A compact incoherent broadband cavity-enhanced absorption ...

Atmos Meas Tech 13 4317ndash4331 2020httpsdoiorg105194amt-13-4317-2020copy Author(s) 2020 This work is distributed underthe Creative Commons Attribution 40 License

A compact incoherent broadband cavity-enhanced absorptionspectrometer for trace detection of nitrogen oxides iodine oxide andglyoxal at levels below parts per billion for field applicationsAlbane Barbero Camille Blouzon Joeumll Savarino Nicolas Caillon Aureacutelien Dommergue and Roberto GrilliUniversiteacute Grenoble Alpes CNRS IRD Grenoble INP (Institute of Engineering Universiteacute Grenoble Alpes)IGE 38000 Grenoble France

Correspondence Roberto Grilli (robertogrillicnrsfr)

Received 25 March 2020 ndash Discussion started 2 April 2020Revised 9 July 2020 ndash Accepted 14 July 2020 ndash Published 17 August 2020

Abstract We present a compact affordable and robust in-strument based on incoherent broadband cavity-enhancedabsorption spectroscopy (IBBCEAS) for simultaneous de-tection of NOx IO CHOCHO and O3 in the 400ndash475 nmwavelength region The instrument relies on the injectionof a high-power LED source in a high-finesse cavity (F sim33100) with the transmission signal being detected by acompact spectrometer based on a high-order diffraction grat-ing and a charge-coupled device (CCD) camera A minimumdetectable absorption of 20times 10minus10 cmminus1 was achievedwithin sim 22 min of total acquisition corresponding to a fig-ure of merit of 18times 10minus10 cmminus1 Hzminus12 per spectral ele-ment Due to the multiplexing broadband feature of the setupmulti-species detection can be performed with simultaneousdetection of NO2 IO CHOCHO and O3 achieving detec-tion limits of 11 03 10 ppt (parts per trillion) and 47 ppb(parts per billion) (1σ ) within 22 min of measurement re-spectively (half of the time is spent on the acquisition of thereference spectrum in the absence of the absorber and theother half is spent on the absorption spectrum) The imple-mentation on the inlet gas line of a compact ozone generatorbased on electrolysis of water allows for the measurement ofNOx (NO+NO2) and therefore an indirect detection of NOwith detection limits for NOx and NO of 10 and 21 ppt (1σ )respectively The device has been designed to fit in a 19 in3U (525 in) rack-mount case weighs 15 kg and has a totalelectrical power consumption of lt 300 W The instrumentcan be employed to address different scientific objectivessuch as better constraining the oxidative capacity of the at-mosphere studying the chemistry of highly reactive speciesin atmospheric chambers as well as in the field and looking at

the sources of glyoxal in the marine boundary layer to studypossible implications on the formation of secondary aerosolparticles

1 Introduction

Free radicals are controlling the oxidative capacity of theatmosphere and therefore contribute to the upholding of itschemical balance With their unpaired valence electron theyare highly chemically reactive and are therefore consideredthe ldquodetergentsrdquo of the atmosphere (Monks 2005 Monkset al 2009) Even if present at extremely low concentra-tions radicals are constantly formed by photochemical andcombustion processes They may be removed from the atmo-sphere by biological uptakes dry and wet deposition andchemical reactions (Finlayson-Pitts and Pitts 2000) Freeradicals in the troposphere such as nitrogen oxides (NOx)hydroxyl radical (OH) peroxy radicals (HO2 and RO2) andhalogen oxides (BrO and IO) can be found at mixing ra-tios (ie mole fractions) ranging from less than one part pertrillion (10minus12 mol molminus1 or ppt) up to a few parts per mil-lion (10minus6 mol molminus1 or ppm) in the atmosphere (Wine andNicovich 2012) Measuring their concentration and dynamicvariability in different atmospheric environments is key toaddressing specific questions regarding air quality the ox-idative state of the atmosphere the ozone budget aerosol nu-cleation and carbon nitrogen and sulfur cycles The under-standing of the complex interactions involving those specieshas led to numerous investigations during the past decades

Published by Copernicus Publications on behalf of the European Geosciences Union

4318 A Barbero et al IBBCEAS for trace detection of NOx CHOCHO and IO at sub-ppb levels

Especially nitrogen oxides (NOx=NO and NO2) have a di-rect impact on air quality and climate change In the presenceof volatile organic compounds (VOCs) and under solar radi-ation nitrogen oxides stimulate ozone (O3) formation in thetroposphere NOx also plays an important role in rain acid-ification and ecosystem eutrophication by its transformationin nitric acid (HNO3) (Jaworski et al 1997 Vitousek et al1997) Finally NOx contributes to the formation of particu-late matter in ambient air and to the aerosol formation leadingto cloud formation (Atkinson 2000) The NO2 mixing ratioin the troposphere ranges from a few tens of ppt in remote ar-eas to hundreds of parts per billion (10minus9 mol molminus1 or ppb)in urban atmospheres (Finlayson-Pitts and Pitts 2000) Be-ing able to measure such species in situ at low levels andat a timescale compatible with its reactivity (ie in min) ischallenging and puts stringent constraints on the instrumentsensitivity time response energy consumption and compact-ness Among the various techniques that have so far been de-veloped chemiluminescence detection (CLD) (Maeda et al1980 Ryerson et al 2000) long-path differential optical ab-sorption spectroscopy (LP-DOAS) (Lee et al 2005 2008Pikelnaya et al 2007) and multi-axis differential opticalabsorption spectroscopy (MAX-DOAS) (Platt and Perner1980 Sinreich et al 2007 Wagner et al 2010) have beenused to detect nitrogen species and halogen oxides The CLDtechnique using the chemiluminescence reaction occurringbetween O3 and NO after the reduction of NO2 into NOis widely used for air quality measurements with sensitivi-ties better than 100 ppt (Ryerson et al 2000) Neverthelessthe interferences in the reduction of NO2 to NO with otherspecies (ie HONO and HNO3) and the sensitivity to envi-ronmental conditions (temperature and humidity) leave un-certainties about absolute mixing ratio measurements (Gros-jean and Harrison 1985 Williams et al 1998) The MAX-DOAS technique has been used to measure BrO and NO2by making use of the characteristic absorption features ofgas molecules along a path in the open atmosphere (Leseret al 2003) Although MAX-DOAS is relatively simple todeploy the data analysis makes it a complex approach for insitu field measurements due to the influence of clouds on theradiative transfer which alters the path length of light (Wit-trock et al 2004 Rozanov and Rozanov 2010) While inLP-DOAS the optical path length is known the signal degra-dation due to the environment (clouds rain and wind) re-mains of importance for the data retrieval and results areintegrated over the long path leading to a limited spatialresolution (Chan et al 2012 Poumlhler et al 2010) Com-pact highly sensitive and point-source measurements maybe achieved using cavity enhanced techniques such as cav-ity ring-down spectroscopy (CRDS) and cavity-enhanced ab-sorption spectroscopy (CEAS) (Atkinson 2003) The po-tential of the CRDS for accurate sensitive and rapid mea-surements in a compact and transportable instrument has al-ready been demonstrated (Fuchs et al 2009 Brown et al2002) eg Fuchs et al (2009) reached a sensitivity of 22 ppt

for NO2 within 1 s of integration time using the CRDStechnique Incoherent broadband cavity-enhanced absorptionspectroscopy (IBBCEAS) (Fiedler et al 2003) is a simpleand robust technique for in situ field observations Differentsources and wavelength regions have been used for the de-tection of NO2 leading to different levels of performanceVenables et al (2006) were able to detect simultaneouslyNO3 NO2 O3 and H2O in an atmospheric simulation cham-ber with a sensitivity of tens of ppb for NO2 Gherman et al(2008) reached sim 013 and sim 038 ppb for HONO and NO2in a 4 m3 atmospheric simulation chamber between 360 and380 nm Triki et al (2008) used a red LED source centredat 643 nm reaching a sensitivity of 5 ppb Langridge et al(2006) developed an instrument with a blue-light-emittingdiode (LED) centred at 445 nm allowing for detection lim-its ranging from 01 to 04 ppb Ventrillard-Courtillot et al(2010) reached 600 ppt detection limit for NO2 with an LEDcentred at 625 nm and Thalman and Volkamer (2010) re-ported a detection limit of 30 ppt within 1 min of integra-tion time More recently Min et al (2016) proved a preci-sion of 80 ppt in 5 s of integration at 455 nm using a spec-trometer with a thermoelectric-cooled charge-coupled device(CCD) camera and very high-reflectivity mirrors This non-exhaustive list of works underline the need of robust com-pact and transportable instruments also allowing for directmulti-species detection and low detection limits for appli-cations in remote areas such as Antarctica where the ex-pected mixing ratio of NO2 could be as low as a few tens ofppt Fuchs et al (2010) during the NO3Comp campaign atthe SAPHIR (Simulation of Atmospheric PHotochemistry Ina Large Reaction Chamber) atmospheric simulation cham-ber demonstrated the potential of these optical techniquesto compete with the CLD instruments as routine measure-ments of NO2 concentrations in the future The present pa-per describes a compact and affordable instrument based onthe IBBCEAS technique allowing for the simultaneous de-tection of nitrogen dioxide iodine oxide glyoxal and ozone(NO2 IO CHOCHO and O3) with detection limits of 1103 10 ppt and 47 ppb (1σ ) respectively for a measurementtime of 22 min (half of the time is spent on the acquisitionof the reference spectrum in the absence of the absorberand the other half is spent on the absorption spectrum) Thefour species are directly detected by a broadband blue-light-emitting diode centred at 445 nm The wavelength region wasselected in order to optimize the detection of NO2 Direct de-tection of NO is only possible in the ultraviolet region forwavelengths around 226 nm (Dooly et al 2008) or in themid-infrared region at 53 microm (Richard et al 2018) wave-lengths difficult to achieve with LED technology Here anindirect measurement is proposed which relies on the oxi-dation of NO to NO2 under a controlled excess of O3 Thesum of NO and NO2 is therefore measured leading to a sup-plemental indirect measurement of NO if the concentrationof NO2 is also monitored The field deployment for the mea-surements of NO2 and NOx consists of two twin instruments

