Calcite and aragonite dissolution in seawater: kinetics, mechanisms, and fluxes in the North Pacific

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content CALCITE AND ARAGONITE DISSOLUTION IN
SEAW A TER: KINETICS, MECHANISMS, AND FLUXES
IN THE NORTH P ACIFIC

by

Sijia Dong

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the degree
DOCTOR OF PHILOSOPHY
(GEOLOGICAL SCIENCES)

August, 2019

Copyright 2019 Sijia Dong
ii
ABSTRACT
Calcium carbonate minerals play a critical role in regulating geochemical cycles through
dissolution and precipitation in aqueous environments due to the mineral’s wide occurrence and
high reactivity on the earth surface. The formation of carbonate rocks on long timescales is
canonically driven by the interaction of aqueous CO2 and the cations from silicate rock
weathering. Rivers deliver dissolved weathering products to the ocean in the form of alkalinity,
which at steady state is removed via the production and burial of calcite and aragonite minerals.
Estimates of open ocean calcification vary between 0.4~1.8 Gt (1 Gt = 10
15
g) PIC yr
-1
, whereas
only 0.1 Gt PIC yr
-1
is buried in deep-sea sediments. As a result, the majority of CaCO3 produced
in the surface open ocean must be dissolved either in the water column or in deep sea sediments.
The location and rate of CaCO3 dissolution, and thus where and how fast alkalinity returns to the
ocean system, are crucial in determining the response of oceanic system to perturbations in either
alkalinity or CO2 input to the ocean-atmosphere system.
This dissertation makes new measurements of calcite and aragonite dissolution in
seawater, in an attempt to constrain dissolution kinetics and to understand the dissolution
mechanisms in natural ocean environments with strictly controlled variables in the laboratory.
I first identify a kinetic pressure effect on calcite dissolution that cannot be explained by
the influence of pressure on calcite stoichiometric solubility product (K
*
sp). My work aims to
address a simple question: is dissolution rate under the influence of a saturation state controlled
by ion activity product (IAP) the same as controlled by K
*
sp? The answer is, surprisingly, no. The
enhancement in dissolution rate is a factor of 2-4 at 700 dbar compared to dissolution at the same
W under ambient pressure (10 dbar). This kinetic pressure effect points out caution in applying
iii
lab-determined rate laws to ocean dissolution processes that are under high pressure, and implies
that sinking particles would dissolve at shallower depth than previously thought.
Next, I present laboratory and in situ aragonite dissolution rate measurements, and the
particulate inorganic carbon (PIC) and particulate organic carbon (POC) fluxes and
concentrations in sinking and suspended materials along a North Pacific transect. I show that the
measured aragonite flux combined with the inorganic dissolution rate only account for a small
fraction of the excess alkalinity observed in the North Pacific, and respiration-driven dissolution
of PIC or metazoan/zooplankton driven dissolution is more likely the source of excess alkalinity.
For my third PhD project, I utilize Atomic Force Microscopy (AFM) to investigate
calcite dissolution mechanisms in seawater. One key finding is that there is significantly higher
etch pit density in seawater than in freshwater, whereas step velocity in seawater is inhibited at
high and mid W, leading to net lower bulk dissolution rate near equilibrium. I further
demonstrate that a newly defined parameter, edge length density, is a notably better measure of
Angstrom-scale surface roughness than surface area, and is a good indicator of dissolution active
sites.
Finally, the catalysis mechanism of carbonic anhydrase (CA) on calcite dissolution is
investigated with AFM. I find that dissolution can be attributed to the adsorption of CA on the
calcite surface, and the transfer of protons from CA catalytic center to the mineral. CA does not
affect step retreat velocity but creates etch pits by contacting the calcite surface. Therefore, to
what extent can CA enhance PIC dissolution and thus the marine alkalinity cycle highly depends
on local availability of CA in contact with PIC, and should be considered as an important yet still
unconstrained aspect of oceanic carbon cycling.
iv
ACKNOWLEDGEMENTS

Financial support for this study came from the National Science Foundation, USC Dornsife
Doctoral Fellowship, the Elizabeth and Jerol Sonosky Fellowship for Earth and Ocean Sciences,
and USC Sea Grant Fellowship.

I cannot thank my advisor Dr. William Berelson enough for being a tremendous mentor and a
good friend. You have made my journey in pursuing a PhD degree in a foreign country an
extremely pleasant and fulfilling experience. Your advice and encouragement on both research
and my career have been priceless. I would also like to thank Dr. Jess Adkins for all your
brilliant ideas and valuable suggestions for my research.

A special thanks to our lab tech Nick Rollins – an amazing lab-mate, who has made lab and field
work much more fun and has offered great help in almost every piece of work during my
graduate school. I am also grateful to Adam Subhas and John Naviaux for working together to
get our understanding of carbonate dissolution deeper; and to Caty Tems, Abby Lunstrum and
Jaclyn Pittman for being wonderful lab-mates.

I would like to thank Dr. Josh West for being supportive, understanding and helpful. Thanks to
my qualification exam and dissertation committee members: Will Berelson, Jess Adkins, Josh
West, Doug Hammond, Donal Manahan and Moh El-Naggar for your precious comments and
suggestions.

Thanks to Sherwood Liu, Christopher Moore, Loraine Martell-Bonet, Nathan Kemnitz as well as
Kilo Moana captain and crews for their help during CDisK-IV. Thanks to Professor Robert
Byrne for discussions of the cruise results. Aaron Celestian and Nitya Turaga are warmly
thanked for their effort and time spent in analyzing cruise sample mineralogy.

A special gratitude goes to my undergraduate advisor Dr. Henry Teng who has helped me a lot in
both understanding the mechanistic models of dissolution and also operating AFM in the lab. I
v
would also like to thank Sahand Pirbadian for his great patience and tremendous help in setting
up my AFM experiments.

I have so much appreciation for friends who have offered me wisdom in science and academy,
companionship in everyday life and road trips, as well as emotional comfort during the past five
years. Among them, I would like to specially thank Mark Torres, Gen Li, Christine Wu, Tony
Wang, Yi Hou, Nick Rollins, Abby Lunstrum, Jaclyn Pittman, Weiwei Hua, Jotautas Baronas,
Jun Shao, Audra Bardsley, Joyce Yager, Hyejung Lee, Niloufar Abolfathian, Hongrui Qiu, Lei
Qin, Haoran Meng, Jun Hu, Yifang Cheng, Zhefu Dai, and Kai Xie, I could not in any way have
done this without you.

Finally, and most importantly, words cannot express how grateful I am to my father and mother
for respecting and supporting all my choices. I will do my best to prove that they are wise
choices. To my soul mate and hero Jimmy, thank you for creating all the beautiful life memories
with me, and giving me the courage to face any difficulty in life.

Thanks to everyone who makes me feel that my life is so great.

vi
Table of Contents

ABSTRACT .............................................................................................................................. ii
ACKNOWLEDGEMENTS ....................................................................................................... iv
CHAPTER 1: INTRODUCTION ................................................................................................ 9
1.1 Oceanic Carbon Cycle and CaCO 3 Dissolution............................................................................... 9
1.2 Previous Studies on CaCO 3 dissolution ........................................................................................ 10
1.3 Dissertation Chapters and Research Objectives ............................................................................ 12
References ......................................................................................................................................... 15
CHAPTER 2: PRESSURE DEPENDENCE OF CALCITE DISSOLUTION RATE IN
SEAWATER ............................................................................................................................ 20
ABSTRACT ...................................................................................................................................... 20
2.1 INTRODUCTION ....................................................................................................................... 21
2.2 METHODS ................................................................................................................................. 26
2.3 RESULTS ................................................................................................................................... 30
2.3.1 Dissolution rates at changing pressures ............................................................................................... 30
2.3.2 Dissolution rates at constant pressures................................................................................................. 31
2.3.3 Log-log correlation of dissolution rate vs. undersaturation ................................................................... 32
2.4 DISCUSSION ............................................................................................................................. 34
2.4.1 Thermodynamic uncertainty and Partial Molal Volume ....................................................................... 34
2.4.2 Different dissolution mechanisms and critical degree of undersaturation .............................................. 37
2.4.3 Effect of pressure on shallow water column carbonate dissolution ....................................................... 43
2.5 CONCLUSIONS ......................................................................................................................... 45
ACKNOWLEDGEMENTS ............................................................................................................... 46
REFERENCES .................................................................................................................................. 47
Appendix. “Ex situ” and “in situ” rates for dissolution experiments in the pressure case. .................... 53
CHAPTER 3: KINETICS OF ARAGONITE DISSOLUTION IN SEAWATER AND ITS ROLE
IN SHALLOW WATER COLUMN DISSOLUTION IN THE NORTH PACIFIC ................... 55
ABSTRACT ...................................................................................................................................... 55
3.1 INTRODUCTION ....................................................................................................................... 56
3.2 METHODS ................................................................................................................................. 59
3.2.1 Labeled aragonite for dissolution experiments ..................................................................................... 59
3.2.2 Lab and field dissolution experiments ................................................................................................. 60
3.2.3 Sinking flux measurements with sediment traps .................................................................................. 62
vii
3.2.4 Suspended particle concentration measurements with in situ pumps .................................................... 63
3.3 RESULTS ................................................................................................................................... 64
3.3.1 Aragonite dissolution rates.................................................................................................................. 64
3.3.2 Sinking C fluxes captured by sediment traps ....................................................................................... 65
3.3.3 Suspended particulate C concentrations measured with in situ pumps .................................................. 69
3.4 DISCUSSION ............................................................................................................................. 72
3.4.1 Kinetics of aragonite dissolution in the lab versus in the field .............................................................. 72
3.4.2 Total C and PIC export production rates, PIC/POC ratios in sinking fluxes and suspended materials in
the North Pacific ......................................................................................................................................... 75
3.4.3 PIC dissolution and POC remineralization rates in sinking fluxes and shallow depth dissolution .......... 77
3.4.4 Calcite/aragonite ratios in sinking and suspended materials in the North Pacific water column ............. 79
3.4.5 Aragonite dissolution fluxes in the North Pacific water column ........................................................... 80
3.5 CONCLUSIONS ......................................................................................................................... 84
ACKNOWLEDGEMENTS ............................................................................................................... 85
REFERENCES .................................................................................................................................. 86
Appendix. Model of aragonite dissolution fluxes in the North Pacific water column. .......................... 90
CHAPTER 4: AN ATOMIC FORCE MICROSCOPY STUDY OF CALCITE DISSOLUTION
IN SEAWATER AND THE DEPENDENCE OF DISSOLUTION RATE ON SURFACE
PROPERTIES ........................................................................................................................... 92
ABSTRACT ...................................................................................................................................... 92
4.1 INTRODUCTION ....................................................................................................................... 93
4.2 METHODS ................................................................................................................................. 95
4.2.1 Sample and solution preparation ......................................................................................................... 95
4.2.2 In situ dissolution experiment set-up and AFM imaging ...................................................................... 96
4.2.3 Dissolution rate analysis and edge length measurement ....................................................................... 97
4.3 RESULTS ................................................................................................................................. 102
4.3.1 Effect of flow rate on dissolution and the variation of step velocity on the calcite surface....................102
4.3.2 Etch pit morphology in seawater and etch pit density vs. undersaturation ............................................103
4.3.3 The dependence of step velocity and bulk rate on saturation state .......................................................105
4.4 DISCUSSION ........................................................................................................................... 107
4.4.1 Comparison of etch pit morphology, bulk rate, step velocity and etch pit density between seawater and
freshwater ..................................................................................................................................................107
4.4.2 Identification of changes in calcite dissolution mechanisms in seawater ..............................................114
4.4.3 The role of calcite surface properties on step velocity and bulk rate ....................................................117
4.4.4 The role of calcite surface properties on step velocity and bulk rate ....................................................119
4.4.5 The correlation between edge length density and bulk dissolution rate ................................................120
viii
4.5 CONCLUSIONS ....................................................................................................................... 125
ACKNOWLEDGEMENTS ............................................................................................................. 126
REFERENCES ................................................................................................................................ 126
CHAPTER 5: A MECHAMISTIC STUDY OF THE EFFECT OF CARBONIC ANHYDRASE
ON CALCITE DISSOLUTION AND ITS IMPLICATIONS FOR THE OCEAN CARBON
CYCLE ................................................................................................................................... 131
ABSTRACT .................................................................................................................................... 131
5.1 INTRODUCTION ..................................................................................................................... 132
5.2 METHODS ............................................................................................................................... 133
5.3 RESULTS ................................................................................................................................. 134
5.3.1 Effect of dissolved CA on calcite step retreat velocity ........................................................................134
5.3.2 Effect of CA aggregate on dissolution through CA-calcite contact ......................................................136
5.4 DISCUSSION ........................................................................................................................... 140
5.4.1 Interpretations of bulk dissolution rate vs. undersaturation correlation w/wo CA .................................140
5.4.2 Possible mechanism for the CA catalysis on calcite dissolution ..........................................................142
5.4.3 Geological implications of the catalytic effect ....................................................................................144
5.5 CONCLUSIONS ....................................................................................................................... 146
ACKNOWLEGEMENTS ................................................................................................................ 146
REFERENCES ................................................................................................................................ 147
CHAPTER 6: CONCLUSIONS .............................................................................................. 150
REFERENCES ................................................................................................................................ 154

1
List of Figures
Figure 2.1 Separation of effect of water chemistry and pressure on !"#$"%&' in a Pacific section
by Ocean Data View simulation. (a) !('#$ : actual !. (b) !")'*%+&(, : ! calculated by
Ocean Data View if there is no effect of pressure on - ∗+/ (- ∗+/ after Mucci, 1983),
namely ! affected only by water chemistry (DIC, alkalinity, salinity, temperature). (c) 0!:
difference between (b) and (a), ! affected by pressure; up to 1.2 ! units at 5000 m depth.
(d) Defined section (red box) on map. ............................................................................... 23
Figure 2.2 Pressure case diagram. ............................................................................................. 27
Figure 2.3 Moles of labeled carbonate dissolved based on d
13
C measurement from P45 (1050
dbar). Yellow squares are ex situ samples (three before experiment, three after experiment);
grey circles are in situ samples; trend line is fit to all data points. Replicate value
uncertainties are less than the size of the data points. Note that only 4×10
-7
mol calcite was
dissolved over 4 days (0.001% of total calcite powder). ..................................................... 29
Figure 2.4 ! contour plot as a function of pressure and alkalinity. Six sets of experiments were
conducted. Black asterisks were dissolution experiments conducted with the same fill bag
seawater, ! changed by changing pressure (Route 1). White circles were dissolution
experiments conducted under five different constant pressures (Route 2 – 10 dbar; Route 3 –
350 dbar; Route 4 – 700 dbar; Route 5 – 1050 dbar; Route 6 – 2500 dbar). Note that this
diagram is only a rough demonstration of ! and experimental parameters, because DIC in
this plot has to be set to a fixed value (2030 µmol/kg) while actual DIC for different
experiments (different batches of Dickson seawater) vary from 2000 µmol/kg to 2040
µmol/kg. ............................................................................................................................ 30
Figure 2.5 Dissolution rate vs. undersaturation. Yellow squares: changing water chemistry at
atmospheric pressure and 21°C (Route 2 in Figure 2.4). Light grey circles: changing
pressure with the same fill-bag seawater at 21°C (Route 1 in Figure 2.4), ΔV=-37.6 cm
3
mol
-1
at 21°C (Ingle, 1975). Dark grey diamonds: changing pressure with the same fill-bag
seawater, ! recalculated with ΔV=-57.6 cm
3
mol
-1
at 21°C (error bars not shown). Pressures
at which dissolution experiments were conducted (light grey circles) were 350 dbar, 700
dbar, 830 dbar, 1070 dbar, 1100 dbar, 1380 dbar (marked on the plot). Uncertainties in !

