HR-AMS Quantification— Calibrations and Comparisons from Recent Campaignscires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg19/... · 2018. 9. 9. · Jason C. Schroder, Jose L. Jimenez,

AMS Quantification—Calibrations and Comparisons from Recent Campaigns

Benjamin A. Nault, Pedro Campuzano-Jost, Doug A. Day, Hongyu Guo, Jason C. Schroder, Jose L. Jimenez, and the Science Teams from KORUS-AQ and ATom 1

When AMS is fully and carefully calibrated and characterized, the measurements are completely quantitative

2General Calibrations Transmission Curve First Order Intercomparisons Filter EvaporationQuantitative AMS

4

3

2

1

0

Sulfa

te (µ

g sm

−3)

8/1/20

16

8/3/20

16

8/4/20

16

8/7/20

16

8/9/20

16

8/13/2

016

8/15/2

016

8/17/2

016

8/20/2

016

8/22/2

016

8/23/2

016

Time

4

3

2

1

0

AMS

Sulfa

te (µ

g sm

−3)

43210

SAGA MC-IC Sulfate (µg sm−3)

1:1 Line

MC-IC Sulfate Slope: 1.11, r2 = 0.72

AMS Sulfate (IC Resolution, Response Adjusted)

8/2/2016

8/6/2016

8/10/2016

8/14/2016

8/18/2016

8/22/2016

When AMS is fully and carefully calibrated and characterized, the measurements are completely quantitative


14

12

10

8

6

4

2

0

PM1 V

olum

e (µ

m3 c

m−3

)

7/30/2

016

8/2/20

168/4

/2016

8/7/20

168/9

/2016

8/13/2

016

8/15/2

016

8/16/2

016

8/17/2

016

8/20/2

016

8/22/2

016

8/23/2

016

Time

8

6

4

2

0

AMS

Volu

me

(incl

. BC

and

SS)

(µm

3 cm

−3)

86420

UHSAS PM1 Volume (µm3 cm−3)

1:1 Line

UHSAS PM1 Volume Slope: 0.95, r2 = 0.92

AMS Volume (including SS and BC)

7/31/20168/2/20168/4/20168/6/20168/8/20168/10/20168/12/20168/14/20168/16/20168/18/20168/20/20168/22/2016

Calibrations (not including exotic species)

4Transmission Curve First Order Intercomparisons Filter EvaporationQuantitative AMS General Calibrations

Current CU-AMS Aircraft Configuration


General Calibrations


Calibrations of the Ammonium Salts for RIEs and Ion Ratios

7

• See John Jayne’s, Manjula Canagaratna’s, and Phil Croteau’s Talks• Calibrate NO3 IE as often as possible (for aircraft).

• Look for changes in AB to help determine if you should calibrate for ground (40/28 for MCP)

• For ground, calibrate as often as you feel comfortable (maybe 2 – 4 days)

• Calibrate everything before and after a filament change• Calibrate SO4 RIE

• Average observed during one campaign 1.56• Typical value in Batch Table is 1.2

• The NO2:NO ratio, NH4 RIE, and CO2+/NO3 Artifact• NO2:NO ratio needed for particle organic nitrate values• NO2:NO ratio can be used for instrument performance

evaluation• Ideally, calibrate Chl before and after campaign

• Typically not much non-refractory Chl• Chl can linger on vaporizer, causing artifacts

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/ (SI Fig. 3)

Transmission Curve First Order Intercomparisons Filter EvaporationQuantitative AMS General Calibrations

http://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg19/Jayne_CalibrationOverview.pdfhttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg18/Croteau_ACSMAnalysisTutorialPKU2017.pdfhttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg19/Croteau_ACSMCalibrationsSTL2018.pdfhttps://www.atmos-chem-phys-discuss.net/acp-2018-838/

IE, RIE, and Comparisons NOT Always Stable


9

Impact of Mixtures vs Pure Ammonium Salts on the RIEs

NO3/(NO3 + SO4)

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/ SI Fig. 7


• No variation of RIE from pure particles to mixtures of variable composition

• This implies that RIE should be constant for ambient particles as composition varies

https://www.atmos-chem-phys-discuss.net/acp-2018-838/

10

Impact of Mixtures vs Pure Ammonium Salts on the RIEs

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/, SI Fig 6


Lab Calibrations

Field Data


11

Calibrating (e)PToF

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/, SI Fig 5


• PSLs provide calibration of (e)PToF, and a check of the sizing of the DMA calibration system as well

• Check the DMA sizing of NH4NO3 (and its lack of evaporation) by comparing with PSL cals

• Calibrate at beginning of intensive measurements

• Calibration only works for a given pressure inside the lens

• E.g. if flying in aircraft may experience some changes in lens pressure calibrate (e)PToF at those pressures as well!