Atmos Meas Tech 13 4317ndash4331 2020 httpsdoiorg105194amt-13-4317-2020

A Barbero et al IBBCEAS for trace detection of NOx CHOCHO and IO at sub-ppb levels 4319

IBBCEAS-NO2 and IBBCEAS-NOx the later equipped withan ozone generator system

2 Method

In IBBCEAS a broadband incoherent light source is cou-pled to a high-finesse optical cavity for trace gas detectionA picture of the spectrometer and a schematic diagram ofthe setup is shown in Fig 1 In the present study the broad-band light source consisted of a high-power LED LuminusSBT70 allowing sim 1 W of optical power to be injected intothe resonator A thermoelectric (TEC) Peltier cooler (ET-161-12-08-E) and a fanndashheat-sink assembly were used to di-rectly evacuate outside of the instrument up to sim 75 W ofthermal heat from the LED A temperature regulator (RKCRF100) with a PT100 thermistor was used to stabilize theLED temperature atplusmn01 C The LED spectrum was centredat 445 nm with 19 nm FWHM (full width at half maximum)which covers the main absorption features of NO2 IO CHO-CHO and O3 For better collimation of the LED spatially di-vergent emission (7 mm2 surface) a dedicated optic (LEDiLHEIDI RS) was used and coupled with a 25 mm focal lens(L1 Thorlabs LA1951-A) The high-finesse optical cavitywas formed by two 05 in diameter high-reflectivity mirrors(maximum reflectivity at 450 nm gt99990plusmn 0005 LAY-ERTEC 109281) separated by a 417 cm long PFA (perfluo-roalkoxy alkane) tube (14 mm internal diameter 1 mm thick)held by an external stainless-steel tube Both mirrors werepre-aligned and glued with Torr Seal epoxy glue on remov-able stainless-steel supports which were then screwed on thecavity holders This enables the easy cleaning of the mirrorswhen required and also the removal of the cavity tube to per-form open-cavity measurements which is of interest for thedetection of the highly reactive IO radical Behind the cav-ity a Thorlabs FB450-40 filter was used in order to removethe broadband component of the radiation sitting outside thehighly reflective curve of the cavity mirrors The radiationis focused on an optical fibre (FCRL-7UV100-2-SMA-FC)using a 40 mm focal lens (L2 Thorlabs LA1422-A) The op-tical fibre input was composed of 7 cores in a round shapepattern on the collecting side whereas at the fibre end onthe spectrometer side the cores were assembled in a linefor better matching the 100 microm slit at the spectrometer Thespectrometer (Avantes AvaSpec-ULS2048L) was composedof a diffraction grating (2400 lines mmminus1) and 2048-pixelcharge-coupled device (CCD) The resolution of the spec-trometer was 054plusmn 010 nm All the optics including thecavity were mounted on a Z-shaped 8 mm thick aluminiumboard fixed on the rack using cylindrical dampers (Paulstra)On the board four 5 W heating bands and one PT100 sensorwere glued and a second RKC module was used to regu-late its temperature The board therefore acts as a large radi-ator inside the instrument allowing for the minimization ofinternal thermal gradients and thermalization of the instru-

ment Air circulation from outside is ensured by an apertureat the front and a fan placed at the back wall of the instrument(Fig 1) The gas line system was composed of a manual PFAneedle valve MV and a three-way two-position PTFE (poly-tetrafluoroethylene) valve V (NResearch 360T032) at theentrance while a proportional valve PV (Burkert 239083)a flowmeter F (Honeywell HAFUHT0010L4AXT) a pres-sure sensor P (STS ATMECO accuracy plusmn02 ) and a di-aphragm pump (KNF N 816 AV12DC-B) were placed af-ter the cavity The entire line was made of 025 in PFA tub-ing which was found to be least lossy for the transport ofhighly reactive species (Grilli et al 2012) The pump pro-vided a constant flow that can reach 11 L minminus1 at the end ofthe gas line while a constant pressure in the cavity was ob-tained by a PID (proportionalndashintegralndashderivative) regulatoron the proportional valve A data acquisition card (NationalInstruments USB 6000) was interfaced to read the analoguesignal from the pressure sensor while a microcontroller (Ar-duino Due) derived the proportional valve The manual valveat the entrance allowed for tuning the flow rate At the in-let a three-way two-position PTFE valve allowed to switchbetween the gas sample and zero-air mixture for acquiring areference spectra in the absence of absorption Zero air wasproduced by flowing outdoor air through a filtering system(TEKRAN 90-25360-00 Analyzer Zero Air Filter) Particlefilters (Whatmanreg PTFE membrane filters ndash TE 38 5 microm47 mm) were also placed in the inlet lines (reference andsample) for preventing optical signal degradation due to Miescattering as well as a degradation of the mirror reflectiv-ity for long-term deployment The air flow was introducedat the centre of the cavity and extracted at both ends of thecavity The optimal cavity design was selected by runningSolidWorks flow simulations at flow rates between 05 and1 L minminus1 Cavity mirrors were positioned in order to maxi-mize the sample length d and to avoid localized turbulencesin front of the mirrors (see Sect S11 in the Supplement)The LEDrsquos power is monitored over time using a photodiodePD (Hamamatsu S1223-01) allowing for discrimination ifa decreasing of intensity at the CCD (during the acquisitionof the spectra in the absence of the absorber) is due to a de-creasing of the LED intensity or to the mirrorsrsquo cleanlinessdegradation (for more details see Sect S12) All the compo-nents fit in a 19 in 3U (525 in) aluminium rack-mount casehave a total weight of 15 kg and have a total electrical powerconsumption oflt 300 W The instrument interface measure-ments and data analysis are performed automatically withoutthe intervention of an operator by dedicated LabVIEW soft-ware Instrument calibrations however must be performedby an operator on a regular basis

3 Spectral fit

The absorption spectrum is calculated as the ratio betweenthe spectrum of the light transmitted through the cavity with-

httpsdoiorg105194amt-13-4317-2020 Atmos Meas Tech 13 4317ndash4331 2020


Figure 1 (a) A picture of the instrument mounted on a 19 in 3U rack-mount case (b) Schematic of the instrument The LED is protectedby a cap in which a photodiode (PD) is monitoring its power The light from the LED is collimated by lens L1 and injected into the cavityThe exiting light is then collimated with lens L2 and injected into the spectrometer M1 and M2 are steering mirrors and F is an optical filterThe gas line is composed of a pump a pressure sensor P a flowmeter FM and a proportional valve PV At the inlet a three-way two-positionvalve in PTFE V is used to switch between the sample and zero air A manual PFA needle valve MV is used to fix the flow rate An ozonizercan be inserted in the inlet line for NOx measurements

out a sample I0(λ) and with a sample in the cavity I (λ) Itis expressed as the absorption coefficient (in units of cmminus1)by the following equation (Fiedler et al 2003)

α(λ)=

(I0(λ)

I (λ)minus 1

)(1minusR(λ)

d

) (1)

where R(λ) is the wavelength-dependent mirror reflectivityand d the length of the sample inside the cavity Equation (1)is derived from the BeerndashLambert law and applied to light inan optical resonator (Ruth et al 2014) The light transmittedthrough the optical cavity is attenuated by different processessuch as absorption reflection and scattering of the mirrorsubstrates and coating as well as losses due to the mediuminside the cavity The losses of the cavity mirrors are as-sumed to be constant between the acquisition of the referenceand the sample spectrum Mie scattering is minimized witha 5 microm particle filter in the gas inlet while Rayleigh scatter-ing losses were calculated to be 255times10minus7 cmminus1 at 445 nmat 25 C and 1 atm (Kovalev and Eichinger 2004) and thusnegligible with respect to the cavity losses normalized bythe cavity length

(1minusRd= 228times 10minus6 cmminus1

) Therefore the

light transmitted through the cavity is mainly affected by theabsorption of the gas species which leads to well-definedabsorption spectral features αi(λ) that are analysed in realtime by a linear multicomponent fit routine Experimental ab-sorption spectra of the species i (i = NO2 IO CHOCHOand O3) have been compared with literature cross-sectiondata accounted for the gas concentration and experimentalconditions of temperature and pressure and convoluted withthe spectrometer instrumental function Those experimentalspectra are then used as reference spectra for the fit

