Ref. No. [UMCES] CBL 2013-020
ACT VS12-03
protocols is provided below. A complete description of the testing protocols is available in the
report,
Protocols for the ACT Verification of In Situ Hydrocarbon Sensors
(ACT PV11-01) and
can be downloaded from the ACT website
.
Analysis of Reference Samples
Hydrocarbon concentrations
Diesel range hydrocarbons (C10 to C36) and volatile organic hydrocarbons were
analyzed by using GC-FID by the contract laboratory, Test America (West Sacramento Lab),
following their internal SOP’s based on EPA SW846 Method 8015B,C. The Laboratory
provides reporting limits of 50 ppb for this hydrocarbon range. Reference samples were
collected in certified pre-cleaned amber glass bottles supplied by Test America. Bottles were
filled, stored and shipped according to their SOP’s. Reference samples, along with sampling
blanks, were shipped to the contract lab not more than three days after collection to meet their
holding time requirements.
Excitation Emission Matrix Spectroscopy (EEMS)
A SPEX ISA Fluoromax-2 scanning spectrofluorometer, operated in ratio mode, was
used to generate EEM fluorescence spectra for all reference samples. To optimize sample
throughput, fluorescence spectra were determined over an excitation range of 230-500 nm at 5
nm intervals and an emission range of 300 – 600 nm at 3 nm intervals. For each scan, an
integration time of 1 second was used, and bandpass widths were set to 5 nm for both excitation
and emission spectrometers. Xenon lamp intensity as well as emission monochrometer
performance were verified and recalibrated once per day according to the instrument manual.
For all generated EEM’s, dark counts were subtracted and spectra were subsequently
corrected for wavelength-dependent instrument effects using ISA-supplied and user-generated
correction files. Fluorescence spectra intensities were then normalized to the area under the
Raman peak, determined daily using MilliQ water (Murphy, 2011; Murphy et al. 2010). This
value exhibited less than 2% variation over the length of the study period. In addition to daily
Raman scans, daily EEM’s of MilliQ water were generated as background blanks and were
subtracted from all subsequent sample EEM’s. At the beginning and end of each analytical batch
a four-point calibration curve (0-50 ppb) of Quinine Sulfate (QS) in 50 mM H
2
SO
4
was run to
track drift in fluorometer response over time
.
The QS response factor was used to standardize
emission intensities across each analytical batch (Coble et al. 1993). Finally, all sample EEM’s
were corrected for Raman and Rayleigh scattering peaks, following Zepp et al. 2004.
Excitation and emission windows for each instrument (as provided by manufacturers)
were mapped onto corrected EEM space and fluorescence intensity similarly regressed and
compared to regression of instrument reads versus challenge concentration in order to identify
sources of response variation observed.
Colored Dissolved Organic Matter (CDOM)
Approximately 50 ml of the CDOM designated subsample were filtered using 47 mm
GF/F filters with low vacuum pressure and poured into an acid-cleaned, combusted, 60 ml amber
glass bottle. All samples were stored in the dark at 4° C until analysis, within approximately one
6
