Ultra-Low Detection Technique for Hydrophobic Organic Contaminants C.L.A.M. Background, Validation Studies, Strengths and Weaknesses

Transcription

1 Ultra-Low Detection Technique for Hydrophobic Organic Contaminants C.L.A.M. Background, Validation Studies, Strengths and Weaknesses Jamie Aderhold Brent Hepner

2 The C.LA.M. a submersible SPE extractive sampler was developed to provide a time integrative, large volume, quantitative sample extract Active Sampler water is pumped continuously through solid phase extraction media. Uses US EPA Method 3535 SPE to sequester Pesticides, Herbicides, PAHs, TPH and other trace organics from water in-situ. Draws water into the media disk first, preventing adhesion loss from nonpolar contaminates. Small compact convenient 4 AA Batteries Total volume actively extracted known, allowing precise quantitation. (typically Liters) Large volumes extracted provides orders of magnitude lower detection ability. Leaves the water behind saving shipping costs. Increased Holding Times Disks can be frozen for future analysis

3 SPE Disks High Capacity C-18 SPE This media is well established for most non-polar compounds. Many EPA methods have incorporated its use, for Pesticides, PAH's, Semivolatiles, PCB's, Dioxins, Furans, PBDE's and other HRMS methods. EPA1664 gravimetric Oil & Grease may even be run using this media disk. High Capacity HLB SPE This media is best suited for semi-volatile organic analysis, EPA Method 1694 (pharmaceuticals and personal care products), emerging contaminants in wastewater and drinking water. They are self wetting and will not lose functionality. Glass Filtration Disks This filter assembly uses the triple lofted glass fiber filters to remove suspended sediment from the water. It allows for toxological studies for total and dissolved trace organics, when used in a two stage filter.

4 Total Volume Determination Three Techniques Available Start and ending flow rate per unit time estimates total volume 0%-20% Flexible for field screening Empirically collect and weigh/measure the effluent - accurate but cumbersome 100 liters weigh 222 lbs The C.L.A.M. Totalizer (CT) was developed to provide accurate total flow data for high volume long term active extractive events. The CT uses mechanical oval gear technology with a unique positive displacement flow sensor.

5 Laboratory Disk Cleaning and Conditioning C.L.A.M. Disks need to be solvent cleaned and conditioned prior to field deployment. Spiking surrogates / analytes can be added before or after deployment prior to elution by spiking directly into the inlet disk media. Disks are solvent eluted with a syringe or vacuum eluted like current SPE cartridges or disks. Laboratory acceptance criteria can be performed for each method, just as with standard SPE disks or cartridges. MDL and IDC s can be performed prior to field deployment.

6 Comparison and Validation Project funded to compare CLAM to Passive Samplers and Grabs Suite of 69 Trace Organic Compounds Wastewater Dominated River at outfall and 16 km downstream Field work performed April-May, 2012 Peer Reviewed Journal published January 2014

7 Comparison and Validation Science of the Total Environment Sampling trace organic compounds in water: A comparison of a continuous active sampler to continuous passive and discrete sampling methods Alissa L. Coes a.*, Nicholas V. Paretti a, William T. Foreman b, Jana L.lverson b, David A. Alvarez c a U.S. Geological Survey, Arizona Water Science Center, 520 North Park Avenue, Tucson, AZ 85719, USA b U.S. Geological Survey, National Water Quality Laboratory, Denver, CO 80225, USA c U.S. Geological Survey, Columbia Environmental Research Center, Columbia, MO 65201, USA HIGHLIGHTS Continuous active sampling method was compared to continuous passive and discrete sampling methods. Trace organic compounds in surface water were sampled by the three methods. Continuous active sampling method detected the most compounds but at lower concentrations. All three methods detected compounds across a wide polarity range. Results were dependent on discharge, loading, compound type, and method performance

8 1,4-Dichlorobenzene Benzophenone Bromoform Camphor Ethyl citrate Galaxolide (HHCB) Methyl salicylate N,N-diethyl-meta-toluamide (DEET) para-cresol Tonalide (AHTN) Tri(2-chloroethyl) phosphate Tributyl phosphate Tri(dichloroisopropyl) phosphate 3-beta-Coprostanol 3,4-Dichloropheynyl isocyanate * Acetophenone beta-stigmastanol 4-Octylphenol diethoxylate (OP2EO) Tetrachloroethylene 13 COMPOUNDS WERE DETECTED BY ALL 3 SAMPLING METHODS C.L.A.M & Replicate 6 COMPOUNDS WERE DETECTED BY CLAM AND POCIS METHODS ONLY POCIS 9 COMPOUNDS WERE DETECTED BY CLAM METHODS ONLY 0 COMPOUNDS WERE DETECTED BY THE DISCRETE WATER SAMPLE METHOD ONLY 3-tert-Butyl-4-hydroxyanisole (BHA) Anthraquinone beta-sitosterol - P Cumene - P* Indole 4-Octylphenol monoethoxylate (OP1EO) para-nonylphenol-total -W Skatol Triclosan Cholesterol -P Isophorone Tri(2-butoxyethyl) phosphate Triphenyl phospahte 4-tert-Octylphenol -C Diazinon Menthol 4 COMPOUNDS WERE DETECTED BY CLAM AND DISCRETE SAMPLE METHODS ONLY 3 COMPOUNDS WERE DETECTED BY POCIS METHODS ONLY