α(λ)=sumi

σi(λ)ci +p(λ) (2)

A fourth order polynomial function p(λ)= a0+a1λ+a2λ2+

a3λ3+ a4λ

4 is added to the absorption coefficient Eq (2) toadjust the spectral baseline and account for small changes be-tween the reference and the sample spectra The transmitted-light intensity as well as the optical absorption path willbe modulated by the shape of the mirror reflectivity curveTherefore the later should be defined in order to retrieve thecorrect absorption spectrum recorded at the cavity output

4 Calibration performance and multi-species detection

41 Calibration

Washenfelder et al (2008) described a procedure for retriev-ing the mirror reflectivity curve by taking advantage of a dif-ferent Rayleigh scattering contribution to the cavity losseswhile the measuring cell was filled with different bulk gases(eg helium versus air or nitrogen) The Rayleighrsquos empir-ical cross sections used by Min et al (2016) and the the-oretical ones used by Thalman and Volkamer (2010) werefound to disagree leading this work to propose an eas-ier approach consisting of using trace gases at known con-centrations (in this case NO2 produced by the calibratorFlexStreamtrade Gas Standards Generator from KIN-TEK An-alytical Inc and CHOCHO produced by evaporating a gly-oxal solution 40 in water from Sigma-Aldrich since theO3 spectrum is less structured and IO is highly reactive)and their literature cross sections (Vandaele et al 1998 forNO2 and Volkamer et al 2005 for CHOCHO) for retriev-ing the wavelength-dependent reflectivity curve Such an ap-proach to calculate mirror reflectivity has been proposed be-fore (Venables et al 2006) and has been used by previousstudies (Duan et al 2018) The shape of the reflectivity curve



is adjusted to best match the convoluted literature NO2 spec-trum at the concentration given by the KIN-TEK calibrator(for more information see Sect S2) Figure 2a shows theresulting reflectivity curve and the transmitted light throughthe cavity and the optical filter The maximum reflectivityachieved with both mirrors given by the calibration proce-dure is 999905 leading to an effective optical path lengthof sim 44 km and a cavity finesse

(F = π

radicR

(1minusR)

)of sim 33100

While the shape of the mirror reflectivity curve is determinedonce and for all its offset is slightly adjusted after each mir-rorsrsquo cleaning by flushing in the cavity a known concentra-tion of NO2 The spectral emission of the LED centred at445 nm is well suited also for the detection of IO CHOCHOand O3 which are other key species in atmospheric chem-istry For the field measurements of NO2 and NOx two twininstruments named IBBCEAS-NO2 and IBBCEAS-NOx aredeployed with the later equipped with an ozone generatoron the gas inlet line At this wavelength region water vapouralso absorbs and is accounted for in the spectral-fit analysisHowever the absorption of oxygen dimer is not required inthe fit routine since the absorption feature will be presentin the reference (I0) as well as in the absorption (I ) spectraIn Fig 2b simultaneous detection of NO2 IO and O3 is re-ported Ozone at 289 ppm is produced by water electrolysisas described in Sect 44 1918 ppb of NO2 is provided by apermeation tube and 4256 ppt of IO is generated by photo-chemical reaction of sublimated iodine crystals and ozone inthe presence of radiation inside the cavity For this spectrumthe light transmitted is integrated for 350 ms on the CCD andaveraged over 1000 spectra yielding to a 1σ standard devi-ation of the residuals of 4times 10minus8 cmminus1 (Fig 2b bottom) InFig 2c simultaneous detection of NO2 CHOCHO and H2Ois reported CHOCHO at 43 ppb is produced by evaporat-ing a glyoxal solution (40 in water Sigma-Aldrich) at thesample gas inlet of the instrument NO2 at 140 ppb and H2Oat 054 are ambient concentrations observed in the lab-oratory during the experiment For this spectrum the lighttransmitted is integrated for 250 ms on the CCD and aver-aged over 1000 spectra yielding to a 1σ standard deviationof the residuals of 5times10minus9 cmminus1 (Fig 2c bottom) The top ofFig 2b and c shows the experimental spectra (black traces)the fit result (red traces) and contributions from each species(middle) which are included in the spectral fit The concen-trations of the species are retrieved with respect to the lit-erature cross sections using the calibrated reflectivity curvediscussed above

42 Instrumental calibration and intercomparison

A calibrator (FlexStreamtrade Gas Standards Generator KIN-TEK Analytical Inc) was used to produce a stable NO2source The sample was produced using a permeation tube ofNO2 (KIN-TEK ELSRT2W) calibrated at an emission rateof 115 ng minminus1 at 40 C loaded into the calibrator This type

of calibrator is ideally suited for creating trace concentrationmixtures (from ppt to ppm) The instruments were calibratedagainst the calibrator delivering NO2 at 4959plusmn 017 ppbTo confirm the calibration process as well as the stabilityof the instrument within a greater range of concentrationstwo intercomparisons of the IBBCEAS with two differentCLD instruments (Thermo Fishertrade 42i NOx analyser andThermo Fishertrade 42iTL NOx trace analyser) were performedin outdoor air over 39 and 12 h respectively Results are re-ported in Fig 3 The experiments took place at the Insti-tute of Environmental Geosciences (IGE) in Saint-Martin-drsquoHegraveres France The IGE is located in the university cam-pus sim 1 km from the city centre of Grenoble and sim 300 mfrom a highway Ambient air was pumped simultaneouslyfrom the same gas line by the instruments at flow rates of10 and 05 L minminus1 for the IBBCEAS and the CLD instru-ment (Thermo Fishertrade 42i NOx analyser) respectively Themeasurements were conducted in September 2018 On theevening of Saturday 29 September the NO2 peak occurredat a slightly later time than normally expected (from 2000to midnight Fig 3a) This may be due to the fact that on aSaturday night urban traffic can be significant until late butit also may be due to severe weather conditions prevailing atthis time with a storm and lightning known to be a major nat-ural source of NOx (Atkinson 2000) For the second experi-ment shown in Fig 3b ambient air was pumped at flow ratesof 10 and 08 L minminus1 for the IBBCEAS and the CLD traceinstrument (Thermo Fishertrade 42iTL NOx trace analyser) re-spectively The measurements were conducted in July 2019Both instruments showed the expected variability of an ur-ban environment with an increase of NO2 in the evening andmorning due to photochemical processes and anthropogenicactivities (ie mainly urban traffic)

The correlation plot based on data of all instruments inFig 4a shows good linearity with a slope of 1088plusmn 0005and a correlation coefficient of R2

= 0989 with measure-ments averaged over 5 min In order to perform linearitytests the previous NO2 FlexStreamtrade calibrator was usedto produced various concentrations of NO2 covering a largerange of concentrations Figure 4b shows the good linear-ity from the ppt to ppb range of the IBBCEAS instrumentwith a slope of 1015plusmn 0006 and a correlation factor ofR2= 09996 confirming the validity of the calibration ap-

proach The discrepancies observed between the IBBCEASand the CLD techniques might be explained by positive andnegative interferences on the CLD technique While the sys-tem presented here measures NO2 directly the CLD tech-nique applies an indirect measurement of NOx from the ox-idation of NO through a catalyser then in CLD the NO2mixing ratio is obtained by the subtraction of the NO sig-nal to the total NOx signal Villena et al (2012) demonstratethat the interferences on an urban atmosphere for the CLDtechnique implied positive interferences when NOy speciesphotolysis occurred leading to an overestimation of daytimeNO2 levels while negative interferences were attributed to



Figure 2 (a) The mirror reflectivity curve (red) in comparison with the spectrum of the LED light transmitted by the cavity and the opticalfilter for a single acquisition of 250 ms (b) In black an example of an experimental spectrum of NO2 IO and O3 at concentrations of1918 ppb 4256 ppt and 289 ppm respectively in red the multi-species spectral fit and in blue orange and grey the absorptions of thedifferent species At the bottom (in black) the residual of the experimental fit with a 1σ standard deviation of 4times 10minus8 cmminus1 after 1000averages (c) In black an example of an experimental spectrum of NO2 CHOCHO and H2O at concentrations of 140 ppb 43 ppb and054 respectively in red the multi-species spectral fit and in blue orange and green the absorptions of the different species At thebottom (in black) the residual of the experimental fit with a 1σ standard deviation of 5times 10minus9 cmminus1 after 1000 averages

the VOCs photolysis followed by peroxy radical reactionswith NO

43 Instrument sensitivity and long-term stability

In remote areas such as East Antarctica NO2 ranges from afew tens to a few hundreds of ppt (50ndash300 ppt) (Frey et al2013 2015) Due to the low signal-to-noise ratio of thespectrometer a single acquired spectrum (with an integra-tion time ranging between 200 and 350 ms) does not pro-vide the required detection limit However the sensitivitycan be improved by averaging the measurements for longertimes over which the instrument is stable The stability ofthe IBBCEAS system is mainly affected by temperature fluc-tuations mechanical instabilities and pressure drifts In or-der to characterize the long-term stability of the instrument

two different studies were conducted on the IBBCEAS-NO2during the Antarctica field campaign at Dome C in 2019ndash2020 (the same results for the IBBCEAS-NOx can be foundin the Supplement) For both studies the light transmit-ted through the cavity (I ) was integrated at the CCD for250 ms providing a signal-to-noise ratio of 110 for a singlespectrum The reference spectrum (I0) was taken by aver-aging 2000 individual spectra (sim 8 min) while flushing thecavity with zero air Subsequently a 9 h long time serieswas recorded for each instrument maintaining the zero-airflow The instrument was regulated at 120plusmn 02 C witha cavity pressure of 6300plusmn 07 mbar and a gas flow of102plusmn 011 L minminus1 The minimum absorption coefficient(αmin) corresponding to the standard deviation of the resid-ual of the spectrum was deduced for different time aver-