2
represent standard deviation of replicates in DIC and alkalinity measurements; uncertainties
in dissolution rate represent goodness of fits in Figure 2.3. ................................................ 31
Figure 2.6 Dissolution rates versus undersaturation at 5 different pressures (Routes 2-6 in Figure
2.4). Dashed lines are fits to data points of 10 dbar, 350 dbar, 700 dbar, 1050 dbar. Fits to
700 dbar and 1050 dbar overlap such that these data are fitted to the same curve. .............. 32
Figure 2.7 (a) Log-log plot of geometry-normalized dissolution rate versus undersaturation state
at 10, 350, 700, 1050 dbar. Dashed lines are fits to data points in Zone II. For reference,
log(1- !) from -1.5 to -0.3 is equivalent to ! = 0.97 to 0.50. (b) Log-log plot of dissolution
rate versus undersaturation state. White squares are 10 dbar data from this study; grey
diamonds are 10 dbar data from Subhas et al. (2015) using the calcite sample from the same
batch, and the same experimental set-ups; black triangles are 1050 dbar from this study. At
near-equilibrium conditions, where dissolution rates are very low (Zone I), error bars in log
scale plot are large. Irrespective, dissolution rates are clearly off the trend lines in Zone II.
Linear regression fits are given. ......................................................................................... 34
Figure 2.8 Dissolution velocity (m/s) versus 3/5 where 5 = 78 (!) (framework proposed by
Dove et al. (2005)). Higher 3/5 value means closer to saturation (! = 1). Dissolution
velocity are off the linear trend lines in Zone I and start to increase as 3/5 increases,
indicating a change from step retreat dissolution mechanism (Zone I) to defect-assisted
dissolution mechanism (Zone II). Different slopes of different pressure trend lines in Zone
II indicate different local interfacial energy barrier ; under different pressures (Table 3). (a)
Dissolution velocity (m/s) versus 3/5 for 10, 350, 700, 1050, 2500 dbar. (b) Dissolution
velocity (m/s) versus 3/5 for 10 dbar (including data from Subhas et al. (2015)) and 1050
dbar only, to show a clearer difference of slopes in Zone II and a clearer change of
mechanism at around 3/5=9, equivalent to ! = 0.89. The trend lines in Zone 1 are
schematic fits to the data.................................................................................................... 41
Figure 2.9 (a) Surface energy barrier < (calculated from slopes in Figure 8) versus different
pressure. The trend line is a schematic to emphasize the apparent change in slope in < vs.
pressure; (b) Change in the intercepts (proportional to kinetic parameters = and/or >+ )
versus different pressure for defect-assisted dissolution (Zone II). ..................................... 42

3
Figure 3.1 (a) A map of the Northeast Pacific Ocean and station locations for the CDisK-IV
Cruise (August, 2017); (b) !?(#@A>%&' in the upper 1200 m along the transect.
White diamonds are locations where in situ aragonite dissolution experiments were
conducted in Niskin Incubators (Station 2 to 5). There was an additional deployment
at 2020 m, Station 2, (!?(#@A>%&' = 0.57) that is not shown in Figure 1b.
!?(#@A>%&' in Figure 1b is calculated based on alkalinity and pH measurements of
CDisK-IV CTD cast, and is slightly different from !?(#@A>%&' data from GLODAP
v2 ODV collection. This discrepancy will be discussed in detail in a forthcoming
paper of our group. ................................................................................................... 59
Figure 3.2 Aragonite dissolution rates measured in the field along the North Pacific transect
(2~7℃), and in the lab (5℃ and 21℃). (a) Specific dissolution rates (g cm
-2
day
-1
) vs.
(1-!); (b) log-log scale of specific dissolution rates (g cm
-2
day
-1
) vs. (1-!). The
reaction orders (slopes of the fitting lines) of field, lab 5℃, lab 21℃ rate laws at ! =
0.45~0.87 are 1.37 ± 0.18, 1.33 ± 0.56, and 1.59 ± 0.12 respectively. At ! = 0.9~1
[log(1- !)<-1.0], reaction order n is smaller than at lower !. The transition between
the two reaction orders (!"(%&%"#$ ) occurs at !"(%&%"#$ ~ 0.9. ............................... 65
Figure 3.3 Sinking fluxes captured by sediment traps at 100m and 200m along the North Pacific
transect in August 2017. Station 1, 2, 3 were in the North Pacific subtropical gyre,
while Station 4 and 5 were in the subarctic gyre. The solid bar for the 100 m sample at
42°N represent the difference in the duplicate samples at that site (The two ends of the
bars are the duplicate values, and the trend lines are fitted through the average of the
duplicates). (a-d) fluxes vs. latitude; (e-h) fluxes vs. depth. ....................................... 67
Figure 3.4 CaCO3 mineral composition (a, b) and calculated mineral fluxes (c) in the sinking
material along the North Pacific transect at 100m and 200m in August 2017............. 69
Figure 3.5 Suspended PIC and POC captured by in situ pumps along the North Pacific transect in
August 2017. (a) Suspended PIC (Dg/L), (b) Suspended POC (Dg/L), (c) Suspended
calcite concentration (count/L), (d) Suspended aragonite concentration (count/L), (e)
Calcite and aragonite percentages. ............................................................................ 72
Figure 3.6 Comparison of the aragonite dissolution rates in this study and previous studies.
Closed symbols are dissolution rates of synthetic aragonite, whereas open symbols are
rates of pteropods. Note that dissolution rates of synthetic aragonite are ~30 times

4
larger than pteropods in Morse et al. (1979) and Keir (1980). Additionally, previously
published field dissolution rates are orders of magnitude lower than lab dissolution
rates. Dissolution rates reported in this study show a consistency between field and
lab. ........................................................................................................................... 74
Figure 3.7 Amount of PIC dissolved and POC remineralized between 100 m and 200 m in the
sinking fluxes during CDisK-IV (August, 2017). (a) PIC weight loss as a function of
POC weight loss. PIC dissolves and POC remineralizes at a ratio of 0.29. (b) Percent
of PIC and POC lost from 100 m to 200 m. High PIC dissolution is associated with
high POC remineralization. ....................................................................................... 78
Figure 3.8 (a) Aragonite sinking flux below the saturation horizon (430 m) if dissolution only
occurs as abiotic aragonite dissolution; (b) in situ dissolution rate in the water column
(Dmol kg
-1
yr
-1
) if dissolution only occurs as abiotic aragonite dissolution. Different
symbols in (a) and (b) represent different sinking rates of aragonite particles.
Dissolution rates are calculated based on the saturation states measured in Station 3
(35° N, 151° W) and the dissolution rate law determined in this chapter. Aragonite
sinking flux of 0.1 mmol m
-2
day
-1
is assumed to reach 430 m, below which seawater
starts to become under-saturated. (c) PIC sinking flux assuming 40% dissolution every
100 m due to organic matter respiration driven dissolution or metazoan/zooplankton
consumption; (d) in situ dissolution rate generated by respiration driven dissolution or
metazoan/zooplankton consumption. For (c), (d), total PIC sinking flux of 1.2 mmol
m
-2
day
-1
, calcite flux of 0.8 mmol m
-2
day
-1
, aragonite flux of 0.4 mmol m
-2
day
-1
are
assumed at 100 m...................................................................................................... 82

Figure 4.1 AFM in situ dissolution experiment setup. ................................................................ 97
Figure 4.2 Step velocities of individual etch pits during different time periods for two
experiments. Measured step velocities are independent of time and etch pit location. The
variation is largely due to the limited precision in width measurement. .............................. 99
Figure 4.3 Images showing how the MATLAB code determines (a) edges and (b) dissolution
rates (AF-5). (a-1) raw image; (a-2) processed image after binning the convoluted surface to
one monolayer increment; (a-3) detected edges overlying the raw image. (b-1) image at t1

5
during AF-5; (b-2) image at t2 during AF-5; (b-3) number of monolayers dissolved between
t1 and t2. .......................................................................................................................... 100
Figure 4.4 Edge length analysis using MATLAB. (a) to (e) are images of the four experiments in
Table 4.3. (b) and (c) are sub-regions of a larger image. Black curves are edges detected
after smoothing the images. ............................................................................................. 101
Figure 4.5 One example of “bulk” rate determination by MATLAB (AF-3, spot 3). Image area =
107.9 Dm
2
. “Bulk” rate calculated = 1.58×10
-6
g cm
-2
day
-1
. ........................................... 102
Figure 4.6 Average step velocity of acute and obtuse edges at three different flow rates (15 mL h
-
1
, 30 mL h
-1
, 45 mL h
-1
) at Ω = 0.37 ± 0.01. The grey crosses are velocities at different
individual etch pits. The squares are the mean values of the crosses, with the error bars
representing the standard deviation of the population....................................................... 103
Figure 4.7 One example of dissolution on the calcite (104) cleavage surface in seawater (W =
0.46 ± 0.01). 0 min in Figure 4.7a is actually 37 min after the start of the continuous
seawater flow. Field of view is 11×11 Dm. Color scale spans 4 nm in surface height. ..... 104
Figure 4.8 (a) Ω = 0.88 ± 0.04; (b) Ω = 0.50 ± 0.02 (etch pits formed before t=31 min, but
image quality was poor). At Ω = 0.88 ± 0.04, dissolution only happens as step retreat (black
arrows). No etch pit formation was found for 30 min within the total scanned area of 17.04
Fm × 17.04 Fm. At Ω = 0.50 ± 0.02, dissolution happens both at existing step edges
(black arrows) and at newly-formed etch pits (yellow arrows). The highest Ω observed for
etch pit formation is 0.82 ± 0.04 with a pit density of 3.1×10
6
cm
-2
. ............................... 104
Figure 4.9 Average step velocity of acute and obtuse edges vs. undersaturation. The two yellow
diamonds are two dissolution experiments at similar Ω but on calcite crystals with distinct
surface features (see Figure 4.10). ................................................................................... 106
Figure 4.10 (a) AF-18, ! = 0.56, dissolution at point defects; (b) AF-5, ! = 0.58, dissolution at
dislocations or pre-existed holes on calcite surface. (c) Ratios of step velocity, bulk rate and
surface area per unit image area between the two surfaces (surface (a) is considered as 1).
One monolayer = 0.3 nm. Note that in AF-5, “0 min” is not the real start time of dissolution,
but 38 min after seawater was in contact with the calcite sample. “SA” is short for surface
area. ................................................................................................................................ 106
Figure 4.11 Comparisons of bulk dissolution rate, step velocity and pit density between
dissolution in seawater (SW) and in freshwater (FW). Subhas et al. (2017) and this study are

6
in seawater, all others are in freshwater. “Other fresh water studies” in Figure 4.11a and
4.11b include: Shiraki et al., 2000; De Giudici, 2002; Arvidson et al., 2003; Arvidson et al.,
2006; Vinson and Luttge, 2005; Lea et al, 2001; Harstad and Stipp, 2007. Except for Xu et
al. (2010) that has experimental temperatures of 50~70 ℃, all other studies are between 20
℃ and 25 ℃. Experimental details are listed in Table 4.3 and 4.4. ................................... 109
Figure 4.12 (a) Kinetic coefficient G as a function of undersaturation state (G calculated for each
different dissolution experiment). (b) G calculated for different etch pits and at different
time points during one dissolution experiment (AF-20, W = 0.37). As noted, G is slightly
lower (~30%) when an etch pit first forms. ...................................................................... 119
Figure 4.13 Bulk dissolution rate / step velocity vs. edge length density. Circles are six distinct
regions on the calcite surface during four dissolution experiments at W~0.5. Dashed line is
fit to the circles with R
2
=0.97. Yellow line is the theoretical calculation of the correlation
with a slope of )H =8.13×10
-4
g cm
-3
/

Dm
-1
. Note that experimental data points are in very
good agreement with theoretical calculation. ................................................................... 122
Figure 4.14 Schematic illustration of calcite dissolution at etch pit that is 1 monolayer deep vs. 2
monolayers deep. 2D area = A Dm
2
. h = monolayer thickness = 0.3 nm........................... 125

Figure 5.1 (a) Average step velocity of acute and obtuse edges vs. undersaturation with and
without CA. Trend lines are fits to all “no CA” rates. (b) Ratios of step velocity with
dissolved-phase CA to without CA. The absolute step velocities at W ~ 0.86 and 0.74 are
not included in Figure 5.1a, because the step velocities are not the average of acute and
obtuse step retreat velocity. The absolute and ratio of step velocities w/wo CA at W ~ 0.53
are indicated by arrows in the figure. ............................................................................... 136
Figure 5.2 (a) CA aggregates in contact with the calcite surface and their effect on dissolution at
Ω = 0.83. The straight horizontal dark stripes near the CA aggregate are artifacts in the
AFM imaging process that attempt to balance the total grayscale in individual rows, and are
not trenches on the surface. The irregular-shaped dark areas, however, are real etched
features (yellow arrows). (b) Schematic illustration of CA-induced dissolution on the calcite
surface. Note that some CA blobs did not induce dissolution (e.g. the blob that appeared at 2
min next to the black arrow in Figure 5.2a), likely because they did not touch the surface.
........................................................................................................................................ 137

7
Figure 5.3 Dissolution patterns observed by dipping the AFM probe into a CA aggregate and
moving to a CA-free area to scan. “Before” means before dipping the probe into CA; “After
(0 min)” means the image obtained immediately after the probe is coated with CA; “After (x
min)” means x minutes after continuous scanning. Note that Figure 5.3a (before) has a large
scan area, and Figure 5.3a (after) is only a sub-area of the initial surface. ........................ 139
Figure 5.4 The relationship between saturation state, CA concentration, and calcite dissolution
rate in bulk dissolution experiments in seawater reported in Subhas et al. (2017). The
background colors of green, white, and orange indicate the transition of dominating
dissolution mechanisms from step retreat, to defect-assisted, to homogeneous etch pit
spreading as W decreases. ................................................................................................ 141
Figure 5.5 A mechanistic illustration of CA-promoted dissolution on calcite (104) cleavage
surface. (a) the proton generated during the catalysis of CO2 hydration attacks calcite (104)
surface and promotes dissolution; (b) the adsorption of CA on the calcite (104) surface, and
the intramolecular proton transfer from the catalytic center to the mineral surface. .......... 144

8
List of Tables
Table 2.1 Coefficients for the effect of pressure on the dissociation constants of acids in
seawater. ................................................................................................................... 35
Table 2.2 Apparent partial molal volume changes for calcite dissolution ................................... 36
Table 2.3 Fits to Figure 2.8 & 2.9 and calculated local interfacial energy barrier I (mJ/m
2
) under
different pressure ...................................................................................................... 42

Table 3.1 PIC fluxes, carbonate percentages and PIC/POC in sediment trap samples of CDisK-
IV (August, 2017) ..................................................................................................... 68
Table 3.2 Experimental conditions and materials used in published aragonite dissolution studies
................................................................................................................................. 74

Table 4.1 DIC, alkalinity and W of the fill bag, solution in inflow syringe and solution in outflow
syringe. ..................................................................................................................... 97
Table 4.2 A comparison of bulk rate and step velocity between this study and previous
publications. ........................................................................................................... 111
Table 4.3 Measured etch pit density against solution undersaturation. ..................................... 113
Table 4.4 Bulk dissolution rates, step velocities, and edge lengths for six sub-regions on the
calcite (104) surface during dissolution experiments at W~0.5. ................................ 123

9
CHAPTER 1: INTRODUCTION

1.1 Oceanic Carbon Cycle and CaCO3 Dissolution
The ocean contains 38,000 Pg C, which is 50 times the amount of carbon in the
atmosphere, and 20 times more than the reservoirs on land. Consequently, the ocean features
prominently in the global carbon cycle, and has a large impact on atmospheric CO2 concentration
(Sigman et al., 2010).
The atmosphere and the surface ocean exchange carbon in the form of gaseous CO2 via
Henry’s Law. The exchange of carbon between the surface and the interior of the ocean is
mediated by the physical and biological carbon pumps (Volk and Hoffert, 1985). The physical
(or solubility) pump denotes the processes of carbon transport that is associated with deep
vertical mixing. The biological carbon pump encompasses two processes. One is the soft tissue
(or organic carbon) pump, which is driven by phytoplankton that converts DIC (dissolved
inorganic carbon) into POC (particulate organic carbon) by photosynthesis. The other process is
the carbonate counter pump, which describes the production and sinking of CaCO3 by calcifying
organisms. The fate of PIC export is partitioned into dissolution in the water column, dissolution
at the sea floor, and burial into the sediments.
In terms of the alkalinity cycle, rivers deliver dissolved weathering products to the ocean
in the form of alkalinity, which at steady state is removed via the production and burial of calcite
and aragonite minerals. Alkalinity is consumed in the surface ocean by CaCO3 precipitation and
produced in the deep ocean by CaCO3 dissolution. Estimates of open ocean calcification vary
between 0.4~1.8 Gt PIC yr
-1
, whereas only 0.1 Gt PIC yr
-1
is buried in deep-sea sediments
(Berelson et al., 2007). Therefore, the majority of CaCO3 produced in the surface ocean must be
dissolved either in the water column or in ocean sediments. The location and rate of CaCO3

10
dissolution, and thus where and how fast alkalinity returns to the ocean system, has been
underconstrained. As a result, many of the uncertainties in diagnostic and prognostic marine
carbon cycle models arise from an imperfect understanding of the correct formulation for the
dissolution of CaCO3 (Iglesias-Rodriguez et al., 2002).
Excess Ca
2+
and excess alkalinity are terms proposed to evaluate the changes due to
CaCO3 dissolution for a given isopycnal surface (Brewer et al., 1975; Chen, 1978; Sabine et al.,
1995; Gruber et al., 1996; Feely et al., 2002). The determinations of the two terms in the water
column indicate that as much as 60-80% CaCO3 dissolution may happen in the upper 400-1000
m of the ocean, well above the calcite saturation horizon (Milliman et al., 1999; Feely et al.,
2002). The shallow dissolution phenomenon may be due to different processes, such as
dissolution in microenvironments (e.g. guts of zooplankton (Bishop et al., 1980; Harris, 1994);
bacterial organic matter oxidation in sinking particles (Jansen and Wolf-Gladrow, 2001); or
dissolution of more soluble Mg-calcite or aragonite (Byrne et al., 1984)). Quantifying water
column dissolution, especially in the upper ocean, seems crucial to understand alkalinity
dynamics and to determine the response of the oceanic system to perturbations in either
alkalinity or CO2 input to the ocean-atmosphere system.

1.2 Previous Studies on CaCO3 dissolution
Dissolution kinetics of CaCO3 have been investigated extensively with a variety of
focuses. In non-seawater solutions, calcite dissolution is studied to simulate relevant geological
environments, e.g. formation of karst landscapes (Dreybrodt, 1987, 1990) and the evolution of
downstream water chemistry in rivers (Herman, 1982; Dreybrodt et al., 1992); or to explain
dissolution behaviors with basic chemical reactions and adsorption processes of the solute by

11
varying solution components (Plummer et al., 1978; Chou et al., 1989; Arakaki and Mucci, 1995;
Svensson and Dreybrodt, 1992; Pokrovsky and Schott, 2002). Meanwhile laboratory dissolution
studies in natural seawater aim to better constrain dissolution rates in ocean water columns
(Morse and Berner, 1972; Keir, 1980; Walter and Morse, 1985; Gehlen et al., 2005), and in
marine sediments (Sulpis et al., 2017, 2019). In situ measurements (Peterson, 1966; Milliman,
1975; Honjo and Erez, 1978; Boudreau, 2013) further justify the real dissolution rates in natural
marine environments despite potential effects of various uncontrolled variables.
In terms of mineral species, calcite has received the most attention (e.g. Berner, 1967;
Berner and Morse, 1974; Sjöberg, 1976; Chou et al., 1989; Dreybrodt et al., 1996; Arvidson et
al., 2003; Gehlen et al., 2005; Sulpis et al., 2017 etc.). Aragonite dissolution has been less
studied, despite the fact that it is the dominant carbonate mineral in tropical and subtropical
shallow water carbonate-rich sediments (Milliman, 1975; Honjo and Erez, 1978; Morse et al.,
1979; Keir, 1980; Busenberg and Plummer, 1986; Acker et al., 1987; Gutjahr et al., 1996 a, b).
In addition to synthetic calcite and aragonite, dissolution rates of biogenic carbonates have also
been measured and compared (Peterson, 1966; Berger, 1967, 1970; Keir, 1980; Walter, 1985,
1986; Walter and Morse, 1984, 1985; Kennish and Lutz, 1999; Cubillas et al., 2005).
In the past 25 years, an increasing number of dissolution studies have focused on direct
observation and quantification of the kinetics of dissolution processes on calcite surfaces using
Atomic Force Microscopy (AFM) and Vertical Scanning Interferometry (VSI) (e.g. Stipp et al.,
1994; Dove and Platt, 1996; Liang et al., 1996; Liang and Baer, 1997; McCoy and LaFemina,
1997; Shiraki et al., 2000; Lea et al., 2001; De Giudici, 2002; Morse and Arvidson, 2002;
Arvidson et al., 2003, 2006; Teng, 2004; Vinson and Lüttge, 2005; Bisschop et al., 2006; Lüttge
and Arvidson, 2010). Considerations of microscopic scale processes significantly facilitate the

12
interpretation of results obtained from bulk dissolution experiments (Xu et al., 2012; Smith et al.,
2013; Arvidson et al., 2003), yet also leave many questions as to why and whether there is
agreement between lab-based and in situ rates.
Our group developed a novel determination of the dissolution kinetics of calcite in
seawater by dissolving
13
C-labeled calcite in unlabeled seawater, and tracing the evolving δ
13
C
composition of the fluid over time to establish dissolution rates (Subhas et al., 2015, 2017; Dong
et al., 2018; Naviaux et al., 2019). This method provides sensitive determinations of dissolution
rate, which we couple with tight constraints on seawater saturation state. This thesis contains
work that takes advantage of the isotope-tracer technique, and reports dissolution kinetics of
calcite and aragonite both in the lab and in the field. The dissolution mechanism in seawater with
and without carbonic anhydrase (CA) is further investigated with AFM.

