New Compact and Versatile Sampler for Uninterrupted, Time- resolved - PowerPoint PPT Presentation

New Compact and Versatile Sampler for Uninterrupted, Time- resolved Chemical Speciation of Ambient Particulate Matter Arantzazu Eiguren Fernandez S. Hering 1 , G. Lewis 1 , S. Spielman 1 , A. Hecobian 2 , J. Collett 2 1 Aerosol Dynamics Inc., Berkeley, CA 2 Colorado State University, Fort Collins, CO Acknowledgements: 8 th Asian Aerosol Conference National Institutes of Health (ES014997 & ES014997) Sydney, 2-5 December 2013 Carl Camp (ARB – Stockton)

Standard Instrumentation Filter-based methods: SASS, IMPROVE simple field operations, but time-consuming laboratory steps limited data coverage, delayed results SASS Real-time instruments: AMS / PILS no laboratory analysis, but complex and expensive field operation continuous data sets, immediate results AMS Aim: To combine the simplicity of filter sampling with the data completeness and automation of real time instruments. Approach: To provide a collector with directly analyzable samples and an automated interface to lab- based analytical instruments.

New Challenges for New Applications New approach to laminar flow, water-based condensation WCPC output, like mixing instruments, is warm and humid... Collection / concentrator applications want output at ambient temperature & RH Saturation Ratio Saturation Ratio 1.6 1.4 1.2 1.0 Dew Point = 32 30 T or DP (°C) 20 Temperature = 27 10 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 x/Q (s/cm2)

1. Development, Validation and Applications

Sequential Spot Sampler (S 3 ) Conditioner (2-5 o C) Initiator (32-35 o C) Modulator (8-10 o C) Collection chamber

Output Droplet Size Original: 06 / 40 / 40 °C GT 8000 3-Stage: 06 / 40 / 25 °C 6000 dN/dLogDp APS 4000 Directly coupled Growth Tube to APS 2000 0 2 4 6 8 10 Droplet Diameter (µm) Nearly the same as in Original Configuration

Output Droplet Size with Flow Rate 0.4 lpm At Variable Flow 3-stage (5/40/25) 1.0 lpm 0.6 lpm 3000 1.5 lpm GT dN/dLogDp 1.8 lpm 2000 APS 2.5 lpm 1000 With humidified makeup air to Aerodynamic Sizer 0 2 4 6 8 10 Droplet Diameter (µm)

Collection: Collection chamber Dry collection at 26 o C 1. Particles collected as dry deposits 2. Spot size: ~400 µm to ~1 mm Aluminum plate PEEK plate

Bouncing? Arizona Road Dust

1. Performance characterization: laboratory tests PM2.5 ambient air DMA High voltage Upstream S 3 Nebulizer UHASH Downstream

a) Collection efficiency Lab generated aerosols Collection efficiencies >99% for Collection efficiencies >95% concentrations up to 10 6 #/cc for particles larger than 10nm

Ambient PM Ambient air Downstream CPC 1x10 4 Upstream CPC #/cc UPSTREAM CPC 1x10 2 DOWNSTREAM CPC 0 10:45 AM 12:00 PM 1:15 PM 2:30 PM Collection efficiencies were >95% by number for ambient PM

b) Reproducibility and Linearity Lab generated aerosol: Ammonium sulfate and ammonium nitrate Automated analysis by IC: extraction and injection handled by autosampler Plate location in the autosampler Autosampler dispenses the extraction solvent (DDW) and injects the sample

Reproducibility: collection and analysis 5 30 Collection Time (min) (n=6) (n=6) Sulfate (%STDEV) 4.21 3.52 Nitrate (%STDEV) 5.36 4.25 Linearity 0.3 sulfate nitrate Mass on well (ug) 0.2 y = 0.007x - 0.01 R² = 0.99 0.1 y = 0.005x - 0.005 R² = 0.99 0.0 0 5 10 15 20 25 30 Time (min)

c) Volatilization losses Nitrate / Sulfate Ratio Set 1 Set 2 1.0 0.8 0.6 Sulfate: Set 1 Set 2 Collected Anion Mass (ng) 2000 Nitrate: Set 1 Set 2 1500 1000 500 0 0 5 10 15 20 Sample Duration (hrs) Nitrate losses were not significant for collection periods up to 20 hrs.

