SUPPLEMENTARY INFORMATION
Sample Processing
The use of an automated SPE extraction method for the water samples provided better extraction efficiency and consistent data through out the program. The SPE extraction relies on a 50-mm Bakerbond Speedisk DVB to remove and retain the PCB congeners from the water samples. NEA did extensive research when setting up this extraction method. This included how it affects the retention and loss of PCB congeners, particularly the mono-, di-, and trichlorinated congeners. The retention of the congeners on the disk and the subsequent release of the PCBs from the disk have been well documented. NEA tracks each congener recovery/loss through several quality assurance and control steps; these include method detection limit studies (MDL), precision and accuracy studies (P&A), laboratory control spikes (LCS), andStandard Reference Materials (SRM). These data show that the recovery of the mono-, di-, and trichlorinated congeners range from 87% to 96%, with a standard deviation range of 0.0129-0.123 and a RSD of less than 10%. Therefore, we can conclude that the automated SPE extraction, including air-drying of the disk, does not impact the lower-chlorinated PCB congeners with volatile loss.
TABLE S1. DB-1 Column Peaks and Eluting Congeners.
PeakA / PCB CongenerB / Homolog(s) / Peak / PCB Congener / Homolog(s)2 / 1 / 1 / 61C / 77,110,148 / 4, 5, 6
3 / 2 / 1 / 62 / 154 / 6
4 / 3 / 1 / 63 / 82 / 5
5 / 4,10 / 2 / 64 / 151 / 6
6 / 7,9 / 2 / 65 / 124,135 / 5, 6
7 / 6 / 2 / 66 / 144 / 6
8 / 5,8 / 2 / 67 / 107,109,147 / 5, 6
9 / 14 / 2 / 68 / 123 / 5
10 / 19 / 3 / 69 / 106,118,139,149 / 5, 6
11 / 30 / 3 / 70 / 140 / 6
12 / 11 / 2 / 71 / 114,134,143 / 5, 6
13 / 12,13 / 2 / 72C / 122,131,133,142 / 5, 6
14 / 15,18 / 2, 3 / 73 / 146,165,188 / 6, 7
15 / 17 / 3 / 74 / 105,132,161 / 5, 6
16 / 24,27 / 3 / 75 / 153 / 6
17 / 16,32 / 3 / 76 / 127,168,184 / 5, 6, 7
18 / 77 / 141 / 6
19 / 23,34,54 / 3, 4 / 78 / 179 / 7
20 / 29 / 3 / 79 / 137 / 6
21 / 26 / 3 / 80 / 130,176 / 6, 7
22 / 25 / 3 / 81
23 / 31 / 3 / 82 / 138,163,164 / 6
24 / 28,50 / 3, 4 / 83 / 158,160,186 / 6, 7
25 / 20,21,33,53 / 3, 4 / 84 / 126,129 / 5, 6
26 / 22,51 / 3, 4 / 85 / 166,178 / 6, 7
27 / 45 / 4 / 86
28 / 36 / 3 / 87 / 175,159 / 6, 7
29 / 46 / 4 / 88 / 182,187 / 7
30 / 39 / 3 / 89C / 128,162 / 6
31 / 52,69,73 / 4 / 90 / 183 / 7
32 / 43,49 / 4 / 91 / 167 / 6
33 / 38,47 / 4 / 92 / 185 / 7
34 / 48,75 / 4 / 93 / 174,181 / 7
35 / 62,65 / 4 / 94 / 177 / 7
36 / 35 / 3 / 95 / 156,171 / 6, 7
37C / 104,44 / 5, 4 / 96 / 157,202 / 6, 8
38 / 37,42,59 / 3, 4 / 97
39 / 41,64,71,72 / 4 / 98 / 173 / 7
40 / 99 / 201 / 8
41 / 68,96 / 4,5 / 100 / 172,204 / 7, 8
42 / 40 / 4 / 101 / 192,197 / 7, 8
43 / 57,103 / 4, 5 / 102 / 180 / 7
44 / 58,67,100 / 4, 5 / 103 / 193 / 7
45 / 63 / 4 / 104 / 191 / 7
46 / 74,94,61 / 4, 5 / 105C / 200,169 / 6, 8
47 / 70 / 4 / 106 / 170 / 7
48C / 66,76,98,80,93,95,102,88 / 4, 5 / 107 / 190 / 7
49 / 55,91,121 / 4, 5 / 108 / 198 / 8
50 / 56,60 / 4 / 109 / 199 / 8
51 / 84,92,155 / 5, 6 / 110 / 196,203 / 8
52 / 89 / 5 / 111 / 189 / 7
53 / 90,101 / 5 / 112 / 195 / 8
54 / 79,99,113 / 4, 5 / 113 / 208 / 9
55 / 119,150 / 5, 6 / 114 / 207 / 9
56C / 78,83,112,108 / 4, 5 / 115 / 194 / 8
57 / 97,152,86 / 5, 6 / 116 / 205 / 8
58 / 81,87,117,125,115,145 / 4, 5, 6 / 117 / 206 / 9
59 / 116,85,111 / 5 / 118 / 209 / 10
60 / 120,136 / 5, 6
ANote that five DB-1 peaks (PK18, PK40, PK81, PK86, PK97) have been removed from the DB-1 peak numbering scheme. The following low level congeners that were designated as separately eluting peaks have been determined to co-elute with another congener. The DB-1 peak numbers are no longer required for these congeners, but the original DB-1 numbering system has remained intact for all other peaks.
