Wei J Yu L KieseR Eggleston S Butterbach-Bahl K Zeeman M TuzsonB Emmenegger L Six J Henne S and Mohn J At-tribution of N2O sources in a grassland soil with laser spec-troscopy based isotopocule analysis Biogeosciences 16 3247ndash3266 httpsdoiorg105194bg-16-3247-2019 2019

Ishijima K Sugawara S Kawamura K Hashida G MorimotoS Murayama S Aoki S and Nakazawa T Temporal vari-ations of the atmospheric nitrous oxide mole fraction and itsδ15N and δ18O for the latter half of the 20th century recon-structed from firn air analyses J Geophys Res 112 D03305httpsdoiorg1010292006JD007208 2007

Ji Q and Grundle D S An automated laser-based measure-ment system for nitrous oxide isotope and isotopomer ratios atnanomolar levels Rapid Commun Mass Sp 33 1553ndash1564httpsdoiorg101002rcm8502 2019

Johnson J E and Rella C W Effects of variation in back-ground mixing ratios of N2 O2 and Ar on the measure-ment of δ18OminusH2O and δ2HminusH2O values by cavity ring-down spectroscopy Atmos Meas Tech 10 3073ndash3091httpsdoiorg105194amt-10-3073-2017 2017

Keeling C D The mole fraction and isotopic abundances of at-mospheric carbon dioxide in rural areas Geochim CosmochimAc 13 322ndash334 httpsdoiorg1010160016-7037(58)90033-4 1958

Koba K Osaka K Tobari Y Toyoda S Ohte N KatsuyamaM Suzuki N Itoh M Yamagishi H Kawasaki M andKim S J Biogeochemistry of nitrous oxide in groundwa-ter in a forested ecosystem elucidated by nitrous oxide iso-topomer measurements Geochim Cosmochim Ac 73 3115ndash3133 httpsdoiorg101016jgca200903022 2009

Kong X Duan Y Schramm A Eriksen J HolmstrupM Larsen T Bol R and Petersen SO MitigatingN2O emissions from clover residues by 34-dimethylpyrazolephosphate (DMPP) without adverse effects on the earth-worm Lumbricus terrestris Soil Biol Biochem 104 95ndash107httpsdoiorg101016jsoilbio201610012 2017

Koumlster J R Caacuterdenas L Senbayram M Bol R Well R But-ler M Muumlhling K H and Dittert K Rapid shift from denitri-fication to nitrification in soil after biogas residue application asindicated by nitrous oxide isotopomers Soil Biol Biochem 431671ndash1677 httpsdoiorg101016jsoilbio201104004 2011

Koumlster J R Well R Tuzson B Bol R Dittert K GiesemannA Emmenegger L Manninen A Caacuterdenas L and Mohn JNovel laser spectroscopic technique for continuous analysis ofN2O isotopomers ndash application and intercomparison with isotoperatio mass spectrometry Rapid Commun Mass Sp 27 216ndash222 httpsdoiorg101002rcm6434 2013

Lebegue B Schmidt M Ramonet M Wastine B Yver KwokC Laurent O Belviso S Guemri A Philippon C SmithJ and Conil S Comparison of nitrous oxide (N2O) analyzersfor high-precision measurements of atmospheric mole fractionsAtmos Meas Tech 9 1221ndash1238 httpsdoiorg105194amt-9-1221-2016 2016

Lee A Winther M Priemeacute A Blunier T and Chris-tensen S Hot spots of N2O emission move withthe seasonally mobile oxic-anoxic interface in drained

organic soils Soil Biol Biochem 115 178ndash186httpsdoiorg101016jsoilbio201708025 2017

Lewicka-Szczebak D Well R Koumlster J R Fuszlig RSenbayram M Dittert K and Flessa H Experimen-tal determinations of isotopic fractionation factors associ-ated with N2O production and reduction during denitri-fication in soils Geochim Cosmochim Ac 134 55ndash73httpsdoiorg101016jgca201403010 2014

Lewicka-Szczebak D Well R Bol R Gregory A S MatthewsG P Misselbrook T Whalley W R and Cardenas LM Isotope fractionation factors controlling isotopocule sig-natures of soil-emitted N2O produced by denitrification pro-cesses of various rates Rapid Commun Mass Sp 29 269ndash282httpsdoiorg101002rcm7102 2015

Li P Wang S Peng Y Liu Y and He J The syn-ergistic effects of dissolved oxygen and pH on N2O pro-duction in biological domestic wastewater treatment un-der nitrifying conditions Environ Technol 36 1623ndash1631httpsdoiorg1010800959333020141002862 2015

Li X Soslashrensen P Olesen J E and Petersen S O Ev-idence for denitrification as main source of N2O emissionfrom residue-amended soil Soil Biol Biochem 92 153ndash160httpsdoiorg101016jsoilbio201510008 2016

Malowany K Stix J Van Pelt A and Lucic G H2S interfer-ence on CO2 isotopic measurements using a Picarro G1101-icavity ring-down spectrometer Atmos Meas Tech 8 4075ndash4082 httpsdoiorg105194amt-8-4075-2015 2015

Miller J B and Tans P P Calculating isotopic fractionation fromatmospheric measurements at various scales Tellus B 55 207ndash214 httpsdoiorg101034j1600-0889200300020x 2003

