Opto‐Electrochemical Dissolution Reveals Coccolith Calcium Carbonate Content

Abstract Coccoliths are plates of biogenic calcium carbonate secreted by calcifying marine phytoplankton; annually these phytoplankton are responsible for exporting >1 billion tonnes (1015 g) of calcite to the deep ocean. Rapid and reliable methods for assessing the degree of calcification are technically challenging because the coccoliths are micron sized and contain picograms (pg) of calcite. Here we pioneer an opto‐eletrochemical acid titration of individual coccoliths which allows 3D reconstruction of each individual coccolith via in situ optical imaging enabling direct inference of the coccolith mass. Coccolith mass ranging from 2 to 400 pg are reported herein, evidencing both inter‐ and intra‐species variation over four different species. We foresee this scientific breakthrough, which is independent of knowledge regarding the species and calibration‐free, will allow continuous monitoring and reporting of the degree of coccolith calcification in the changing marine environment.


Introduction
Coccolithophores play af undamental role in the carbon cycle,producing over 1billion tonnes (10 15 g) of calcite and an estimated 6 10 25 individual coccoliths each year. [1] Coccoliths dominate the calcareous pelagic sediments [2] and are responsible for approximately half of open ocean calcite precipitation. [1] Thec alcite mass of each individual lith represents the intensity and/or rate of calcite production by coccolithophore cells,a nd is an important biogeochemical parameter in terms of the impact of coccolithophore production on the alkalinity budget of the surface ocean, [1] which dictates the air-sea flux of CO 2 . [3] Thec alcite mass of individual liths also contributes to the export efficiencyo f organic carbon to depth, the dense calcite mineral aids the ballast of buoyant organic matter and prevents the remineralization of the latter by bacteria as it traverses down the water column. [4] In addition to being af undamental component of the geological cycle of carbon, the mass of secreted coccoliths is also abiologically important characteristic of the cell, [5] and species [6] potentially yielding information about cellular adaptation to growth conditions such as nutrient availability and carbonate chemistry. [7] Thes ecretion of coccoliths onto the surface of ac occolithophore to form an inter-locking mineralized layer can be at arate of up to 1-2 per hour. [8] These elaborate bioengineered coccoliths generally,but not exclusively,consist of alternating nano-units of calcite arranged with the optical axis radial (R units) and vertical (V units) to the plane of the coccolith. [9] In the lifecycle of acoccolithophore,such as the most abundant species Emiliania huxleyi,many coccoliths become detached, resulting in ar atio of detached-coccolith vs.i ntact-coccospheres as high as % 100:1 in surface waters. [4b] Due to the scale of phytoplankton blooms and the light-scattering properties of calcite,awhitening of surface waters can be observed from space. [10] Satellite studies have revealed that the global coverage of this enhanced scattering in surfacewater, as ar esult of dense detached liths,i s1 .4 10 6 km 2 (averaged from [1979][1980][1981][1982][1983][1984][1985]. [11] On ag lobal scale,a lthough satellite studies have attempted to convert the remotelysensed optical images to maps of calcite mass,t his required relating the light-scattered images remotely acquired from space to local conditions in the surface waters,s uch as the number and density of coccolithophores and coccolith mass. [12] More generally,t oe stimate the mass of coccoliths in the open ocean or sediment samples,m any use values statistically averaged over large sample sizes,f or example, shape factor, [13] which do not account for the natural variability in particulate inorganic carbon per coccolith that inevitably results from changes in coccolithophore diversity and environmental stressors in the marine environment. [5a,14] Mass estimation of individual coccolith, and the CaCO 3 content, in the local marine environment will provide as olution to this global challenge which serves as ab eacon reporting changes in the marine carbonate chemistry. Bio-accumulated calcium carbonates in the form of coccoliths excreted by phytoplankton vary markedly between species exhibiting significant intra-species diversity,asshown in the SEM images in Figure 1. This image depicts loose/ detached coccoliths from four globally significant species of coccolithophores (E. huxleyi, Calcidiscus leptoporus, Gephyrocapsa oceanica and Coccolithus pelagicus subsp.b raaudii).