PK 18 (23) now elutes in PK 19 (23,34,54)
PK 40 (68) now elutes in PK 41 (68,96)
PK 86 (166) now elutes in PK 85 (166,178)
PK 97 (157) now elutes in PK 96 (157,202)
BPCB congener numbers listed in bold were found to be present in at lease one of the Aroclors at or above 0.05 weight percent. These congeners should be considered the primary congeners existing in a peak composed of coeluting congeners. Congener numbers listed in italics were absent or present below 0.05 weight percent. CDB-1 peaks may include one or more coeluting PCB congeners. In the case of some peaks, the congeners assigned to the peak consist of coeluting congeners and a congener that is resolved or is just slightly out of the normal retention time window of ± 0.07 minutes. If detection of one of the resolved congeners occurs, a comment will be included in the report narrative indicating the assigned DB-1 peak includes the presence of the resolved congener. The DB-1 peaks consisting of coeluting congeners and a congener that is resolved are as follows:
DB-1 Peak / Resolved PCB Congener37 (44, 104) / 104
48 (66,76,98,80,93,95,102,88) / 80,88,93
56 (78,83,112,108) / 108
61 (77,110,148) / 77
72 (122,131,133,142) / 122
89 (128, 162) / 162
105 (200, 169) / 169
Aroclor Mix Standard
For the past 20 years, NEA has maintained a Monsanto Aroclor batch library. The batches of Aroclor mixtures NEA possesses have been well documented. Each Aroclor stock standard is made yearly and checked against second source of neat material from AccuStandard (New Haven, CT) and Chem Services (West Chester, PA). These standards are also checked against NEA’s congener specific methodology SW-846 Method 8082 With Comprehensive Quantitative Congener-Specific Analysis (CQCS). This method uses the 209 individual congener mix in which each congener is at the same concentration to calibrate and measure. If our Aroclor mix standards show any changes from our library, the batch is rejected and remade.
Although we are aware of the variations in Aroclor batches, NEA has used the same batch library for the mGBM method and the Hudson River Programs for the past 18 years. This Aroclor-based standard is also very comparable to past methods run and data collected for the Hudson River. The comparison of the same sample analyzed on both GC/ECD methods employed at NEA (mGBM and CQCS) have comparable results of congener versus congener, homolog groups, and total PCBs of +/- 10%.
Linearity Verification and Initial Calibration
The Instrument Calibration Standard (ICAL) is a Mixed Aroclor Calibration Standard. This stock standard is prepared by combining A1232, A1248, and A1262 in a 25:18:18 ratio with a final mixture concentration of 25.7 ug/mL, 18.6 ug/mL, and 18.4 ug/mL respectively (total = 62.7 ug/mL). These ratios are strictly maintained so that the percent composition data remains applicable, since it was developed for use under these fixed mixture parameters. The final concentration of the mixed standard may vary to accommodate instrument sensitivity or more closely represent sample concentrations, but the same ratio values must be maintained. Prior to sample analysis, a four-point calibration for each separable (single PCB congener or congener mixture) chromatographic peak was performed. There are 22 DB-1 peaks that cannot be measured in Calibration Standard 4 (low-calibration standard), and only a three-point calibration curve was generated for the following peaks: DB-1 peaks 10, 13, 20, 44, 52, 55, 56, 67, 72, 79, 84, 87, 89, 91, 98, 101, 104, 108, 111, 113, 114, and 118. The ICAL was analyzed at the following levels: 6.25ng/mL, 12.5 ng/mL, 25 ng/mL, 314 ng/mL and 627 ng/mL. Relative response factors the laboratory limit for relative standard deviation (RSD) for each GC peak of less than 20%. The low-level initial linearity is comprised of the following Mixed Aroclor Calibration standards: 6.27 ug/mL, 1.27 ug/mL, and 0.127 ug/mL. The correlation coefficient of the three- or four-point calibration curve was ≥ 0.995 for each quantified peak using a linear equation with 1/X weighting. The three- or four-point calibration curve met a %RSD of ≤ 20% for each quantified peak.
