NO X AT CAPE VERDE (CVO) Chris Reed, Katie Read, Luis Mendes, James Lee, Lucy Carpenter

Transcription

1 NO X AT CAPE VERDE (CVO) Chris Reed, Katie Read, Luis Mendes, James Lee, Lucy Carpenter

2 The Cape Verde Atmospheric Observatory (CVO) Observatory established in 2006 as global GAW station. Only GAW global station continuously measuring NO & NO 2 in the tropics & at a background site. Other measurements: O 3, CO, CH 4, CO 2, N 2 O, NMHCs, DMS, O-VOCs, Halocarbons, Mercury, Meteorology, Aerosol, Spec-Rad.

3 NO x measurements at Cape Verde Custom dual channel NO chemiluminescence instrument. Channel 1: NO and NO x by selective photolytic dissociation of NO 395nm or 385nm (since March 2015). Nafion dryer on sample. Channel 2: NO y 4 channel thermal dissociation NO y (ΣPANs, ΣANs, HNO 3, Σreactive nitrogen.) LODs; NO ~0.3 ppt, NO 2 ~ 0.35 ppt, NOy ~ 5 ppt / hour. Calibrated: Sensitivity, NO 2 converter efficiency, artifact NO/NO 2 signals on both channels. Pre-reactor zero signal every 5 minutes NO x measuring near continuously since October NO y until 2009, Speciated NO y since 2015 only one anywhere! Original single channel instrument from , upgraded to dual channel 2009 present. Air Quality Design, inc. Golden, Colorado, USA

4 2014 & 2015 NO x time series Typically very low mixing ratios of NO x ; pptv leading to net ozone destruction Replace UV-LEDs --- NO NO New Lab Manifold pump changed Left: Daytime (11:00 15:00) averages of NO x, N, and NO 2 respectively. Data taken from October 13 September 14 Represents time between replacing sample MFC and moving the instrument to the new labs. *June is exceptional as data coverage is poor during this month Ozoniser failure New NO 2 converter

5 2015 NO x time series Changing NO 2 converter showed us that the zero air was bad ~13 ppt NO 2! Observed artifacts should be stable, assume ZA doesn t change. Simultaneous equations 2015 AAAAAA 1 + AAAAAA ZZZZ --- NO NO SSSSSSSS 1 CCCC 1 = Art obs 1 AAAAAA 2 + AAAAAA ZZZZ SSSSSSSS 2 CCCC 2 = Art obs 2 New NO 2 converter 30 90% Still uncertain however.

6 NO 2 measurement uncertainty: What is the artifact? Treatment of the artifact signal is key at these low levels data has no NO 2 artifact correction. Lets see why!

7 NO 2 measurement uncertainty: What is the artifact? Subtract the NO 2 artifact assuming signal is NO i.e. 100% conversion. Looks reasonable until June, after which NO 2 is too low

8 NO 2 measurement uncertainty: What is the artifact? Subtract the NO 2 artifact assuming signal is NO 2 i.e. correcting for NO 2 CE. Oh dear... This tells us something though.

9 NO 2 measurement uncertainty: What is the artifact? Artifact is biggest uncertainty in our measurements Need better zero air or a better technique.

10 Conclusions and outlook? NO measurement appears robust NO 2 artifact is biggest uncertainty in our measurements NO 2 levels appear to decrease with greater NO 2 CE This will be addressed in a talk later. Acknowledgements: Dr Marty Buhr, AQD/NOxWerx Dr Tom Ryerson, NOAA/CSD

11 AQD BLUE LIGHT CONVERTERS Chris Reed, Mat Evans

12 BLCs: Use ultra violet LEDs either 385 or 395nm to photolyse NO 2 to NO. Manufactured by Sonoma Technology/Droplet Measurement Technologies. Succeeded by Air Quality Design inc. CCCC = jjjj jjjj + kk Ox tt 1 eeeeee ( jjjj kk Ox tt) Selective to compounds which absorb and specific UV wavelengths. Depend on: lamp power, j; residence time, t; oxidant concentration [Ox]. UV LEDs improved over time - less heat, more light Nichia 2w, Norlux 1w, Engin 3(6)w. Possibly others. From 50 97% conversion at 1 s -1 residence time.