These four species reflect the differing sizes,s hapes and morphology of the bio-accumulated calcium carbonate.D ue to the small coccolith sizes (Figure 1), the mass of each individual is too small for accurate weight measurement (pico-to nano-grams) using traditional methods.Accordingly, the mass of individual liths is such ab iogeochemically and biologically important parameter that an umber of methods have been developed for this measurement. One approach is high-resolution x-ray nanotomography that allows coccoliths to be 3D reconstructed. [15] Alternatively,c ircular polarised light techniques,w hich have become popular over the past decade,a re restricted to species which have am aximum calcite thickness of 1.56 mmf or imaging with ab lack-andwhite camera or < 4.5 mmi nc olour. [16] These techniques utilize the birefringence property of calcite crystalline such that retardation of the polarised light emerging from the crystal provides information on crystal thickness. [16a] Prior to measurement, ac alibration for pixel-intensity to calcite thickness is necessary which requires as pecial cylindricalshaped coccolith called rhabdolith. [16b,17] Moreover,f or birefringence to occur, the optical axis of the calcite crystalline has to be perpendicular to the incident polarised ray.T his is problematic for coccoliths composed of entirely Vunits (or am ixture of Ra nd Vu nits), as the coccolith can appear entirely (or partially) optically isotropic to the incident ray. [16a] An alternative approach to image particles with lightscattering properties is dark-field microscopy.Whencoupled with in situ electrochemical techniques this optical approach can provide real time visualization of particles of nanometre size undergoing dynamic reactions triggered by as uitable electrochemical potential. [18] Our study uses electrochemistry to trigger the calcite dissolution of individual coccolith which is monitored in situ via dark-field optical measurements allowing 3D reconstruction of the initial coccolith volume, and the CaCO 3 content, prior to dissolution. This volume measurement is independent of knowledge of the crystalline orientation of the calcite in the coccolith, and applicable to coccoliths of any mass or thickness. Analytical "titration" of the calcium carbonate content of individual liths is achieved within tens of seconds by the controlled dissolution of asingle coccolith with acid electrochemically generated within an opto-electrochemical cell.

Results and Discussion
In the following,asuspension of detached coccoliths is placed in an opto-electrochemical cell [19] which allows the generation of at iny amount of acid local to some of the coccoliths as shown in the schematic in Figure 2. Theresulting controlled dissolution of coccoliths is monitored optically, allowing their size to be measured as the dissolution proceeds to complete "titration" of the calcium carbonate content. A series of images recorded during the course of dissolution gives shape and size information, related to the aggregate of bio-mineralized calcite crystals which constitute the coccolith.
Specifically,aplankton sample containing coccoliths was mixed with at iny quantity of electroactive acid precursor (10 mM 1,4-dihydroxybenzene,H 2 BQ), the sample was placed in at hin-layer electrochemical cell and the individual coccoliths monitored by dark-field optical microscopy.T he thin-layer electrochemical cell consists of at hree-electrode setup, [20] including ac arbon fibre working electrode (diameter, 7 mm) in ac ell volume of approximately 1cm 1cm 100 mm( Figure S2). Theo ptical image acquisition of the coccolith was obtained using a4 0 objective lens and was digitally synchronized with the electrochemical system. SI sections 1a nd 2p rovide further details regarding the experimental design and setup.During the experiment, oxidation of the acid precursor (+ 1.2 Vvs. Ag wire) leads to adecrease in the pH local to the working electrode and this drives the dissolution of calcium carbonate.
where BQ is benzoquinone.SIsection 3provides more details on the electrochemistry of H 2 BQ.F igure 3a depicts the temporal evolution of ar epresentative coccolith detached from C. braaudii and imaged using darkfield microscopy.The electrolyte is the K/2 growth medium [21] in which the cells were grown. This coccolith was located at ad istance of 45.5 mma way from the electrode measured from electrode edge to lith centre and undergoes acid dissolution;t he top row of images are obtained directly from am onochrome camera and the bottom images are after image thresholding. Thet hickness of placoliths are known to be heterogeneous across the lith and are generally the thinnest at the centre. [9a,15] This pattern is replicated in the light scattered by the coccolith. At the start of the experiment, aw eak but nonzero pixel intensity varying radially from the centre is seen. By thresholding the optical images,the projection area of the coccolith is obtained irrespective of its thickness nor light intensity.Asthe experiment proceeds,the coccolith is seen to dissolve over the course of tens of seconds.F or coccoliths at adistance to the electrode that is comparable to the coccolith length, that is,tens of microns,the near side of the lith to the electrode is seen to dissolve faster than the far side due to the high concentration gradient of acid near the electrode.Inthe following,w ed emonstrate how quantitative kinetic and hence volume information can be obtained from the analysis of these image series.