Minamikawa K Nishimura S Nakajima Y Osaka KISawamoto T and Yagi K Upward diffusion of nitrous oxideproduced by denitrification near shallow groundwater table in thesummer a lysimeter experiment Soil Sci Plant Nutr 57 719ndash732 httpsdoiorg101080003807682011625556 2011

Mohn J Guggenheim C Tuzson B Vollmer M K Toy-oda S Yoshida N and Emmenegger L A liquid nitrogen-free preconcentration unit for measurements of ambient N2Oisotopomers by QCLAS Atmos Meas Tech 3 609ndash618httpsdoiorg105194amt-3-609-2010 2010

Mohn J Tuzson B Manninen A Yoshida N Toy-oda S Brand W A and Emmenegger L Site selec-tive real-time measurements of atmospheric N2O isotopomersby laser spectroscopy Atmos Meas Tech 5 1601ndash1609httpsdoiorg105194amt-5-1601-2012 2012

Mohn J Wolf B Toyoda S Lin C T Liang C M Bruumlgge-mann N Wissel H Steiker A E Dyckmans J Szwek LOstrom N E Casciotti K L Forbes M Giesemann AWell R Doucett R R Yarnes C T Ridley A R KaiserJ and Yoshida N Inter-Laboratory assessment of nitrous ox-ide isotopomer analysis of isotopomer analysis by isotope ra-tio mass spectrometry and laser spectroscopy current statusand perspectives Rapid Commun Mass Sp 28 1995ndash2007httpsdoiorg101002rcm6982 2014

Mosier A Kroeze C Nevison C Oenema O SeitzingerS and Van Cleemput O Closing the global N2O bud-get nitrous oxide emissions through the agriculturalnitrogen cycle Nutr Cycl Agroecosys 52 225ndash248httpsdoiorg101023A1009740530221 1998



Murray R Erler D Rosentreter J Maher D and EyreB A seasonal source and sink of nitrous oxide in man-groves Insights from mole fraction isotope and isotopomermeasurements Geochim Cosmochim Ac 238 169ndash192httpsdoiorg101016jgca201807003 2018

Nara H Tanimoto H Tohjima Y Mukai H Nojiri YKatsumata K and Rella C W Effect of air composition(N2 O2 Ar and H2O) on CO2 and CH4 measurement bywavelength-scanned cavity ring-down spectroscopy calibrationand measurement strategy Atmos Meas Tech 5 2689ndash2701httpsdoiorg105194amt-5-2689-2012 2012

Nikolenko O Orban P Jurado A Morana C Jamin PRobert T Knoumlller K Borges A V and Brouyegravere S Dy-namics of greenhouse gases in groundwater hydrogeologicaland hydrogeochemical controls Appl Geochem 105 31ndash44httpsdoiorg101016japgeochem201904009 2019

NOAAESRL Combined Nitrous Oxide data NOAAESRL GlobalMonitoring Division available at ftpaftpcmdlnoaagovproductstrendsn2o last access 11 November 2019

Ogawa M and Yoshida N Intramolecular distribution of sta-ble nitrogen and oxygen isotopes of nitrous oxide emit-ted during coal combustion Chemosphere 61 877ndash887httpsdoiorg101016jchemosphere200504096 2005

Ostrom N E and Ostrom P H The Isotopomers of Nitrous Ox-ide Analytical Considerations and Application to Resolution ofMicrobial Production Pathways in Handbook of EnvironmentalIsotope Geochemistry edited by Baskaran M Springer BerlinHeidelberg Berlin Heidelberg 1 453ndash476 2011

Ostrom N E and Ostrom P H Mining the isotopic complex-ity of nitrous oxide a review of challenges and opportunitiesBiogeochemistry 132 359ndash372 httpsdoiorg101007s10533-017-0301-5 2017

Ostrom N E Pitt A Sutka R Ostrom P H GrandyA S Huizinga K M and Robertson G P Isotopo-logue effects during N2O reduction in soils and in purecultures of denitrifiers J Geophys Res 112 G02005httpsdoiorg1010292006JG000287 2007

Ostrom N E Gandhi H Coplen T B Toyoda S BoumlhlkeJ K Brand W A Casciotti K L Dyckmans J Giese-mann A Mohn J and Well R Preliminary assessment ofstable nitrogen and oxygen isotopic composition of USGS51and USGS52 nitrous oxide reference gases and perspectives oncalibration needs Rapid Commun Mass Sp 32 1207ndash1214httpsdoiorg101002rcm8157 2018

Pang J Wen X Sun X and Huang K Intercomparison oftwo cavity ring-down spectroscopy analyzers for atmospheric13CO2

12CO2 measurement Atmos Meas Tech 9 3879ndash3891 httpsdoiorg105194amt-9-3879-2016 2016

Park S Peacuterez T Boering K A Trumbore S E GilJ Marquina S and Tyler S C Can N2O stable iso-topes and isotopomers be useful tools to characterize sourcesand microbial pathways of N2O production and consump-tion in tropical soils Global Biogeochem Cy 25 GB1001httpsdoiorg1010292009GB003615 2011

Pataki D E Bowling D R Ehleringer J R and Zob-itz J M High resolution atmospheric monitoring of ur-ban carbon dioxide sources Geophys Res Lett 33 L03813httpsdoiorg1010292005gl024822 2006