1.3 Dissertation Chapters and Research Objectives
This dissertation summarizes my research on the kinetics and mechanisms of calcite and
aragonite dissolution in seawater. There are four separate chapters focusing on different
components under this general theme, and each chapter is written as a unit submitted or to be
submitted for publication. The research objectives and main conclusions from each chapter are
listed as below.
Chapter 2. Pressure dependence of calcite dissolution rate in seawater
This chapter provides laboratory data for the calcite dissolution rate under variable
pressure. Oceanographers have assumed that carbonate dissolution under pressure follows the
same dissolution rate law obtained at atmospheric pressure, with pressure only changing the
stoichiometric solubility product (K
*
sp). The objective of this chapter is to distinguish the

13
pressure effect on calcite dissolution kinetics from the effect of water chemistry. Results show
that dissolution rates are enhanced by a factor of 2-4 at 700 dbar compared to dissolution at the
same W under ambient pressure (10 dbar). The observed enhancement is well beyond the
uncertainty associated with the thermodynamic properties of calcite under pressure, and thus
should be interpreted as a kinetic pressure effect on calcite dissolution. We also find that the
pressure effect does not scale with increasing pressure, and dissolution rates become independent
of pressure above 700 dbar (up to 2500 dbar). This pressure effect implies that sinking particles
would dissolve at shallower depth than previously thought.

Chapter 3. Kinetics of aragonite dissolution in seawater and its role in shallow water
column dissolution in the North Pacific
This chapter aims to evaluate the role of aragonite dissolution as the source of shallow
depth excess alkalinity which has intrigued chemical oceanographers for decades (Feely et al.,
2002). I present a comparison of aragonite dissolution rates measured in the lab vs. in situ rates
along a transect between Hawaii and Alaska, as well as the particulate inorganic carbon (PIC)
and particulate organic carbon (POC) fluxes and concentrations in sinking and suspended
materials along the North Pacific transect. Using a simple 1-D model and the measured aragonite
flux combined with the measured dissolution rate, we find that aragonite dissolution can only
account for a small fraction of the excess alkalinity observed in the North Pacific. However,
respiration-driven dissolution or metazoan/zooplankton consumption, indicated by the
simultaneous attenuation of PIC and POC in sediment traps, is proposed as being capable of
generating the magnitude of dissolution suggested by the observed excess alkalinity.

14
Chapter 4. An atomic force microscopy study of calcite dissolution in seawater and the
dependence of dissolution rate on surface properties
This chapter presents the first examination of calcite dissolution in seawater using
Atomic Force Microscopy (AFM). Our goal is to relate changes in dissolution kinetics in bulk
water experiments as a function of water chemistry to some surface property changes on the
calcite crystal. I quantify step retreat velocity, etch pit density, and introduce a new parameter,
“the edge length density” in order to compare calcite dissolution under AFM conditions to those
conducted in bulk solution experiments. The effects of undersaturation on step velocity and etch
pit density have been separated and compared between seawater and freshwater. The tight
correlation between dissolution rate and the “edge length density” suggests that this new
parameter captures an Angstrom-scale roughness that BET surface area does not capture and is
essential in determining bulk dissolution rates.

Chapter 5. A mechanistic study of the effect of carbonic anhydrase on calcite dissolution
and its implications for the ocean carbon cycle
This chapter examines the catalytic mechanism of carbonic anhydrase (CA) on calcite
dissolution and discusses its implications for the ocean carbon cycle. CA has been reported to
promote calcite dissolution in seawater (Subhas et al., 2017). This was a shocking discovery and
lead to many conversations about how CA actually enhances dissolution kinetics. In this chapter,
I use AFM to directly observe calcite dissolution in CA-bearing seawater. No significant
difference in calcite step velocity is observed with vs. without CA. However, CA is found to
enhance dissolution when in contact or is very close (~nm) to the calcite surface. The irregular
morphology of etch pits (curved steps) suggests that the catalysis is likely a surface process that

15
includes the adsorption of CA on the calcite surface and proton transfer from the catalytic center
to the mineral surface. These results recommend caution in estimating carbonate rock weathering
and oceanic PIC dissolution with bulk CA concentrations in natural environments because the
key to enhanced dissolution may be the contact that occurs between CA and the mineral surface.

References
Acker, J. G., Byrne, R. H., Ben-Yaakov, S., Feely, R. A., & Betzer, P. R. (1987). The effect of
pressure on aragonite dissolution rates in seawater. Geochimica et Cosmochimica
Acta, 51(8), 2171-2175.
Arakaki, T., & Mucci, A. (1995). A continuous and mechanistic representation of calcite
reaction-controlled kinetics in dilute solutions at 25 C and 1 atm total pressure. Aquatic
Geochemistry, 1(1), 105-130.
Arvidson, R. S., Collier, M., Davis, K. J., Vinson, M. D., Amonette, J. E., & Lüttge, A. (2006).
Magnesium inhibition of calcite dissolution kinetics. Geochimica et Cosmochimica
Acta, 70(3), 583-594.
Arvidson, R. S., Ertan, I. E., Amonette, J. E., & Luttge, A. (2003). Variation in calcite
dissolution rates: A fundamental problem?. Geochimica et cosmochimica acta, 67(9),
1623-1634.
Berelson, W. M., Balch, W. M., Najjar, R., Feely, R. A., Sabine, C., & Lee, K. (2007). Relating
estimates of CaCO3 production, export, and dissolution in the water column to
measurements of CaCO3 rain into sediment traps and dissolution on the sea floor: A
revised global carbonate budget. Global Biogeochemical Cycles, 21(1).
Berger, W. H. (1967). Foraminiferal ooze: solution at depths. Science, 156(3773), 383-385.
Berger, W. H. (1970). Planktonic foraminifera: selective solution and the lysocline. Marine
Geology, 8(2), 111-138.
Berner, R. A. (1967). Comparative dissolution characteristics of carbonate minerals in the
presence and absence of aqueous magnesium ion. American Journal of Science, 265(1),
45-70.
Berner, R. A., & Morse, J. W. (1974). Dissolution kinetics of calcium carbonate in sea water; IV,
Theory of calcite dissolution. American Journal of Science, 274(2), 108-134.
Bishop, J. K., Collier, R. W., Kettens, D. R., & Edmond, J. M. (1980). The chemistry, biology,
and vertical flux of particulate matter from the upper 1500 m of the Panama Basin. Deep
Sea Research Part A. Oceanographic Research Papers, 27(8), 615-640.
Bisschop, J., Dysthe, D. K., Putnis, C. V., & Jamtveit, B. (2006). In situ AFM study of the
dissolution and recrystallization behaviour of polished and stressed calcite
surfaces. Geochimica et cosmochimica acta, 70(7), 1728-1738.
Boudreau, B. P. (2013). Carbonate dissolution rates at the deep ocean floor. Geophysical
Research Letters, 40(4), 744-748.
Brewer, P. G., Wong, G. T., Bacon, M. P., & Spencer, D. W. (1975). An oceanic calcium
problem?. Earth and Planetary Science Letters, 26(1), 81-87.

16
Busenberg, E. U. R. Y. B. I. A. D. E. S., Plummer, L. N., & Mumpton, F. A. (1986). A
comparative study of the dissolution and crystal growth kinetics of calcite and
aragonite. Studies Diagenesis USGS Bull, 1578, 139-168.
Byrne, R. H., Acker, J. G., Betzer, P. R., Feely, R. A., & Cates, M. H. (1984). Water column
dissolution of aragonite in the Pacific Ocean. Nature, 312(5992), 321.
Chen, C. T. A. (1978). Decomposition of calcium carbonate and organic carbon in the deep
oceans. Science, 201(4357), 735-736.
Chou, L. E. I., Garrels, R. M., & Wollast, R. (1989). Comparative study of the kinetics and
mechanisms of dissolution of carbonate minerals. Chemical geology, 78(3-4), 269-282.
Cubillas, P., Köhler, S., Prieto, M., Chaïrat, C., & Oelkers, E. H. (2005). Experimental
determination of the dissolution rates of calcite, aragonite, and bivalves. Chemical
Geology, 216(1-2), 59-77.
De Giudici, G. (2002). Surface control vs. diffusion control during calcite dissolution:
Dependence of step-edge velocity upon solution pH. American Mineralogist, 87(10),
1279-1285.
Dong, S., Subhas, A. V., Rollins, N. E., Naviaux, J. D., Adkins, J. F., & Berelson, W. M. (2018).
A kinetic pressure effect on calcite dissolution in seawater. Geochimica et Cosmochimica
Acta, 238, 411-423.
Dove, P. M., & Platt, F. M. (1996). Compatible real-time rates of mineral dissolution by atomic
force microscopy (AFM). Chemical Geology, 127(4), 331-338.
Dreybrodt, W. (1987). The kinetics of calcite dissolution and its consequences to karst evolution
from the initial to the mature state. Natl. Speleol. Soc. Bull., 49: 31-49.
Dreybrodt, W. (1990). The role of dissolution kinetics in the development of karst aquifers in
limestone: a model simulation of karst evolution. The Journal of geology, 98(5), 639-655.
Dreybrodt, W., Buhmann, D., Michaelis, J., & Usdowski, E. (1992). Geochemically controlled
calcite precipitation by CO2 outgassing: Field measurements of precipitation rates in
comparison to theoretical predictions. Chemical Geology, 97(3-4), 285-294.
Dreybrodt, W., Lauckner, J., Zaihua, L., Svensson, U., & Buhmann, D. (1996). The kinetics of
the reaction CO2+ H2O→ H++ HCO3− as one of the rate limiting steps for the
dissolution of calcite in the system H2O-CO2-CaCO3. Geochimica et Cosmochimica
Acta, 60(18), 3375-3381.
Feely, R. A., Sabine, C. L., Lee, K., Millero, F. J., Lamb, M. F., Greeley, D., ... & Ono, T.
(2002). In situ calcium carbonate dissolution in the Pacific Ocean. Global
Biogeochemical Cycles, 16(4), 91-1.
Gehlen, M., Bassinot, F. C., Chou, L., & McCorkle, D. (2005). Reassessing the dissolution of
marine carbonates: II. Reaction kinetics. Deep Sea Research Part I: Oceanographic
Research Papers, 52(8), 1461-1476.
Gruber, N., Sarmiento, J. L., & Stocker, T. F. (1996). An improved method for detecting
anthropogenic CO2 in the oceans. Global Biogeochemical Cycles, 10(4), 809-837.
Gutjahr, A., Dabringhaus, H., & Lacmann, R. (1996a). Studies of the growth and dissolution
kinetics of the CaCO3 polymorphs calcite and aragonite I. Growth and dissolution rates
in water. Journal of crystal growth, 158(3), 296-309.
Gutjahr, A., Dabringhaus, H., & Lacmann, R. (1996b). Studies of the growth and dissolution
kinetics of the CaCO3 polymorphs calcite and aragonite II. The influence of divalent
cation additives on the growth and dissolution rates. Journal of Crystal Growth, 158(3),
310-315.

17
Harris, R. P. (1994). Zooplankton grazing on the coccolithophore Emiliania huxleyi and its role
in inorganic carbon flux. Marine Biology, 119(3), 431-439.
Herman, J.S., 1982. The dissolution kinetics of calcite, dolomite, and dolomitic rocks in the
carbon dioxidewater system. Ph.D. Dissertation, The Pennsylvania State University,
University Park, Pa.
Honjo, S., & Erez, J. (1978). Dissolution rates of calcium carbonate in the deep ocean; an in-situ
experiment in the North Atlantic Ocean. Earth and Planetary Science Letters, 40(2), 287-
300.
Iglesias-Rodriguez, M. D., Armstrong, R., Feely, R., Hood, R., Kleypas, J., Milliman, J. D., ... &
Sarmiento, J. (2002). Progress made in study of ocean's calcium carbonate budget. Eos,
Transactions American Geophysical Union, 83(34), 365-375.
Jansen, H., & Wolf-Gladrow, D. A. (2001). Carbonate dissolution in copepod guts: a numerical
model. Marine Ecology Progress Series, 221, 199-207.
Keir, R. S. (1980). The dissolution kinetics of biogenic calcium carbonates in
seawater. Geochimica et Cosmochimica Acta, 44(2), 241-252.
Kennish, M. J., & Lutz, R. A. (1999). Calcium carbonate dissolution rates in deep-sea bivalve
shells on the East Pacific Rise at 21 N: results of an 8-year in-situ
experiment. Palaeogeography, Palaeoclimatology, Palaeoecology, 154(4), 293-299.
Lea, A. S., Amonette, J. E., Baer, D. R., Liang, Y., & Colton, N. G. (2001). Microscopic effects
of carbonate, manganese, and strontium ions on calcite dissolution. Geochimica et
Cosmochimica Acta, 65(3), 369-379.
Liang, Y., & Baer, D. R. (1997). Anisotropic dissolution at the CaCO3 (101̄4)—water
interface. Surface Science, 373(2-3), 275-287.
Liang, Y., Baer, D. R., McCoy, J. M., Amonette, J. E., & Lafemina, J. P. (1996). Dissolution
kinetics at the calcite-water interface. Geochimica et Cosmochimica Acta, 60(23), 4883-
4887.
Lüttge, A., & Arvidson, R. S. (2010). Reactions at surfaces: A new approach integrating
interferometry and kinetic simulations. Journal of the American Ceramic Society, 93(11),
3519-3530.
McCoy, J. M., & LaFemina, J. P. (1997). Kinetic Monte Carlo investigation of pit formation at
the CaCO3 (1014) surface-water interface. Surface Science, 373(2-3), 288-299.
Milliman, J. D. (1975). Dissolution of aragonite, Mg-calcite, and calcite in the North Atlantic
Ocean. Geology, 3(8), 461-462.
Milliman, J. D. (1993). Production and accumulation of calcium carbonate in the ocean: budget
of a nonsteady state. Global Biogeochemical Cycles, 7(4), 927-957.
Milliman, J. D., & Droxler, A. W. (1995). Calcium carbonate sedimentation in the global ocean:
linkages between the neritic and pelagic environments. Oceanography, 8(3), 92-94.
Milliman, J. D., Troy, P. J., Balch, W. M., Adams, A. K., Li, Y. H., & Mackenzie, F. T. (1999).
Biologically mediated dissolution of calcium carbonate above the chemical
lysocline?. Deep Sea Research Part I: Oceanographic Research Papers, 46(10), 1653-
1669.
Morse, J. W., & Arvidson, R. S. (2002). The dissolution kinetics of major sedimentary carbonate
minerals. Earth-Science Reviews, 58(1-2), 51-84.
Morse, J. W., & Berner, R. A. (1972). Dissolution kinetics of calcium carbonate in sea water; I,
A kinetic origin for the lysocline. American Journal of Science, 272(9), 840-851.