A Supplemental Congener Standard was analyzed along with the Mixed Aroclor Calibration Standard. This standard contains congeners that exist at low levels in the mixed Aroclor standard and comprises congeners that are not typically found in Aroclor formulations, but could become important in by-product PCB analysis or the study of model experiments that use unusual PCB congeners. This standard is analyzed to supply accurate retention time information and response factors for quantification. The supplemental stock standard consisted of IUPAC congeners 2, 14, 30, 11, 34, 36, 39, 65, 35, 96, 103, 154, 123, 140, and 127. The standard concentration was 2.00 ug/mL for 3-chlorobiphenyl and 0.050 ug/mL for all other congeners in the standard. Continuing calibration check standards (CCCS) are also analyzed at 0.122 ug/mL and at a frequency of one every 10 samples analyzed. The CCCS was prepared from Mixed Aroclor solutions (A1232, A1248, and A1262 in a 25:18:18 ratio) obtained from a different source (ULTRA Scientific) than the calibration standard.
A surrogate compound was added to each sample, matrix spike, matrix spike duplicate, duplicate, method blank, and laboratory control spike at the time of extraction. The surrogate compound chosen for this methodology was PCB 2,2’,3,3’,4,4’,5,6,6’ – nonachlorobiphenyl, IUPAC congener 207 (AccuStandard Cat. No. C-206S-TP). This congener is not present or found at trace amounts in Aroclor formulations. The surrogate stock standard is prepared from 5.0 mg of the neat standard material and is dissolved into 50 mL of hexane, for a concentration of 100 ug/mL. A volume of 0.50 mL at 0.20 ug/mL 2,2’,3,3’,4,4’,5,6,6’-nonachlorobiphenyl was added to the water sample prior to extraction.
Octachloronaphthalene (OCN) was used as an internal standard (IS). The surrogate stock standard was prepared from 10.1 mg of solid OCN (primary: Ultra Scientific, Hope, RI; secondary: Accustandard, New Haven, CT) dissolved into 50 mL of toluene for a concentration of 202 ug/mL. The IS was added to all calibration standards, CCCS, blanks, samples, and QC samples prior to GC analysis. In most cases, this was achieved by spiking 4.5 uL of OCN IS solution to 5 mL of sample extract to give a solution concentration of 0.1818 ug/mL. Thus, the IS was used as a quantification spiking standard and eliminated sample injection volume variations, but did not correct for analytical losses during sample preparation. A portion of the IS was transferred to a 5-mL reacti-vial with a Teflon syringe-valve cap to use on a daily basis. The reacti-vial minimizes evaporation since the cap does not have to be removed.
Laboratory Control Spike (LCS), also referred to as a QC reference check standard, was extracted with each batch of samples at a rate of one per 20 samples. One liter of laboratory organic free water was spiked with 200 µL of A1242 at 1.0 ug/mL (at 200 ng/L total PCB) extracted and analyzed.
GC Data System
NEA uses a chromatography software module from Waters (Waters, Milford, MA). The Empower GC workstation/database is interfaced to the GC and an in-house laboratory information management system (LIMS) is utilized for data handling, reporting, and data validation package generation. The Empower workstation acquires and processes the analog detector signal and converts it to a digital format. Empower then stores the digitized chromatograms in a database on a designated computer server. All phases of integration of peak areas, quantitation and production of chromatograms and raw data reports are performed within the Empower software. This processed data is then uploaded to the LIMS for further data handling, scoring, reporting, validation packaging, and data distribution.
Due to the complex nature of the PCB patterns encountered and the range of peak intensities that can occur in the sample chromatograms, manual peak integration was performed to accurately integrate the samples. Manual peak integration was also performed on standards to best address the changing signal intensities. Manual integration provides for better peak start and peak end positioning, better control of peak baselines and more accurate data. The congener analytes were identified by matching retention time to the calibrated peak in the initial calibration standard that are within the retention time window of ±0.07 min. The capillary column GC analysis is performed by an internal standard calibration technique.