Peng L Ni BJ Erler D Ye L and Yuan Z The effect of dis-solved oxygen on N2O production by ammonia-oxidizing bac-teria in an enriched nitrifying sludge Water Res 66 12ndash21httpsdoiorg101016jwatres201408009 2014

Peacuterez T Trumbore S E Tyler S C Matson P A Ortiz-Monasterio I Rahn T and Griffith D W T Identi-fying the agricultural imprint on the global N2O budgetusing stable isotopes J Geophys Res 106 9869ndash9878httpsdoiorg1010292000JD900809 2001

Peacuterez T Garcia-Montiel D Trumbore S Tyler S Ca-margo P D Moreira M Piccolo M and CerriC Nitrous oxide nitrification and denitrification 15Nenrichment factors from Amazon forest soils EcolAppl 16 2153ndash2167 httpsdoiorg1018901051-0761(2006)016[2153NONADN]20CO2 2006

Picarro Inc Cavity Ring-Down Spectroscopy (CRDS) available athttpswwwpicarrocomcompanytechnologycrds last access15 August 2019

Potter K E Ono S and Prinn R G Fully automatedhigh-precision instrumentation for the isotopic analysis oftropospheric N2O using continuous flow isotope ratio massspectrometry Rapid Commun Mass Sp 27 1723ndash1738httpsdoiorg101002rcm6623 2013

Prokopiou M Martinerie P Sapart C J Witrant E Mon-teil G Ishijima K Bernard S Kaiser J Levin I Blu-nier T Etheridge D Dlugokencky E van de Wal R SW and Roumlckmann T Constraining N2O emissions since 1940using firn air isotope measurements in both hemispheres At-mos Chem Phys 17 4539ndash4564 httpsdoiorg105194acp-17-4539-2017 2017

R Core Team R A language and environment for statistical com-puting R Foundation for Statistical Computing Vienna Austriaavailable at httpswwwR-projectorg (last access 28 March2019) 2017

Ravishankara A R Daniel J S and Portmann R W Nitrousoxide (N2O) the dominant ozone-depleting substance emitted inthe 21st century Science 326 123ndash125 2009

Rella C W Chen H Andrews A E Filges A Gerbig CHatakka J Karion A Miles N L Richardson S J Stein-bacher M Sweeney C Wastine B and Zellweger C Highaccuracy measurements of dry mole fractions of carbon diox-ide and methane in humid air Atmos Meas Tech 6 837ndash860httpsdoiorg105194amt-6-837-2013 2013

Roumlckmann T and Levin I High-precision determinationof the changing isotopic composition of atmospheric N2Ofrom 1990 to 2002 J Geophys Res 110 D21304httpsdoiorg1010292005JD006066 2015

Rothman L S Jacquemart D Barbe A Benner D C BirkM Brown L R Carleer M R Chackerian Jr C Chance KCoudert L E A and Dana V The HITRAN 2004 molecularspectroscopic database J Quant Spectrosc Ra 96 139ndash204httpsdoiorg101016jjqsrt200410008 2005

Soto D X Koehler G and Hobson K A Combining den-itrifying bacteria and laser spectroscopy for isotopic analyses(δ15N δ18O) of dissolved nitrate Anal Chem 87 7000ndash7005httpsdoiorg101021acsanalchem5b01119 2015

Sutka R L Ostrom N E Ostrom P H Gandhi Hand Breznak J A Nitrogen isotopomer site preference ofN2O produced by Nitrosomonas europaea and Methylococ-



cus capsulatus bath Rapid Commun Mass Sp 17 738ndash745httpsdoiorg101002rcm968 2003

Sutka R L Ostrom N E Ostrom P H Breznak J A GandhiH Pitt A J and Li F Distinguishing nitrous oxide pro-duction from nitrification and denitrification on the basis ofisotopomer abundances Appl Environ Microb 72 638ndash644httpsdoiorg101128AEM721638-6442006 2006

Tian H Chen G Lu C Xu X Ren W Zhang B Banger KTao B Pan S Liu M Zhang C Bruhwiler L and WofsyS Global methane and nitrous oxide emissions from terres-trial ecosystems due to multiple environmental changes EcosystHealth Sustain 1 1ndash20 httpsdoiorg101890ehs14-001512015

Toyoda S and Yoshida N Determination of nitrogen iso-topomers of nitrous oxide on a modified isotope ra-tio mass spectrometer Anal Chem 71 4711ndash4718httpsdoiorg101021ac9904563 1999

Toyoda S Mutobe H Yamagishi H Yoshida N andTanji Y Fractionation of N2O isotopomers during pro-duction by denitrifier Soil Biol Biochem 37 1535ndash1545httpsdoiorg101016jsoilbio200501009 2005

Toyoda S Yamamoto S I Arai S Nara H YoshidaN Kashiwakura K and Akiyama K I Isotopomericcharacterization of N2O produced consumed and emittedby automobiles Rapid Commun Mass Sp 22 603ndash612httpsdoiorg101002rcm3400 2008

Toyoda S Yano M Nishimura S Akiyama H Hayakawa AKoba K Sudo S Yagi K Makabe A Tobari Y OgawaN O Ohkouchi N Yamada K and Yoshida N Charac-terization and production and consumption processes of N2Oemitted from temperate agricultural soils determined via iso-topomer ratio analysis Globl Biogeochem Cy 25 GB2008httpsdoiorg1010292009GB003769 2011