18
Morse, J. W., de Kanel, J. O. H. N., & Harris, K. A. R. E. N. (1979). Dissolution kinetics of
calcium carbonate in seawater; VII, The dissolution kinetics of synthetic aragonite and
pteropod tests. American Journal of Science, 279(5), 488-502.
Naviaux, J. D., Subhas, A. V., Rollins, N. E., Dong, S., Berelson, W. M., & Adkins, J. F. (2019).
Temperature dependence of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 246, 363-384.
Peterson, M. N. A. (1966). Calcite: rates of dissolution in a vertical profile in the central
Pacific. Science, 154(3756), 1542-1544.
Plummer, L. N., Wigley, T. M. L., & Parkhurst, D. L. (1978). The kinetics of calcite dissolution
in CO 2-water systems at 5 degrees to 60 degrees C and 0.0 to 1.0 atm CO 2. American
journal of science, 278(2), 179-216.
Pokrovsky, O. S., & Schott, J. (2002). Surface chemistry and dissolution kinetics of divalent
metal carbonates. Environmental science & technology, 36(3), 426-432.
Sabine, C. L., Mackenzie, F. T., Winn, C., & Karl, D. M. (1995). Geochemistry of carbon
dioxide in seawater at the Hawaii Ocean Time series station, ALOHA. Global
Biogeochemical Cycles, 9(4), 637-651.
Shiraki, R., Rock, P. A., & Casey, W. H. (2000). Dissolution kinetics of calcite in 0.1 M NaCl
solution at room temperature: an atomic force microscopic (AFM) study. Aquatic
Geochemistry, 6(1), 87-108.
Sigman, D. M., Hain, M. P., & Haug, G. H. (2010). The polar ocean and glacial cycles in
atmospheric CO2 concentration. Nature, 466(7302), 47.
Sjöberg, E. L. (1976). A fundamental equation for calcite dissolution kinetics. Geochimica et
Cosmochimica Acta, 40(4), 441-447.
Smith, M. E., Knauss, K. G., & Higgins, S. R. (2013). Effects of crystal orientation on the
dissolution of calcite by chemical and microscopic analysis. Chemical Geology, 360, 10-
21.
Stipp, S. L. S., Eggleston, C. M., & Nielsen, B. S. (1994). Calcite surface structure observed at
microtopographic and molecular scales with atomic force microscopy
(AFM). Geochimica et Cosmochimica Acta, 58(14), 3023-3033.
Subhas, A. V., Adkins, J. F., Rollins, N. E., Naviaux, J., Erez, J., & Berelson, W. M. (2017).
Catalysis and chemical mechanisms of calcite dissolution in seawater. Proceedings of the
National Academy of Sciences, 114(31), 8175-8180.
Subhas, A. V., Rollins, N. E., Berelson, W. M., Dong, S., Erez, J., & Adkins, J. F. (2015). A
novel determination of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 170, 51-68.
Sulpis, O., Lix, C., Mucci, A., & Boudreau, B. P. (2017). Calcite dissolution kinetics at the
sediment-water interface in natural seawater. Marine Chemistry, 195, 70-83.
Sulpis, O., Mucci, A., Boudreau, B. P., Barry, M. A., & Johnson, B. D. (2019). Controlling the
diffusive boundary layer thickness above the sediment–water interface in a thermostated
rotating-disk reactor. Limnology and Oceanography: Methods.
Svensson, U., & Dreybrodt, W. (1992). Dissolution kinetics of natural calcite minerals in CO2-
water systems approaching calcite equilibrium. Chemical Geology, 100(1-2), 129-145.
Teng, H. H. (2004). Controls by saturation state on etch pit formation during calcite
dissolution. Geochimica et Cosmochimica Acta, 68(2), 253-262.
Vinson, M. D., & Lüttge, A. (2005). Multiple length-scale kinetics: an integrated study of calcite
dissolution rates and strontium inhibition. American Journal of Science, 305(2), 119-146.

19
Volk, T., & Hoffert, M. I. (1985). Ocean carbon pumps: Analysis of relative strengths and
efficiencies in ocean-driven atmospheric CO2 changes. The carbon cycle and
atmospheric CO2: natural variations Archean to present, 99-110.
Walter, L. M. (1985). Relative reactivity of skeletal carbonates during dissolution: implications
for diagenesis.
Walter, L. M. (1986). Relative efficiency of carbonate dissolution and precipitation during
diagenesis: a progress report on the role of solution chemistry.
Walter, L. M., & Morse, J. W. (1984). Reactive surface area of skeletal carbonates during
dissolution; effect of grain size. Journal of Sedimentary Research, 54(4), 1081-1090.
Walter, L. M., & Morse, J. W. (1985). The dissolution kinetics of shallow marine carbonates in
seawater: A laboratory study. Geochimica et Cosmochimica Acta, 49(7), 1503-1513.
Xu, J., Fan, C., & Teng, H. H. (2012). Calcite dissolution kinetics in view of Gibbs free energy,
dislocation density, and pCO2. Chemical Geology, 322, 11-18.

20
CHAPTER 2: PRESSURE DEPENDENCE OF CALCITE
DISSOLUTION RATE IN SEAWATER

This chapter was published in 2018 as:
Dong, S., Subhas, A. V., Rollins, N. E., Naviaux, J. D., Adkins, J. F., & Berelson, W. M. (2018).
A kinetic pressure effect on calcite dissolution in seawater. Geochimica et Cosmochimica
Acta, 238, 411-423.

ABSTRACT
This chapter provides laboratory data of calcite dissolution rate as a function of seawater
undersaturation state (1- Ω) under variable pressure.
13
C-labeled calcite was dissolved in
unlabeled seawater and the evolving d
13
C composition of the fluid was monitored over time to
evaluate the dissolution rate. Results show that dissolution rates are enhanced by a factor of 2-4
at 700 dbar compared to dissolution at the same Ω under ambient pressure (10 dbar). This
dissolution rate enhancement under pressure applies over an Ω range of 0.65 to 1 between 10
dbar and 700 dbar. Above 700 dbar (up to 2500 dbar), dissolution rates become independent of
pressure. The observed enhancement is well beyond the uncertainty associated with the
thermodynamic properties of calcite under pressure (partial molar volume ΔV), and thus should
be interpreted as a kinetic pressure effect on calcite dissolution. Dissolution at ambient pressure
and higher pressures yield non-linear dissolution kinetics, the pressure effect does not
significantly change the reaction order n in Rate = k(1- M)
n
, which is shown to vary from 3.1±0.3
to 3.8±0.5 from 10 dbar to 700 dbar over Ω = 0.65 to 0.9. Furthermore, two different dissolution
mechanisms are indicated by a discontinuity in the rate-undersaturation relationship, and seen at
both ambient and higher pressures. The discontinuity, Ω
NOPQPNRS
= 0.87±0.05 and 0.90±0.03 at 10

21
dbar and 1050 dbar respectively, are similar within error. The reaction order, n, at Ω>0.9 is
0.47±0.27 and 0.46±0.15 at 10 dbar and 700 dbar respectively. This Ω
NOPQPNRS
is considered to be
the threshold between step retreat dissolution and defect-assisted dissolution. The kinetic
enhancement of dissolution rate at higher pressures is related to a decrease in the interfacial
energy barrier at dissolution sites. The impact of pressure on the calcite dissolution kinetics
implies that sinking particles would dissolve at shallower depth than previously thought.

2.1 INTRODUCTION
The investigation of calcium carbonate dissolution in the ocean is critical in constructing
global carbon budgets and understanding the ocean’s role in neutralizing atmospheric CO2 by
dissolving CaCO3. Most kinetics studies (Morse, 1978; Morse et al., 1979; Keir, 1980; Byrne et
al., 1984; Walter and Morse, 1985; Hales and Emerson, 1997; Gehlen et al., 2005; Subhas et al.,
2015) of the dissolution behavior of calcium carbonates in seawater have expressed the
dependence of dissolution rate on seawater saturation state through an empirical equation of the
form:
TUVW= X(1−M)
[
(1)
where k is a rate constant, n is a positive constant known as the “order” of the reaction, Ω
is the saturation state defined as the ion concentration product over the stoichiometric solubility
product (\
∗
]^
):
M =
_`R
ab
c[`e
f
ag
]
i
∗
jk
(2)
Ω varies either by changing the numerator – ion concentrations (water chemistry), or by
changing the denominator – \
∗
]^
. In deep oceanic waters, which are undersaturated with respect
to calcite, salinity varies only slightly (34 to 36), the water masses are nearly isothermal (3±2°C)

22
and pressure is the most important influence on \
∗
]^
. At 2°C and a pressure corresponding to a
depth of about 6500 m, \
∗
]^
is about 3.7 times greater than at atmospheric pressure (Ingle,
1975). Another factor that decreases Ω in the deep ocean is that respiration of sinking organic
matter releases CO2, decreases pH and [CO3
2-
], and thus Ω. The relative importance of these two
factors (pressure and microbial respiration of organic matter) in modulating Ω was investigated
in a North Pacific Ocean Section (Figure 2.1). Ω
OlRS
is the actual Ω in the ocean. Ω
NmlnP]QOo
is Ω
calculated by Ocean Data View if there is no effect of pressure on \
∗
]^
(Ω affected only by DIC,
alkalinity, salinity and temperature). Ω
^Ol]]pOl
is Ω afftected by pressure, which is shown as the
difference of Ω
OlRS
and Ω
NmlnP]QOo
here.
Above 1000 m, Figure 2.1a (Ω
OlRS
) generally follows the trend of 2.1b (Ω
NmlnP]QOo
),
meaning that water chemistry dominates changes in Ω. Ω
NmlnP]QOo
varies between 6 at the ocean
surface (not shown in the figure) and 1. Below 1000 m, Ω
OlRS
deviates from Ω
NmlnP]Q Oo
due to
the effect of pressure on Ω (Ω
^Ol]]pOl
). At 5000 m depth, Ω can be up to 1.2 units lower than
predicted by Ω
NmlnP]QOo
(Figure 2.1c). The impact of Ω
^Ol]]pOl
is greater in the South Pacific,
somewhat less in the north, which reflects the North Pacific oxygen minimum zone’s strong
influence on the saturation state.

23

Figure 2.1 Separation of effect of water chemistry and pressure on Ω
NRSNPQl
in a Pacific section by Ocean Data View simulation.
(a) Ω
OlRS
: actual Ω. (b) Ω
NmlnP]QOo
: Ω calculated by Ocean Data View if there is no effect of pressure on \
∗
]^
(\
∗
]^
after Mucci,
1983), namely Ω affected only by water chemistry (DIC, alkalinity, salinity, temperature). (c) ΔΩ: difference between (b) and (a),
Ω affected by pressure; up to 1.2 Ω units at 5000 m depth. (d) Defined section (red box) on map.

(a)
(b)
(c)
(d)

24
Oceanographers have assumed that carbonate dissolution under pressure follows the same
dissolution rate law (same k and n for Eq. (2)) obtained at atmospheric pressure, with pressure
only changing the stoichiometric solubility product (\
∗
]^
). In other words, a change in the
denominator in Eq. (1) has the same effect as a change in the numerator in determining
dissolution rates (Eq. (2)). This implies that dissolution will be kinetically the same regardless of
pressure. Whether this assumption is true is the question I address in this chapter.
The only study, as far as I know, that examined the effect of pressure on carbonate
dissolution rates versus saturation state in seawater was that of Acker et al. (1987). Their
shipboard experiments measured pteropod (aragonite) dissolution rates in natural seawater at
various depths between 100 and 5000 meters, based on spectrophotometric examination of
seawater pH using a pH-sensitive dye, phenol red. They concluded that aragonite dissolution in
seawater at variable pressures was well described by the equation Rate=k’ ([CO3
2-
]s-[CO3
2-
])
n
,
where [CO3
2-
]s is the carbonate ion concentration at saturation, [CO3
2-
] is the observed carbonate
ion concentration, and k’ and n are empirical constants; and that their measurements were
consistent with the estimate ΔV=-36.5cm
3
/mole for the volume change accompanying the
dissolution of aragonite. However, two factors limited the measurement precision and thus could
change the interpretation of data in their study. First, dissolution rates determined from ΔpH may
have a limited precision. The largest amount of dissolution in Acker et al. (1987) was only
equivalent to a ΔpH~0.05; at Ω = 0.8~1, ΔpH was < 0.01. Even though the error on dissolution
rates were not discussed in the paper, pH determination based on the spectrophotometric method
has an error on the order of 0.002 (Robert-Baldo et al., 1985), which is greater than 20% of the
ΔpH at Ω = 0.8~1. This precision limit also arises in other seawater carbonate dissolution studies
in which dissolution rates were estimated from ΔpH (e.g. Byrne et al., 1984; Gehlen et al., 2005),

25
acid additions to keep pH constant (e.g. Walter and Morse, 1985), alkalinity change (e.g. Keir,
1980), or CaCO3 dry weight loss (e.g. Peterson, 1966; Berger, 1967; Honjo and Erez, 1978; Keir,
1980; Thunell et al.,1981; Fukuhara et al., 2008). Second, shipboard experiments performed at
pressures which corresponded to the depths from which the seawater samples were obtained
lacked control in other key variables. Specifically, dissolution experiments were conducted under
different water chemistry (carbonate system parameters) and other factors that may affect
dissolution (e.g. microbial activity, soluble reactive phosphate concentrations, carbonic
anhydrase activity etc.). The problem of uncontrolled variables also exists in other in situ
seawater dissolution studies (e.g. Peterson, 1966; Berger, 1967; Milliman, 1977; Honjo and Erez,
1978; Thunell et al., 1981; Metzler et al., 1982; Fukuhara et al., 2008). The value of these earlier
studies is unquestionable, yet I sought to conduct more controlled experiments to constrain the
effect of pressure on calcite dissolution kinetics, focusing on Ω values closest to those most
commonly encountered in the modern ocean (0.65-1).
Our goal is to distinguish the pressure effect on carbonate dissolution kinetics from the
effect of water chemistry. I performed dissolution experiments in acidified Dickson standard
seawater using
13
C labeled inorganic calcite following methodologies described in Subhas et al.
(2015). This
13
C labeled method, coupled with tight constraints on seawater saturation state
(from measurements of dissolved inorganic carbon (DIC) and total alkalinity (TA)), provides
much more sensitive determinations of dissolution rates compared to previously employed
methods. Desired Ω values were obtained by changing experimental pressure and/or adding HCl
to the same standard Dickson seawater, so that the effect of changing pressure versus changing
alkalinity on dissolution rates could be distinguished and compared.

26
2.2 METHODS
Synthetic inorganic Ca
13
CO3 (calcite) purchased from Sigma Aldrich (SKU 492027, ³99
atom%) was wet-sieved to a grain size of 70-100 µm using 18.2 MW Milli-Q water adjusted to a
pH of ~8 using ammonium hydroxide. The powders were then dried at 60°C in oven overnight.
The mineralogy was confirmed to be 100% calcite via XRD. Approximately 4 mg of
13
C labeled
calcite was placed in a Supel Inert Foil Gas Sampling Bag (SUPELCO 30226-U) with 300 mL of
standard reference Dickson seawater
(https://www.nodc.noaa.gov/ocads/oceans/Dickson_CRM/batches.html). The Dickson standard
seawater used for the dissolution experiments has a phosphate concentration of 0.28~0.58
µmol/kg. The
13
C calcite was placed inside a polyester mesh (41 µm mesh size) bag (Component
Supply Co. /U-CMY-41-B)) to allow interaction of the solids with seawater while preventing
them from leaving the system during sampling. The evolution of seawater d
13
C was traced over
time (hours to days). d
13
C gain during the experiments varied from 5‰ to 50‰. d
13
C signals
were converted into mass loss, and the dissolution rate was obtained based on the slope of moles
dissolved over time. Slopes, intercepts, and the goodness of fits were obtained using the LINEST
function in Microsoft Excel. Carbonate system parameters were determined by measurement of
DIC and TA. d
13
C and DIC were determined using a Picarro Cavity Ring-Down Spectroscopy
(G2131-i) coupled to a Liaison interface and a modified AutoMate autosampler. Alkalinity was
determined by open-system Gran titration, performed on a custom-built instrument. All DIC and
TA measurements were standardized with Dickson seawater. An in-house standard was also used
to track long-term drift of the TA titration system. Characterization of materials and
determination of carbonate system parameters were discussed in our previous paper in detail
(Subhas et al., 2015). Ω was calculated from the CO2SYS program using the calcite partial molal

27
volume (ΔV) reported by Ingle (1975), -37.6 cm
3
mol
-1
at the experimental T=21°C; K’1 K’2
(apparent dissociation constants of carbonic acid in seawater) reported by Mehrbach et al. (1973)
and refit by Dickson and Millero (1987); K
tu
v
reported by Dickson et al. (1990). Replicate
precision (standard error) for DIC and TA measurements were ±1.9 µm/kg and ±1.0 µmol/kg
(1s) respectively, which allow Ω to be defined to ± 0.02.

Figure 2.2 Pressure case diagram.

Dissolution experiments were conducted in a custom-made pressure vessel (Figure 2.2)
capable of holding a sample bag at constant pressure while allowing for aliquots of the
experimental solution to be withdrawn. The pressure vessel was placed on a shaker table (VWR
Standard Analog Shaker) and set to a rate of 60 rpm for all experiments. The pressure case was
filled with hydraulic fluid (Lubriplate Hydraulic Jack Oil L0768-054, Lot: M0226) that was
manually pumped into the pressure case around the sample bag or released from the pressure
case until the system was at the desired pressure (upper limit was 5500 dbar). The sample bag

28
was connected through a flex tubing (Altaflo PVDF Flex Tubing 14C8101, Lot: W5520-06) to a
spring-loaded check valve that could “crack” at pre-set pressures, which allowed samples to be
removed from the bag while pressure remained relatively constant. During sampling, a 20 mL
syringe was connected to the check valve, and hydraulic fluid was slowly pumped into the
pressure case to squeeze the aliquot out of the bag by increasing the pressure 20-50 dbar above
the set value. Extrusion of seawater through the check valve reduced the system’s volume,
allowing pressure to return to the set value. Prior to sampling, the tubing and valve assembly
(volume < 5 mL) was flushed by extruding 20 mL of seawater. An 8 mL sample was then
injected directly into an evacuated sample vial to avoid degassing. Samples collected from a
pressurized bag are referred here as in situ, while samples collected directly from the sample bag
outside the pressure case (before and after a run) are called ex situ. Ex situ samples were taken to
quantify the effect of in situ sampling through the tubing and check valve as artifacts may arise
due to: mixing in the tubing, potential contamination from the tubing and check valve, degassing
through the tubing etc. Ex situ rates are preferentially used because of less potential
contamination, while in situ rate errors are used only when no obvious contamination was
observed (see discussion in Supplementary Materials). The agreement between these two
methods (±5%) provides confidence that our system and derived rates are free of artifact. Total
moles dissolved based on both in situ and ex situ samples for one experiment (P45) are shown in
Figure 2.3 as an example. Potential degassing during sampling would result in lower DIC values
and biased d
13
C (isotope fractionation during degassing) for in situ samples. Good agreement
between in situ and ex situ DIC and d
13
C values suggests no apparent degassing while sampling.
Nevertheless, for the determination of Ω based on DIC and TA, values of ex situ samples are
used to avoid the potential contamination mentioned above.

29

Figure 2.3 Moles of labeled carbonate dissolved based on d
13
C measurement from P45 (1050 dbar). Yellow squares are ex situ
samples (three before experiment, three after experiment); grey circles are in situ samples; trend line is fit to all data points.
Replicate value uncertainties are less than the size of the data points. Note that only 4×10
-7
mol calcite was dissolved over 4 days
(0.001% of total calcite powder).