NE273_01 Aroclor Peak Quantitation
For each Aroclor, five peaks were selected to quantify the amount of that Aroclor. The peaks selected from the multi-component Aroclor formulations were based on maximizing the separation for each Aroclor (i.e., minimizing peak overlap in retention time). Consideration was also given to selecting peaks that normally did not have problems with co-elution with interfering peaks or possible co-elution with organochlorine pesticides. The determined area of the five peaks selected for calibration is processed by the data workstation as a group, combining the area for calculations of the calibration factors.
Water Treatment Plant Operations
Stillwater
Treatment consists of injection of NaOCl (prechlorination), and KMnO4 (iron precipitation), addition of soda ash (pH adjustment), and anthracite/manganese green sand for filtration and iron removal. Zinc orthophosphate is added for corrosion control, and NaOCl for residual disinfection. The system produces between 0.3 – 0.45 mgd, and serves 3200 people.
Halfmoon
Trident-Actifloc (conventional) filtration as follows: raw water is injected with polyaluminum chloride coagulant (PCH180) and powdered activated carbon (Calgon WPH 1000), followed by flocculation, upflow clarification, bed filtration (anthracite over sand), and NaOCl addition for residual disinfection. This system produces 1.5 MGD and serves 13000 people.
Waterford
Conventional treatment as follows: raw water is aerated then injected with polyaluminum chloride coagulant (PCH180), and powdered activated carbon (Calgon WPH 1000) is added via a feed hopper and auger. This is followed by flocculation, sedimentation, chlorine gas addition, and mixed media filtration (anthracite/sand). Sodium fluorosilicate is added (fluoridation), soda ash (pH adjustment), and chlorine gas (residual disinfection). The system produces an average of 1.1 MGD and serves 9800 people.
Green Island
Conventional filtration as follows: raw water is injected with KMnO4 (iron oxidation) and coagulant, followed by flocculation, sedimentation, and mixed media filtration (anthracite/sand). NaOCl is added for residual disinfection. This system produces about 0.7 MGD and serves 3800 people.
Rhinebeck
Conventional filtration is primarily as follows: raw water injected with coagulant (PCH180) and floc enhancer, flocculation, sedimentation, and rapid sand filtration. Chlorine gas (residual disinfection) and polyphosphate inhibitor (corrosion control) are added prior to distribution. The system produces an average of 0.55 MGD and serves 4300 people.
Port Ewen
Conventional filtration as follows: coagulant addition using alum injection; upflow clarification; and mixed-media/rapid sand filtration (anthracite and sand). Chlorine gas (residual disinfection) and lime slurry (pH adjustment, corrosion control) are added prior to distribution. Average daily production is 0.5 MGD and serves 4800 people.
Poughkeepsie
Conventional treatment as follows: raw water injection with polyaluminum chloride coagulant and polymer in rapid mix tank; solids contact/flocculation; sodium hypochlorite application to control bioactivity/virus inactivation; sedimentation; aeration; dual-media (anthracite-sand) rapid-rate filtration; UV disinfection; and clearwell addition of sodium hypochlorite and caustic soda (pH adjustment). Orthophosphate inhibitor (corrosion control) and ammonia (chloramine residual disinfection) are added prior to distribution. The system produces an average of 10.5 MGD and serves 75,000 people.
FIGURE S1. Congener profiles (percent distribution) for raw and finished water at Halfmoon (a, b), Rhinebeck (c, d), and Port Ewen (e, f). Peaks 97 to 118 are not shown.
TABLE S2. Mean Percent Distribution of Homologs in Raw Water Samples analyzed by the mGBM.