Toyoda S Kuroki N Yoshida N Ishijima K Tohjima Yand Machida T Decadal time series of tropospheric abun-dance of N2O isotopomers and isotopologues in the North-ern Hemisphere obtained by the long-term observation atHateruma Island Japan J Geophys Res-Atmos 118 3369ndash3381 httpsdoiorg101002jgrd50221 2013

Toyoda S Yoshida N and Koba K Isotopocule anal-ysis of biologically produced nitrous oxide in vari-ous environments Mass Spectrom Rev 36 135ndash160httpsdoiorg101002mas21459 2017

Tuzson B Mohn J Zeeman M J Werner R A EugsterW Zahniser M S Nelson D D McManus J B andEmmenegger L High precision and continuous field measure-ments of δ13C and δ18O in carbon dioxide with a cryogen-freeQCLAS Appl Phys B 92 451 httpsdoiorg101007s00340-008-3085-4 2008

Verhoeven E Decock C Barthel M Bertora C SaccoD Romani M Sleutel S and Six J Nitrificationand coupled nitrification-denitrification at shallow depthsare responsible for early season N2O emissions underalternate wetting and drying management in an Ital-ian rice paddy system Soil Biol Biochem 120 58ndash69httpsdoiorg101016jsoilbio201801032 2018

Verhoeven E Barthel M Yu L Celi L Said-Pullicino DSleutel S Lewicka-Szczebak D Six J and Decock CEarly season N2O emissions under variable water manage-

ment in rice systems source-partitioning emissions using iso-tope ratios along a depth profile Biogeosciences 16 383ndash408httpsdoiorg105194bg-16-383-2019 2019

Vogel F R Huang L Ernst D Giroux L Racki S and WorthyD E J Evaluation of a cavity ring-down spectrometer for insitu observations of 13CO2 Atmos Meas Tech 6 301ndash308httpsdoiorg105194amt-6-301-2013 2013

Waumlchter H Mohn J Tuzson B Emmenegger L and Sigrist MW Determination of N2O isotopomers with quantum cascadelaser based absorption spectroscopy Opt Express 16 9239ndash9244 httpsdoiorg101364OE16009239 2008

Wassenaar L I Douence C Altabet M A and Aggarwal P KN and O isotope (δ15Nα δ15Nβ δ18O δ17O) analyses of dis-solved NOminus3 and NOminus2 by the Cd-azide reduction method andN2O laser spectrometry Rapid Commun Mass Sp 32 184ndash194 httpsdoiorg101002rcm8029 2018

Wei J Zhou M Vereecken H and Bruumlggemann N Largevariability in CO2 and N2O emissions and in 15N site prefer-ence of N2O from reactions of nitrite with lignin and its deriva-tives at different pH Rapid Commun Mass Sp 31 1333ndash1343httpsdoiorg101002rcm7912 2017

Well R and Flessa H Isotopologue signatures of N2O pro-duced by denitrification in soils J Geophys Res 114 G02020httpsdoiorg1010292008JG000804 2009

Well R Flessa H Jaradat F Toyoda S and YoshidaN Measurement of isotopomer signatures of N2Oin groundwater J Geophys Res 110 G02006httpsdoiorg1010292005JG000044 2005

Well R Flessa H Xing L Xiaotang J and RoumlmheldV Isotopologue ratios of N2O emitted from microcosmswith NH+4 fertilized arable soils under conditions fa-voring nitrification Soil Biol Biochem 40 2416ndash2426httpsdoiorg101016jsoilbio200806003 2008

Well R Eschenbach W Flessa H von der Heide C and Wey-mann D Are dual isotope and isotopomer ratios of N2O use-ful indicators for N2O turnover during denitrification in nitrate-contaminated aquifers Geochim Cosmochim Ac 90 265ndash282 httpsdoiorg101016jgca201204045 2012

Wen X-F Meng Y Zhang X-Y Sun X-M and Lee XEvaluating calibration strategies for isotope ratio infrared spec-troscopy for atmospheric 13CO2

12CO2 measurement AtmosMeas Tech 6 1491ndash1501 httpsdoiorg105194amt-6-1491-2013 2013

Werle P O Muumlcke R and Slemr F The limits of signal averag-ing in atmospheric trace-gas monitoring by tunable diode-laserabsorption spectroscopy (TDLAS) Appl Phys B 57 131ndash139httpsdoiorg101007BF00425997 1993

Werner R A and Brand W A Referencing strategies and tech-niques in stable isotope ratio analysis Rapid Commun Mass Sp15 501ndash519 httpsdoiorg101002rcm258 2001

Winther M Balslev-Harder D Christensen S Priemeacute A El-berling B Crosson E and Blunier T Continuous measure-ments of nitrous oxide isotopomers during incubation experi-ments Biogeosciences 15 767ndash780 httpsdoiorg105194bg-15-767-2018 2018

Wolf B Merbold L Decock C Tuzson B Harris E Six JEmmenegger L and Mohn J First on-line isotopic character-ization of N2O above intensively managed grassland Biogeo-



sciences 12 2517ndash2531 httpsdoiorg105194bg-12-2517-2015 2015

Wunderlin P Mohn J Joss A Emmenegger L and Siegrist HMechanisms of N2O production in biological wastewater treat-ment under nitrifying and denitrifying conditions Water Res46 1027ndash1037 httpsdoiorg101016jwatres2011110802012