Vacuum


Calibrating vaporizer temperature for non-refractory compounds and to minimize refractory compounds

12

Leah Williamshttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg11/WilliamsAMSUsersMtg_2010_VapT.pdfhttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg12/Presentations/Williams_VapT.pdfhttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg13/AMSUsersMtg_2012_VapT.pdf


http://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg11/WilliamsAMSUsersMtg_2010_VapT.pdfhttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg12/Presentations/Williams_VapT.pdfhttp://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg13/AMSUsersMtg_2012_VapT.pdf

Select power setting prior to sharp NaNO3 PToF peak


Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/acp-2018-838-supplement.pdf, SI Fig 8

• INSTRUMENT DEPENDENT!

• Depends on boards and vaporizer

• Only needs to be calibrated at beginning of intensive measurements, UNLESS:

• Any boards are changed• Vaporizer was changed• Wires were changed

https://www.atmos-chem-phys-discuss.net/acp-2018-838/acp-2018-838-supplement.pdf

Check MCP voltage not too low & calibrating single ion - Instrument Specific

Air Amm MSA


• Ratio of air ions shows that for big ions, the stable setting is ~1.8 mV ns, or ~2100 V

• Air ions show that for small ones, the stable setting is ~1.3 mV ns, or ~2000 V

• For aerosol, same trend

• INSTRUMENT & MCP SPECIFIC, CHANGES IN TIME! NEED TO CALIBRATE

Conclusion on Calibrations• Calibrate NO3 IE and NH4+ RIE frequently (for aircraft) and as often as you feel confident (for aircraft, laboratory, ground)

• Provides insight into instrumental performance• Prevents headaches later if the instrument values drastically change• Provides insight into NO2+/NO+ ratio• But… problems can be introduced when trying to calibrate (need a lot of practice)

• Calibrate SO42- RIE frequently as well• Our group has consistently observed the calibrated RIE is not the same as the default

value in Squirrel/PIKA• Only calibrate halogens beginning and end of campaign• RIEs not impacted by mixtures• Calibrate (e)PToF sizing at beginning of measurements• Calibrate vaporizer temperature once a campaign/intensive laboratory

study (more often if parts or electronics are replaced)• Check MCP voltage is appropriate before and during the campaign


Transmission Curve

16General Calibrations First Order Intercomparisons Filter EvaporationQuantitative AMS Transmission Curve

Differences in two different literature transmission curves

120

100

80

60

40

20

0

Tran

smis

sion

(%)

2 3 4 5 6 7 8 91000

2

Aerodynamic Diameter (nm)

Liu et al. (2007) Hu et al. (2017b)


• Ideally, calibrate at least once prior to intensive measurements

• If in a moving environment (e.g., airplane, mobile lab), may need to calibrate more often during intensive measurements (it can change, although unusual)

• BE CAREFUL! Mass based transmission curves can be challenging due to doubly-charged particles at low concentrations. (ET cal can be found at this talk).

• Need to know the curve to do quantitative comparisons with other measurements or with models

• MAYBE instrument specific

http://cires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg17/2016-10-19_Nault_AMS_Users_Meeting_2016_v4_online.pdf

CU AMS, beginning of one campaign120

100

80

60

40

20

0

Tran

smis

sion

(%)

2 3 4 5 6 7 8 91000

2


18

Campaign 1a

General Calibrations First Order Intercomparisons Filter EvaporationQuantitative AMS Transmission Curve

CU AMS, end of campaign120

100

80

60

40

20

0

Tran

smis

sion

(%)

2 3 4 5 6 7 8 91000

2


19

Campaign 1a

Campaign 1b


CU AMS, campaign 2 years later120

100

80

60

40

20

0

Tran

smis

sion

(%)