Figure 3 (a) A 39 h long intercomparison of the IBBCEAS instrument and the commercial CLD instrument (Thermo Fishertrade 42i analyser)on the NO2 detection in an outdoor urban area performed in September 2018 (dates in the format of two-digit yearmonthday) The plotreports continuous (dashed lines) and hourly (dots) average data for both techniques The grey area corresponds to the nighttime period (b) A12 h long intercomparison of the IBBCEAS instrument and the commercial CLD trace instrument (Thermo Fishertrade 42iTL trace analyser)on the NO2 detection in an outdoor urban area performed in July 2019 The plot reports continuous (dashed lines) and 30 min (dots) averagedata for both techniques The grey area corresponds to the nighttime period

Figure 4 (a) A linear correlation was obtained with a slope of 1088plusmn 0005 and a correlation coefficient of R2= 0989 between the

IBBCEAS system and the Thermo Fisher instruments (b) Results of the systemrsquos calibration using an NO2 FlexStreamtrade calibrator Alinear correlation was obtained with a slope of 1015plusmn 0005 and a correlation coefficient of R2

= 09996



Figure 5 The minimum absorption coefficient αmin versus thenumber of spectral average for the IBBCEAS-NO2 instrument Forthese measurements the cell was continuously flushed with a flowof 102 L minminus1 of zero air and the αmin was calculated from thestandard deviation of the residual of the spectra at different timeaverages

ages The results are shown in the logndashlog plot of Fig 5where the dots are the data and the dashed line indicatesthe trend in the case of a pure white-noise regime Fromthe graph one can see that the instrument follows the white-noise trend for about 22 min (5200 averages) afterwards thebaseline noise starts to deviate due to the rise of frequency-dependent noise The chosen αmin value corresponds to 20times10minus10 cmminus1 within sim 22 min (5200 spectra) of measurementduring which a reference spectrum in the absence of ab-sorbers and the absorption spectrum are acquired The cor-responding figure of merit (noise-equivalent absorption sen-

sitivity ndash NEAS ndash or αmin(BW) = αmintimes

radictintM

)is therefore

18times 10minus10 cmminus1 Hzminus12 per spectral element (with tint be-ing the integration time ofsim 11 min andM being the numberof independent spectral elements here 800 spectral elementsare considered for the spectral fit)

For the same time series an AllanndashWerle (AW) statisti-cal method on the measured concentrations was employed(Werle et al 1993) In this case spectra were averaged ina block of 10 and analysed by the fit routine The resultsof the fit are reported (Fig 6a) For an acquisition time of25 s corresponding to 10 averaged spectra the AW stan-dard deviation σAW-SD was 175 48 125 ppt and 725 ppbfor NO2 IO CHOCHO and O3 respectively By increas-ing the integration time the σAW-SD decreased following thewhite-noise trend the coloured dashed line of Fig 6b witha characteristic

radicN slope (where N is the number of aver-

aged spectra) Because a reference spectrum in the absence

of absorbers is required by this CEAS technique depend-ing on the shape of the AW plot different strategies may befollowed In our case the AW trends continue to decreasefor all species for sim 22 min (5200 averages) this means thatone can spend 11 min acquiring the reference spectrum anda further 11 min for the absorption spectrum leading to lim-its of detection (LODs) of 11 03 10 ppt and 47 ppb (1σ )for NO2 IO CHOCHO and O3 respectively In our case wechose to divide the measurement times by 2 (ie sim 11 minand 2600 averages for acquiring both the reference and theabsorption spectra) offering equally interesting LODs 1604 12 ppt and 72 ppb for NO2 IO CHOCHO and O3 (1σ )respectively and allowing us to stay within the white-noiseregime

The long-term stability and repeatability of the instru-ment were further studied by taking regular reference spectrawithin the optimum integration time of the instrument whilecontinuously flushing the instrument with the zero-air mix-ture In this casesim 5 andsim 3 min intervals were chosen cor-responding to 1024 and 580 averages (for 350 and 300 msintegration time respectively) and a sensitivity on the NO2concentrations of 55 and 53 ppt (1σ ) respectively The re-sults are reported in Fig 7 Test1 (1024 averages) was runfor 9h and Test2 (580 averages) was run for 15 h Thesetests highlight the reliability of the measurement protocolwith the long-term measurement well distributed within the3σ level of the measurements precision (165 and 159 ppt re-spectively) A boxplot is also reported representing the aver-age values (green triangles) and medians quartiles and min-imum and maximum values The histogram analysis of thosetests show a good distribution of the measurements withinthe 3σ level

Table 1 shows a comparison between the instrument pre-sented in this work and other recently developed IBBCEASsystems The detection limits are given in ppt minminus1 (1σ )with the normalization time that accounts for the acquisi-tion of the reference (without absorption) and sample spec-tra to allow for a better comparison It should be noticedthat all the other developments took advantage of an opti-cal spectrometer with a cooled CCD device to reduce darknoise A more compact and affordable spectrometer was pre-ferred in this work The cooling at the CCD would allowfor a gain of up to a factor of 10 on the signal-to-noise ra-tio which would directly apply to the achievable detectionlimits Furthermore a CCD with a higher sensitivity wouldallow for selecting higher-reflectivity mirrors and increasethe optical pathlength Noteworthy the optimum integrationtime corresponding to a minimum of the σAW-SD is at 1300 s(sim 22 min) allowing for the achievement of low detectionlimits even without a cooled CCD

44 Indirect measurement of NO

Measuring NO and NO2 simultaneously is important to studythe NOx budget in the atmosphere In the selected blue vis-



Figure 6 (a) Mixing ratios of the target species NO2 O3 IO and CHOCHO measured during a 9 h AllanndashWerle variance statistical experi-ment flowing zero air through the cavity on the IBBCEAS-NO2 instrument (b) The logndashlog AllanndashWerle standard deviation plot illustratingthat the instrument performance follows the white-noise regime up to a certain extent identified by the dashed lines This represents theoptimum integration time after which instrumental instabilities start to dominate

Figure 7 (a b) Time series of two long-term stability tests with the black dashed lines representing the 3σ level (c d) boxplot of thestability test while continuously flushing zero air in the cavity with means over 46 and 148 measurements (for Test1 and Test2 respectively)shown as green triangles and dots representing the outliers (e f) histogram analysis showing well distributed measurements within the 3σlevel (black dashed lines)



Table 1 Performance comparisons with other recently developed IBBCEAS systems

Centred Source NO2 detection Sample Mirror Optical Mirrors CCD MinimumReferences wavelength FWHM limit path length reflectivity length purged cooled σAW-SD

(nm) (nm) (ppt minminus1) (cm) () (km) (C) deviation (s)

Min et al (2016) 455 18 16 48 999973 178 no minus70 100Jordan et al (2019) 505 30 200 102 9998 51 yes minus80 300Liu et al (2019) 455 18 33 84 99993 103 yes minus70 100Liang et al (2019) 448 15 15 589 999942 117 yes minus10 3500This work 450 19 40 417 999905 44 no no 1300

ible region there are no NO absorption features for directoptical measurements and optical absorption detection ofNO is typically done in the infrared region (Richard et al2018) However its detection can be performed by indi-rectly measuring NO2 after chemical conversion of NO toNO2 in a controlled O3 excess environment This will lead tothe measurement of NOx which coupled with a simultane-ous detection of NO2 will provide the concentration of NO([NO]=[NOx]minus[NO2]) (Fuchs et al 2009)

NO+O3rarr NO2+O2

k1 = 180times 10minus14 cm3 molecminus1 sminus1 at 25 C (R1)