Wunderlin P Lehmann M F Siegrist H Tuzson B Joss AEmmenegger L and Mohn J Isotope signatures of N2O in amixed microbial population system constraints on N2O produc-ing pathways in wastewater treatment Environ Sci Technol 471339ndash1348 httpsdoiorg101021es303174x 2013

Yamamoto A Uchida Y Akiyama H and Nakajima Y Con-tinuous and unattended measurements of the site preference ofnitrous oxide emitted from an agricultural soil using quantumcascade laser spectrometry with intercomparison with isotope ra-tio mass spectrometry Rapid Commun Mass Spectr 28 1444ndash1452 httpsdoiorg101002rcm6916 2014

Yoshida N and Toyoda S Constraining the atmospheric N2Obudget from intramolecular site preference in N2O isotopomersNature 405 330ndash334 httpsdoiorg10103835012558 2000

Yue F J Li S L Liu C Q Mostofa K M Yoshida NToyoda S Wang S L Hattori S and Liu X L Spa-tial variation of nitrogen cycling in a subtropical stratified im-poundment in southwest China elucidated by nitrous oxideisotopomer and nitrate isotopes Inland Waters 8 186ndash195httpsdoiorg1010802044204120181457847 2018


Abstract

Introduction

Materials and methods

Analytical techniques

OA-ICOS (ABB ndash Los Gatos Research Inc)

CRDS (Picarro Inc)

QCLAS (Aerodyne Research Inc)

TREX-QCLAS

GC-IRMS

Sample and reference gases

Matrix and interference test gases

Reference gases (S1 S2) and pressurized air (PA1 PA2)

Laboratory setup measurement procedures and data processing

Laboratory setup

Measurement procedures data processing and calibration

Testing of instruments

Allan precision

Temperature effects

Repeatability

N2O mole fraction dependence

Gas matrix effects (O2 and Ar)

Trace gas effects (CO2 CH4 CO and H2O)

CO2 and CO removal using NaOH (Ascarite) and Sofnocat

Two-end-member mixing

Results

Allan precision

Temperature effects

Repeatability

Dependence of isotopic measurements on N2O mole fraction


Gas matrix effects at ambient N2O mole fractions

Continuity of gas matrix corrections at higher N2O mole fractions

Trace gas effects (H2O CO2 CH4 and CO)

Trace gas effects at ambient N2O mole fractions

Continuity of trace gas corrections at higher N2O mole fractions


Comparison with expected [N2O] and values

Source identification using Keeling analysis

Discussion

Factors affecting the precision and accuracy of N2O isotopocule measurements using laser spectroscopy

Pre-measurement considerations

Measurement workflow

What degree of accuracy can be achieved using this workflow

Conclusions

Data availability

Supplement

Author contributions

Competing interests

Acknowledgements

Financial support

Review statement

References


future analyzer models and instruments focusing on isotopicspecies of other molecules

1 Introduction

Nitrous oxide (N2O) is a long-lived greenhouse gas with a100-year global warming potential nearly 300 times that ofcarbon dioxide (CO2 Forster et al 2007) and is the largestemission source of ozone-depleting nitrogen oxides in thestratosphere (Ravishankara et al 2009) In 2019 the glob-ally averaged [N2O] reached approximately 332 ppb com-pared to the pre-industrial level of 270 ppb (NOAAESRL2019) While this increase is known to be linked primarily toincreased fertilizer use in agriculture (Bouwman et al 2002Mosier et al 1998 Tian et al 2015) understanding the un-derlying microbial processes producing and consuming N2Ohas proved more challenging and individual source contri-butions from sectors such as agricultural soils wastewatermanagement and biomass burning to global bottom-up esti-mates of N2O emissions have large uncertainties (Denmanet al 2007) Stable isotopes are an effective tool for dis-tinguishing N2O sources and determining production path-ways which is critical for developing appropriate mitigationstrategies (Baggs 2008 Ostrom and Ostrom 2011 Toyodaet al 2017)

The N2O molecule has an asymmetric linear structure(NNO) with the following most abundant isotopocules14N15N16O (15NαminusN2O) 15N14N16O (15NβminusN2O)14N14N18O (18OminusN2O) and 14N14N16O (Yoshida andToyoda 2000) The terms 15Nα and 15Nβ refer to therespective central and terminal positions of nitrogen (N)atoms in the NNO molecule (Toyoda and Yoshida 1999)Isotopic abundances are reported in δ notation whereδ15N= R(15N14N)sampleR(

15N14N)referenceminus 1 denotesthe relative difference in per mil (permil) of the sample versusatmospheric N2 (AIR-N2) The isotope ratio R(15N14N)equals x(15N)x(14N) with x being the absolute abun-dance of 14N and 15N respectively Similarly ViennaStandard Mean Ocean Water (VSMOW) is the interna-tional isotope-ratio scale for δ18O In practice the isotopeδ value is calculated from measurement of isotopoculeratios of sample and reference gases with the latter beingdefined on the AIR-N2 and VSMOW scales By extensionδ15Nα denotes the corresponding relative difference ofisotope ratios for 14N15N16O14N14N16O and δ15Nβ for15N14N16O14N14N16O The site-specific intramoleculardistribution of 15N within the N2O molecule is termed 15N-site preference (SP) and is defined as SP= δ15Nα minus δ15Nβ

(Yoshida and Toyoda 2000) The term δ15Nbulk is usedto express the average δ15N value and is equivalent toδ15Nbulk

