June 2025

Welcome to interactive presentation, created with Publuu. Enjoy the reading!

Oceanography | Vol. 38, No. 2

68

is used for solutions of PC (Jaeschke et al., 2021) and PE (Ardiles

et al., 2020), and deionized water is used for RWT. The PIXIE is

compatible with the above solvents but cannot tolerate the sul­

furic acid that dissolves quinine sulfate, the stand-in for crude

oil calibration. Sulfuric acid is used almost exclusively in lit­

erature to prepare fluorescence standards of quinine sulfate

(Kristoffersen et al., 2018).

Sensor Design

Design notes for the PIXIE, available in the PIXIE Complete

User Guide document provided at the GitHub link, are described

briefly here for context. The PIXIE was designed with a focus

on user configurability, and customization if desired. Users can

simply acquire the hardware and assemble the default configura­

tion or use the PIXIE’s documentation if more detailed custom­

ization is needed.

The PIXIE performs fluorometry using standard optics

through an O-ring-sealed glass window. The user configures the

targeted fluorophore for each of the four channels by selecting

the appropriate optical filters and excitation LEDs. The PIXIE

described in this work was equipped with hardware to target

PC, RWT, Chl-a, and crude oil, though only the RWT and Chl-a

channels were used. The PIXIE cannot measure using more than

one channel at a time, but it can cycle between channels fast

enough to achieve quasi-simultaneous measurements.

To extract only the fluorescence excited by the device itself,

the PIXIE modulates the brightness of its excitation LEDs sinu­

soidally. While the change in brightness is imperceptible to the

eye, the resulting fluorescence will have a synchronous bright­

ness that can be distinguished from other sources of light. This

allows PIXIE measurements even in bright laboratory or sun­

lit outdoor conditions. The process of measuring the sinusoidal

fluorescence and converting it to a measure of fluorophore con­

centration, or lock-in amplification, is implemented digitally and

can therefore be adjusted by more technically inclined users. The

PIXIE detects the sinusoidal fluorescence using an AC-coupled

transimpedance amplifier with a software-​configurable gain of

400 MΩ, within a sample volume of 0.1 mL. More technical

details can be found on our GitHub page.

Calibration

The PIXIE’s RWT and Chl-a channels were calibrated in the

laboratory using a set of temperature-controlled standards.

Fluorescence is known to depend strongly on temperature

(Smart and Laidlaw, 1977), so the PIXIE was calibrated across

a range of temperatures and concentrations. The set of tempera­

tures and RWT concentrations were chosen to parallel previ­

ous work in this area (Park et al., 2023) for comparison’s sake.

The protocols used were adapted from a US Environmental

Protection Agency Method 445 on Chl-a fluorometer calibra­

tion (Arar and Collins, 1997). An effort was made to adapt the

protocol to keep the number of expensive lab instruments and

equipment to a minimum, though the protocol applied to Chl-a

required a fume hood. The PIXIE was suspended above a beaker

such that its sensing end was submerged without overflowing

or trapping air bubbles (see Figure 1, bottom inset). A complete

description of the calibration protocols is available in the PIXIE

Complete User Guide document on GitHub.

Field Deployment

A 10 L Niskin bottle (General Oceanics, Miami, FL, USA) was

prepared for use during an RWT tracer release experiment in

Halifax Harbor, as depicted in Figure 1. The bottle was equipped

with a proprietary datalogger that powered an array of exter­

nal sensors, including for temperature and depth. The PIXIE

was also mounted to the Niskin bottle, pointed downward, and

constantly streamed its fluorometry data to the logger via the

RS-232 protocol at a frequency of 16 Hz.

On August 10, 2023, a pre-set amount of RWT was released

from the Tufts Cove Power Generation Station to study the

dye plume. Multiple sensors and techniques were used, includ­

ing the PIXIE-integrated Niskin bottle and an ecoCTD (Dever

FIGURE 2. The PIXIE, in assem­

bled view (left) and in labeled

exploded view (right). For clar­

ity, only one channel is popu­

lated with optics. Cable, exci­

tation LEDs, and some O-rings

are omitted.

Made with Publuu - flipbook maker