O3 was produced by electrolysis of water using commercialozone micro cells (INNOVATEC) allowing for the genera-tion of O3 without nitrogen oxide impurities and without theneed of an oxygen gas bottle The cells were mounted in ahomemade plastic container offering a 200 cm3 water reser-voir With a miniaturized design (15 cm times 15 cm times 15 cm)ozone production can be controlled upon injection into theinlet line The sample air flow to be analysed works as car-rier gas for flushing the ozone-enriched surface water Thisdesign also prevents the production of unwanted oxidizingagents such as peroxides as well as sample dilution caus-ing a signal degradation and requiring precise flow measure-ments for quantitative analysis The production of O3 is con-trollable by the amount of electrolytic cells used and the sup-plied current offering a dynamic range of 0ndash50 ppm of O3for a 1 L minminus1 total flow rate A diagram and details of thesystem can be found in Sect S32 For long-term use of theinstrument the overall water consumption should be consid-ered Losses due to evaporation were estimated to be between7 and 30 cm3 dminus1 at 10 and 30 C respectively for a flow rateof 1 L minminus1 while losses due to electrolysis are negligiblewith only 0024 cm3 dminus1 of consumption The other parame-ter to consider is the mixing time between the ozone gener-ator and the measurement cell with respect to the O3 excessFor instance the calculated production rate of NO2 from Re-action (R1) (ie reaction speed or conversion rate of NO) isv = 420times 1011 molec cm3 sminus1 for 5 ppb of NO and 8 ppmof O3 Under these conditions a mixing time of 029 s is re-quired for completing the conversion With an air flow of 1L minminus1 a 40 cm long 4 mm internal-diameter tube is there-

fore required between the ozone generator and the measure-ment cell The performance of the ozone generation systemwas tested on the IBBCEAS instrument with a nitrogen oxidestandard gas bottle containing sim 195 ppb of NO in air (AirLiquide) Kinetic simulations using Tenua software (Wachs-stock 2007) were made in order to establish the O3 excessconcentrations needed to achieve the complete conversion ofNO to NO2 which along with its detection was tested withthe IBBCEAS instrument by varying the excess concentra-tion of O3 until complete conversion of NO was achievedat different flows (ie different reaction times before reach-ing the measurement cell) The experimental results were ingood agreement with the simulations as reported in Fig S9In addition the instrument was found to have a linear re-sponse regarding the detection of the produced O3 The de-tection limit for the NOx measurement was found to be sim-ilar to the one of NO2 10 ppt (1σ ) in 22 min of integrationtime while for NO retrieved as the difference between theNOx and the NO2 concentrations the detection limit esti-mated from the error propagation corresponds to 21 ppt

5 Possible chemical and spectral interferences

Further possible interferences on NO2 detection in the pres-ence of high levels of O3 were also studied since a largeexcess of O3 could trigger the following reactions with rateconstants that are a few orders of magnitude lower than k1(from the NIST Kinetics Database)

NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2


To study those possible interferences 100 ppb of NO2 pro-duced by a permeation tube were pumped through theozonizer and the spectrometer at a flow rate of 1 L minminus1

while varying the concentration of O3 from 0 to 10 ppm



The NO2 concentration was stable at low ozone concentra-tions while a drop of 14 was observed at high levels of O3(gt8 ppm) Kinetics simulations showed that the NO2 con-sumption in favour of the NO3 production (NO2+O3rarr

NO3+O2) was kinetically possible under those conditionsThe consumption of NO2 is strongly dependent on the reac-tion time and the concentration of O3 The later should beselected according to the reaction time imposed by the vol-ume of the inlet line and the flow rate therefore making thisinterference negligible Other chemical reactions could led toan overestimation of NO2 mixing ratios

HONOrarr NO+HO


HO2NO2rarr NO2+HO2


Couach et al (2002) estimated the background levels ofHONO and HO2NO2 in Grenoble to be 4 and 2 ppq(or 10minus15 mol molminus1 parts per quadrillion) respectively(Couach et al 2002) With such low concentrations andkinetic constant rates interferences due to Reactions (R5)and (R6) can be neglected in an urban environment How-ever in remote areas such as the East Antarctic PlateauHO2NO2 levels were estimated by indirect measurements tobe around 25 ppt (Legrand et al 2014) Because the lifetimeof HO2NO2 decreases with temperature (τHO2NO2 = 86 h atminus30 C and 645 mbar) its measurement using an instrumentstabilized at higher temperature would lead to an overesti-mation of the NO2 due to the thermal degradation of theHO2NO2 However this interference can be minimized byworking at low temperatures at 10 C and 1 L minminus1 flowin our IBBCEAS instrument the NO2 signal would be over-estimated by only 1 ppt which is below the detection limitof the sensor The instruments were therefore designed forworking at low temperatures (up to a few degrees Celsius)One more reaction Reaction (R7) may also lead to possibleinterferences for NO2 detection

HONO+OHrarr NO2+H2O


In urban environments OH radicals of up to 4times 106 OH rad-icals cmminus3 can be observed (Heard 2004 Mauldin et al2001) With background levels of HONO such as 4 ppqin the city of Grenoble and around 30 ppt in Dome CAntarctica (Legrand et al 2014) very low mixing ratiosof NO2 (lt few ppq) would be produced by Reaction (R7)in less than 8 s (residence time of the molecules in the in-strument at 1 L minminus1) Therefore contribution from thisinterference can be neglected Previous works also high-lighted possible artefacts through the heterogenous reactionof NO2 and H2O occurring in thin films on surfaces theapproximate rate production of HONO plus NO calculated

in their study was reported to be between 4times 10minus2 and8times 10minus2 ppb minminus1 ppmminus1 of NO2 (Finlayson-Pitts et al2003) Assuming linearity between production rates and con-centrations this would represent a range of 8 to 16 ppq for200 ppt of NO2 in remote areas such as the East AntarcticPlateau The losses that may occur on the thin films on sur-faces through the heterogeneous reaction of NO2 and H2Oare therefore negligible Finally detection of NO2 CHO-CHO and IO may be affected by spectral interferences Forinstance water vapour also shows an absorption signature atthis wavelength region which was included in the fit routineIts spectral fit is important particularly for the measurementof NOx where the inlet sampling line gets saturated in wa-ter vapour while passing through the water reservoir of theozone generator In addition artefacts on the signal and thespectral fit were studied by varying the O3 NO2 or NO mix-ing ratios in the cavity Small imperfections of the fit couldlead to large effects on the NO2 retrieved mixing ratio partic-ularly at sub-ppb concentrations and in the presence of largeamounts of ozone However no appreciable effects of possi-ble artefacts were observed while O3 concentrations of up to8 ppm were used These performance studies and the simplic-ity of the ozone generator compact and fully controllablemake it suitable for field applications

6 Conclusions

A compact robust affordable and highly sensitive IB-BCEAS instrument for direct detection of NO2 IO CHO-CHO and O3 and indirect detection of NO is reported inthis work The instrument relies on the injection of incoher-ent radiation from a compact high-power and low-cost LEDsource into a high-finesse optical cavity The instrument pro-vides a minimum detectable absorption of 20times10minus10 cmminus1

corresponding to a figure of merit (noise-equivalent ab-sorption sensitivity NEAS) of 18times 10minus10 cmminus1 Hzminus12 perspectral element Due to the multiplexing broadband fea-ture of the setup multi-species detection can be performedwith simultaneous detection of NO2 IO CHOCHO and O3achieving detection limits of 11 03 10 ppt and 47 ppb (1σ )within 22 min of measurement (which account for the ref-erence and absorption spectra acquisition) respectively De-tection limits for the indirect measurement of NOx and NOare 10 and 21 ppt (1σ ) respectively The instrument has beendesigned to fit in a 19 in 3U rack-mount case weighs 15 kgand has a total electrical power consumption oflt 300 W Thedetection limits could be further improved by replacing theAvantes ULS2048L spectrometer which offers at this work-ing wavelength a signal-to-noise ratio on a single acquisitionof 110 and a sensitivity of 172 000 counts microWminus1 msminus1 witha spectrometer with an integrated cooled CCD The cool-ing would allow to gain up to a factor of 10 on the signal-to-noise ratio which would directly apply to the detectionlimits Better sensitivity of the CCD would also allow for



the use of higher-reflectivity mirrors as done by Min et al(2016) providing an effective optical path length ofsim 18 km(with a similar cavity length) sim 4 times higher than the oneobtained in this work The dynamic ranges detection limitsand multi-species detection character make this instrumentwell suited for measurements in different environments fromhighly polluted to very remote areas such as polar regionsThe instruments can be used in the future to address differ-ent scientific questions related to the oxidative capacity atparticular regions (ie inland and coastal polar atmospheres)where variability of NOx and IO would provide key informa-tion for understanding the mechanisms taking place in suchremote areas The detection of the α-dicarbonyl CHOCHOmay have applications at the marine boundary layer whereits source remains unknown and its contribution to secondaryaerosol particle formation may be relevant (Ervens et al2011 Volkamer et al 2007 Fu et al 2008)

Data availability The data are available upon request

Supplement The supplement related to this article is available on-line at httpsdoiorg105194amt-13-4317-2020-supplement

Author contributions Grants obtained by JS AD and RG fundedthe project The IBBCEAS instruments were designed and devel-oped by CB under the supervision of RG AB developed and val-idated the ozone generation The instruments were optimized andvalidated by CB and AB who also did the instrumental intercom-parison the measurements for the long-term stability and the sim-ulations and study of the possible chemical interferences RG wasthe principal supervisor of the project JS and AD contributed withtheir knowledge in atmospheric sciences and they closely followedthe project with regular meetings JS and RG are the supervisorsof ABrsquos PhD thesis under which the instruments are deployed NCprovided technical and engineering input particularly at the begin-ning of the project The paper was written by AB CB and RG withcontributions from all authors