= (δ15Nα + δ15Nβ)2Extensive evidence has shown that SP δ15Nbulk and δ18O

can be used to differentiate N2O source processes and

biogeochemical cycling (Decock and Six 2013 Denk etal 2017 Heil et al 2014 Lewicka-Szczebak et al 20142015 Ostrom et al 2007 Sutka et al 2003 2006 Toy-oda et al 2005 2017 Wei et al 2017) Isotopocule abun-dances have been measured in a wide range of environ-ments including the troposphere (Harris et al 2014 Roumlck-mann and Levin 2005 Toyoda et al 2013) agriculturalsoils (Buchen et al 2018 Ibraim et al 2019 Koumlster etal 2011 Mohn et al 2012 Ostrom et al 2007 Park etal 2011 Peacuterez et al 2001 2006 Toyoda et al 2011 Ver-hoeven et al 2018 2019 Well et al 2008 Well and Flessa2009 Wolf et al 2015) mixed urbanndashagricultural environ-ments (Harris et al 2017) coal and waste combustion (Har-ris et al 2015b Ogawa and Yoshida 2005) fossil fuel com-bustion (Toyoda et al 2008) wastewater treatment (Harriset al 2015a b Wunderlin et al 2012 2013) groundwa-ter (Koba et al 2009 Minamikawa et al 2011 Nikolenkoet al 2019 Well et al 2005 2012) estuaries (Erler etal 2015) mangrove forests (Murray et al 2018) stratifiedwater impoundments (Yue et al 2018) and firn air and icecores (Bernard et al 2006 Ishijima et al 2007 Prokopiouet al 2017) While some applications like laboratory incu-bation experiments allow for analysis of the isotopic signa-ture of the pure source most studies require analysis of thesource diluted in ambient air This specifically applies to ter-restrial ecosystem research since N2O emitted from soils isimmediately mixed with background atmospheric N2O Tounderstand the importance of soil emissions for the globalN2O budget two-end-member mixing models commonly in-terpreted using Keeling or MillerndashTans plots are frequentlyused to back-calculate the isotopic composition of N2O emit-ted from soils (Keeling 1958 Miller and Tans 2003)

N2O isotopocules can be analyzed by isotope-ratio massspectrometry (IRMS) and laser spectroscopic techniqueswith currently available commercial spectrometers operatingin the mid-infrared (MIR) region to achieve highest sensitiv-ities IRMS analysis of the N2O intramolecular 15N distri-bution is based on quantification of the fragmented (NO+mz 30 and 31) and molecular (N2O+ mz 44 45 and46) ions to calculate isotope ratios for the entire molecule(15N14N and 18O16O) and the central (Nα) and terminal(Nβ ) N atom (Toyoda and Yoshida 1999) The analysis ofN2O SP by IRMS is complicated by the rearrangement of Nα

and Nβ in the ion source while analysis of δ15Nbulk (4544)involves correction for NN17O (mass 45) IRMS can achieverepeatability as good as 01 permil for δ15N δ18O δ15Nα andδ15Nβ (Potter et al 2013 Roumlckmann and Levin 2005) butan interlaboratory comparison study showed substantial de-viations in measurements of N2O isotopic composition inparticular for SP (up to 10 permil) (Mohn et al 2014)

The advancement of mid-infrared laser spectroscopic tech-niques was enabled by the invention and availability of non-cryogenic light sources which have been coupled with dif-ferent detection schemes such as direct absorption quantumcascade laser absorption spectroscopy (QCLAS Aerodyne































Tabl

e1

Ove

rvie

wof

the

wav

elen

gth

regi

ons

line

posi

tions

and

line

stre

ngth

sof

N2O

isot

opoc

ules

and

key

oper

atin

gpa

ram

eter

sfo

reac

hla

sers

pect

rom

eter

test

edin

this

stud

yN

TP

stan

dsfo

rnor

mal

tem

pera

ture

and

pres

sure

Det

ectio

nsc

hem

eN

2OW

aven

umbe

rIs

otop

ocul

esL

ine

posi

tions

Flow

rate

Cel

lC

ell

Inte

rnal

Eff

ectiv

eM

easu

rem

ent

Ref

eren

ces

(mod

el

rang

ere

gion

(cmminus

1 )li

ne(c

m3

minminus

1 )te

mpe

ratu

repr

essu

repl

umbi

ngvo

lum

efr

eque

ncy

man

ufac

ture

r)(p

pb)

(cmminus

1 )st

reng

th(c

mminus

1 (

C)

(hPa

)vo

lum

eat

NT

P(s

)(m

olec

ule

cmminus

2 ))

(cm

3 )(c

m3 )

OA

-IC

OS

I30

0ndash10

000

021

921

ndash219

25

14N

14N

16O

21924

04

919times

10minus

2060

436

6193

060

50

100

Bae

reta

l(2

002)

(N

2OIA

-30e

-EP

21924

44

919times

10minus

20A

BB

ndashL

osG

atos

AB

Bndash

Los

Gat

os21

924

83

375times

10minus

19R

esea

rch

Inc

Res

earc

h14

N15

N16

O21

923

13

31times

10minus

21(2

019)

Inc

)15

N14

N16

O21

923

32

968times

10minus

21

14N

14N

18O

21921

31

113times

10minus

21

CR

DS

IampII

300ndash

1500

2195

7ndash2

196

314

N14

N16

O21

962

15

161times

10minus

2025

2(C

RD

SI)