13 Photolytic artifact sources: Overlap with HONO, nitrate and BrONO 2 absorbance bands. Can exploit interferences to our advantage i.e HONO

14 HONO conversion: Overlap with HONO 30% better at 385 nm than 395 nm. Conversion of HONO still significant at 395 nm 25% or more.

15 HONO conversion: Constant difference of HONO conversion over a range of HONO calibration concentrations. Absolute conversion less obvious Comparison with FT-IR shows linearity up to 150 ppb [HONO] Good if you want to quantify HONO, less so if you need accurate NO 2. Variable interference!

16 Thermal NO 2 measurement error: Modelled PAN thermal decomposition Measured NO 2 signal from PAN Lamp temperatures >70 o C contacting sample gas lead to thermal artifact from NO y NO 2 * more easily photolysed. PAN Heat! Peroxy radical + (*NO 2 NO 2 ) Proportionately greater contribution to NO 2 signal. Higher conversion efficiency and lower temperature to reduce the error.

17 2014 & 2015 NO x time series NO 2 levels appear to decrease with increasing converter efficency. Replace UV-LEDs --- NO NO Left: Daytime (11:00 15:00) averages of NO x, N, and NO 2 respectively. Data taken from October 13 September 14 Represents time between replacing sample MFC and moving the instrument to the new labs. *June is exceptional as data coverage is poor during this month New NO 2 converter

18 Thermal artifact applicability: Cold, remote, and high altitude observations possibly biased True for any in-situ measurement which raises the sample temperature.

19 Conclusions and outlook? BLCs a significant improvement over past technology However, not a true NO 2 measurement Should not be assumed or treated as such. Careful treatment of NO 2 data required Especially in fundamental questions e.g. Leighton ratio. Individuals must determine the applicability of BLC system Good for trends, air quality etc Antarctica and high altitude less so Acknowledgements: Dr Lisa Whalley, Leeds Dr Marty Buhr, AQD/NOxWerx Dr Tom Ryerson, NOAA/CSD References: Reed et al., Atmos. Chem. Phys. Discuss., 15, , doi: /acpd , Reed et al., Atmos. Meas. Tech. Discuss., doi: /amt , in review, 2016.

20 NO Y AT CAPE VERDE (CVO) Chris Reed, Mat Evans, Katie Read, Luis Mendes, James Lee, Lucy Carpenter

21 NOy: NO y species a reservoir for NO x Likely source of NO 2 at Cape Verde Partitioning between gas and particulate phase necessitates speciation measurement. NO y inlet mounted on 10m tower 3 heated quartz furnaces 1 molybdenum catalyst Switchable cyclone 1

22 Why measure Noy?: NO x lower than modelled by global chemical transport model GEOS-Chem. NO 2 peaks during the day contrary to steady state chemistry. Indicative of a source of NO 2 during the day.

23 NOy speciation measurements NO x from thermal decomposition of NO y species detected quantitatively. NO y = NO + NO PANs+ XONO 2 + NO 3 + N 2 O 5 + HONO + HO 2 NO 2 + RO 2 NO 2 PANs + CH 3 ONO 2 + C 2 H 5 ONO RONO2 ANs + HNO 3 + p-no 3 + NPN NO and NO 2 observed HNO 3 Inter-conversion conserves NO x Need to measure both HNO 3

24 Surprisingly high labile NOy: 100s of ppt of most thermally labile class of compounds observed. Much higher than model predictions ~ 10s ppt. Likely the driver of daytime NO 2 production observed.

25 Conclusions and Outlook Conclusions: Surprisingly large contribution of thermally labile compounds. May also be readily photolysed producing NO 2 during the day. NO 2 diurnal necessitates daytime production of NO 2 in-situ. Nitrate particulate consistent with modelled HNO 3 + other nitrate. Gas phase nitric acid consistent with previous measurements. Alkyl nitrates consistently low with modelled. NOy still dominated by particulate. Outlook Unique long-term and ongoing dataset. Used to constrain NO x production models.

26 Extra slides 1 Atmos Chem (2010) 67: DOI /s

27 Extra slides 4 2 DO 3 / ppbv day NO / pptv

28 Extra slides Spring Summer

29 Extra slides 7 distinct air masses derived from NAME back trajectories

30 Extra slides

31 Extra slides Pros HNO 3, Tot-Nitrogen, P-Nitrogen, NO Cons Slow time resolution Completely PFA/Teflon inlet and cyclone Switching box and NO 2 converter. High surface area completely quartz ovens with cooling region