Performance Verification Statement for the Campbell Scientific Instruments pH Sensor

8 / 59

Ref. No. [UMCES] CBL 2015-008
ACT VS15-01
8
HI. The manufacturer representatives and the ACT Chief Scientist verified that all staff were
trained in both instrument and sample collection protocols.
This performance verification report presents instrument output in derived pH values reported
over time as directly downloaded from the test instruments or captured through independent
dataloggers. A summary of the testing protocols is provided below. A complete description of
the testing protocols is available in the report, Protocols for the Performance Verification of In
Situ pH Sensors (ACT PV12-01) and can be downloaded from the ACT website
.
Analysis of Reference Samples
The reference pH measurements were the pH of discrete water samples as determined by
pH indicator dyes, either, meta-cresol purple or phenol red (freshwater) (Liu et al. 2011, Yao and
Byrne 2001 respectively). All dyes used throughout the Verification were prepared and
characterized by Bob Byrne at the University of South Florida. An initial reading for
background correction was taken on each cell and then dye (from Byrne's lab) was introduced
into each cell (10 µl of purified 10 mM meta-cresol purple for samples with salinity > 5 and 10
µl of 10 mM phenol red for freshwater samples), mixed thoroughly, and the dye R ratio
measured on an Agilent 8453 spectrophotometer (the R ratio is the ratio of absorbance measured
at wavelengths corresponding to the peak absorbance of the acid and base forms of the dye).
The R value of each cell was measured minimum of 5 times in succession in the field test
and 10 times in succession for the lab test. pH was calculated from the R ratio, temperature and
salinity according to the equations published in Liu et al., 2011 (meta-cresol purple) or Yao and
Byrne 2001 (phenol red); both pH total scale and pH free scale are reported (Liu et al. 2011, Eq.
12). The actual sample pH at ambient temperature was calculated using CO2-Sys with
dissociation constants from Millero et al., (2006) and an estimated value of the alkalinity that is
based on the average of the samples analyzed by SIO. The anticipated level of accuracy of these
measurements is better than 0.01 pH for seawater (S ≥ 20), but actual laboratory based accuracy
and precision levels were quantified and reported for the brackish and freshwater environments.
Laboratory Test
Two thermally insulated, covered, 4.5 m3 “ source-water tanks” were filled with 1 µm
filtered seawater, or a mixture of freshwater and filtered seawater. The two tanks were then
isolated so that each could be maintained at a specific temperature. The source-water tanks were
used to supply water into a third smaller "test tank" (capacity - .8 m3), where all instrument
performance measurements took place. One tank with source-water was used to continuously
flush the third test tank (containing the in-situ instruments); and water from the second source
tank was used to create a quick transition to a new temperature condition within the test-tank.
Test conditions within the source-water tank were set and equilibrated for several days prior to
delivery into the test tank. The large volume of equilibrated water in the source-water tank
allowed for a rapid transition (10-15 minutes) of temperature and salinity conditions in the test
tank. Temperature was maintained within the source water tank to ± 1oC using an AquaLogic
MT-3 circulating heat exchanger. Water in both of the tanks was mixed continuously with
several submerged bilge pumps. Evaporation and heat exchange through the water surface was