402

100

404

223

41(C

RD

SI)

Pica

rro

(G51

31-i

21

962

45

161times

10minus

2012

5(C

RD

SII

)2

54(C

RD

SII

)In

c(2

019)

Pica

rro

14N

15N

16O

21957

62

734times

10minus

21

Inc

)15

N14

N16

O21

958

02

197times

10minus

21

14N

14N

18O

21959

51

431times

10minus

21

QC

LA

SI

IIamp

III

300ndash

9000

021

877

ndash218

815

14N

14N

16O

21880

42

601times

10minus

2113

0a20

a53

3a

2100

104a

100

aW

aumlcht

eret

al

(CW

-QC

-TIL

DA

S-SC

-D

and

14N

15N

16O

21879

43

294times

10minus

21(2

008)

Aer

odyn

e22

031

ndash220

34

15N

14N

16O

21878

53

274times

10minus

21H

arri

set

al

Res

earc

h14

N14

N18

O22

032

81

794times

10minus

21(2

014)

Inc

)

TR

EX

-QC

LA

SI

300ndash

1500

ab

2203

1ndash2

203

4a14

N14

N16

O22

031

02

710times

10minus

21ndash

b20

a35

6a

620

20a

100

aIb

raim

etal

(20

18)

(mod

ified

22031

11

435times

10minus

21

CW

-QC

-TIL

DA

S-76

-CS

14N

15N

16O

22033

69

798times

10minus

22

Aer

odyn

eR

esea

rch

15N

14N

16O

22032

07

016times

10minus

22

Inc

)14

N14

N18

O22

032

81

794times

10minus

21

aD

ualQ

C-T

ILD

AS

and

min

iQC

-TIL

DA

Sar

efle

xibl

esp

ectr

omet

erpl

atfo

rms

whi

chca

nbe

used

with

diff

eren

tpar

amet

erse

tting

sT

hein

dica

ted

num

bers

wer

ech

osen

fort

hede

scri

bed

expe

rim

ents

bT

hem

iniQ

C-T

ILD

AS

spec

trom

eter

isus

edin

com

bina

tion

with

apr

econ

cent

ratio

nde

vice

(Ibr

aim

etal

20

18)

the

indi

cate

dN

2Oco

ncen

trat

ion

rang

eis

prio

rto

prec

once

ntra

tion

cT

hepr

econ

cent

ratio

nndash

min

iQC

-TIL

DA

Ssy

stem

isus

edin

are

petit

ive

batc

hcy

cle

with

outa

cont

inuo

ussa

mpl

ega

sflo

w(I

brai

met

al

2018

201

9)







214 TREX-QCLAS






215 GC-IRMS















b CH4b COb N2Ob n


Matrix gases


O2a Ara CO2

a CH4a COa N2Oa n

() () () (ppm) (ppm) (ppb)





















minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)


x

(1δ(1[x]GδmeasG

))(2)

and


2 (3)





)+δrefS1 (4)















Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References






























Tabl

e1

Ove

rvie

wof

the

wav

elen

gth

regi

ons

line

posi

tions

and

line

stre

ngth

sof

N2O

isot

opoc

ules

and

key

oper

atin

gpa

ram

eter

sfo

reac

hla

sers

pect

rom

eter

test

edin

this

stud

yN

TP

stan

dsfo

rnor

mal

tem

pera

ture

and

pres

sure

Det

ectio

nsc

hem

eN

2OW

aven

umbe

rIs

otop

ocul

esL

ine

posi

tions

Flow

rate

Cel

lC

ell

Inte

rnal

Eff

ectiv

eM

easu

rem

ent

Ref

eren

ces

(mod

el

rang

ere

gion

(cmminus

1 )li

ne(c

m3

minminus

1 )te

mpe

ratu

repr

essu

repl

umbi

ngvo

lum

efr

eque

ncy

man

ufac

ture

r)(p

pb)

(cmminus

1 )st

reng

th(c

mminus

1 (

C)

(hPa

)vo

lum

eat

NT

P(s

)(m

olec

ule

cmminus

2 ))

(cm

3 )(c

m3 )

OA

-IC

OS

I30

0ndash10

000

021

921

ndash219

25

14N

14N

16O

21924

04

919times

10minus

2060

436

6193

060

50

100

Bae

reta

l(2

002)

(N

2OIA

-30e

-EP

21924

44

919times

10minus

20A

BB

ndashL

osG

atos

AB

Bndash

Los

Gat

os21

924

83

375times

10minus

19R

esea

rch

Inc

Res

earc

h14

N15

N16

O21

923

13

31times

10minus

21(2

019)

Inc

)15

N14

N16

O21

923

32

968times

10minus

21

14N

14N

18O

21921

31

113times

10minus

21

CR

DS

IampII

300ndash

1500

2195

7ndash2

196

314

N14

N16

O21

962

15

161times

10minus

2025

2(C

RD

SI)

402

100

404

223

41(C

RD

SI)

Pica

rro

(G51

31-i

21

962

45

161times

10minus

2012

5(C

RD

SII

)2

54(C

RD

SII

)In

c(2

019)