Placoliths are quasi-spherical to elliptical in shape,where the shape is specific to the speciation of the coccolithophore [22] as seen in Figure 1. Theo verall rate of dissolution was inferred by an assessment of the changes in the "effective radius" with time,w here pr eff 2 is the equivalent projection area of the coccolith. This effective radius was determined from the thresholded images and is plotted against time in Figure 3b). Thec hanges in the effective radius of the coccolith with time,a sinferred from the projection area, provides ad issolution rate averaged over all sides of the coccolith in spite of the acid concentration gradient across the coccolith. Twor egimes for the coccolith dissolution in Figure 3b)are outlined by red lines;aslow initial rate (dr eff /dt = 0.056 mms À1 )followed by a"rapid" loss of material (dr eff /dt = 0.95 mms À1 )asthe reaction completes (r eff = 0). This change in the optically measured dissolution rate was observed to be ag eneral feature of coccolith dissolution under these conditions and reflects the fact that the coccoliths are disklike and are significantly thinner in the dimension perpendicular to the image plane.S Is ection 4p resents further example dissolution transients all sharing the same qualitative dissolution behaviour. Hence,asthe coccolith is dissolved, the  . Opto-electrochemical dissolution of arepresentative coccolith from C. braaudii. Ap otential of + 1.2 V(vs. Ag wire) was applied to the working electrode at t on = 0s.a )T emporal evolution of the coccolith optically imaged via dark-field scatter-top:r aw image, bottom:image after auto-threshold. The coccolith is situated at 45.5 mmfrom the carbon fibre electrode, measured from the lith centre to the electrode edge. The electrolyte solution was K/2 culture medium with 10 mM H 2 BQ(aq). The time interval of the images is 2seconds and the scale bar is 5 mm. b) Ap lot of the effective radius of the coccolith versus time. c) Dissolution rates measured in different electrolyte medium relative to that predicted for smooth and nonporous calcite particles (see text). Bars:black À0.7 MKNO 3 ,o range À0.7 MKNO 3 + 54.6 mM Mg 2+ ,blue À0.7 MKNO 3 + 2.4 mM HCO 3 À and red-K/2 culture medium.

Angewandte Chemie
Research Articles apparent dissolution rate increases rapidly towards the end of the reaction as the coccolith thickness decreases to apoint at which it is no longer identifiable in the optical image.T he time taken for the dissolution to occur is therefore more related to the calcite disc thickness,r ather than volume. However,p rior to discussing the method of deducing the coccolith volume,wefirst seek to answer the following:what are the dominant chemical factors controlling the coccolith dissolution rate under strong acid attack?

CaCO3 Dissolution Kinetics in Strong Acid
In multiple opto-electrochemical experiments,t he initial dissolution rate of coccoliths was seen to decrease with an increase in distance of the lith from the electrode,asshown in SI section 4. Close to the electrode (ca. 70 mm) the dissolution starts almost immediately after the start of the proton generation (onset/potential switched on) and the dissolution rate dr eff /dt is linear over the time range of 0-3 s. Forl arger electrode to lith distances,progressively longer "lag time (s)" are observed. Ac a2 .5 sd elay before the onset of coccolith dissolution was seen at ac occolith distance of 120 mmf rom the electrode ( Figure S5).
Oxidation of 10 mM of H 2 BQ leads to the formation of ac a. millimolar concentration of protons generated at the electrode interface,r esulting in ah ighly acidic chemical environment (pH < 3) local to the wire electrode.T he acid, H + ,d iffuses radially outward. Thep roton concentration profile varies both as afunction of time t and distance x from the electrode [H + ](x,t)a nd is modelled in SI section 5. Importantly,f or the electrochemical cell geometry used, in the vicinity ( % 50 mm) of the electrochemical interface,anear steady-state mass-transport regime is established within % 1second, shown in Figure S6. Thet ime required to approach this regime increases progressively with the distance away from the electrode resulting in adelay in the dissolution onset. This explains experimental observations of a"time lag" in the dissolution of coccoliths at al ong distance from the electrode (see Figure S5), in which, as noted above,afinite time is required for protons to diffuse hundreds of microns from the electrode.H ence,u nder the present conditions and to ag ood approximation, although the proton concentration varies as af unction of the distance from the wire,w ithin ad istance of 70 mmf rom the electrode,a se videnced in Figure S6 and Figure S10, the unperturbed acid concentration profile can be considered essentially constant over the experimental time of interest.