Temporal variability of sulfate and nitrate in ambient PM (Berkeley) 1.0 sulfate (µg/m3) 0.5 0.0 1.0 nitrate (µg/m3) 0.5 0.0 08/20/11 08/21/11 08/22/11 08/23/11 08/24/11 08/25/11

2. Field Performance

1. San Bernardino Mountains Deployment at San Gorgonio, IMPROVE site (7,000 ft) June 12 – July 5, 2012 Anion measurements: sulfate and nitrate Instruments Spot Samplers: mostly 1-hr resolution PILS: ~17 min resolution URG Denuder-Filter Packs: 12-hr samples URG-filters Analysis: Dionex IC with modified autosampler PILS Spot Samplers Denuders

a) S 3 vs PILS: sulfate Sulfate Spot Sampler PIILS PILS Sulfate (µg/m 3 ) 10 5 0 6/16/2012 6/21/2012 6/26/2012 7/1/2012

b) S 3 vs PILS: nitrate 30 Nitrate Spot Sampler PIILS PILS Nitrate (µg/m 3 ) 20 10 0 6/16/2012 6/21/2012 6/26/2012 7/1/2012

c) S 3 vs URG: sulfate Sulfate (µg/m 3 ) 10 Sulfate Spot Sampler Filter 5 0 6/16/2012 6/21/2012 6/26/2012 7/1/2012

d) S 3 vs URG: nitrate 30 Nitrate Spot Sampler Filter Nitrate (µg/m 3 ) 20 10 0 6/16/2012 6/21/2012 6/26/2012 7/1/2012

2. Stockton, CA Deployment at Stockton, CA - ARB monitoring station Nov 2011 – Feb 2012 Instruments Spot samplers in triplicate 12hr samples @ 1.5 lpm for PAHs 6 hr samples @ 1.0 lpm for trace elements Parallel 48-hr filter collection (25mm, 9 lpm) Run unattended for weeks Analysis CTC-PAL autosampler: automated extraction/injection Agilent HPLC-FL: 15 EPA-PAHs 22 min analysis

a) Automated extraction and analysis CTC- PAL autosampler Sonication bath

b) Optimization of the extraction method soaking vs. mechanical agitation vs. sonication 1200 y = 1.76x + 41.76 R² = 0.81 1000 1:1 Sonication (pg/m3) 800 600 Soaking the sample for 30 min 400 was not enough to extract PAHs at high concentrations 200 0 0 200 400 600 800 1000 1200 Soaking (pg/m3) 2500 y = 1.36x - 14.9 R² = 0.91 2000 Sonication (pg/m3) 1500 Sonication for 20 sec was more efficient than mechanical 1000 agitation for 1 min 500 0 0 500 1000 1500 2000 2500 Mechanical agitation (pg/m3)

c) Comparison with filter collection 2.0 PHE FLT PYR BAA CRY BBF BKF BAP DBA BGP IND 1.5 Plate (ng/m3) 1.0 0.5 Individual PAH concentrations measured on filters and averaged of 12-hr spot samples 0.0 0.0 0.5 1.0 1.5 2.0 Filter (ng/m3) Good correlation between filters and S 3

d) Spot Sampler intercomparison: 1600 1400 All correspond to the same sample 1200 PHE ANT System 2 (pg/m3) 1000 FLT PYR 800 BAA CHY BBF BKF 600 BAP DBA 400 BGP IND 200 System reproducibility for individual PAHs (n=26) 0 0 400 800 1200 1600 System 1 (pg/m3) Excellent reproducibility between samplers for all PAHs

e) Temporal variability: PM 2.5 in Stockton, CA 1800 PHE BAA 1600 BBF BKF 1400 BAP BGP 1200 IND PAH (pg/m3) 1000 800 600 400 200 0 ΣPAHs reached concentrations up to 9.53 ng/m 3 • BBF (401 pg/m 3 ) and BGP (380 pg/m 3 ) showed the highest concentrations • • PHE highest among the semivolatile

f) Trace Elements by LIBS Intensity (a.u.) Wavelength (nm) Obtain elemental spectra in minutes • • Toxic elements such as iron and copper can be identified

Summary • Collection efficiency >95% for a broad range of particle types, sizes and concentrations • Reproducibility & precision between 3% - 5 % • Systems can run unattended for several weeks • Collection plate can be extracted and analyzed directly (automated extraction and analysis) • Good correlation with standard anion measurements (PILS and URG-filters) • Sequential spot sampler is capable of measuring peak concentration missed by filter collection systems • PAH - Filter Comparison: → 15% for individual PAHs • Spots allow fast trace elements analysis using LIBS

The Sequential Spot Sampler is a simple and reliable collection system that can be streamlined with lab equipment for un- interrupted, time-resolved chemical characterization of ambient PM.

Theodore Roosevelt National Park (ND) Organic speciation: TD-GCxGC-HRTOFMS February 18, 2013 I.S. PAHs I.S. n-alkanes Polarity separation n-alkanes Volatility separation

Chemical Mass Balance (CMB8.2) Vehicular emission profile from Eiguren-Fernandez and Miguel (2005) Wood burning emission profile from Li and Kamens (1993) 100 wood 90 diesel 80 gasoline Source Contribution 70 60 50 40 30 20 10 0