Competing interests The authors declare that they have no conflictof interest

Acknowledgements The authors would like to thank the LabexOSUG2020 for funding the Thermo Fishertrade 42i NOx analyserand IPEV for funding the Thermo Fisher 42iTLtrade NOx trace anal-yser that were used during the development of the IBBCEAS in-struments presented here The authors thank Guillaume Meacutejean andDaniele Romanini from LiPhy France for the very useful exchangeof information regarding IBBCEAS techniques as well as Andy Al-bert Ruth from the University of Cork Ireland for his very use-ful feedback on the paper The authors thank Stephan Houdier fordiscussions about glyoxal measurement And finally the authors

greatly thank the technical staff of the IGE for their technical sup-port

Financial support The research leading to these results has re-ceived funding from the PARCS project (Pollution in the Arc-tic System a project of the CNRS French Arctic Initiative athttpsparcsaeris-datafr last access 27 March 2020) the LabexOSUG2020 (ldquoInvestissements drsquoavenirrdquo ANR10 LABX56) theFrench national programme LEFE (Les Enveloppes Fluides etlrsquoEnvironnement) the Agence Nationale de la Recherche (ANRcontract no ANR-16-CE01-0011-01 EAIIST) the BNP ParibasFoundation through its Climate amp Biodiversity Initiative and theFrench Polar Institute (IPEV) through programmes 1177 (CAPOXI35ndash75) and 1169 (EAIIST)

Review statement This paper was edited by Hendrik Fuchs and re-viewed by two anonymous referees

References

Atkinson D B Solving chemical problems of environmental im-portance using Cavity Ring-Down Spectroscopy Analyst 128117ndash125 httpsdoiorg101039b206699h 2003

Atkinson R Atmospheric chemistry of VOCs and NOV At-mos Environ 34 2063ndash2101 httpsdoiorg101016S1352-2310(99)00460-4 2000

Brown S S Stark H Ciciora S J McLaughlin R J and Rav-ishankara A R Simultaneous in situ detection of atmosphericNO3 and N2O5 via Cavity Ring-Down Spectroscopy Rev SciInstrum 73 3291ndash3301 httpsdoiorg101063114992142002

Chan K L Poumlhler D Kuhlmann G Hartl A Platt U andWenig M O NO2 measurements in Hong Kong using LEDbased long path differential optical absorption spectroscopy At-mos Meas Tech 5 901ndash912 httpsdoiorg105194amt-5-901-2012 2012

Couach O Balln J Jimenez R Rjstorj P Simeonov VQuaglla P and Clappier A Etude drsquoun eacutepisode photochimiqueagrave lrsquoalde drsquoun modegravele meacuteso-eacutechelle et de mesures intensives surla reacutegion de Grenoble ndash Study of a photochemical episode overthe Grenoble area using a mesoscale model and intensive mea-surements Pollution atmospheacuterique 174 277ndash295 2002

Dooly G Fitzpatrick C and Lewis E Deep UV based DOASsystem for the monitoring of nitric oxide using ratiometric sep-aration techniques Sensor Actuat B Chem 134 317ndash323httpsdoiorg101016jsnb200805011 2008

Duan J Qin M Ouyang B Fang W Li X Lu KTang K Liang S Meng F Hu Z Xie P Liu W andHaumlsler R Development of an incoherent broadband cavity-enhanced absorption spectrometer for in situ measurementsof HONO and NO2 Atmos Meas Tech 11 4531ndash4543httpsdoiorg105194amt-11-4531-2018 2018

Ervens B Turpin B J and Weber R J Secondary organicaerosol formation in cloud droplets and aqueous particles (aq-SOA) a review of laboratory field and model studies Atmos



Chem Phys 11 11069ndash11102 httpsdoiorg105194acp-11-11069-2011 2011

Fiedler S E Hese A and Ruth A A Incoherent Broad-BandCavity-Enhanced Absorption Spectroscopy Chem Phys Lett371 284ndash294 httpsdoiorg101016S0009-2614(03)00263-X2003

Finlayson-Pitts B J and Pitts J N Chapter 4 ndash Photochem-istry of important atmospheric species in Chemistry of the Up-per and Lower Atmosphere edited by Finlayson-Pitts B Jand Pitts J N Academic Press San Diego USA 86ndash129httpsdoiorg101016B978-012257060-550006-X 2000

Finlayson-Pitts B J Wingen L M Sumner A L Syomin Dand Ramazan K A The heterogeneous hydrolysis of NO2 inlaboratory systems and in outdoor and indoor atmospheres anintegrated mechanism Phys Chem Chem Phys 5 223ndash242httpsdoiorg101039b208564j 2003

Frey M M Brough N France J L Anderson P S TraulleO King M D Jones A E Wolff E W and Savarino JThe diurnal variability of atmospheric nitrogen oxides (NO andNO2) above the Antarctic Plateau driven by atmospheric stabil-ity and snow emissions Atmos Chem Phys 13 3045ndash3062httpsdoiorg105194acp-13-3045-2013 2013

Frey M M Roscoe H K Kukui A Savarino J France J LKing M D Legrand M and Preunkert S Atmospheric ni-trogen oxides (NO and NO2) at Dome C East Antarctica dur-ing the OPALE campaign Atmos Chem Phys 15 7859ndash7875httpsdoiorg105194acp-15-7859-2015 2015

Fu T-M Jacob D J Wittrock F Burrows J P Vrekous-sis M and Henze D K Global budgets of atmosphericglyoxal and methylglyoxal and implications for formation ofsecondary organic aerosols J Geophys Res 113 D15303httpsdoiorg1010292007JD009505 2008

Fuchs H Dubeacute W P Lerner B M Wagner N L WilliamsE J and Brown S S A sensitive and versatile detector foratmospheric NO2 and NOx based on blue diode laser Cav-ity Ring-Down Spectroscopy Environ Sci Technol 43 7831ndash7836 httpsdoiorg101021es902067h 2009

Fuchs H Ball S M Bohn B Brauers T Cohen R C DornH-P Dubeacute W P Fry J L Haumlseler R Heitmann U Jones RL Kleffmann J Mentel T F Muumlsgen P Rohrer F RollinsA W Ruth A A Kiendler-Scharr A Schlosser E ShillingsA J L Tillmann R Varma R M Venables D S Vil-lena Tapia G Wahner A Wegener R Wooldridge P J andBrown S S Intercomparison of measurements of NO2 con-centrations in the atmosphere simulation chamber SAPHIR dur-ing the NO3Comp campaign Atmos Meas Tech 3 21ndash37httpsdoiorg105194amt-3-21-2010 2010

Gherman T Venables D S Vaughan S Orphal J andRuth A A Incoherent Broadband Cavity-Enhanced Ab-sorption Spectroscopy in the near-Ultraviolet application toHONO and NO2 Environ Sci Technol 42 890ndash895httpsdoiorg101021es0716913 2008

Grilli R Meacutejean G Kassi S Ventrillard I Abd-Alrahman CFasci E and Romanini D Trace measurement of BrO at theppt level by a transportable Mode-Locked Frequency-DoubledCavity-Enhanced Spectrometer Appl Phys B 107 205ndash212httpsdoiorg101007s00340-011-4812-9 2012

Grosjean D and Harrison J Response of chemilumines-cence NOx analyzers and ultraviolet ozone analyzers to or-

ganic air pollutants Environ Sci Technol 19 862ndash865httpsdoiorg101021es00139a016 1985

Heard D E High levels of the hydroxyl radical in thewinter urban troposphere Geophys Res Lett 31 L18112httpsdoiorg1010292004GL020544 2004

Jaworski N A Howarth R W and Hetling L J At-mospheric deposition of nitrogen oxides onto the land-scape contributes to coastal eutrophication in the north-east United States Environ Sci Technol 31 1995ndash2004httpsdoiorg101021es960803f 1997

Jordan N Ye C Z Ghosh S Washenfelder R A BrownS S and Osthoff H D A broadband cavity-enhanced spec-trometer for atmospheric trace gas measurements and Rayleighscattering cross sections in the cyan region (470ndash540 nm) At-mos Meas Tech 12 1277ndash1293 httpsdoiorg105194amt-12-1277-2019 2019

Kovalev V A and Eichinger W E Elastic Lidar theory practiceand analysis methods John Wiley amp Sons Inc Hoboken NJUSA httpsdoiorg1010020471643173 2004

Langridge J M Ball S M and Jones R L A compact Broad-band Cavity Enhanced Absorption Spectrometer for detection ofatmospheric NO2 using light emitting diodes Analyst 131 916ndash922 httpsdoiorg101039b605636a 2006

Lee J S Kim K-H Kim Y J and Lee J Applicationof a Long-Path Differential Optical Absorption Spectrome-ter (LP-DOAS) on the measurements of NO2 SO2 O3 andHNO2 in Gwangju Korea J Environ Manag 86 750ndash759httpsdoiorg101016jjenvman200612044 2008

Lee J S Kim Y J Kuk B Geyer A and Platt U Simultane-ous measurements of atmospheric pollutants and visibility witha Long-Path DOAS System in urban areas Environ Monit As-sess 104 281ndash293 httpsdoiorg101007s10661-005-1616-62005