Dissolution rates were normalized to geometric surface area (Subhas et al., 2015), which
was calculated using the mean sieving size and assuming cubic geometry:
w.y.
zl{n
=
|
}∙
Ä
(3)
where Å = 2.63 g/cm
3
is the density of calcite and Ç
̅ is the mean grain diameter of the
sieved fraction (70-100 µm).
Dissolution experiments were conducted by (i) varying pressure at fixed water chemistry
(black asterisks in Figure 2.4, Route 1); and (ii) varying water chemistry at fixed pressure (white
circles in Figure 2.4, Routes 2-6). For (ii), five sets of experiments at pressures of 10 dbar
(ambient pressure), 350 dbar, 700 dbar, 1050 dbar, and 2500 dbar were conducted. 1 dbar is
equivalent to 1 m ocean depth.

30

Figure 2.4 Ω contour plot as a function of pressure and alkalinity. Six sets of experiments were conducted. Black asterisks were
dissolution experiments conducted with the same fill bag seawater, Ω changed by changing pressure (Route 1). White circles
were dissolution experiments conducted under five different constant pressures (Route 2 – 10 dbar; Route 3 – 350 dbar; Route 4
– 700 dbar; Route 5 – 1050 dbar; Route 6 – 2500 dbar). Note that this diagram is only a rough demonstration of Ω and
experimental parameters, because DIC in this plot has to be set to a fixed value (2030 µmol/kg) while actual DIC for different
experiments (different batches of Dickson seawater) vary from 2000 µmol/kg to 2040 µmol/kg.

2.3 RESULTS
2.3.1 Dissolution rates at changing pressures
Dissolution rate measurements were made through a range of Ω values for which the
change of Ω was achieved in one of two ways: by changing water chemistry (alkalinity) at
atmospheric pressure (yellow squares in Figure 2.5) and by changing pressure under which
dissolution experiments occurred while keeping alkalinity and DIC constant (light grey circles in
Figure 2.5, experimental pressures are marked next to data points). In theory, both should yield
similar results, but it is clear that the data are offset from one another. To get a sense of how
thermodynamic uncertainty can affect the determination of Ω, the Ω’s for the pressure
experiments were recalculated using ΔV values that are more negative than the accepted value of
-37.6 cm
3
mol
-1
. A change of ΔV from -37.6 cm
3
mol
-1
(light grey circles) to -57.6 cm
3
mol
-1
(dark
grey diamonds) would make the changing pressure curve agree with the atmospheric pressure

31
curve, but by using the conventional value of ΔV, for the same saturation state, the dissolution
rate under pressure is approximately 2-4 times greater at typical oceanic Ω values of 0.8~1.0.

Figure 2.5 Dissolution rate vs. undersaturation. Yellow squares: changing water chemistry at atmospheric pressure and 21°C
(Route 2 in Figure 2.4). Light grey circles: changing pressure with the same fill-bag seawater at 21°C (Route 1 in Figure 2.4),
ΔV=-37.6 cm
3
mol
-1
at 21°C (Ingle, 1975). Dark grey diamonds: changing pressure with the same fill-bag seawater, Ω
recalculated with ΔV=-57.6 cm
3
mol
-1
at 21°C (error bars not shown). Pressures at which dissolution experiments were conducted
(light grey circles) were 350 dbar, 700 dbar, 830 dbar, 1070 dbar, 1100 dbar, 1380 dbar (marked on the plot). Uncertainties in Ω
represent standard deviation of replicates in DIC and alkalinity measurements; uncertainties in dissolution rate represent
goodness of fits in Figure 2.3.

2.3.2 Dissolution rates at constant pressures
To further explore the impact of pressure on dissolution kinetics, five sets of experiments
at different pressures were conducted (Figure 2.6). For a given pressure, I adjusted the water
chemistry so that a range of saturation states could be achieved. Dissolution rates increased for
the same Ω value as pressure increased from 10 dbar to 700 dbar. However, no obvious kinetic
enhancement was seen when pressure was increased beyond 700 dbar up to 2500 dbar. Rate
enhancement from 10 dbar to 350 dbar scaled similarly across a range of Ω from 0.9-0.65 (by a
factor of approximately 1.5). The increase in rate found from 350 dbar to 700 dbar was similar
(also approximately a factor of 1.5), yet, the effect of pressure on dissolution kinetics is not

32
constant with increased pressure. This discontinuity also suggests that the pressure effect on
dissolution rate is not a systematic thermodynamic error but rather a kinetic effect. This pressure
effect enhances dissolution rate up to a certain pressure (approximately 700 dbar), and then
ceases exerting further impact above that pressure, at least up to 2500 dbar.

Figure 2.6 Dissolution rates versus undersaturation at 5 different pressures (Routes 2-6 in Figure 2.4). Dashed lines are fits to
data points of 10 dbar, 350 dbar, 700 dbar, 1050 dbar. Fits to 700 dbar and 1050 dbar overlap such that these data are fitted to the
same curve.

2.3.3 Log-log correlation of dissolution rate vs. undersaturation
The formulation of Eq. 1 lends itself to an examination of values of k and n by plotting
dissolution rates versus undersaturation. I have shown, however, in work by Subhas et al. (2015);
Subhas et al. (2017); and later in this manuscript, that the values of n and k do not provide a
mechanistic framework by which I can examine dissolution kinetics. Nevertheless, I provide this
analysis to allow comparison with previous studies (Keir et al.,1980; Subhas et al., 2015).
The data for a given pressure (Figure 2.7) are linearly correlated in log-log plot at lower
Ω within a limited range (Zone II), but follow a different trend line closer to saturation (Zone I).

33
The zones of different slopes are separated roughly at Ω = 0.90. Slopes of the fitting lines for 10,
350, 700, and 1050 dbar in Zone II (Ω from 0.90 to 0.50) are 3.6±0.2, 3.9±0.5, 3.1±0.3, 3.1±0.1
respectively. The slope of the log-log transformation represents the value of the empirical
reaction order, n (Eq. 1). The calculated n for pressures < 1050 dbar in Zone II indicates that the
observed pressure enhancement of dissolution rate does not significantly nor systematically
change the reaction order.

34
Figure 2.7 (a) Log-log plot of geometry-normalized dissolution rate versus undersaturation state at 10, 350, 700, 1050 dbar.
Dashed lines are fits to data points in Zone II. For reference, log(1- Ω) from -1.5 to -0.3 is equivalent to Ω = 0.97 to 0.50. (b) Log-
log plot of dissolution rate versus undersaturation state. White squares are 10 dbar data from this study; grey diamonds are 10
dbar data from Subhas et al. (2015) using the calcite sample from the same batch, and the same experimental set-ups; black
triangles are 1050 dbar from this study. At near-equilibrium conditions, where dissolution rates are very low (Zone I), error bars
in log scale plot are large. Irrespective, dissolution rates are clearly off the trend lines in Zone II. Linear regression fits are given.

Although dissolution kinetics are enhanced at higher pressures, there are many similarities in
ambient and high-pressure dissolution phenomena. The change of slope (critical Ω) was found
between Zone I and II at both ambient and high pressure. The reaction order n (slopes) for 10
dbar and 1050 dbar between Ω= 0.5 to Ω
NOPQ
were 3.8±0.2, 3.1±0.1 respectively; slopes for Ω >
Ω
NOPQ
were 0.47±0.27 and 0.46±0.15. The critical point for both pressure conditions was within
the error of each other; Ω
NOPQ
is 0.87±0.05 at 10 dbar; and 0.90±0.03 at 1050 dbar. I note that this
Ω
NOPQ
is a different critical undersaturation point for calcite dissolution in seawater from that
reported by Subhas et al. (2017) (Ω=0.7) and other researchers (e.g. Berner and Morse, 1974,
Ω=0.67). This study is the first study to get close enough to equilibrium, because of the high
sensitivity of rate measurements using our
13
C labeling technique, to see a Ω
NOPQ
at 0.87~0.90.
The implicit meaning of the two different Ω
NOPQ
will be discussed in section 4.2.

2.4 DISCUSSION
2.4.1 Thermodynamic uncertainty and Partial Molal Volume
According to our results, a significant difference in dissolution rate vs. undersaturation
exists between varying water chemistry and varying pressure (Figure 2.5). This discrepancy can
either be due to uncertainty in values of undersaturation, or real differences in rate at fixed
saturation state. If the former is true, the degree of undersaturation under pressure may have an
error related to the uncertainty in the Partial Molal Volume, ΔV, of the dissolution reaction. If

35
the latter is true, our results can be interpreted as a real pressure effect on dissolution rate, above
and beyond its effect on thermodynamics.
The effect of pressure on thermodynamic constants can be determined in two ways
(Millero, 1995): (1) using direct measurements of the constants (Culberson and Pytkowicz, 1968)
and (2) using partial molal volume and compressibility data (Millero, 1979, 1983). The two
methods have been shown to be in good agreement (Millero, 1979). The effect of pressure on the
dissociation constants of acids (Ki) can be made from (Millero, 1979) equations of the form:
ÑÖÜ
i
á
à
i
á
â
ä= −ã
åç
á
éè
êë+ã0.5
åï
á
éè
êë
ñ
(4)
where P is the applied pressure in bars, and ΔVi and Δki are the molal volume and
compressibility change for the association or dissociation reactions. The values of ΔVi and Δki
for the ionization of acids have been fit to equations of the form for seawater of S=35:
ΔVi=a0+a1t+a2t
2
; (5)
Δki=b0+b1t+b2t
2
. (6)
where t is temperature in °C. The coefficients in Eq. (5) and (6) for the dissociation of
H2CO3, HCO3
-
, CaCO3 (calcite), CaCO3 (aragonite) are given in Table 2.1 (Culberson and
Pytkowicz, 1968; Ingle, 1975).

Table 2.1 Coefficients for the effect of pressure on the dissociation constants of acids in seawater.
Acid -a0 a1 a2 -b0 b1
H2CO3 25.50 0.1271 0 3.08 0.0877
HCO3
-
15.82 -0.0219 0 -1.13 -0.1475
CaCO3 (calc.) 48.76 0.5304 0 11.76 0.3692
CaCO3 (arag.) 35 0.5304 0 11.76 0.3692

36
The effect of pressure on the solubility of calcite (\
∗
]^
) was measured and reviewed in
Ingle (1975) (Table 2.2). At 25°C, reported ΔVcalc. varied by about 20% (-39.4 to -31.8 cm
3
mol
-
1
, ΔV calc. obtained as pressure was increased from 1 to 1000 atm). Due to the small discrepancies
in ΔVcalc. values reported by different researchers, these ΔVcalc. expressions and values have been
widely used since 1970s. CO2SYS (v1.1, 2011), the program I used to calculate carbonate
system variables, adopts parameters from Ingle (1975) and Millero (1979), and determines
calcite ΔVi with the equation:
ΔVi =-48.76+0.5304t (7)
which yielded -35.5 cm
3
mol
-1
at 25°C, and -37.6 cm
3
mol
-1
at 21°C.

Table 2.2 Apparent partial molal volume changes for calcite dissolution
Substance T (°C) ΔV (cm
3
mol
-1
) Author
Calcite 25 -34.4 Ingle (1975)
Iceland spar 25 -35.5 Ingle (1975)
Calcite 25 -39.4 Millero and Berner (1969)
Calcite 20 -43.4 Duedall (1972)
Foraminifera 22 -30.7 Pytkowicz and Fowler (1967)
Foraminifera 25 -31.8 Pytkowicz and Fowler (1967)
Calcite 2 -42.3 Ingle (1975)
Calcite 2 -43.8 Sayles (1980)
Oolite 2 -31.8 Ingle (1975)
Oolite 2 -33.1 Hawley and Pytkowicz (1969)

* ΔV values in white boxes are under 20-25°C, temperature comparable to this study
(21°C). ΔV values in grey boxes are under 2°C for reference.

37
If the difference between ambient and higher pressure experimental dissolution rates is
due to an inaccurate ΔV calc. value, then an adjustment of ΔVcalc. at 21°C from -37.6 cm
3
mol
-1
to -
57.6 cm
3
mol
-1
is necessary to make the two results agree (Figure 2.5). The uncertainty (±20%)
in partial molar volume based on previous research does not support the almost factor of two
change in ΔVcalc. necessary to align our ambient and high-pressure data sets.
This change in calcite ΔVcalc. is also approximately equivalent to a change of ΔV(H2CO3)
(pressure effect on K1) at 21°C from -22.8 cm
3
mol
-1
to -59.3 cm
3
mol
-1
; or a change of
ΔV(HCO3
-
) (pressure effect on K2) at 21°C from -16.3 cm
3
mol
-1
to 12.0 cm
3
mol
-1
, both of
which seem unreasonable. Therefore, I believe that dissolution under higher pressure is a true
kinetic effect. Previous work by Subhas et al. (2015) yield similar ambient pressure data as I
obtained here, hence I believe the low-pressure work is reproducible and the difference I see
between different pressures is real.

2.4.2 Different dissolution mechanisms and critical degree of undersaturation
Dissolution and crystal growth are usually considered complementary processes, and are
thus linked in terms of mechanisms. The earliest mechanistic studies on mineral dissolution can
be traced back to the well-established BCF theory (Burton et al., 1951; Cabrera et al., 1954;
Cabrera and Levine, 1956). According to the BCF theory, a critical free energy value is required
to open etch pits at line defects such as screw dislocations. Only if this critical value is reached
can pre-existing hollow cores (Frank, 1951) open up into etch pits. Lasaga and Lüttge (2001,
2003) further introduced the stepwave model, explaining that the pit walls are the source for
steps that emanate from the outskirts of the pits and travel across the crystal surface. In contrast,
at near-equilibrium conditions, the difference in free energy is insufficient for hollow cores to

38
open into pits; the critical source for stepwaves is therefore missing. Near equilibrium, the
dissolution mechanism is driven primarily by point defects and dissolution on pre-existing edges
and corners, and advances as step edge retreats, a much slower process than the stepwave
mechanism (Arvidson and Lüttge, 2010).
Dissolution experiments of various minerals have shown discontinuities in the function of
dissolution rate versus the distance from equilibrium. These minerals include carbonates (Berner
and Morse, 1974; Pokrovsky and Schott, 1999; Pokrovsky and Schott, 2001; Teng, 2004;
Pokrovsky et al., 2009; Lüttge and Arvidson, 2010; Xu et al., 2012; Subhas et al., 2017), silicates
(Knauss and Wolery, 1986; Nagy et al., 1991; Burch et al., 1993; Gautier et al., 1994; Ganor et
al., 1995; Oelkers et al., 1994; Devidal et al., 1997; Oelkers and Schott, 1999; Taylor et al., 2000;
Dove et al., 2005; Hellmann and Tisserand, 2006; Arvidson and Lüttge, 2010), and others (Nagy
and Lasaga, 1992; Pokrovsky and Schott, 2004). For calcite dissolution in seawater, Berner and
Morse (1974) related the “near-equilibrium criticality” to the abundance of phosphate ion
absorbed to the mineral surface. They hypothesized that these adsorbed ions prevent dissolution
steps from propagating. Once below a certain threshold saturation state (ΔpH=0.10, Ω=0.67 for
calcite in seawater), monomolecular steps on the crystal surface can penetrate between adsorbed
inhibiting phosphate ions. The presence of the chemical lysocline in the oceans was believed to
be due to this mechanism. Direct observations of crystal surfaces with Vertical Scanning
Inferometry and Atomic Force Microscopy have linked observed rate discontinuities with the
activation of different crystal surface features (Teng, 2004; Beig and Lüttge, 2006; Lüttge, 2006;
Xu et al., 2010; Xu et al., 2012; Schott et al., 2012). Based on AFM observations, Teng (2004)
reported a Ω
NOPQ
= 0.54 for calcite step-edge propagation (opening up of etch pits) in freshwater,
whereas Xu et al. (2012) observed the first appearance of etch pits at Ω = 0.904. This Ω
NOPQ
was

39
correlated with an abrupt change in velocity of obtuse steps instead of acute steps (Xu et al.,
2010, Ω ≈ 0.8). Subhas et al. (2017) reported a Ω
NOPQ
= 0.70~0.75 in seawater, between defect-
assisted dissolution mechanism and the homogeneous spreading of etch pits, by plotting lab
dissolution rate of calcite in seawater within the mechanistic framework proposed by Dove et al.
(2005), as briefly described below.
Our work confirms a discontinuity of dissolution behavior much closer to equilibrium, as
defined by the change of slope in log(rate) vs. log(undersaturation) plots (Figure 2.7). Linear fits
change slope at log(1- Ω) > -0.9, which is equivalent to Ω
NOPQ
of around 0.9. As noted previously,
the influence of pressure does not change the value of Ω
NOPQ
. Below, I investigate the significance
of dissolution rates under ambient and higher pressures by organizing the data, rather than by n
and k, but within the mechanistic framework proposed by Dove et al. (2005). By reorganizing
equations in Dove et al. (2005) to describe dissolution, for homogenous or defect-assisted
dissolution I find:
ÑÖò
|TUVW|
(1−Ω)
ñ
ö|õ|
ú
|
ù= ÑÖÜℎü†
l
(°
ñ
ℎÖ
]
U)
ú
ö
ä−
¢I
ñ
°ℎ
3(X
§
•)
ñ
¶
1
õ
¶
(8)
where Rate is the normalized dissolution velocity (m/s), |õ| = ln (Ω), h is the step height
(nm), ü is the step kinetic coefficient (cm/s), †
l
is the equilibrium concentration of dissolved
species in solution (molecules/m
3
), ° is the molecular volume (cm
3
), Ö
]
is the density of pit
nucleation sites, a is lattice spacing (m), I is the interfacial free energy (mJ/m
2
), X
§
is the
Boltzman Constant, T is absolute temperature (K). As given by this formulation, the dissolution
rate should be a linear function of ©
ú
™
© and the slope should be negative.