9 Discussion C.L.A.M. events were hours once per week (4 weeks) captured on standard laboratory SPE. The POCIS passive sampler was 29 days, and the Grab sample once per week. C.L.A.M. deployed in 3 disk chain. Pre-Filter (GFF) HLB - HLB C.L.A.M. method detected the most compounds but at lower concentrations. TOC concentrations from POCIS samples and filtered discrete samples were compared with TOC concentrations from the CLAM front HLB disks. TOC concentrations from the back HLB disk were not considered in the comparative data analysis. TOC concentrations from the GFF disks and back HLB disks, however, were considered in the QC data analysis to investigate the TOC retention on the disks under field-sampling conditions.

10 GFF= CLAM Pre-filter Discussion For many TOCs, the lower concentrations detected by the CLAM are most likely related to partial TOC non-retention by the front CLAM HLB disk. For the less polar TOCs, lower CLAM concentrations may also be related to the sorbing of dissolved-phase TOCs onto particle material as material accumulates on the GFF over the sampling period,

11 Discussion The number of CLAM detections at Site 1 remained relatively consistent over the 4 sampling periods at 21 to 24 detections, which was similar to the integrated POCIS sample (23 detections). The variability of the CLAM concentrations and the discrete detections suggests that the variability of TOCs in the samples from Site 1 may have been affected by factors other than stream discharge. The most likely cause is changes in the discharge from the wastewater treatment plant. Seasonal changes in wastewater constituent composition with seasonal local population shifts have previously been documented (Walker et al., 2009).

12 Discussion CLAM SPE Disks Blank data and field data showed evidence of blank contamination with high levels of low molecular weight PAH's and other volatiles. This is usually caused by excessive drying of the disk by drawing air through the media bed until the residual water is removed. Avoiding this step using methanolic dewatering instead has since been established, and greatly improves blank cleanliness.

13 Example Detection Limits a greater number of TOCs were detected in the CLAM samples than in the discrete or POCIS samples. Detection Limits (Partial List) - ng/l CLAM Discrete Pocis Caffeine Phenol Acetophenone Isophorone Tetrachloroethylene Triclosan Triphenyl Phosphate Differences in method reporting levels (RL) for various TOCs (Supplemental Table 2) were primarily because of differences in actual (discrete and CLAM) or estimated (POCIS) sample volumes, and secondarily by differences in analyte method recoveries.

14 City of Los Angeles Organochlorine Pesticide Investigation Ballona Creek Los Angeles, CA

15 Project Description Duplicate CLAMs deployed July/August 2011 off of bridge hour deployments gave liter extractions Grab Samples taken from bridge All Grab Samples came back non-detect for full suite LA Lab did Disk Conditioning, Surrogate Addition, Elution and Analysis LA lab had no prior solid phase extraction experience recoveries improved with practice of elution technique

16 Calculated Pesticide Results based on Water Volume Extracted Sample received date: 7/12/11 7/12/11 8/23/11 8/23/11 Sample preparation: 7/25/11 7/25/11 9/1/11 9/1/11 9/1/11 analysis date: 8/18/11 8/18/11 9/2/11 9/2/11 9/2/11 Disk 1 Disk 2 Disk 4 Disk 5 Liter Grab BCE-1A BCE-1A BCE-1A BCE-1A RL, ug/l ug/l ug/l ug/l ug/l ug/l A-BHC nd>0.01 G-BHC nd nd>0. 01 Heptachlor nd nd nd nd>0.01 Aldrin nd nd nd nd>0.01 B-BHC nd nd>0.01 D-BHC nd nd nd nd>0.01 Heptachlor Epoxide nd nd nd nd>0.01 2,4'-DDE nd nd nd>0.01 Endosulfan I nd nd>0.01 4,4'-DDE nd nd nd nd>0.01 Dieldrin nd>0.01 2,4'-DDD nd nd nd nd nd>0.01 Endrin nd nd nd nd nd>0.01 2,4'-DDT nd nd nd nd nd>0.01 4,4'-DDD nd nd>0.01 Endosulfan II nd nd nd nd nd>0.01 4,4'-DDT nd nd nd nd nd>0.01 Endrin Aldehyde nd nd nd nd nd>0.01 Mirex nd nd nd nd nd>0.01 Endosulfan II Sulfate nd nd nd nd nd>0.01 Methoxychlor nd nd nd nd nd>0.01 TCMX-SURR # % 23% 32% 50% DBC-SURR # % 28% 41% 62% Sample volume, Liters 89.4L 69.7L 79.2L 70.2L 1.0L Values are based on 10ml extract Spike std conc. in 10ml extract: Pesticides (30ppb), surrogates (40ppb) nd=not detected, NA=not analyzed