Legrand M Preunkert S Frey M Bartels-Rausch Th KukuiA King M D Savarino J Kerbrat M and Jourdain BLarge mixing ratios of atmospheric nitrous acid (HONO) atConcordia (East Antarctic Plateau) in summer a strong sourcefrom surface snow Atmos Chem Phys 14 9963ndash9976httpsdoiorg105194acp-14-9963-2014 2014

Leser H Houmlnninger G and Platt U MAX-DOAS measurementsof BrO and NO2 in the marine boundary layer Geophys ResLett 30 1537ndash1541 httpsdoiorg1010292002GL0158112003

Liang S Qin M Xie P Duan J Fang W He Y Xu J LiuJ Li X Tang K Meng F Ye K Liu J and Liu W Devel-opment of an incoherent broadband cavity-enhanced absorptionspectrometer for measurements of ambient glyoxal and NO2 ina polluted urban environment Atmos Meas Tech 12 2499ndash2512 httpsdoiorg105194amt-12-2499-2019 2019

Liu J Li X Yang Y Wang H Wu Y Lu X Chen M HuJ Fan X Zeng L and Zhang Y An IBBCEAS system foratmospheric measurements of glyoxal and methylglyoxal in thepresence of high NO2 concentrations Atmos Meas Tech 124439ndash4453 httpsdoiorg105194amt-12-4439-2019 2019

Maeda Y Aoki K and Munemori M Chemiluminescencemethod for the determination of nitrogen dioxide Anal Chem52 307ndash311 httpsdoiorg101021ac50052a022 1980

Mauldin R L Eisele F L Tanner D J Kosciuch E Shetter RLefer B Hall S R Nowak J B Buhr M Chen G Wang



P and Davis D Measurements of OH H2SO4 and MSA at theSouth Pole during ISCAT Geophys Res Lett 28 3629ndash3632httpsdoiorg1010292000GL012711 2001

Min K-E Washenfelder R A Dubeacute W P Langford A OEdwards P M Zarzana K J Stutz J Lu K Rohrer FZhang Y and Brown S S A broadband cavity enhancedabsorption spectrometer for aircraft measurements of glyoxalmethylglyoxal nitrous acid nitrogen dioxide and water vaporAtmos Meas Tech 9 423ndash440 httpsdoiorg105194amt-9-423-2016 2016

Monks P S Granier C Fuzzi S Stohl A Williams M Aki-moto H Amann M Baklanov A Baltensperger U BeyI Blake N Blake R Carslaw K Cooper O DentenerF Fowler D Fragkou E Frost G Generoso S GinouxP Grewe V Guenther A Hansson H Henne S HjorthJ Hofzumahaus A Huntrieser H Isaksen I Jenkin MKaiser J Kanakidou M Klimont Z Kulmala M Laj PLawrence M Lee J Liousse C Maione M McFiggansG Metzger A Mieville A Moussiopoulos N Orlando JOrsquoDowd C Palmer P Parrish D Petzold A Platt U PoumlschlU Preacutevocirct A Reeves C Reimann S Rudich Y Selle-gri K Steinbrecher R Simpson D ten Brink H ThelokeJ van der Werf G Vautard R Vestreng V VlachokostasC and von Glasow R Atmospheric composition change ndashglobal and regional air quality Atmos Environ 43 5268ndash5350httpsdoiorg101016jatmosenv200908021 2009

Monks P S Gas-phase radical chemistry in the troposphereChem Soc Rev 34 376 httpsdoiorg101039b307982c2005

Pikelnaya O Hurlock S C Trick S and Stutz J In-tercomparison of MultiAXis and Long-Path Differential Op-tical Absorption Spectroscopy measurements in the marineboundary layer J Geophys Res-Atmos 112 D10S01httpsdoiorg1010292006JD007727 2007

Platt U and Perner D Direct measurements of atmosphericCH2O HNO2 O3 NO2 and SO2 by Differential Optical Ab-sorption in the near UV J Geophys Res-Oceans 85 7453ndash7458 httpsdoiorg101029JC085iC12p07453 1980

Poumlhler D Vogel L Friess U and Platt U Obser-vation of halogen species in the Amundsen Gulf Arc-tic by active Long-Path Differential Optical AbsorptionSpectroscopy P Natl Acad Sci USA 107 6582ndash6587httpsdoiorg101073pnas0912231107 2010

Richard L Romanini D and Ventrillard I Nitric oxideanalysis down to ppt levels by Optical-Feedback Cavity-Enhanced Absorption Spectroscopy Sensors 18 1997httpsdoiorg103390s18071997 2018

Rozanov V V and Rozanov A V Differential optical ab-sorption spectroscopy (DOAS) and air mass factor conceptfor a multiply scattering vertically inhomogeneous mediumtheoretical consideration Atmos Meas Tech 3 751ndash780httpsdoiorg105194amt-3-751-2010 2010

Ruth A A Dixneuf S and Raghunandan R Broadband Cavity-Enhanced Absorption Spectroscopy with Incoherent Lightin Cavity-Enhanced Spectroscopy and Sensing edited byGagliardi G and Loock H-P Springer Berlin HeidelbergBerlin Heidelberg 179 485ndash517 httpsdoiorg101007978-3-642-40003-2_14 2014

Ryerson T B Williams E J and Fehsenfeld F CAn efficient photolysis system for fast-response NO2 mea-surements J Geophys Res-Atmos 105 26447ndash26461httpsdoiorg1010292000JD900389 2000

Sinreich R Volkamer R Filsinger F Frieszlig U Kern C PlattU Sebastiaacuten O and Wagner T MAX-DOAS detection of gly-oxal during ICARTT 2004 Atmos Chem Phys 7 1293ndash1303httpsdoiorg105194acp-7-1293-2007 2007

Thalman R and Volkamer R Inherent calibration of a blueLED-CE-DOAS instrument to measure iodine oxide glyoxalmethyl glyoxal nitrogen dioxide water vapour and aerosol ex-tinction in open cavity mode Atmos Meas Tech 3 1797ndash1814httpsdoiorg105194amt-3-1797-2010 2010

Triki M Cermak P Meacutejean G and Romanini D Cavity-Enhanced Absorption Spectroscopy with a red LED sourcefor NOx trace analysis Appl Phys B 91 195ndash201httpsdoiorg101007s00340-008-2958-x 2008

Vandaele A Hermans C Simon P Carleer M Colin R FallyS Meacuterienne M Jenouvrier A and Coquart B Measure-ments of the NO2 absorption cross-section from 42 000 cmminus1

to 10 000 cmminus1 (238ndash1000 nm) at 220 K and 294 K J QuantSpectrosc Ra 59 171ndash184 httpsdoiorg101016S0022-4073(97)00168-4 1998

Venables D S Gherman T Orphal J Wenger J C and RuthA A High sensitivity in situ monitoring of NO3 in an atmo-spheric simulation chamber using Incoherent Broadband Cavity-Enhanced Absorption Spectroscopy Environ Sci Technol 406758ndash6763 httpsdoiorg101021es061076j 2006

Ventrillard-Courtillot I Sciamma OrsquoBrien E Kassi S MeacutejeanG and Romanini D Incoherent Broad-Band Cavity-EnhancedAbsorption Spectroscopy for simultaneous trace measurementsof NO2 and NO3 with a LED source Appl Phys B 101 661ndash669 httpsdoiorg101007s00340-010-4253-x 2010

Villena G Bejan I Kurtenbach R Wiesen P and KleffmannJ Interferences of commercial NO2 instruments in the urbanatmosphere and in a smog chamber Atmos Meas Tech 5 149ndash159 httpsdoiorg105194amt-5-149-2012 2012

Vitousek P M Aber J D Howarth R W Likens G E MatsonP A Schindler D W Schlesinger W H and Tilman D GHuman alteration of the global nitrogen cycle sources and conse-quences Ecol Appl 7 737ndash750 httpsdoiorg1018901051-0761(1997)007[0737HAOTGN]20CO2 1997

Volkamer R Spietz P Burrows J and Platt U High-resolution absorption cross-section of glyoxal in the UVndashvisand IR spectral ranges J Photoch Photobio A 172 35ndash46httpsdoiorg101016jjphotochem200411011 2005

Volkamer R San Martini F Molina L T Salcedo D JimenezJ L and Molina M J A missing sink for gas-phase glyoxal inMexico City formation of secondary organic aerosol GeophysRes Lett 34 L19807 httpsdoiorg1010292007GL0307522007

Wachsstock D Tenua the kinetics simulator for Java available athttpbililitecomtenua (last access 7 May 2020) 2007

Wagner T Ibrahim O Shaiganfar R and Platt U MobileMAX-DOAS observations of tropospheric trace gases AtmosMeas Tech 3 129ndash140 httpsdoiorg105194amt-3-129-2010 2010

Washenfelder R A Langford A O Fuchs H and Brown SS Measurement of glyoxal using an incoherent broadband cav-



ity enhanced absorption spectrometer Atmos Chem Phys 87779ndash7793 httpsdoiorg105194acp-8-7779-2008 2008