Pica

rro

14N

15N

16O

21957

62

734times

10minus

21

Inc

)15

N14

N16

O21

958

02

197times

10minus

21

14N

14N

18O

21959

51

431times

10minus

21

QC

LA

SI

IIamp

III

300ndash

9000

021

877

ndash218

815

14N

14N

16O

21880

42

601times

10minus

2113

0a20

a53

3a

2100

104a

100

aW

aumlcht

eret

al

(CW

-QC

-TIL

DA

S-SC

-D

and

14N

15N

16O

21879

43

294times

10minus

21(2

008)

Aer

odyn

e22

031

ndash220

34

15N

14N

16O

21878

53

274times

10minus

21H

arri

set

al

Res

earc

h14

N14

N18

O22

032

81

794times

10minus

21(2

014)

Inc

)

TR

EX

-QC

LA

SI

300ndash

1500

ab

2203

1ndash2

203

4a14

N14

N16

O22

031

02

710times

10minus

21ndash

b20

a35

6a

620

20a

100

aIb

raim

etal

(20

18)

(mod

ified

22031

11

435times

10minus

21

CW

-QC

-TIL

DA

S-76

-CS

14N

15N

16O

22033

69

798times

10minus

22

Aer

odyn

eR

esea

rch

15N

14N

16O

22032

07

016times

10minus

22

Inc

)14

N14

N18

O22

032

81

794times

10minus

21

aD

ualQ

C-T

ILD

AS

and

min

iQC

-TIL

DA

Sar

efle

xibl

esp

ectr

omet

erpl

atfo

rms

whi

chca

nbe

used

with

diff

eren

tpar

amet

erse

tting

sT

hein

dica

ted

num

bers

wer

ech

osen

fort

hede

scri

bed

expe

rim

ents

bT

hem

iniQ

C-T

ILD

AS

spec

trom

eter

isus

edin

com

bina

tion

with

apr

econ

cent

ratio

nde

vice

(Ibr

aim

etal

20

18)

the

indi

cate

dN

2Oco

ncen

trat

ion

rang

eis

prio

rto

prec

once

ntra

tion

cT

hepr

econ

cent

ratio

nndash

min

iQC

-TIL

DA

Ssy

stem

isus

edin

are

petit

ive

batc

hcy

cle

with

outa

cont

inuo

ussa

mpl

ega

sflo

w(I

brai

met

al

2018

201

9)







214 TREX-QCLAS






215 GC-IRMS















b CH4b COb N2Ob n


Matrix gases


O2a Ara CO2

a CH4a COa N2Oa n

() () () (ppm) (ppm) (ppb)





















minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)


x

(1δ(1[x]GδmeasG

))(2)

and


2 (3)





)+δrefS1 (4)















Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References






214 TREX-QCLAS






215 GC-IRMS















b CH4b COb N2Ob n


Matrix gases


O2a Ara CO2

a CH4a COa N2Oa n

() () () (ppm) (ppm) (ppb)





















minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)


x

(1δ(1[x]GδmeasG

))(2)

and


2 (3)





)+δrefS1 (4)















Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




215 GC-IRMS















b CH4b COb N2Ob n


Matrix gases


O2a Ara CO2

a CH4a COa N2Oa n

() () () (ppm) (ppm) (ppb)





















minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)


x

(1δ(1[x]GδmeasG

))(2)

and


2 (3)





)+δrefS1 (4)















Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




b CH4b COb N2Ob n


Matrix gases


O2a Ara CO2

a CH4a COa N2Oa n

() () () (ppm) (ppm) (ppb)





















minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)


x

(1δ(1[x]GδmeasG

))(2)

and


2 (3)





)+δrefS1 (4)















Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References









minus

sumx

(1 [N2O]

(1[x]G [N2O]measG

))(1)


x

(1δ(1[x]GδmeasG

))(2)

and


2 (3)





)+δrefS1 (4)















Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References











Pressurized air












TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References









TREX-QCLAS I




functions










241 Allan precision







243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




243 Repeatability





























3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




















3 Results


31 Allan precision
















(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References








(r2= 085) and δ15Nβ (r2

= 096)

33 Repeatability













(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




(300 s) (600 s) (300 s) (600 s) (300 s) (600 s) (300 s) (600 s)

3265 ppb N2O


sim 1000 ppb N2O


sim 10000 ppb N2O




















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References
















)=

(A middot [N2O]2


)middot1[O2]A (5)


)=

(a middot [N2O]2





[N2O]mcG

=





δmcG = δmeasG

minus

(a middot [N2O]mcG


)middot1[O2]G (8)



















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References


















)=

(Ax

1[N2O]expmix

+Bx

)middot1[x]A (9)


)=

(ax middot

1[N2O]expmix

+ bx

)middot1[x]A (10)





minus

sumx

((Ax

1[N2O]measG

+Bx

)middot1[x]G

)(11)

δtcG = δmeasG

minus

sumx

((ax

1[N2O]measG

+ bx

)middot1[x]G

) (12)






OA-ICOS I









CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References







CRDS I


CRDS II



QCLAS I






TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




TREX-QCLAS I















4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References












4 Discussion












































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References





































5 Conclusions



















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References
















References

































































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References























































































































Abstract

Introduction




CRDS (Picarro Inc)


TREX-QCLAS

GC-IRMS





Laboratory setup



Allan precision

Temperature effects

Repeatability






Results

Allan precision

Temperature effects

Repeatability











Discussion





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References

n o isotopocule measurements using laser …...revised: 15 april 2020 – accepted: 21 april 2020...

Documents