When ac occolith is exposed to acid, the calcium carbonate is dissolved in accordance with the following reaction; where k 1 is the heterogeneous rate constant as defined by flux = k 1 [H + ]. Av alue of k 1 equal to 0.043 cm s À1 has been reported for the dissolution of am acro-sized Icelandic Spar (calcite) crystal at pH < 4, measured in ab uffer-free electrolyte solution. [23] Fora ni solated CaCO 3 particle the ratedetermining step for the dissolution reaction may either be the rate of diffusion of protons to the mineral interface or the surface reaction rate depending on k 1 and the size of the CaCO 3 particle.F or a smooth and solid (non-porous) calcium carbonate particle,this switch in the kinetic regime occurs at ap article radius of % 20 mm; as estimated via numerical simulation and is discussed in more detail in SI section 6. For as mooth calcite particle placed in as olution containing ah omogeneous acid concentration, [H + ] solution ,t he mass transport of H + from the bulk solution to calcite particles of radius < 20 mmisfast as compared to the rate of consumption at the particle interface,a se videnced in Figure S7 where H owever,a sp article sizes shift to much bigger than 20 mm, the mass transport of H + is insufficient to replenish the acid consumed at the calcite interface.Consequently,the rate of acid dissolution is limited by the mass transport of protons.
It is important to recognize that the surface area of ac occolith is higher compared to as mooth, non-porous calcite particle of the same radius.T his non-unity in the surface roughness will affect the particle size at which the calcite dissolution rate switches from mass-transport to surface-area limitation. Figure S8 shows ap lot of [H + ] surface at steady-state when ac alcite particle with non-unity surface roughness factors is exposed to strong acid solution. The effect of ahigh surface roughness causes the switch in kinetic regimes to occur at as lightly lower particle radius.I nt his study,the largest coccoliths are produced by C. braaudii,with atypical thickness of 1-2 mmand r eff % 5 mm. Due to the small dimensions of the coccolith, 1-5 mmwhen viewed as acalcite disc, the reaction kinetics remains in the surface reaction limited regime even with asurface-roughness factor of 4-8 so that the mass-transport of protons to the particle is fast and not rate-determining,t hus in this case [H + ] surface % [H + ] solution . Forthe other speciation of coccoliths,which have dimensions smaller than the C. braaudii (Figure 1), the acid induced dissolution process is again limited by the surface-area.
Thel iterature reporting the value of k 1 were conducted under idealized conditions (1.0 MK Cl) which do not reflect the ionic composition of seawater. [23] Thef ollowing seeks to elucidate to what extent other components of seawater alter the dissolution kinetics.