2 3 4 5 6 7 8 91000

2

Aerodynamic Diameter (nm)20

Campaign 1a

Campaign 1b

Campaign 2


CU AMS, campaign 1 year later120

100

80

60

40

20

0

Tran

smis

sion

(%)

2 3 4 5 6 7 8 91000

2

Aerodynamic Diameter (nm)21

Campaign 1a

Campaign 1b

Campaign 2

Campaign 3


Using Beam Width Probes to “Probe” Transmission


• Use BWP to spot check center of air and particle beam every day (flightcalibrate every day)

• Changes in location of center of particle beam full transmission curve calibration realignment full transmission curve

• For aircraft measurements, have not seen center of beam move after each flight (and some VERY rough landings)

Quantitative AMS

Comparison of two AMSs during 1 campaign

120

100

80

60

40

20

0

Tran

smis

sion

(%)

2 3 4 5 6 7 8 91000


1.0

0.8

0.6

0.4

0.2

0.0

Norm

alized PToF Distribution

AMS Lens Transmission curve > 200 nm from Hu et al. (2017b) CU-AMS Ammonium Nitrate Mass Transmission Calibration CU-AMS Ammonium Nitrate Event Trigger Transmission Calibration K-AMS Ammonium Nitrate Mass Transmission Calibration Fit to K-AMS Transmission Calibration

Average CU-AMS BL PToF

23

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/acp-2018-838-supplement.pdf SI Fig 26

General Calibrations First Order Intercomparisons Filter EvaporationTransmission Curve


Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/acp-2018-838-supplement.pdf SI Fig 26

Average size distribution of species peakedabove size cut-off for one AMS

24

120

100

80

60

40

20

0Tr

ansm

issi

on (%

)2 3 4 5 6 7 8 9

1000Aerodynamic Diameter (nm)

1.0

0.8

0.6

0.4

0.2

0.0

Norm

alized PToF Distribution

AMS Lens Transmission curve > 200 nm from Hu et al. (2017b) CU-AMS Ammonium Nitrate Mass Transmission Calibration CU-AMS Ammonium Nitrate Event Trigger Transmission Calibration K-AMS Ammonium Nitrate Mass Transmission Calibration Fit to K-AMS Transmission Calibration

Average CU-AMS BL PToF OA pNO3

SO4

NH4

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018General Calibrations First Order Intercomparisons Filter EvaporationQuantitative AMS Transmission Curve

Ambient size distributions


Quantitative AMS

Very different fits and R2 between periods with high and low aerosol sizes

25

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/acp-2018-838-supplement.pdf, SI Fig. 27

Capt

ure

Vapo

rizer

AM

S (µ

g sm

–3)

Standard Vaporizer AMS (µg sm–3)

Periods when ambient PM1below size cut-off

Period when ambient PM1above size cut-off



Quantitative AMS

The fit is more exponential due to transmission curve

26

Capt

ure

Vapo

rizer

AM

S (µ

g sm

–3)

Standard Vaporizer AMS (µg sm–3)

Periods when PM1below size cut-off

Period when PM1above size cut-off




Conclusions on transmission curve calibration• This should be done prior to start campaign/laboratory study

• Spot checks throughout study maybe necessary, as alignment/performance may change

• Calibrated transmission curve is NECESSARY to compare against volume and other aerosol measurements

• Calibrated transmission curve is NECESSARY to know what is being measured and in order to be able to provide insight on the chemistry

27Quantitative AMS General Calibrations First Order Intercomparisons Filter EvaporationTransmission Curve

First order comparisons

28General Calibrations Transmission Curve Filter EvaporationQuantitative AMS First Order Intercomparisons

Total nr-PM1 measured by capture and standard vaporizer agreed across different collection efficiencies

29


General Calibrations Transmission Curve Filter EvaporationQuantitative AMS First Order Intercomparisons

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Stan

dard

Vap

oriz

er /

Cap

ture

Vap

oriz

er

1.00.90.80.70.60.5Calculated Dry SV CE


Comparison of calculated PM1 volume (AMS species + . . . ) versus any optical particle counter (OPC)