40
For the step retreat mechanism, which occurs closer to equilibrium, the dissolution rate
function is not linear with ©
ú
™
©:
ÑÖò
|TUVW|
(1−M)
ñ
ö|õ|
ú
|
ù= ÑÖÜ
°ü†
l
´ℎ
ë
ä+ÑÖ¨(1−M)
ú
ö
¶
1
õ
¶
ú
|
≠−ÑÖ (1+8Ü
°I
ëX
§
•
ä¶
1
õ
¶)
(9)
Therefore, by plotting lab-derived dissolution rate data in this framework and
interrogating its trends, I can demonstrate a shift from the step-retreat dissolution mechanism
near equilibrium to the defect-assisted mechanism farther from equilibrium for calcite
dissolution in seawater (Figure 2.8). In Zone II (©
ú
™
©<9), ÑÖØ
|éRQl |
(ú∞±)
a
f|™|
≤
≥
¥ decreases as ©
ú
™
©
increases, and the relationship is linear, as predicted by Eq. (8). In Zone I (©
ú
™
©>9), data points fall
off the linear trends with negative slopes and ÑÖØ
|éRQl |
(ú∞±)
a
f|™|
≤
≥
¥ increases slightly with increasing
©
ú
™
©, as predicted by Eq. (9). A closer examination of the Zone II data reveals that there may be
two different negative slopes representing homogenous and defect-assisted dissolution
respectively, but our data are not dense enough, especially far from equilibrium (low ©
ú
™
©) to see
this conclusively. Closer to equilibrium, the transition between defect-assisted and step-retreat
dissolution mechanisms (boundary of Zone II and I) occurs at ©
ú
™
© = 8~9, which is equivalent to
M= 0.88~0.89, consistent with Ω
NOPQ
determined in log-log plots (Figure 2.7). Although, the Ω
NOPQ

does not appear to be sensitive to pressure, the slopes at different pressures in Zone II (Figure
2.8) indicate a change of the local interfacial energy barrier, I (Figure 2.9a and Table 2.3). The
energy barrier I decreases as pressure increases, but this decrease appears to level off at ~700

41
dbar. The intercept of the Zone II slope in Figure 2.8 provides information about two other
parameters in Eq. 8. Increasing pressure does not seem to significantly change the intercept
(Figure 2.9b), which means that the density of nucleation sites (Ö
]
) and/or the velocity of step
retreat (ü) are similar.

Figure 2.8 Dissolution velocity (m/s) versus |1/σ| where |σ|= ln (Ω) (framework proposed by Dove et al. (2005)). Higher |1/σ|
value means closer to saturation (Ω = 1). Dissolution velocity are off the linear trend lines in Zone I and start to increase as |1/σ|
increases, indicating a change from step retreat dissolution mechanism (Zone I) to defect-assisted dissolution mechanism (Zone
II). Different slopes of different pressure trend lines in Zone II indicate different local interfacial energy barrier α under different
pressures (Table 3). (a) Dissolution velocity (m/s) versus |1/σ| for 10, 350, 700, 1050, 2500 dbar. (b) Dissolution velocity (m/s)
versus |1/σ| for 10 dbar (including data from Subhas et al. (2015)) and 1050 dbar only, to show a clearer difference of slopes in
Zone II and a clearer change of mechanism at around |1/σ|=9, equivalent to Ω = 0.89. The trend lines in Zone 1 are schematic
fits to the data.

42

Figure 2.9 (a) Surface energy barrier I (calculated from slopes in Figure 8) versus different pressure. The trend line is a
schematic to emphasize the apparent change in slope in I vs. pressure; (b) Change in the intercepts (proportional to kinetic
parameters ü and/or Ö
]
) versus different pressure for defect-assisted dissolution (Zone II).

Table 2.3 Fits to Figure 2.8 & 2.9 and calculated local interfacial energy barrier I (mJ/m
2
) under different pressure
Pressure
(dbar)
Defect-Assisted Dissolution
Note Intercept
ln (ℎü†
l
(°
ñ
ℎÖ
]
U)
≤
f
)
Slope
−
¢I
ñ
°ℎ
3(X
§
•)
ñ

Energy Barrier
I (mJ m
-2
)
10 -21.5±0.2 -0.67±0.05 24±1
10 -21.7±0.3 -0.69±0.07 24±1 *
350 -21.8±0.4 -0.48±0.07 20±1
700 -20.9±0.4 -0.54±0.09 22±1
1050 -21.5±0.2 -0.42±0.03 19±1
2500 -21.5±0.1 -0.39±0.02 18±1

* Including data from Subhas et al. (2015) (revised).

These results suggest that increased pressure may lower the interfacial energy barrier at
the calcite-seawater interface, resulting in enhanced dissolution kinetics. This effect may result
from pressure-dependent speciation changes in seawater surrounding the mineral surface, and/or
changes in bonding energy between ions in seawater and the mineral surface. The changes in
speciation and/or bonding energy take place as pressure increases from ambient pressure to 700

43
dbar, but become complete at ~700 dbar. Conversely, the number of nucleation sites for
dissolution and/or the velocity of step retreat do not appear to change with pressure. How exactly
pressure affects the surface energy of calcite is still an open question, yet a deeper understanding
of calcite dissolution kinetics will require delving into this particular phenomenon.
Combined with the results from Subhas et al. (2017), I report distinct and robust
mechanism transitions for calcite dissolution in seawater from step-retreat to defect-assisted
dissolution, and further to an etch pit spreading mechanism. The two M
NOPQ
values were shown to
be ~0.9 for the former transition, and ~0.75 for the latter transition. Our
13
C labeling method has
allowed us high enough sensitivity at slow dissolution rates to determine the M
NOPQ
between the
step retreat mechanism and the defect-assisted mechanism, and both mechanisms are in evidence
at ambient pressure and as pressure increases.

2.4.3 Effect of pressure on shallow water column carbonate dissolution
Although estimates of carbonate production in the surface, open ocean vary (0.4-2 Gt/yr)
(e.g. Milliman, 1993; Milliman, 1999; Balch and Kilpatrick, 1996; Balch et al., 2007; Dunne et
al., 2007; Dunne et al., 2012 etc.), it has been recognized that carbonate production in the ocean
well exceeds the amount of carbonate burial (Sarmiento and Gruber, 2006; Berelson et al., 2007).
Battaglia et al. (2015) reported median global CaCO3 export to be 0.82 Gt PIC/yr (within the
lower half of previously published estimates (0.4-1.8 Gt PIC/yr), where PIC is Particulate
Inorganic Carbon); and that only 29% reaches the seafloor. These studies of global carbonate
flux have set up a problem in global geochemistry which has yet to be resolved—how much
carbonate is dissolving in the open ocean water column and what factors control the dissolution
rate.

44
Water column shallow-depth CaCO3 dissolution (SDCCD) has been recognized by
several independent studies (e.g. Anderson and Sarmiento, 1994; Lohmann, 1995; Milliman and
Droxler, 1996; Milliman et al., 1999; Chen et al., 2002; Schiebel, 2002; Gangstø et al., 2008),
and could account for some of the discrepancy between carbonate production and sediment-trap
flux estimates mentioned above. Conventional wisdom is that calcium carbonate shells are
dissolved when they reach the saturation horizon of aragonite or calcite during sinking, which
provides the alkalinity that can be supplied back to the surface ocean via upwelling/diapycnal
mixing. However, researchers have shown significant dissolution above the “saturation horizons”
(Feely et al., 2002; Gangstø et al., 2008). The impact of the SDCCD finding is that it implies a
shorter time scale for the cycling of calcium carbonate within the ocean. This would have
implications for feedbacks between climate, atmospheric CO2, and the marine carbon cycle.
Possible mechanisms responsible for SDCCD were proposed to include (1) dissolution of
CaCO3 particles in the guts of zooplankton (Takahashi, 1975; Bishop et al., 1980, 1986; Van der
Wal et al., 1995; White et al., 2016), even though others provide some skepticism regarding this
mechanism (Harris, 1994; Pond et al., 1995); (2) dissolution of CaCO3 particles in
microenvironments where bacterial oxidation of organic matter can enhance the dissolution
process (Jansen et al., 2002); (3) dissolution of the more soluble forms of CaCO3 in shallow
waters, including aragonite and high-Mg calcites (Byrne et al., 1984; Morse and Mackenzie,
1990); (4) the source of excess alkalinity for subsurface waters in the open ocean is from the
decomposition of organic matter occurring in shelf sediments (Chen, 2002; Burt et al., 2016), or
transport/mixing subsequent to dissolution (Friis et al., 2006, 2007), and hence there is not
evidence of SDCCD.

45
Whereas the pressure effect on inorganic calcite dissolution described in this chapter is
beyond doubt, the direct application of the rate law to illustrate carbonate dissolution patterns in
the ocean remains to be tested. Sinking calcite particles are “ballasted” within fecal pellets,
“snow” and other aggregates (Honjo, 1980, 1995; Honjo et al., 1982; Armstrong et al., 2001;
Klaas and Archer, 2002), which will complicate the process of dissolution while sinking by (1)
enhancing settling velocity by orders of magnitude (Armstrong et al., 2009), (2) lowering PIC
reactivity by providing physical protection (Armstrong et al., 2001), (3) providing locally acidic
environments through bacterial oxidation of organic matter (Jansen et al., 2002), etc. Rates of
carbonate dissolution at the seafloor, on the other hand, are further complicated by boundary
layer processes between sediments and the overlying seawater (Boudreau, 1982, 2001, 2013;
Sulpis et al., 2017). Nevertheless, the revised dissolution rate formulation as a function of
pressure should be taken into consideration in future marine carbon cycle studies. It will also be
interesting to investigate the effect of pressure on other minerals, including the more soluble
forms of CaCO3 (e.g. aragonite and high-Mg calcite). I believe that previous water column
dissolution rate measurements of CaCO3 (calcite and possibly other carbonate minerals) were
underestimated, and this pressure effect can account, at least partially, for the observed ocean
production-burial discrepancy.

2.5 CONCLUSIONS
I have presented experiments demonstrating elevated dissolution rates of calcite in
seawater under high pressure compared to equivalent undersaturation states under ambient
pressure. This enhancement was observed over an Ω range of 0.65 to 1. The discrepancy of
dissolution behavior under different pressures is greater than the uncertainty of Ω based on

46
uncertainties in DV, and is therefore attributed to a pressure effect on the dissolution rate itself.
This kinetic pressure enhancement effect takes place relatively evenly between ambient pressure
to 700 dbar, but ceases at pressures above 700 dbar (pressures higher than 700 dbar yield
dissolution rates similar to rates at 700 dbar, at least up to 2500 dbar). The calcite dissolution
rates in seawater at 700 dbar is 2-4 times faster than at ambient pressure. Dissolution rates at
different pressures share the same reaction order n in Rate=k(1- Ω)
n
. Furthermore, a discontinuity
in the functional relationship between dissolution rate and undersaturation indicates a change of
the rate-controlling dissolution mechanism at values of Ω
∑∏π∫π∑ªº
. I interpret this Ω
∑∏π∫π∑ªº
to be the
threshold between a step-retreat dissolution mechanism and a defect-assisted mechanism.
Ω
∑∏π∫π∑ªº
at 10 dbar and 1050 dbar are similar, 0.87±0.05 and 0.90±0.03 respectively. The reaction
order n above Ω
∑∏π∫π∑ªº
is ~0.5, and ~3.4 below Ω
∑∏π∫π∑ªº
. By transforming dissolution rate vs.
saturation state into an equation (Eq. 8) that includes mechanistic information, I find that the
effect of pressure is to lower the surface energy barrier to dissolution. The enhanced pressure-
related dissolution of carbonate minerals may explain excess alkalinity distributions in the ocean
as this would effectively decrease the depth at which rapid dissolution occurs.

ACKNOWLEDGEMENTS

I would like to acknowledge the work by USC machine shop machinists (Don Wiggins
and colleagues) who built the pressure vessel and undergraduate student Laura Morine for her
help running alkalinities. I also thank Dr. Alfonso Mucci, Dr. Bernard Boudreau, and two
anonymous reviewers for their valuable and constructive comments on the manuscript submitted
to GCA.

47
REFERENCES

Acker, J. G., Byrne, R. H., Ben-Yaakov, S., Feely, R. A., & Betzer, P. R. (1987) The effect of
pressure on aragonite dissolution rates in seawater. Geochimica et Cosmochimica
Acta, 51(8), 2171-2175.
Anderson, L. A. and Sarmiento, J. L. (1994) Redfield ratios of remineralization determined by
nutrient data analysis, Global Biogeochem. Cycles, 8, 65–80.
Armstrong, R. A., Lee, C., Hedges, J. I., Honjo, S., & Wakeham, S. G. (2001). A new,
mechanistic model for organic carbon fluxes in the ocean based on the quantitative
association of POC with ballast minerals. Deep Sea Research Part II: Topical Studies in
Oceanography, 49(1-3), 219-236.
Armstrong, R. A., Peterson, M. L., Lee, C., & Wakeham, S. G. (2009). Settling velocity spectra
and the ballast ratio hypothesis. Deep Sea Research Part II: Topical Studies in
Oceanography, 56(18), 1470-1478.
Arvidson, R. S., & Lüttge, A. (2010) Mineral dissolution kinetics as a function of distance from
equilibrium–New experimental results. Chemical geology, 269(1), 79-88.
Balch, W. M., and K. A. Kilpatrick (1996) Calcification rates in the equatorial Pacific along
140_W, Deep Sea Res., Part II, 43, 971– 993.
Balch, W. M., D. T. Drapeau, B. C. Bowler, and E. Booth (2007) Prediction of pelagic
calcification rates using satellite-measurements, Deep Sea Res., Part II, Vol. 54, Issues 5-
7, Pages 478-495, doi: 10.1016/j.dsr2.2006.12.006
Battaglia, G., Steinacher, M., & Joos, F. (2015) A probabilistic assessment of calcium carbonate
export and dissolution in the modern ocean. Biogeosciences Discussions, 12(24).
Beig, M. S., & Lüttge, A. (2006) Albite dissolution kinetics as a function of distance from
equilibrium: Implications for natural feldspar weathering. Geochimica et Cosmochimica
Acta, 70(6), 1402-1420.
Berelson W.M., Balch W.M., Najjar R., Feely R.A., Sabine C., Lee K. (2007) Relating estimates
of CaCO3 production, export, and dissolution in the water column to measurements of
CaCO3 rain into sediment traps and dissolution on the sea floor: A revised global
carbonate budget. Global Biogeochem. Cycles, Vol. 20, GB1024,
doi:10.1029/2006GB002803.
Berger, W. H. (1967) Foraminiferal ooze: solution at depths. Science, 156(3773), 383-385.
Berner, R. A., & Morse, J. W. (1974) Dissolution kinetics of calcium carbonate in sea water; IV,
Theory of calcite dissolution. American Journal of Science, 274(2), 108-134.
Bishop, J. K. B., J. C. Stepien, and P. H. Weibe (1986) Particulate matter distributions, chemistry
and fluxes in the Panama Basin: Response to environmental forcing, Prog. Oceanogr.,
17, 1– 59.
Bishop, J. K. B., R. W. Collier, D. R. Kettens, and J. M. Edmond (1980) The chemistry, biology
and vertical flux of particulate matter from the upper 1500 m of the Panama Basin, Deep
Sea Res., Part I, 27, 615– 640.
Boudreau, B. P. (1982). The influence of a diffusive sublayer on accretion, dissoluton, and
diagenesis at the sea floor. The dynamic environment of the ocean floor.
Boudreau, B. P. (2001). Solute transport above the sediment-water interface. The Benthic
boundary layer: Transport processes and biogeochemistry, 104-126.
Boudreau, B. P. (2013). Carbonate dissolution rates at the deep ocean floor. Geophysical
Research Letters, 40(4), 744-748.

48
Burch, T. E., Nagy, K. L., & Lasaga, A. C. (1993) Free energy dependence of albite dissolution
kinetics at 80 C and pH 8.8. Chemical Geology, 105(1-3), 137-162.
Burt, W. J., Thomas, H., Hagens, M., Pätsch, J., Clargo, N. M., Salt, L. A., ... & Böttcher, M. E.
(2016). Carbon sources in the North Sea evaluated by means of radium and stable carbon
isotope tracers. Limnology and Oceanography, 61(2), 666-683.
Burton, W. K., Cabrera, N., & Frank, F. C. (1951) The growth of crystals and the equilibrium
structure of their surfaces. Philosophical Transactions of the Royal Society of London A:
Mathematical, Physical and Engineering Sciences, 243(866), 299-358.
Byrne, R. H., J. G. Acker, P. R. Betzer, R. A. Feely, and M. H. Cates (1984) Water column
dissolution of aragonite in the Pacific Ocean, Nature, 312, 321– 326.
Cabrera, N., & Levine, M. M. (1956) XLV. On the dislocation theory of evaporation of
crystals. Philosophical Magazine, 1(5), 450-458.
Cabrera, N., Levine, M. M., & Plaskett, J. S. (1954) Hollow dislocations and etch pits. Physical
Review, 96(4), 1153.
Chen, C.-T. A. (2002) Shelf-vs. dissolution-generated alkalinity above the chemical lysocline,
Deep Sea Res., Part II, 49, 5365– 5375.
Culberson, C., & Pytkowicx, R. M. (1968) Effect of pressure on carbonic acid, boric acid, and
the pH 1N seawater. Limnology.
Devidal, J. L., Schott, J., & Dandurand, J. L. (1997) An experimental study of kaolinite
dissolution and precipitation kinetics as a function of chemical affinity and solution
composition at 150 C, 40 bars, and pH 2, 6.8, and 7.8. Geochimica et Cosmochimica
Acta, 61(24), 5165-5186.
Dickson, A.G., Millero, F.J. (1987) A comparison of the equilibrium constants for the
dissociation of carbonic acid in seawater media. Deep Sea Research Part A,
Oceanographic Research Papers. 34, 1733-1743.
Dickson A. G., Wesolowski D. J., Palmer D. A. and Mesmer R. E. (1990) Dissociation constant
of bisulfate ion in aqueous sodium chloride solutions to 250 _C. J. Phys. Chem. 94(20),
7978–7985.
Dove, P. M., Han, N., & De Yoreo, J. J. (2005) Mechanisms of classical crystal growth theory
explain quartz and silicate dissolution behavior. Proceedings of the National Academy of
Sciences, 102(43), 15357-15362.
Duedall, I. W. (1972) The partial molal volume of calcium carbonate in sea water. Geochimica et
Cosmochimica Acta, 36(7), 729-734.
Dunne, J. P., B. Hales, and J. R. Toggweiler (2012) Global calcite cycling constrained by
sediment preservation controls, Global Biogeochem. Cycles, 26, GB3023,
doi:10.1029/2010GB003935.
Dunne, J. P., J. L. Sarmiento, and A. Gnanadesikan (2007) A synthesis of global particle export
from the surface ocean and cycling through the ocean interior and on the seafloor, Global
Biogeochem. Cycles, 21, GB4006, doi:10.1029/2006GB002907.
Feely R.A., Sabine C.L., Lee K., Millero F.J., Lamb M.F., Greeley D., Bullister J.L., Key R.M.,
Peng T.-H., Kozyr A., Ono T., Wong C.S. (2002) In situ calcium carbonate dissolution in
the Pacific Ocean. Global Biogeochemical Cycles. Vol. 16, No. 4, 1144,
doi:10.1029/2002GB001866, 2002.
Frank, F. C. (1951) Capillary equilibria of dislocated crystals. Acta Crystallographica, 4(6), 497-
501.