Effects of the Chemical Components of Seawater on Dissolution Rates
Thedata presented in Figure 2b)was measured in the K/2 growth medium. This growth medium contains av ariety of salts,t race metals,m inerals and ac arbonate buffer as tabulated in the SI. In this section, we seek to understand to what extent these constituents might influence the coccolith acid dissolution rate.T ot his end, the coccoliths were transferred to electrolytes of differing composition. Tw o primary factors that must be considered when assessing the calcite dissolution kinetics in seawater are i) the influence of Mg 2+ on the surface reaction rate [24] and ii)t he role of the carbonate buffer. [25] Accordingly,i ndividual detached coccoliths (C. braaudii)w ere additionally optical monitored in solutions separately containing i) 0.7 MKNO 3 ,ii) 0.7 MKNO 3 and 54.6 mM Mg 2+ and iii)0.7 MKNO 3 and 2.4 mM HCO 3 À . From the thresholded images an effective particle radius was extracted as afunction of time.Under all conditions,the plot of particle effective radius versus time again exhibited two distinct linear regimes where the initial rate (mms À1 )d irectly reflects the calcite dissolution kinetics under the prevailing conditions.O nt he basis of the known distance of the coccolith from the electrode it is possible to compare this initial dissolution rate to that calculated for as mooth, solid and non-porous calcite particle of the same size dissolved in an ionic solution of 0.7 MK NO 3 .F igure 2c)p lots the measured initial coccolithophore dissolution rates relative to the calculated rate for as mooth and solid particle.O ptoelectrochemical dissolution of detached C. braaudii coccoliths in 0.7 MKNO 3 ,revealed an initial dissolution rate 6.5(AE 2.5) times faster than that predicted for asimilarly sized "smooth" calcite particle at equal distances from the electrode (see Figure 3c). This increase of % 6i nt he dissolution rate as compared to that seen for smooth-surfaced particles reflects the surface roughness of the bio-excreted coccolith;t he dissolution rate is proportional to the calcite surface area and this enhancement of % 6i sc onsistent with reports in the literature for the roughness of comparable coccoliths. [15] Performing asimilar dissolution experiment with the addition of 54.6 mM Mg 2+ ,w hich correspond to the Mg 2+ level in seawater, leads to the dissolution rate changing to ar elative dissolution rate of 4.5(AE 1.7). Mg 2+ cations are known to adsorb on to the calcite surface and inhibit the rate of calcite dissolution and precipitation. [24] Furthermore,t he separate addition of bicarbonate (2.4 mM) to the system decreases the dissolution rate to give arelative dissolution rate of 3.8 (AE 1.7). Clearly,the presence of the bicarbonate anion in solution, pK a (H 2 CO 3 *) = 6.3, serves to partially "titrate" away afraction of the electrogenerated protons to form its conjugate acid H 2 CO 3 (aq) (see above) which chemically decomposes to form H 2 O(l) and CO 2 (aq). Hence the removal of the protons from the system leads to ac oncomitant decrease in the dissolution rate of coccoliths.F inally,experiments conducted in the phytoplankton culture medium, K/2, shows ar elative dissolution rate of 2.9(AE 1.0), which is approximately afactor of two lower than that compared for acoccolith in 0.7 KNO 3 . Since the two ions,H CO 3 À and Mg 2+ ,s erve to decrease the dissolution kinetics via non-competing mechanisms (solution phase titration and surface inhibition), within error, the decreased dissolution rate in the K/2 medium predominantly reflects the presence of the bicarbonate and magnesium ions with the effects adding to each other.
We conclude,the initial calcite dissolution rate is sensitive to the composition of the ionic solution used, due to the size of the coccoliths,the reaction is limited by the kinetics of the surface reaction. Ther eaction rate dr eff /dt is essentially constant over the course of the dissolution reaction for agiven solution composition. Thef ollowing section demonstrates how this reaction rate can be used to infer the thickness of the coccolithophore and hence provide am easurement of the volume of individual coccoliths.

Extracting Coccolith Volume from Dissolution Kinetics
In the opto-electrochemical cell, the coccolith is exposed to an electrochemically generated acid environment H þ ½ x; t ðÞ and the dissolution reaction occurs almost uniformly across the surface of the coccolith. In the surface-area controlled kinetic regime,t he rate of mass transport is fast and at an electrode distance larger than the coccolith length, the proton concentration at the side/rim of the coccolith is no different to that on the top of the coccolith. Therefore,the initial average dissolution rate,v ,i so btained from the orthographic projection of the coccolith on the 2D image plane from dr eff /dt. We assume this rate controls the dissolution of the shortest dimension of the lith which controls the time for complete dissolution by acid titration. Thetime to dissolve completely is expected to vary between liths (both intra-and interspecies) and to further be sensitive to the prevailing chemical environment. However,f or the purpose of lith volume measurement ak ey fact is that the dissolution rate is limited by the rate of the surface reaction (as opposed to masstransport). Consequently,t his rate is essentially constant during the course of the dissolution, due to the quasi-steadystate proton concentration arising from the electrochemical cell geometry-see SI Section 5a nd 6.
