Rapid Opto-electrochemical Differentiation of Marine Phytoplankton

The use of electro-generated oxidants in seawater facilitates the discrimination of different plankton groups via monitoring the decay in real time of their chlorophyll-a (chl-a) fluorescence signals following potentiostatic initiation of electrolysis in their vicinity (YangM.Chem. Sci.2019, 10( (34), ), 7988−799331588335). In this paper, we explore the sensitivity of phytoplankton to different chemical species produced at various potentials in seawater. At low potentials, the oxidation of ca. millimolar bromide naturally present in seawater to hypobromous acid ‘switches-off’ the chl-a signal of individual Chlamydomonas concordia cells (green algae) located on the electrode surface within tens of seconds of the potential onset. At higher oxidative potentials, the oxidation of chloride and water produces oxidants (Cl2, OH, H2O2, etc.) that are also lethal to the plankton. To deconvolute the contributions to the response from the chemical identity of the oxidant and the amount of charge delivered to ‘titrate’ the individual living plankton using the loss of fluorescence as the ‘end point’, we introduce a ramped galvanostatic method. This approach enabled the controlled injection of charge applied to a bespoke electrochemical cell in which the plankton are immobilized on an electrode surface for rapid and sensitive measurement. It is shown that the number of moles (charge) of oxidants required to react leading to chl-a switch-off is independent of the chemical identity of the electro-generated oxidant(s) among hypobromous acid, chlorine, or water-derived oxidants. Comparative experiments between C. concordia and Emiliania huxleyi (where the latter are encapsulated by extracellular plates of calcium carbonate) show that significantly different amounts of absolute charge (moles of electro-generated oxidants) are required in each case to ‘switch-off’ the chl-a signal. The method provides the basis for a tool that could distinguish between different plankton cells within ca. 2 min including the setup time.


Table of Contents
. Specifically, the oxidizable species on a carbon electrode present at sufficient concentrations in seawater at potentials near or below water breakdown are 0.84mM Br -(aq), 0.56M Cl -(aq) and 55M H 2 O (l). Section 2: Pourbaix diagram: bromine Figure S1 shows a Pourbaix diagram of bromine in an aqueous solution at 25°C. As can be seen, near 1.0V vs SCE (+0.244V vs SHE at 25°C), the thermodynamic product as a result of the oxidation of bromide at pH 8.2 is hypobromous acid (HOBr) and its conjugate base (BrO -). 4 Since the concentration of bromide in seawater is 0.84 mM, the upper limit of HOBr formed in the vicinity of the electrode interface, assuming equal diffusion coefficients, is 0.84mM. Figure S1. Pourbaix diagram of bromide as a function of pH and potential. Calculated using Hydra/Medusa. 5 Figure S2 shows the chl-a fluorescence of the C. concordia sample measured in a fluorometer after the addition of bromine water. The experimental procedure is reported elsewhere. 6 The bromine water as supplied by the manufacturer is highly acidic with a pH value of 1, to mitigate any pH effects on the chl-a fluorescence of the C. concordia sample the bromine water was adjusted to pH 8.2 prior to the experiment. After the addition of the bromine water, the sample is stirred and the fluorescence signal is measured immediately afterwards with a timescale of tens of seconds of chemical exposure similar to that generated in situ via electrochemistry. As can be seen in Figure   S2, addition of pH adjusted bromine water to a final concentration of sub-millimolar (0.42 and 0.84mM) results in a catastrophic drop in the measured chl-a fluorescence of the C. concordia. This is in excellent agreement with that observed with in situ electrochemistry.  Figure S3 shows the cyclic voltammograms of 0.42M KNO 3 (aq) and 0.42M NaCl(aq)