17 Port of Los Angeles Analyte Units CS-1 CS-1-DUP 5/9/2012 5/9/2012 Chlorinated Pesticides 2,4'-DDD ng/l ,4'-DDE ng/l <0.52 <0.31 2,4'-DDT ng/l <0.52 <0.31 4,4'-DDD ng/l ,4'-DDE ng/l ,4'-DDT ng/l Alpha Chlordane ng/l Cis-nonachlor ng/l Dieldrin ng/l <0.52 <0.31 Gamma Chlordane ng/l Heptachlor ng/l <0.52 <0.31 Heptachlor Epoxide ng/l <0.52 <0.31 Oxychlordane ng/l <0.52 <0.31 Trans-nonachlor ng/l PCB Congeners PCB003 ng/l <0.52 <0.31 PCB008 ng/l <0.52 <0.31 PCB018 ng/l <0.52 <0.31 PCB027 ng/l <0.52 <0.31 PCB028 ng/l <0.52 <0.31 PCB029 ng/l <0.52 <0.31 PCB031 ng/l <0.52 <0.31 PCB033 ng/l <0.52 <0.31 PCB037 ng/l <0.52 <0.31 PCB044 ng/l PCB049 ng/l PCB052 ng/l PCB056 ng/l <0.52 <0.31 PCB060 ng/l <0.52 <0.31 PCB066 ng/l <0.52 <0.31 PCB070 ng/l PCB074 ng/l <0.52 <0.31 PCB077 ng/l <0.52 <0.31 PCB081 ng/l <0.52 <0.31 PCB087 ng/l <0.52 <0.31 PCB095 ng/l PCB097 ng/l <0.52 <0.31 PCB099 ng/l PCB101 ng/l PCB105 ng/l <0.52 <0.31 PCB110 ng/l PCB114 ng/l <0.52 <0.31 PCB118 ng/l PCB119 ng/l <0.52 <0.31 PCB123 ng/l <0.52 <0.31 PCB126 ng/l <0.52 <0.31 PCB128 ng/l <0.52 <0.31 PCB132 ng/l <0.52 <0.31 PCB137 ng/l <0.52 <0.31 PCB138/158 ng/l PCB141 ng/l <0.52 <0.31 PCB149 ng/l PCB151 ng/l <0.52 <0.31 PCB153 ng/l PCB156 ng/l <0.52 <0.31 PCB157 ng/l <0.52 <0.31 PCB167 ng/l <0.52 <0.31 PCB168 ng/l <0.52 <0.31 PCB169 ng/l <0.52 <0.31 PCB170 ng/l PCB174 ng/l PCB177 ng/l <0.52 <0.31 PCB180 ng/l PCB183 ng/l <0.52 <0.31 PCB184 ng/l <0.52 <0.31 PCB187 ng/l PCB189 ng/l <0.52 <0.31 PCB194 ng/l <0.52 <0.31 PCB195 ng/l <0.52 <0.31 PCB200 ng/l <0.52 <0.31 PCB201 ng/l <0.52 <0.31 PCB203 ng/l <0.52 <0.31 PCB206 ng/l <0.52 <0.31 PCB209 ng/l <0.52 <0.31 Total PCBs ng/l < = results less than the reporting limit. volume 38.7L 65.2L

18 Study of C.L.A.M ability to analyze for ultra trace levels of PCB S, and PBDE s Spokane River Validation Study Traditional grab samples and hour time integrative extractions were obtained and compared The deployed disks and one liter grab water samples were submitted to the WDOE lab in Manchester WA. and the HRMS contract laboratory for elution, extraction and analysis The Disks were eluted and analyzed for PCB s PBDE s and Dioxins using standard and HRMS methods All three methods were performed on one disk extract Total and dissolved extractions were taken with staged disks