Knowledge of the initial rate enables the thickness of the calcite to be determined on ap er-pixel basis.T he volume of individual coccoliths can be determined by iterating through the stack of images,building on apixel-by-pixel basis,toyield the initial volume of coccolith prior to dissolution. Figure 4a) depicts the image analysis procedure leading to a3Dmodel of an individual coccolith, where the volume was reconstructed from as eries of images taken during the course of the dissolution process,s hown in Figure 3a). Thei mage analysis procedure is as follows,f or more see SI section 2.4. In step one,a sa bove,w ei mage an individual coccolith during controlled acid dissolution. In step two,a sd escribed earlier, the images are analysed to give the projection area of the coccolith as af unction of time.N ext, step three,t he initial dissolution rate (v = dr eff /dt)i sm easured.
Step four, the time required to completely titrate the calcite, t(i, j), is measured for each and every pixel coordinates (i, j).
Step five,the calcite height, which is normal to the 2D image plane,isestimated at each pixel coordinate (i,j)b ym ultiplying t(i,j)b yv .L ast, by doing this on ap er pixel basis the initial thickness of the coccolith can be inferred as afunction of pixel position hence yielding a3Dreconstruction of the original coccolith shown in Figure 4a).
Them ethod of volume extraction used is independent of the conditions under which it has been measured, as the initial dissolution rate is used to directly infer the particle thickness from the time to full dissolution. Table S2 contains optical images of four C.braaudii coccoliths obtained prior to acid dissolution in different electrolytes and their 3D reconstructed images after controlled acid dissolution. Irrespective of the electrolyte medium used, the reconstructed images of the coccolith resemble that prior to dissolution. Note that the resolution of the 3D coccolith reconstructed via this method is not as high as obtained via atomic force microscopy, [26] electron microscope [13a] or 3D X-ray nanotomorgraphy [15] due to the nature of optical limitation and the nature of reconstruction from astack of projection images;the general shape of the reconstructed coccoliths is however fully consistent with the shape of placoliths reported elsewhere in the literature.F rom the reconstructed individual coccolith volumes,itispossible,assuming agiven lith calcite density,to convert the volume to ac occolith mass.F igure 4b)p lots the mass of individual C. braaudii coccoliths as measured in the four different electrolyte solutions (separated by colour); different symbols of the same colour represent data obtained from repeated experiments.N otice that the reconstruction is independent of the chemical composition of the electrolyte as diverse solutions are used to generate Figure 4b). Thei nitial dissolution rate of acoccolith is dependent on the difference in electrolyte chemical composition, the distance of coccoliths from the electrode,a nd less importantly,t he intra-species variation in coccolith surface roughness and morphology. Since the volume reconstruction process uses the initial dissolution rate as measured, this internally "calibrates" for all of the effects discussed above so that the result is independent of the numerous variables.H owever,o fc ourse, the faster the initial dissolution rate,the quicker the coccolith dissolves and the fewer images there are for the volume reconstruction. Them ean mass of C. braaudii coccoliths across all four of the electrolyte studied is 0.122(AE 0.064) ng, sample size = 81, which is in excellent agreement with literature values. [13a] Thel atter was overlaid as green shade and the dotted black line in Figure 4b).
Having evidenced the technique using coccoliths detached from C. braaudii we next extend this study to other coccoliths. Figure 5s hows the inferred coccolith mass from three additional species of coccolithophore-E. huxleyi, C. leptoporus and G. oceanica. Theo pto-electrochemical dissolutions were conducted in their corresponding "seawater-mimicking" culturing media, with the addition of 10 mM H 2 BQ acid precursor prior to the experiment. Thecoccolith mass is seen to generally increase with coccolith length inter-species.T he average coccolith mass for E. huxleyi, C. leptoporus and G. oceanica are 10.2(AE 6.5), 23.6(AE 12.1) and 37.0(AE 17.8) picograms (pg) per lith, respectively. E. huxleyi (RCC1212) and G. oceanica (RCC1314) were recently studied via birefringence polarised light approach and 3D X-ray coherent diffraction imaging;t he reported mass of E. huxleyi coccoliths were ca. 1-6 pg for coccolith lengths between 2t o4mm, and the G. oceanica coccoliths are ca. 5-30 pg for coccolith lengths between 4t o6 mm. [15] Compared to the coccolith mass obtained via image reconstruction, good agreement are seen within the overlapping range of coccolith lengths.W