30

“I believe you’re right in concluding that the differences at high concentrations are caused by saturation or particle coincidence in the LAS. We’ve been talking to another collaborator about deploying a sample dilution system during FIREX-AQ so that we can avoid these types of problems when flying through fresh biomass burning plumes.” (PI of LAS OPC team)

Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/acp-2018-838-supplement.pdf


Air flow

AMSOPC

Settings for Laser Aerosol Spectrometer


Comparison of total PM1 versus optical volume measurement

31




Investigating whether Organic RIE caused the discrepancy

1.6

1.2

0.8

0.4

3000200010000LAS Number Concentration

(particles cm–3

)

1.6

1.2

0.8

0.4(A

MS+

BC)/L

AS V

olum

e R

atio

1.00.80.60.40.20.0OA/Total AMS Mass

6040200AMS Total Mass

(µg sm–3

)

(a)

(b) (c)

Ratio, All Data 1 Line

32




Instead, it appears that optical saturation of the LAS is causing the discrepancy




Improved slopes when saturation is taken into account

80

60

40

20

0

AMS

+ BC

Vol

ume

(µm

3 cm

–3)

806040200LAS Volume (µm3 cm

–3)

35

30

25

20

15

10

5

0

Subm

icro

n Vo

lum

e (µ

m3 c

m–3

)

9:00 AM22-May-2016

12:00 PM 3:00 PM

Local Time

(a)

(b) Legend

LAS Volume AMS + BC Volume 1:1 Line All Data; y = 1.56(±0.01)*x–1.88(±0.13); R

2 = 0.86

Filtered Data by # Conc (1600 part./cm−3

); y = 1.25(±0.01)*x–0.20(±0.06); R2 = 0.85

Filtered Data by Mass Conc. (40 μg sm−3

); y = 1.21(±0.01)*x–0.32(±0.08); R2 = 0.87

Filtered Data by # Conc (1600 part./cm−3 + < 0.35 μg sm−3 Ca2+);

y = 1.19(±0.01)*x+0.12(±0.03); R2 = 0.91

Filtered Data by Mass Conc. (20 μg sm−3

); y = 1.00(±0.01)*x+0.19(±0.03); R2 = 0.79

50403020100AMS Total Mass (µg sm

–3)

34

Filtered by #: Slope = 1.25, R2 = 0.85Filtered by Mass (40): Slope = 1.21, R2 = 0.87Filtered by Mass (20): Slope = 1.00, R2 = 0.79




Conclusions for first order intercomparisons


Nault, Campuzano-Jost, Jimenez et al., ACPD, 2018https://www.atmos-chem-phys-discuss.net/acp-2018-838/

• Be aware of size cut-offs prior to comparisons• E.g., optical particle counters have a lower size cut that’s often larger than the

small size cut for the AMS• OPCs measure particles above the AMS size range

• Be aware of other instrumental limitations• OPCs can easily saturate under moderately polluted conditions. OPC

operators are not always aware of this, and it is easy to blame the AMS


Filters


Air flow

Filters collected and handledfor ~3-15 minutes

PAM

AMS

Aircraft OFR Operation

PAM Residence Time: 2.5 minutes

Sample Residence Time to AMS: 0.4 sec

General Calibrations Transmission Curve First Order Intercomparisons Filter EvaporationQuantitative AMS

38Kim, Campuzano-Jost, Jimenez et al., ACP, 2015


Losses of organic aerosol mass during hot periods due to evaporation

Measured in situ every 1 hr

Sit i

n th

e fie

ld a

t am

bien

t T fo

r few

day

s aft

er

sam

plin

g, sh

ippe

d w

ithou

t ref

riger

atio

n

Comparison of AMS and filter NO3 measured on airplane changed with date

39Heim, Dibb, Nault, Campuzano-Jost, Jimenez et al., JGR, in review

30

25

20

15

10

5

0

AMS

NO

3 (μg

sm

−3)

302520151050

Filter NO3 (μg sm−3

)

5/11/2016

5/21/2016

5/31/2016

6/10/2016

Date

Slope = 1.21, R2 = 0.69

1:1 Line


40

However, total nitrate (PM1 NO3 + HNO3) agreed throughout campaign

Heim, Dibb, Nault, Campuzano-Jost, Jimenez et al., JGR, in review


Ammonium nitrate shows large losses in non-chemistry OFR as there was a large temperature difference