Homolog / Stillwater(n=12) / Halfmoon
(n=12) / Waterford
(n=11) / Green Island
(n=7) / Rhinebeck
(n=7) / Port Ewen
(n=7) / Poughkeepsie
(n=7)
1 / 8.5 / 8.0 / 9.6 / 0 / 0 / 0 / 0
2 / 32.6 / 45.4 / 43.8 / 4.6 / 18.1 / 22.7 / 15.0
3 / 38.7 / 27.3 / 26.3 / 48.0 / 36.1 / 34.0 / 37.0
4 / 18.4 / 14.1 / 13.7 / 34.6 / 26.3 / 29.2 / 27.6
5 / 1.7 / 5.1 / 6.5 / 12.8 / 17.4 / 13.2 / 17.4
6 / 0 / 0.2 / 0.1 / 0 / 2.0 / 0.8 / 2.6
7 / 0 / 0 / 0 / 0 / <0.1 / 0 / 0.2
8 / 0 / 0 / 0 / 0 / 0 / 0 / 0.1
9 / 0 / 0 / 0 / 0 / 0 / 0 / <0.1
10 / 0 / 0 / 0 / 0 / 0 / 0 / 0
TABLE S3. Mean Percent Distribution of Homologs in Finished Water Samples analyzed by the mGBM.
Homolog / Stillwater(n=12) / Halfmoon
(n=12) / Waterford
(n=11) / Green Island
(n=7) / Rhinebeck
(n=7) / Port Ewen
(n=7) / Poughkeepsie
(n=7)
1 / 8.8 / 0.0 / 0.0 / 0.0 / 0.0 / 0.0% / 0.0%
2 / 33.8 / 58.9 / 49.1 / 8.7 / 39.9 / 43.2% / 20.2%
3 / 37.4 / 20.8 / 25.9 / 54.3 / 26.4 / 34.8% / 41.6%
4 / 18.0 / 7.0 / 15.7 / 23.8 / 19.0 / 15.1% / 27.8%
5 / 1.9 / 12.2 / 8.5 / 13.2 / 14.5 / 6.5% / 10.0%
6 / 0.1 / 1.0 / 0.7 / 0.0 / 0.2 / 0.4% / 0.5%
7 / 0 / 0 / 0 / 0 / 0 / 0 / 0
8 / 0 / 0 / 0 / 0 / 0 / 0 / 0
9 / 0 / 0 / 0 / 0 / 0 / 0 / 0
10 / 0 / 0 / 0 / 0 / 0 / 0 / 0
TABLE S4. Mean Raw water homolog concentrations (in nanograms per liter).
Homolog / Stillwater(n=12) / Halfmoon
(n=12) / Waterford
(n=11) / Green Island
(n=7) / Rhinebeck
(n=7) / Port Ewen
(n=7) / Poughkeepsie
(n=7)
1 / 10.94 / 2.16 / 2.89 / 0 / 0 / 0 / 0
2 / 42.46 / 11.08 / 12.38 / 0.10 / 4.87 / 5.23 / 6.42
3 / 50.51 / 7.25 / 7.30 / 0.50 / 9.90 / 8.15 / 16.20
4 / 24.05 / 3.65 / 3.82 / 0.42 / 7.18 / 7.09 / 12.53
5 / 2.32 / 1.24 / 1.76 / 0.17 / 4.75 / 3.38 / 8.30
6 / 0.03 / 0.03 / 0.02 / 0 / 0.57 / 0.24 / 1.46
7 / 0 / 0 / 0 / 0 / <0.01 / 0 / 0.12
8 / 0 / 0 / 0 / 0 / 0 / 0 / 0.05
9 / 0 / 0 / 0 / 0 / 0 / 0 / 0.02
10 / 0 / 0 / 0 / 0 / 0 / 0 / 0
TABLE S5. Mean Finished water homolog concentrations (in nanograms per liter).
Homolog / Stillwater(n=12) / Halfmoon
(n=12) / Waterford
(n=11) / Green Island
(n=7) / Rhinebeck
(n=7) / Port Ewen
(n=7) / Poughkeepsie
(n=7)
1 / 11.95 / 0 / 0 / 0 / 0 / 0 / 0
2 / 44.67 / 6.80 / 4.92 / 0.15 / 5.39 / 4.11 / 1.05
3 / 49.74 / 2.35 / 4.07 / 0.49 / 3.38 / 3.23 / 2.48
4 / 24.14 / 0.53 / 2.61 / 0.30 / 2.48 / 1.55 / 1.61
5 / 2.60 / 1.09 / 1.13 / 0.16 / 1.67 / 0.50 / 0.55
6 / 0.09 / 0.03 / 0.07 / 0 / 0.03 / 0.04 / 0.02
7 / 0 / 0 / 0 / 0 / 0 / 0 / 0
8 / 0 / 0 / 0 / 0 / 0 / 0 / 0
9 / 0 / 0 / 0 / 0 / 0 / 0 / 0
10 / 0 / 0 / 0 / 0 / 0 / 0 / 0