49
Friis K., Najjar R.G., Follows M.J., Dutkiewicz S. (2006) Possible overestimation of shallow-
depth calcium carbonate dissolution in the ocean. Global Biogeochem. Cycles, 20,
GB4019, doi:10.1029/2006GB002727.
Friis, K., R. G. Najjar, M. J. Follows, S. Dutkiewicz, A. Kortzinger, and K. M. Johnson (2007)
Dissolution of calcium carbonate: Observations and model results in the North Atlantic,
Biogeosciences, 4. pp. 205-213. DOI 10.5194/bg-4-205-2007.
Fukuhara, T., Tanaka, Y., Ioka, N., & Nishimura, A. (2008) An in situ experiment of calcium
carbonate dissolution in the central Pacific Ocean. International Journal of Greenhouse
Gas Control, 2(1), 78-88.
Gangstø, R., Gehlen, M., Schneider, B., Bopp, L., Aumont, O., & Joos, F. (2008) Modeling the
marine aragonite cycle: changes under rising carbon dioxide and its role in shallow water
CaCO3 dissolution. Biogeosciences,5(4), 1057-1072.
Ganor, J., Mogollón, J. L., & Lasaga, A. C. (1995) The effect of pH on kaolinite dissolution rates
and on activation energy. Geochimica et Cosmochimica Acta, 59(6), 1037-1052.
Gautier, J. M., Oelkers, E. H., & Schott, J. (1994) Experimental study of K-feldspar dissolution
rates as a function of chemical affinity at 150 C and pH 9. Geochimica et Cosmochimica
Acta, 58(21), 4549-4560.
Gehlen, M., Bassinot, F. C., Chou, L., & McCorkle, D. (2005) Reassessing the dissolution of
marine carbonates: II. Reaction kinetics. Deep Sea Research Part I: Oceanographic
Research Papers, 52(8), 1461-1476.
Hales B. and Emerson S. (1997) Evidence in support of first-order dissolution kinetics of calcite
in seawater. Earth Planet. Sci. Lett. 148(1–2), 317–327.
Harris, R. P. (1994) Zooplankton grazing on the coccolithophore Emiliani huxleyi and its role in
the inorganic carbon flux, Mar. Biol., 119, 431–439.
Hawley, J., & Pytkowicz, R. M. (1969) Solubility of calcium carbonate in seawater at high
pressures and 2 C. Geochimica et cosmochimica acta,33(12), 1557-1561.
Hellmann, R., & Tisserand, D. (2006) Dissolution kinetics as a function of the Gibbs free energy
of reaction: An experimental study based on albite feldspar. Geochimica et
Cosmochimica Acta, 70(2), 364-383.
Honjo, S. (1980). Material fluxes and modes of sedimentation in the mesopelagic and
bathypelagic zones. J. mar. Res., 38, 53-97.
Honjo, S., & Erez, J. (1978) Dissolution rates of calcium carbonate in the deep ocean; an in-situ
experiment in the North Atlantic Ocean. Earth and Planetary Science Letters, 40(2), 287-
300.
Honjo, S., Dymond, J., Collier, R., & Manganini, S. J. (1995). Export production of particles to
the interior of the equatorial Pacific Ocean during the 1992 EqPac experiment. Deep Sea
Research Part II: Topical Studies in Oceanography, 42(2-3), 831-870.
Honjo, S., Manganini, S. J., & Cole, J. J. (1982). Sedimentation of biogenic matter in the deep
ocean. Deep Sea Research Part A. Oceanographic Research Papers, 29(5), 609-625.
Ingle, S. E. (1975) Solubility of calcite in the ocean. Marine Chemistry, 3(4), 301-319.
Jansen H., Zeebe R.E., Wolf-Gladrow D.A. (2002) Modeling the dissolution of settling CaCO3 in
the ocean. Global Biogeochemical Cycles, Vol. 16, No. 2, 1027, 10.1029/2000GB001279,
2002.
Keir, R. S. (1980) The dissolution kinetics of biogenic calcium carbonates in
seawater. Geochimica et Cosmochimica Acta, 44(2), 241-252.

50
Klaas, C., & Archer, D. E. (2002). Association of sinking organic matter with various types of
mineral ballast in the deep sea: Implications for the rain ratio. Global Biogeochemical
Cycles, 16(4).
Knauss, K. G., & Wolery, T. J. (1986) Dependence of albite dissolution kinetics on pH and time
at 25 C and 70 C. Geochimica et Cosmochimica Acta, 50(11), 2481-2497.
Lasaga, A. C., & Lüttge, A. (2001) Variation of crystal dissolution rate based on a dissolution
stepwave model. Science, 291(5512), 2400-2404.
Lasaga, A. C., & Lüttge, A. (2003) A model for crystal dissolution. European Journal of
Mineralogy, 15(4), 603-615.
Lohmann, G. P. (1995) A model for variation in the chemistry of planktonic foraminifera due to
secondary calcification and selective dissolution, Paleoceanography, 10, 445–457.
Lüttge, A. (2006) Crystal dissolution kinetics and Gibbs free energy. Journal of Electron
Spectroscopy and Related Phenomena, 150(2), 248-259.
Lüttge, A., & Arvidson, R. S. (2010). Reactions at surfaces: A new approach integrating
interferometry and kinetic simulations. Journal of the American Ceramic Society, 93(11),
3519-3530.
Mehrbach, C., Culberson, C., Hawley J.E., Pytkowicz R.M. (1973) Measurement of the apparent
dissociation constants of carbonic acid in seawater at atmospheric pressure. Limnology
and Oceanography, 18, 897–907.
Metzler, C. V., Wenkam, C. R., & Berger, W. H. (1982) Dissolution of foraminifera in the
eastern equatorial Pacific; an in situ experiment. Journal of Foraminiferal
Research, 12(4), 362-368.
Millero, F. J. (1979) The thermodynamics of the carbonate system in seawater. Geochimica et
Cosmochimica Acta, 43(10), 1651-1661.
Millero, F. J. (1983) The estimation of the pK HA of acids in seawater using the Pitzer
equations. Geochimica et Cosmochimica Acta, 47(12), 2121-2129.
Millero, F. J. (1995) Thermodynamics of the carbon dioxide system in the oceans. Geochimica et
Cosmochimica Acta, 59(4), 661-677.
Millero, F. J., & Berner, R. A. (1972) Effect of pressure on carbonate equilibria in
seawater. Geochimica et Cosmochimica Acta, 36(1), 92-98.
Milliman, J. D. (1978) Dissolution of calcium carbonate in the Sargasso Sea (Northwest
Atlantic). Woods Hole Oceanographic Institution.
Milliman, J. D. (1993) Production and accumulation of calcium carbonate in the ocean: Budget
of a nonsteady state, Global Biogeochem. Cycles, 7, 927– 957.
Milliman, J. D. and Droxler, A. W. (1996) Neritic and pelagic carbonate sedimentation in the
marine environment: Ignorance is not bliss, Geologische Rundschau, 85(3), 496–504.
Milliman, J. D., P. J. Troy, W. M. Balch, A. K. Adams, Y.-H. Li, and F. T. MacKenzie (1999)
Biologically-mediated dissolution of calcium carbonate above the chemical lysocline?
Deep Sea Res., Part I, 46, 1653–1669.
Morse, J. W. (1978) Dissolution kinetics of calcium carbonate in sea water: VI. The near-
equilibrium dissolution kinetics of calcium carbonate-rich deep sea sediments. Am. J.
Sci.;(United States), 278(3).
Morse, J. W., and F. T. Mackenzie (1990) Geochemistry of sedimentary carbonates, 707 pp.,
Elsevier Sci., New York.

51
Morse, J. W., de Kanel, J. O. H. N., & Harris, K. A. R. E. N. (1979) Dissolution kinetics of
calcium carbonate in seawater; VII, The dissolution kinetics of synthetic aragonite and
pteropod tests. American Journal of Science, 279(5), 488-502.
Mucci, A. (1983) The solubility of calcite and aragonite in seawater at various salinities,
temperatures, and one atmosphere total pressure. American Journal of Science, 283(7),
780-799.
Nagy, K. L., & Lasaga, A. C. (1992) Dissolution and precipitation kinetics of gibbsite at 80 C
and pH 3: The dependence on solution saturation state. Geochimica et Cosmochimica
Acta, 56(8), 3093-3111.
Nagy, K. L., Blum, A. E., & Lasaga, A. C. (1991) Dissolution and precipitation kinetics of
kaolinite at 80 degrees C and pH 3; the dependence on solution saturation state. American
Journal of Science, 291(7), 649-686.
Oelkers, E. H., & Schott, J. (1999) Experimental study of kyanite dissolution rates as a function
of chemical affinity and solution composition. Geochimica et Cosmochimica Acta, 63(6),
785-797.
Oelkers, E. H., Schott, J., & Devidal, J. L. (1994) The effect of aluminum, pH, and chemical
affinity on the rates of aluminosilicate dissolution reactions. Geochimica et
Cosmochimica Acta, 58(9), 2011-2024.
Peterson, M. N. A. (1966) Calcite: rates of dissolution in a vertical profile in the central
Pacific. Science, 154(3756), 1542-1544.
Pokrovsky, O. S., & Schott, J. (1999) Processes at the magnesium-bearing carbonates/solution
interface. II. kinetics and mechanism of magnesite dissolution. Geochimica et
cosmochimica acta, 63(6), 881-897.
Pokrovsky, O. S., & Schott, J. (2001) Kinetics and mechanism of dolomite dissolution in neutral
to alkaline solutions revisited. American Journal of Science, 301(7), 597-626.
Pokrovsky, O. S., & Schott, J. (2004) Experimental study of brucite dissolution and precipitation
in aqueous solutions: surface speciation and chemical affinity control. Geochimica et
Cosmochimica Acta, 68(1), 31-45.
Pokrovsky, O. S., Golubev, S. V., Schott, J., & Castillo, A. (2009) Calcite, dolomite and
magnesite dissolution kinetics in aqueous solutions at acid to circumneutral pH, 25 to 150
C and 1 to 55 atm pCO2: New constraints on CO2 sequestration in sedimentary
basins. Chemical geology, 265(1), 20-32.
Pond, D. W., R. P. Harris, and C. A. Brownlee (1995) Microinjection technique using a pH
sensitive dye to determine the gut pH of Calanus helgolandicus. Mar. Biol., 123, 75–79.
Pytkowicz, R. M., & Fowler, G. A. (1967) Solubility of foraminifera in seawater at high
pressures. Geochemical Journal, 1(4), 169-182.
Robert-Baldo, G. L., Morris, M. J., & Byrne, R. H. (1985). Spectrophotometric determination of
seawater pH using phenol red. Analytical Chemistry, 57(13), 2564-2567.
Sarmiento, J.L., Gruber, N. (2006) Ocean Biogeochemical Dynamics. Princeton: Princeton
University Press.
Sayles F. L. (1980) The solubility of CaCO3 in seawater at 2 _C based upon in-situ sampled pore
water composition. Mar. Chem. 9(4), 223–235.
Schiebel, R. (2002) Planktonic foraminiferal sedimentation and the marine calcite budget, Global
Biogeochem. Cycles, 16, 1065, doi:10.1029/2001GB001459.

52
Schott, J., Oelkers, E. H., Bénézeth, P., Goddéris, Y., & François, L. (2012) Can accurate kinetic
laws be created to describe chemical weathering? Comptes Rendus Geoscience, 344(11),
568-585.
Subhas, A. V., Adkins, J. F., Rollins, N. E., Naviaux, J., Erez, J., & Berelson, W. M. (2017)
Catalysis and chemical mechanisms of calcite dissolution in seawater. Proceedings of the
National Academy of Sciences, 114(31), 8175-8180.
Subhas, A. V., Rollins, N. E., Berelson, W. M., Dong, S., Erez, J., & Adkins, J. F. (2015) A
novel determination of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 170, 51-68.
Sulpis, O., Lix, C., Mucci, A., & Boudreau, B. P. (2017). Calcite dissolution kinetics at the
sediment-water interface in natural seawater. Marine Chemistry, 195, 70-83.
Takahashi, T. (1975) Carbonate chemistry of seawater and the calcite compensation depth in the
oceans: Special publication of the Cushman Foundation, Foraminiferal Res., 13, 11– 126,
1975.
Taylor, A. S., Blum, J. D., & Lasaga, A. C. (2000) The dependence of labradorite dissolution and
Sr isotope release rates on solution saturation state. Geochimica et Cosmochimica
Acta, 64(14), 2389-2400.
Teng, H. H. (2004) Controls by saturation state on etch pit formation during calcite
dissolution. Geochimica et Cosmochimica Acta, 68(2), 253-262.
Thunell, R. C., Keir, R. S., & Honjo, S. (1981) Calcite dissolution: An in situ study in the
Panama Basin. Science, 212(4495), 659-661.
Van der Wal, P., R. S. Kempers, and M. J. W. Veldhuis (1995) Production and downward flux of
organic matter and calcite in a North Sea bloom of the coccolithophore Emiliania
huxleyi. Mar. Ecol. Prog. Ser., 126, 247– 265.
Walter, L. M., & Morse, J. W. (1985) The dissolution kinetics of shallow marine carbonates in
seawater: a laboratory study. Geochimica et Cosmochimica Acta, 49(7), 1503-1513.
White, M. M. et al. (2016) Calcium Carbonate Dissolution Above the Lysocline: Implications of
Copepod Grazing on Coccolithophores. Ocean Sciences Meeting Abstract.
Xu, J., Fan, C., & Teng, H. H. (2012) Calcite dissolution kinetics in view of Gibbs free energy,
dislocation density, and pCO2. Chemical geology, 322, 11-18.
Xu, M., Hu, X., Knauss, K. G., & Higgins, S. R. (2010). Dissolution kinetics of calcite at 50–70°
C: An atomic force microscopic study under near-equilibrium conditions. Geochimica et
Cosmochimica Acta, 74(15), 4285-4297.

53
Appendix. “Ex situ” and “in situ” rates for dissolution experiments in the pressure
case.
Two possible reasons may lead to a discrepancy between ex situ and in situ rates. The
first is an inadequate rinsing volume before sampling. If remaining seawater in the sample tubing
(~5 mL) from the previous sampling time is not fully replaced before taking new samples, there
will be a mixing between new sample and old sample, and thus leading to lower in situ
dissolution rate. This factor can be ruled out if the ex situ rate is consistent with the in situ rate
(Figure A1a & b); or replicate in situ samples (8 mL each) are in good agreement with each other
(Figure A1c). When replicate in situ samples are not in agreement (the second sample of the
replicate always has higher d
13
C than the first sample, as shown in Figure A1d), then only the
second sample is used to calculate dissolution rate to reduce rate error. Another reason that
causes biased in situ rate data is contamination in the sample tubing and valves. Even though the
whole sampling line was washed with deionized water and dried by blowing air through the line
after each experiment, contamination may still exist from an improperly cleaned check valve,
13
C
residual from previous experiments (especially for a slow dissolution experiment following a far-
from-equilibrium dissolution experiment), carbonate precipitation in the tubing, and oil intrusion.
These kinds of contamination will lead to a lower R
2
in the linear fitting of in situ samples
(Figure A1e). Both scenarios, however, have no impact on ex situ rates. Therefore, ex situ rates
are always considered as the “real” rates. Ex situ rate uncertainty is based on the difference
between the steepest slope and the shallowest slope of the initial and end ex situ samples, usually
2 initial samples + 4 end samples, or 3 initial samples + 3 end samples. In situ rate uncertainty is
based on the goodness of fit in the linear fitting of in situ samples calculated with the LINEST
function in Microsoft Excel. Ex situ uncertainties are usually smaller than in situ uncertainties
because I am deriving a rate from two points. Therefore, even though “real” rates are considered
to be ex situ rates, errors are taken from in situ errors if there’s not an obvious contamination
observed. The average ratio of ex situ rate / in situ rate is 1.097 for all experiments with both
rates available. When the experiments with in situ contamination are excluded, the average ratio
is 1.031, with ex situ rates 3.1% higher than in situ rates, possibly still due to inadequate rinsing.

54

Figure A1 Moles of labeled carbonate dissolved based on d
13
C measurement from 6 different experiments. White squares are ex
situ samples; black circles are in situ samples. Slopes of trend lines represent dissolution rates. (a) P45 (also shown in Figure 2.3).
Trend line is fitting to ex situ samples. (b) P19. Trend line is fitting to ex situ samples. Ex situ rates in (a) and (b) are in agreement
with in situ rates. (c) P4. Only in situ samples were taken. However, replicate in situ samples agree with each other, thus
disproving the inadequate rinsing problem. Trend line is fitting to all in situ samples. (d) P5. Only in situ samples were taken and
replicate in situ samples were not in good agreement, with the second sample always higher than the first. So only the second
sample of each replicate was used to calculate dissolution rate (trend line is fitting to the second sample of each replicate). (e)
P33. Contamination in the sample tubing/valves was indicated by elevated and nonlinear in situ signals. Indeed, the previous
experiment P32 was far from equilibrium and must have left very high d
13
C contamination in the sample tubing despite DI water
washing. Ex situ samples include 2 initial samples and 4 end samples, and are in perfect agreement with each other (replicate
value uncertainties are less than the size of the data points). Trend line is fitting to ex situ samples. (f) P46. Only ex situ samples
were taken, including 3 initial samples and 3 end samples. Replicate value uncertainties are less than the size of the data points.
Trend line is fitting to all ex situ samples.