Werle P 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

Williams E J Baumann K Roberts J M Bertman S B Nor-ton R B Fehsenfeld F C Springston S R NunnermackerL J Newman L Olszyna K Meagher J Hartsell B Edger-ton E Pearson J R and Rodgers M O Intercomparison ofground-based NOy measurement techniques J Geophys Res-Atmos 103 22261ndash22280 httpsdoiorg10102998JD000741998

Wine P H and Nicovich J M Atmospheric RadicalChemistry in Encyclopedia of Radicals in Chemistry Bi-ology and Materials edited by Chatgilialoglu C andStuder A John Wiley amp Sons Ltd Chichester UKhttpsdoiorg1010029781119953678rad015 2012

Wittrock F Oetjen H Richter A Fietkau S Medeke TRozanov A and Burrows J P MAX-DOAS measurementsof atmospheric trace gases in Ny-Aringlesund ndash Radiative transferstudies and their application Atmos Chem Phys 4 955ndash966httpsdoiorg105194acp-4-955-2004 2004


Abstract

Introduction

Method

Spectral fit

Calibration performance and multi-species detection

Calibration

Instrumental calibration and intercomparison

Instrument sensitivity and long-term stability

Indirect measurement of NO

Possible chemical and spectral interferences

Conclusions

Data availability

Supplement

Author contributions

Competing interests

Acknowledgements

Financial support

Review statement

References


Especially nitrogen oxides (NOx=NO and NO2) have a di-rect impact on air quality and climate change In the presenceof volatile organic compounds (VOCs) and under solar radi-ation nitrogen oxides stimulate ozone (O3) formation in thetroposphere NOx also plays an important role in rain acid-ification and ecosystem eutrophication by its transformationin nitric acid (HNO3) (Jaworski et al 1997 Vitousek et al1997) Finally NOx contributes to the formation of particu-late matter in ambient air and to the aerosol formation leadingto cloud formation (Atkinson 2000) The NO2 mixing ratioin the troposphere ranges from a few tens of ppt in remote ar-eas to hundreds of parts per billion (10minus9 mol molminus1 or ppb)in urban atmospheres (Finlayson-Pitts and Pitts 2000) Be-ing able to measure such species in situ at low levels andat a timescale compatible with its reactivity (ie in min) ischallenging and puts stringent constraints on the instrumentsensitivity time response energy consumption and compact-ness Among the various techniques that have so far been de-veloped chemiluminescence detection (CLD) (Maeda et al1980 Ryerson et al 2000) long-path differential optical ab-sorption spectroscopy (LP-DOAS) (Lee et al 2005 2008Pikelnaya et al 2007) and multi-axis differential opticalabsorption spectroscopy (MAX-DOAS) (Platt and Perner1980 Sinreich et al 2007 Wagner et al 2010) have beenused to detect nitrogen species and halogen oxides The CLDtechnique using the chemiluminescence reaction occurringbetween O3 and NO after the reduction of NO2 into NOis widely used for air quality measurements with sensitivi-ties better than 100 ppt (Ryerson et al 2000) Neverthelessthe interferences in the reduction of NO2 to NO with otherspecies (ie HONO and HNO3) and the sensitivity to envi-ronmental conditions (temperature and humidity) leave un-certainties about absolute mixing ratio measurements (Gros-jean and Harrison 1985 Williams et al 1998) The MAX-DOAS technique has been used to measure BrO and NO2by making use of the characteristic absorption features ofgas molecules along a path in the open atmosphere (Leseret al 2003) Although MAX-DOAS is relatively simple todeploy the data analysis makes it a complex approach for insitu field measurements due to the influence of clouds on theradiative transfer which alters the path length of light (Wit-trock et al 2004 Rozanov and Rozanov 2010) While inLP-DOAS the optical path length is known the signal degra-dation due to the environment (clouds rain and wind) re-mains of importance for the data retrieval and results areintegrated over the long path leading to a limited spatialresolution (Chan et al 2012 Poumlhler et al 2010) Com-pact highly sensitive and point-source measurements maybe achieved using cavity enhanced techniques such as cav-ity ring-down spectroscopy (CRDS) and cavity-enhanced ab-sorption spectroscopy (CEAS) (Atkinson 2003) The po-tential of the CRDS for accurate sensitive and rapid mea-surements in a compact and transportable instrument has al-ready been demonstrated (Fuchs et al 2009 Brown et al2002) eg Fuchs et al (2009) reached a sensitivity of 22 ppt

for NO2 within 1 s of integration time using the CRDStechnique Incoherent broadband cavity-enhanced absorptionspectroscopy (IBBCEAS) (Fiedler et al 2003) is a simpleand robust technique for in situ field observations Differentsources and wavelength regions have been used for the de-tection of NO2 leading to different levels of performanceVenables et al (2006) were able to detect simultaneouslyNO3 NO2 O3 and H2O in an atmospheric simulation cham-ber with a sensitivity of tens of ppb for NO2 Gherman et al(2008) reached sim 013 and sim 038 ppb for HONO and NO2in a 4 m3 atmospheric simulation chamber between 360 and380 nm Triki et al (2008) used a red LED source centredat 643 nm reaching a sensitivity of 5 ppb Langridge et al(2006) developed an instrument with a blue-light-emittingdiode (LED) centred at 445 nm allowing for detection lim-its ranging from 01 to 04 ppb Ventrillard-Courtillot et al(2010) reached 600 ppt detection limit for NO2 with an LEDcentred at 625 nm and Thalman and Volkamer (2010) re-ported a detection limit of 30 ppt within 1 min of integra-tion time More recently Min et al (2016) proved a preci-sion of 80 ppt in 5 s of integration at 455 nm using a spec-trometer with a thermoelectric-cooled charge-coupled device(CCD) camera and very high-reflectivity mirrors This non-exhaustive list of works underline the need of robust com-pact and transportable instruments also allowing for directmulti-species detection and low detection limits for appli-cations in remote areas such as Antarctica where the ex-pected mixing ratio of NO2 could be as low as a few tens ofppt Fuchs et al (2010) during the NO3Comp campaign atthe SAPHIR (Simulation of Atmospheric PHotochemistry Ina Large Reaction Chamber) atmospheric simulation cham-ber demonstrated the potential of these optical techniquesto compete with the CLD instruments as routine measure-ments of NO2 concentrations in the future The present pa-per describes a compact and affordable instrument based onthe IBBCEAS technique allowing for the simultaneous de-tection of nitrogen dioxide iodine oxide glyoxal and ozone(NO2 IO CHOCHO and O3) with detection limits of 1103 10 ppt and 47 ppb (1σ ) respectively for a measurementtime of 22 min (half of the time is spent on the acquisitionof the reference spectrum in the absence of the absorberand the other half is spent on the absorption spectrum) Thefour species are directly detected by a broadband blue-light-emitting diode centred at 445 nm The wavelength region wasselected in order to optimize the detection of NO2 Direct de-tection of NO is only possible in the ultraviolet region forwavelengths around 226 nm (Dooly et al 2008) or in themid-infrared region at 53 microm (Richard et al 2018) wave-lengths difficult to achieve with LED technology Here anindirect measurement is proposed which relies on the oxi-dation of NO to NO2 under a controlled excess of O3 Thesum of NO and NO2 is therefore measured leading to a sup-plemental indirect measurement of NO if the concentrationof NO2 is also monitored The field deployment for the mea-surements of NO2 and NOx consists of two twin instruments




2 Method



3 Spectral fit






α(λ)=

(I0(λ)

I (λ)minus 1

)(1minusR(λ)

d

) (1)



) Therefore the


α(λ)=sumi



a3λ3+ a4λ



41 Calibration





(F = π

radicR

(1minusR)

)of sim 33100




















= 09996






radictintM

)is therefore





















NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References



2 Method



3 Spectral fit






α(λ)=

(I0(λ)

I (λ)minus 1

)(1minusR(λ)

d

) (1)



) Therefore the


α(λ)=sumi



a3λ3+ a4λ



41 Calibration





(F = π

radicR

(1minusR)

)of sim 33100




















= 09996






radictintM

)is therefore





















NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




α(λ)=

(I0(λ)

I (λ)minus 1

)(1minusR(λ)

d

) (1)



) Therefore the


α(λ)=sumi



a3λ3+ a4λ



41 Calibration





(F = π

radicR

(1minusR)

)of sim 33100




















= 09996






radictintM

)is therefore





















NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References



(F = π

radicR

(1minusR)

)of sim 33100




















= 09996






radictintM

)is therefore





















NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References












= 09996






radictintM

)is therefore





















NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References





radictintM

)is therefore





















NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References











NO+O3rarr NO2+O2






NO2+O3rarr NO3+O2


NO2+O3rarr 2O2+NO


NO3+O3rarr 2O2+NO2








HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References




HONOrarr NO+HO


HO2NO2rarr NO2+HO2



HONO+OHrarr NO2+H2O




6 Conclusions














References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References











References








































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References































































Abstract

Introduction

Method

Spectral fit


Calibration





Conclusions

Data availability

Supplement


Competing interests

Acknowledgements

Financial support

Review statement

References

A compact incoherent broadband cavity-enhanced absorption ...

Documents

Transcript of A compact incoherent broadband cavity-enhanced absorption ...