50

40

30

20

10

0

OFR

Mas

s C

once

ntra

tion,

Lig

hts

Off

(µg

sm−3

)

50403020100Ambient Mass Concentration (µg sm−3)

1.0

0.8

0.6

0.4

0.2

0.0

Nor

mal

ized

Dis

tribu

tion

2520151050Cabin Temperature − Ambient Temperature (K)

(b)(a) SO4 Slope = 0.90, R

2 = 0.96 NO3 Slope = 0.18, R

2 = 0.71 NH4 Slope = 0.51, R

2 = 0.90 1:1 Line

41



Differences between OFR and Ambient are same as differences between Filter and Ambient

25

20

15

10

5

0

Ambi

ent −

OFR

NO

3 (μg

sm

−3)

2520151050-5-10AMS − Filter NO3 (μg sm

−3)

14

12

10

8

6

4

2

0

Ambi

ent −

OFR

NH

4 (μg

sm

−3)

151050-5AMS − Filter NH4 (μg sm

−3)

(a) Data

y = 3.66 + 1.10*x, R2 = 0.46

(b) Data

y = 1.31 + 0.97*x, R2 = 0.62

42



Conclusions for comparisons with filter measurements


• Though not shown here, be aware what size cut-off is for filters vs AMS

• BE CAREFUL:• Be aware sampling time and handling of filters. May cause evaporation of aerosol species.• Also, handling of filters may lead to other forms of biases (positive or negative)

• E.g. if particles are very acidic, filters will pick up NH3(g) from human breath etc., and show particles that are more neutralized than what the AMS sees

AMS Quantification—�Calibrations and Comparisons �from Recent CampaignsWhen AMS is fully and carefully calibrated and characterized, �the measurements are completely quantitativeWhen AMS is fully and carefully calibrated and characterized, �the measurements are completely quantitativeCalibrations �(not including exotic species)Current CU-AMS Aircraft ConfigurationGeneral CalibrationsCalibrations of the Ammonium Salts for RIEs and Ion RatiosIE, RIE, and Comparisons NOT Always StableImpact of Mixtures vs Pure Ammonium Salts on the RIEsImpact of Mixtures vs Pure Ammonium Salts on the RIEsCalibrating (e)PToFCalibrating vaporizer temperature for non-refractory compounds and to minimize refractory compoundsSelect power setting prior to sharp NaNO3 PToF peakSlide Number 14Conclusion on CalibrationsTransmission CurveDifferences in two different literature transmission curvesCU AMS, beginning of one campaignCU AMS, end of campaignCU AMS, campaign 2 years laterCU AMS, campaign 1 year laterUsing Beam Width Probes to “Probe” TransmissionComparison of two AMSs during 1 campaignAverage size distribution of species peaked�above size cut-off for one AMSVery different fits and R2 between periods �with high and low aerosol sizesThe fit is more exponential due to transmission curveConclusions on transmission curve calibrationFirst order comparisonsTotal nr-PM1 measured by capture and standard vaporizer agreed across different collection efficienciesComparison of calculated PM1 volume (AMS species + . . . ) �versus any optical particle counter (OPC)Comparison of total PM1 versus optical volume measurementInvestigating whether Organic RIE caused the discrepancyInstead, it appears that optical saturation �of the LAS is causing the discrepancyImproved slopes when saturation is taken into accountConclusions for first order intercomparisonsFiltersAircraft OFR OperationLosses of organic aerosol mass �during hot periods due to evaporationComparison of AMS and filter NO3 measured on airplane changed with dateHowever, total nitrate (PM1 NO3 + HNO3) agreed throughout campaignAmmonium nitrate shows large losses in non-chemistry OFR as there was a large temperature differenceDifferences between OFR and Ambient are same as �differences between Filter and AmbientConclusions for comparisons with filter measurements

Documents

HR-AMS Quantification— Calibrations and Comparisons from Recent Campaignscires1.colorado.edu/jimenez-group/UsrMtgs/UsersMtg19/... · 2018. 9. 9. · Jason C. Schroder, Jose L. Jimenez,