en ote that the variation in the coccolith length range in the sample, and the proportion of malformed/broken coccolith, may be due to variations in different culturing conditions and/or experiments conducted at different stages during the coccolithophore lifecycle.T he size range of the C. leptoporus coccoliths herein (4-7 mm) is small as compared to those generally reported by Young et al. from sediment samples (5-11 mm, average mass = 74.1 pg). Moreover,a ss hown in Figure S1, SEM images revealed al arge proportion of the C. leptoporus coccoliths in this study are either malformed or broken, which may concatenate with the small size distribution leading to an underweight average coccolith mass of 23.6 pg.F igure 5b)s hows the collective coccolith mass data showing the relationship of coccolith mass versus coccolith length that exists in both intra-species and inter-species, leading to al inear logarithmic plot as shown in Figure 5c) with as lope equal to 2.8(AE 0.1). From this,o ne can infer the coccolith mass,b oth intra-and inter-species,v aries broadly with the coccolith length cubed. However,a si sc onsistent with the literature and can be seen from the inset of Figure 5b), the correlation between the measured thickness and the coccolith length is low (Pearsons r value of 0.58). [27]

Conclusion
In situ electrochemical dissolution of coccoliths with simultaneous in situ optical image analysis allows the volume and mass of individual coccoliths to be estimated. The dissolution of coccoliths in as trongly acidic environment occurs under ak inetic regime controlled by the particle surface-area;w ithin this regime,t he rate of change in the projection area of the coccolith during dissolution is directly proportional to the rate of change in the coccolith thickness. This allows the coccolith to be reconstructed from the timestacked 2D images,t op rovide an estimate of coccolith volume on ap ixel-by-pixel basis.S umming across all of the pixels allows the initial pre-dissolution total coccolith volume to be determined and hence this yields am easure of the CaCO 3 mass.
Theopto-electrochemical approach uses the initial rate of coccolith dissolution, inferred from d(r eff )/dt. It therefore internally calibrates all factors that may affect the dissolution rate;these include:the surface roughness of the coccolith, the distance of the coccolith from the electrode,t he presence of inhibitors for calcite dissolution. Since the analysis is performed on an individual coccolith basis,i td oes not rely on using aquantity statistically averaged over,for example,large sediment samples (e.g.s hape factor, k s ). Therefore,t he coccolith mass estimated herein could in principle account for sample abnormalities such as those that include al arge proportion of deformed, partially dissolved, or broken coccoliths.T he opto-electrochemical method can discriminate coccoliths from suspended sediments,w hich cannot be done in an operational manner using conventional remote sensing approaches and thus holds the potential to provide new insight into the presence of coccolithophores in the carbon pool of coastal waters,using in situ samples.
Unlike the birefringence methods which have adetection limit in the greyscale limited to at heoretical maximum of 1.56 mma nd require ac ylindrical rhabdolith to calibrate intensity with thickness;the approach herein provides afacile alternative due to the calibration-free approach and has no inprinciple limitation to the thickness of calcite particle nor the crystalline orientation of calcites in the coccolith. This allows such opto-electrochemical method to probe volumes of larger coccoliths,a se videnced by the C. braaudii coccoliths as presented herein, and with the possibility to extend the study to living coccolith-bearing coccospheres and/or study of sediment samples without prior knowledge of the coccolith species nor thickness of coccoliths within the sample.T he scientific breakthroughs presented herein may find use in Figure 5. a) Coccolith mass estimated for individual coccoliths from four different coccolithophore species. b) Collective estimated coccolith mass against coccolith length. Inlay:t he maximum thicknesso fcoccolith from image reconstruction, Pearson's r = 0.58. c) Collective mass plotted on alogarithmicscale. Circles:g rey-E. huxleyi (RCC1216), blue-C. leptoporus (RCC1130), red-G. oceanica (RCC1314)a nd C. braaudii (RCC1198). Line of best fit over all data:s lope = 2.78 AE 0.09 and Pearson's r = 0.94. The data for E. huxleyi, G. oceanica and C. leptoporus were measured directed from their respective culturing medium with added 10 mM H 2 BQ acid precursor,whereas those for C. braaudii is the collective data from all four electrolytes as seen in Figure 2d). marine environments reporting changes in the carbonate biogeochemistry as inferred from local changes in coccolith mass,p roviding as olution to the global challenge.