Section 3: Oxidation of Cland H 2 O
recorded on a glassy carbon electrode at a voltage scan rate of 0.1 Vs -1 . In the NaCl solution, an exponential rise in current can be seen near +1.2 V. In the KNO 3 solution, however, a higher cathodic potential is required to pass a similar magnitude of current.
In the absence of chloride, in the KNO 3 Figure S4 shows the average chl-a response of C. concordia as a function of potentials in a high-salt electrolyte solution containing a) 0.42M KNO 3 (aq) and b) 0.42M KCl. In the control experiment, the opto-electrochemical cell is disconnected from the potentiostat and the fluorescence signal of the plankton sample is measured as a function of time. As can be seen, in the presence and absence of chloride, a different threshold potential is required to result in a substantial drop in the chl-a intensity over the tens of seconds of the experimental timescale. In both cases, as the potential becomes more cathodic, the time required to switch-off the chl-a fluorescence decreases. In the case of chloride-containing medium (0.42M NaCl (aq)), a threshold potential of +1.2 V vs SCE is required to drive a substantial decrease in the chl-a fluorescence. In the absence of Cl -, a higher threshold potential of 1.4 V is required. In both cases, the threshold potential is in full consistency with that seen in the respective voltammograms.

Derivation: constant current
The steady-state mass transport flux, (mol s -1 ), to a spherical particle on a flat

Equation S1
= (2)4 , ℎ where is the radius of the spherical particle (m) on the electrode and is gives shown in the main text providing an estimated number of moles of oxidant that reacts with phytoplankton under mass-transport control.

Derivation: ramping current
For galvanostatic experiments with a linear ramp in current, , where the is ( ) = the rate of current ramp (As -1 ), the interfacial concentration of oxidants is given by 8,10 Equation S3 (0, ) = where is the mathematical gamma function. Substituting Equation S3 into (5/2) Equation S1 gives

Validation: comparison of analytical expression versus numerical simulation
The experiment presented in the main body of the article can be approximately considered as a sphere on a 'plate', where the sphere is the phytoplankton and the supporting plate is the electrode. Here the electrode generates material that is consumed at the particle surface. First, we numerically consider the solution to this mass-transport problem and second, we compare the numerical result to the analytically approximate solution used in the main body of the text.
In this model system species, A is generated at a surface with a fixed current density.
Species A can then diffuse away from the surface and is irreversibly consumed at the surface of the particle, where the rate of consumption at the particle surface is simply limited by the rate of mass-transport. Initially there is no species A present in the solution phase. Further, due to the symmetry of the system, this particle on a plate problem only needs to be considered in two dimensions (using cylindrical coordinates).

Figure S6
Schematic of the mass-transport problem that needs solving to ascertain the diffusion limited flux to a particle on a generating surface where the flux at the electrode surface is a constant (X). The nomenclate used in this figure follows that used in ref. 11 Axes are labelled as R and Z and the origin is defined as the point of contact between particle and the electrode surface. Figure S4 presents a schematic summary of the differential equation and boundary conditions that needs to be solved for. To numerically solve this diffusion-only problem we use a fully implicit central finite difference scheme. 11 The numerical solution of the resulting system of simultaneous equation was achieved using a GPU optimised iterative solver. 12 Figure S7 presents a series of example concentration profiles obtained from this simulation, showing how the concentration profile evolves as a function of dimensionless time (τ = Dt/r 2 ). In this figure, only half of the particle is shown, further, the x and y coordinates are normalised to the size of the particle where the particle has a radius of one. At short times (τ = 0.1) the diffusion layer is small as compared to the size of the particle. As the time increases the diffusion layer increases and at τ = 10 it expands beyond the dimensions of the particle. From these concentration profiles it is possible to calculate the expected flux to the particle. Figure S8    huxleyi cells. As can be seen, due to the relatively low concentration of bromide, the depletion of bromide at the electrode interface occurs before the switch-off of either of the two phytoplankton species studied. The time required to deplete the interfacial concentration of chloride, however, is longer than that required to switch-off off the chla fluorescence of the two phytoplankton species. Therefore, we conclude that in the galvanostatic experiments, in seawater and at low applied currents, initially bromide is oxidised to form oxidants at the electrode interface, it depletes quickly due to the submillimolar concentration and it is the oxidants produced by oxidation of chloride (and water at higher cathodic potentials) that is responsible for the chl-a switch off.