19 Comparison of Grab Samples to C.L.A.M Time Integrative In-Situ Extractive Samples. Upriver Nine mile Blank Parameter Grab ng/l C.L.A.M. ng/l Grab ng/l C.LA.M. ng/l Grab ng/l C.L.A.M. ng/l Monochlorobiphenyls 10 UI UL u UJ Dichlorobiphenyls 89 B B Trichlorobiphenyls 37.1 UJ B Tetrachlorobiphenyls 38.2 B B Pentachlorobiphenyls 26.9 B B Hexachlorobiphenyls 54 B B Heptachlorobiphenyls 10 UJ UL J Octachlorobiphenyls 10 UJ UL UJ Nonachlorobiphenyls 10 UJ UJ 10 UL UJ DecachlorobiphenyL 10 UJ UL J Total PCB 245 B B Volume Extracted 1.0 L 87.5 L 1.0 L 58.0 L 1.0 L 60.0 L The C.L.A.M extracts represented 58 L to 87.5 L The large volume extracted provided signal to noise peak ratios much higher than the 1 liter grab ratios The C.L.A.M time integrative extracts provided useable real data with no B flags, where the UJ detection levels were ultra low.

20 PCB Congeners PCB congeners in the CLAM samples gave a clear environmental signal with results that were 1-2 orders of magnitude higher than the laboratory method blank concentration. Precision of samples deployed in triplicate was excellent.

21 PBDEs Detection limits for many of the individual PBDE congeners reached down into the sub pg/l range for EPA 1614 and down into the sub ng/l range for EPA Even with this difference in detection levels and the difference in number of congeners analyzed between the two analytical methods, results between the two sets of triplicate data were highly comparable. Figure 6. CLAM Field Triplicate Results for Total PBDEs for Two Analytical Methods (pg/l, ppq).

22 Matrix Analysis 1. Was there a clear environmental signal above the analytical background noise (this was based on laboratory method blank and transfer blank contamination)? 2. Was the variability of field replicates and split samples of acceptable quality? 3. Is the field collection method easily reproducible on a larger scale? 4. Were detection limits low enough to evaluate applicable water quality standards? Surface Water CLAM Suspended Sedimets PCB congeners (1668c) PCB Aroclors (8082) Parameter & Analytical Method PBDEs (8270) PBDEs (1614) Dioxins & Furans (1613b) Metals Cd, Pb, Zn (200.8) 1. No 1. No 2. No P 2. No P 3. Yes Yes Yes 4. NA 1. Yes 1. No 1. Yes 1. Yes 1. No 2. Yes G 2. No P 2. Yes G 2. Yes G 2. No P 3. Yes 3. Yes 3. Yes 3. Yes 3. Yes Yes 4. No 4. NA 4. NA 4. Yes 1. Yes 1. Yes 1. Yes 1. Yes 2. Yes G 2. Yes G 2. No 2. Yes 3. Yes Yes 3. YesOK 3. Yes 4. NA 4. NA 4. NA 4. NA G

23 WA DOE Toxaphene ng/l SPMD concentration is the solid red line, CLAMs are the solid black bars, and grab samples are the grey bars (detection limits). Estimated concentrations of toxaphene in water using an active sampler (CLAM) and passive sampler (SPMD) were very similar.

24 Strengths Simple measured average concentration over deployment time just small Disk to Lab Clam Disks can be used with other pumps New hardware design anodized aluminum more robust and good to 30 Meters depth Standard Laboratory Methods achieve very low detections Multiple method analysis can be performed on a single solvent extract Disks can be frozen for months before analysis

25

26 Weaknesses High sediment loaded water creates challenges Lower volumes Complete clogging resulting in unknown volume On-Board Volume totalizer will solve the known volume issue, but sediment will still reduce the sample volume Storm Drain deployments or other Deployment issues Need 2 inches of constant depth unless tube is used and included in laboratory elution Only deploy to depths of 30 meters Disks can go lower with pump at surface Security must be considered when leaving equipment overnight

27 Costs CLAM $2500 purchase On-Board Totalizer will probably increase cost to $3500 $210-$500 per unit per month lease CLAM SPE Disks C-18 - $69-$89 HLB - $79-$99 Pre-Filter - $44-$54 Laboratory Costs Conditioning Disks pre-deployment Standard Analysis water extraction step already done in field Longer Holding Times Shipping Savings

28 Ongoing Projects and Reports LSU Wastewater Studies USGS 3 new projects in 2016 CADPR CA Water Board Lake Tahoe Other University Projects

29 Questions & Answers Jamie Aderhold