55
CHAPTER 3: KINETICS OF ARAGONITE DISSOLUTION IN
SEAWATER AND ITS ROLE IN SHALLOW WATER COLUMN
DISSOLUTION IN THE NORTH PACIFIC

This chapter was published in 2019 as:
Dong, S., Berelson, W. M., Rollins, N. E., Subhas, A. V., Naviaux, J. D., Celestian, A. J., Liu,
X., Turaga, N., Kemnitz, N. J., Byrne, R. H., & Adkins, J. F. (2019). Aragonite dissolution
kinetics and calcite/aragonite ratios in sinking and suspended particles in the North Pacific. Earth
and Planetary Science Letters, 515, 1-12.

ABSTRACT
The lack of consensus on CaCO3 dissolution rates and calcite to aragonite production and
export ratios in the ocean poses a significant barrier for the construction of global carbon
budgets. I present here a comparison of aragonite dissolution rates measured in the lab vs. in situ
along a transect between Hawaii and Alaska using a
13
C labeling technique. Our results show a
general agreement of aragonite dissolution rates in the lab versus in the field, and demonstrate
that aragonite, like calcite, shows a non-linear response of dissolution rate as a function of
saturation state (Ω). Total carbon fluxes along the N. Pacific transect in August 2017, as
determined using sediment traps, account for 11 ~ 23 weight % of total mass fluxes in the upper
200m, with a PIC (particulate inorganic carbon) /POC (particulate organic carbon) mole ratio of
0.2 ~ 0.6. A comparison of fluxes at depths of 100 m and 200 m indicates that 30~60% PIC
dissolves between these depths with 20~70% attenuation in POC fluxes. The molar ratio of PIC
to POC loss is 0.29. The simultaneous loss of PIC and POC in the upper 200 m potentially
indicates PIC dissolution driven by organic matter respiration, or metazoan/zooplankton
consumption. The calcite/aragonite ratio in trap material is significantly lower in the subtropical

56
gyre than in the subarctic gyre. Aragonite fluxes vary from 0.07 to 0.38 mmol m
-2
day
-1
at 100 m,
and 0.06 to 0.24 mmol m
-2
day
-1
at 200 m along the North Pacific transect, with no specific trend
over latitude. The identification of suspended PIC mineral phases by Raman spectroscopy shows
the presence of aragonite below 3000 m in the subtropical gyre, but none in the subpolar gyre.
These multiple lines of evidence suggest that predictions based on a strictly thermodynamic view
of aragonite dissolution, combined with measured aragonite fluxes, underestimate observed
alkalinity excess and measured PIC attenuation in sinking particles. Our measured aragonite flux
combined with our inorganic dissolution rate only account for 9% and 0.2% of the excess
alkalinity observed in the North Pacific (Feely et al., 2004), assuming aragonite sinking rates of 1
m day
-1
and 100 m day
-1
, respectively. However, respiration-driven dissolution or
metazoan/zooplankton consumption, indicated by the simultaneous attenuation of PIC and POC
in sediment traps, is able to generate the magnitude of dissolution suggested by observed excess
alkalinity.

3.1 INTRODUCTION
The size of the total carbon inventory and its distribution within the ocean has a major
influence on the partitioning of carbon between the ocean and the atmosphere, the net uptake of
anthropogenic CO2 by the ocean, and thus global climate. Calcium carbonate dissolution is a key
process in ocean carbon cycling and many of the uncertainties in diagnostic and prognostic
marine carbon cycle models arise from an imperfect understanding of the correct formulation for
the dissolution of CaCO3 (Iglesias-Rodriguez et el., 2002). An outstanding problem is the relative
roles of water column and sedimentary dissolution in setting the ultimate burial rates of CaCO3
in the ocean.

57
Shallow-depth water column CaCO3 dissolution where calcite is supersaturated has been
documented by several independent studies (Anderson and Sarmiento, 1994; Lohmann, 1995;
Schiebel, 2002), and the dissolution of more soluble forms of CaCO3 including aragonite and
high-Mg calcites has been suggested as one possible mechanism for the observed alkalinity
enrichments (Byrne et al., 1984; Iglesias-Rodriguez et el., 2002; Honjo et al., 2008). Specifically,
the export of pteropod shells (aragonite) to the mesopelagic water column offers investigators a
tenable source for elevated alkalinity in the upper mesopelagic zone (Betzer et al., 1984; Byrne
et al., 1984; Feely et al., 2004). However, compared to the numerous studies addressing the
dissolution kinetics of calcite in seawater (Honjo and Erez, 1978; Keir, 1980; Gehlen et al.,
2005; Subhas et al., 2015; Dong et al., 2018; Naviaux et al., 2019), only a modest number of
investigations focused on the dissolution behavior of aragonite (Milliman, 1975; Honjo and Erez,
1978; Morse et al., 1979; Keir, 1980; Acker et al., 1987). As a result, studies of the marine
aragonite cycle are limited by the lack of available information on aragonite dissolution kinetics
(Gangstø et al., 2008), as well as well-quantified aragonite production and export rates (Honjo et
al., 2008).
Most kinetics studies of the dissolution behavior of CaCO3 in seawater have expressed
the dependence of dissolution rate on seawater saturation state through the empirical equation
Rate = k(1−Ω)
[
, where k is the rate constant, and n is the pseudo reaction order.
Discontinuities in the function of calcite dissolution rate vs. under-saturation in seawater, defined
as Ω
NOPQPNRS
, have been documented in recent studies (Subhas et al., 2017; Dong et al., 2018;
Naviaux et al., 2019). These authors assert that the discontinuities are related to changes in
dissolution mechanisms, transforming from step retreat to defect-assisted dissolution, and to the
homogeneous spreading of etch pits, as seawater gets more under-saturated. Reported dissolution

58
rates for aragonite in seawater are different by as much as 4 orders of magnitude, depending on
the experimental conditions (lab vs. field) and sample type (synthetic aragonite vs. pteropods),
making the estimate of aragonite dissolution fluxes in the ocean highly uncertain.
Pteropods are the major pelagic species in the ocean that make their shells out of
aragonite. These relatively large (1~20 mm) biogenic particles sink after death and may dissolve
while sinking, providing alkalinity that can be supplied back to the surface ocean via upwelling
and diapycnal mixing. Among the previous studies that quantified the biogenic inorganic carbon
export fluxes (Honjo et al., 1982; Berelson, 2001; Klaas and Archer, 2002; Balch et al., 2005;
Berelson et al., 2007; Boyd and Trull, 2007; Bishop and Wood, 2008), estimates of aragonite
production and fluxes are scarce, and span a range from 1 to 50% of the total global CaCO3 flux
(Berger, 1978; Berner and Honjo, 1981; Fabry, 1989; Fabry and Deuser, 1991; Hong and Chen,
2002; Mohan et al., 2006). Because of the significant difference between the solubility and
dissolution rates of calcite and aragonite, it is impossible to quantify CaCO3 dissolution from the
sinking biogenic fluxes without knowing the calcite/aragonite ratio of the sinking materials.
I present here a high-resolution map of aragonite saturation in a North Pacific transect
(Figure 3.1), and aragonite dissolution rate measurements from both laboratory experiments and
the North Pacific transect. I also show calcite/aragonite ratios in sinking and suspended particles
along the transect. These data are used to address whether aragonite dissolution in the water
column can account for the observed excess alkalinity reported for North Pacific Intermediate
Water (600 m to 900 m). I also investigate dissolution that is driven by organic matter
respiration or metazoan/zooplankton grazing by analyzing the attenuation of sinking PIC and
POC through the water column. Our new constraints on the export fluxes of calcite and aragonite
highlight the role of aragonite dissolution in ocean carbon and alkalinity budgets.

59

Figure 3.1 (a) A map of the Northeast Pacific Ocean and station locations for the CDisK-IV Cruise (August, 2017); (b)
Ω
¬ORz{[PQl
in the upper 1200 m along the transect. White diamonds are locations where in situ aragonite dissolution experiments
were conducted in Niskin Incubators (Station 2 to 5). There was an additional deployment at 2020 m, Station 2, (Ω
¬ORz{[PQl
=
0.57) that is not shown in Figure 1b. Ω
¬ORz{[PQl
in Figure 1b is calculated based on alkalinity and pH measurements of CDisK-IV
CTD cast, and is slightly different from Ω
¬ORz{[PQl
data from GLODAP v2 ODV collection. This discrepancy will be discussed
in detail in a forthcoming paper of our group.

3.2 METHODS
3.2.1 Labeled aragonite for dissolution experiments
13
C labeled aragonite used for all the dissolution experiments was synthesized in our
laboratory using a gel-diffusion method first described by Nickl and Henisch (1969). A 120 mL
glass U-shaped tube was filled with 30 mL hydrous gel (0.17 M sodium metasilicate, adjusted to

60
pH 8), separating ~40 mL reservoirs of Na2
13
CO3 (0.20 M) in one arm of the tube, and 40 mL
CaCl2 (0.10 M) + MgCl2 (0.50 M) in the other arm of the tube. The ends of the tube were sealed
using Parafilm and rubber stoppers. Nucleation of aragonite crystals was limited by diffusion in
the hydrous gel and the gel pore spacing, allowing for slow growth of large crystals. Grains were
harvested after 3-6 months of reaction time by pouring off the spent reservoir solutions followed
by physical break-up, sonication, and decantation of the less dense gel matrix from the aragonite
grains. Grains were then triply washed in Milli-Q water. The mineral grains were wet-sieved to a
grain size of 250-495 µm and dried at 60 ° C overnight.
Mineralogy was confirmed with XRD and Raman spectroscopy. SEM imagery was
obtained using a Hitachi TM-1000 environmental SEM. δ
13
C of the synthetic aragonite was
determined using a Picarro Cavity Ring-Down Spectroscopy (G2131-i) by sequentially diluting
small aliquots of materials (0.2~0.8 mg) into natural abundance optical calcite and measuring the
isotope composition of the mixtures. The degree of
13
C isotope labeling of the synthetic
aragonite was determined to be 100 ± 5%. Specific surface area was measured using Kr
adsorption isotherms, fitting the curves following the BET method. BET surface area was
determined to be 2.46 ± 0.05 m
2
/g. Kr has been demonstrated to give accurate surface areas
down to 0.05 m
2
total area (Subhas et al., 2015). Therefore, the specific surface area determined
for the U-tube
13
C aragonite, with a sample mass of 0.54 g and a total area of 1.3 m
2
, was
considered well-constrained.

3.2.2 Lab and field dissolution experiments
Lab dissolution experiments were conducted following the method reported in earlier
publications of our group (Subhas et al., 2015). Briefly,
13
C labeled aragonite was dissolved in

61
acidified Dickson standard reference seawater
(https://www.nodc.noaa.gov/ocads/oceans/Dickson_CRM/batches.html), and d
13
C of seawater
DIC was traced over time to establish dissolution rates. DIC and alkalinity were measured to
determine Ω before and after each 3 to 5-days experiment. Dissolution experiments were
conducted on a shaker table at 60 rpm, at ambient room temperature (21 ℃) and a temperature
most representative of in situ ocean temperature at 200~2000 m depth (5 ℃). Because the mass
of solid was small (1 mg) compared to seawater (300 g), Ω never changed by more than 0.03
during an experiment. The errors in alkalinity and DIC were propagated to Ω by a Monte Carlo
approach (Subhas et al., 2015). The errors in dissolution rates were calculated based on the
goodness of fits in d
13
C-time correlation using the LINEST function in Microsoft Excel. The
stoichiometric solubility product of aragonite (K
*
sp) used to calculate Ω was reported by Mucci
(1983).
Field dissolution rates were measured at 4 stations along a North Pacific transect between
Honolulu, Hawaii and Seward, Alaska (Figure 3.1) in August, 2017. The cruise crossed from the
subtropical gyre into the subarctic gyre, passing through the transition zone – a large-scale
frontal system between the two gyres that has high productivity and particle export. Dissolution
experiments were conducted in custom-built Niskin Incubators. These devices were modified
General Oceanics 1.7 L Niskins. Approximately 1 mg of
13
C labeled aragonite powders were
placed in a membrane bag of 8 µm pore size (Whatman Nuclepore Track-Etched Membranes,
WHA111114), and inserted into a closed “toaster” chamber that was connected to the Niskin
bottle through MasterFlex tubing (Tygon Fuel & Lubricant Tubing, 06401-82 and 06401-17). A
Seabird SBE 5T/5P Submersible Pump was connected to each Niskin Incubator to generate
circulation inside the reactor so that sufficient mixing was provided as water continuously passed

62
through the “toaster” and through the Niskin. This occurred while Niskins traveled down the
water column to the target depth. Messengers then triggered Niskin closure and carbonate
dissolution proceeded at in situ conditions until the Niskins were recovered and sampled.
Discrete samples were taken from the CTD (conductivity, temperature, and depth) cast and
incubator cast. d
13
C of DIC obtained from the CTD cast and from the incubator cast was
compared to determine dissolution rates. The uncertainty of dissolution rates depends on the
standard deviation of the replicate samples. pH (total scale) and alkalinity were measured by the
group in Byrne lab, University of South Florida to determine Ω in the water column. The
determination of Ω by the DIC-alkalinity pair (lab method), and the pH-alkalinity pair (field
method) agreed with each other.

3.2.3 Sinking flux measurements with sediment traps
An array of surface-tethered sediment traps was deployed on a single line; one at 100 m
and the other at 200 m depth at the 5 stations in the North Pacific transect. Traps were deployed
as free-floating arrays for 52 to 78 hours. The traps were polycarbonate particle interceptor tubes
(PIT) that were 70 cm long, 10 cm diameter (12 tubes per trap) with funnels inserted to guide
particles into a Falcon tube attached to the end of the funnel. Falcon tubes were pre-filled with
HgCl2 poison in brine solutions to inhibit diffusive loss of poison during deployment. The
poison-brine solution was made from seawater collected at 150 m with NaCl added to increase
the salinity by ~ 5, and sodium borate was added to increase alkalinity by ~ 2000 µM (US
JGOFS protocol). Samples from 6 arbitrarily-chosen tubes among the 12 tubes at the same depth
were combined and ‘swimmers’ were manually picked out. The samples were then filtered onto
a pre-weighed glass fiber filter (Whatman glass microfiber filters, Grade GF/F, 1825-047) and,

63
after being returned to the lab, were reweighed to calculate sinking mass flux. The solid
materials on the filters were then collected and analyzed with XRD for mineralogy, and with the
Picarro for PIC and total C. The uncertainty associated with the sediment trap method was
evaluated by comparing the duplicate samples at 100 m, Station 4.
Calcite/aragonite ratios were analyzed based on the relative peak intensity of the
strongest calcite (104) and aragonite (111) peak in XRD. Standards of 6 different ratios of
calcite-aragonite mixtures were measured to generate a calibration curve of peak intensity ratio
versus concentration.

3.2.4 Suspended particle concentration measurements with in situ pumps
Suspended particles were captured on an Advantec GC-50 Glass Fiber Filter (LOT No.
70207718, pore size 0.5 µm, diameter = 142 mm) through filtration of a dual-flowpath in situ
pump (McLane WTS-6-1-142LVUP). At targeted depths, approximately 1400 L of seawater
were pumped through the filter. Upon recovery, the filter was sub-sampled into smaller circles
by using two different sized punches. These sub-samples were used for PIC, total C and
calcite/aragonite ratio measurements. Circles of 26 mm diameter and 6.75 mm diameter were
taken for Picarro measurements of PIC and total C, respectively. PIC and total C were
determined by measuring the amount of CO2 released after acid treatment and combustion. A
circle of 26 mm diameter was analyzed for mineral composition on a Horiba ExploRa+
dispersive Raman microscope after being treated with 50% bleach to remove organic matter
which otherwise would inhibit the Raman spectroscopy laser (532 nm). A 1g L
-1
sodium borate
buffer was added to the bleach to preclude potential CaCO3 dissolution during the bleaching
process. The Raman analysis was programmed as an 8mm × 8mm scan of each bleached filter,

64
and a map of calcite and aragonite distribution was generated. Calcite/aragonite ratios were
calculated by ImageJ based on two different counting methods. First, total counts for each
mineral were identified by the software. The minimum carbonate size detected (30×30 µm) was
limited by the size of a single image pixel. Second, the total area occupied by each mineral was
analyzed. The two methods were in general agreement, except for filters that included a single
large calcite or aragonite fragment. Mineral percentages calculated by total area took into
account the size differences of particles. However, random inclusion of a single large particle
could potentially bias the calcite/aragonite ratio. Therefore, the analysis of suspended
calcite/aragonite percentages and concentrations in this chapter was based on the counts of the
two minerals.

3.3 RESULTS
3.3.1 Aragonite dissolution rates
Both aragonite lab and field dissolution rates show a non-linear relationship with
seawater under-saturation (1-Ω) (Figure 3.2a). In situ aragonite dissolution experiments
conducted during our cruise in the North Pacific have a temperature range of 2~7 ℃, with an
average value of 4.8 ℃. Depths at which the experiments were conducted vary from 100 m to
2000 m. Phosphate concentration varies from 0.06 to 3.24 µmol/kg. Lab experiments conducted
with Dickson seawater have a constant phosphate concentration of 0.43 µmol/kg. Despite the
discrepancies in the experimental conditions, in situ dissolution rates generally lie on the same
curve, with a reaction order n of 1.37 ± 0.18 over the Ω range of 0.45 to 0.87 (Figure 3.2b). Lab
dissolution rates in acidified Dickson standard seawater at 5 ℃ show good agreement with field
rates, while lab rates at 21 ℃ are two times faster than at 5 ℃. The reaction orders of lab

65
dissolution rate laws at 5℃ and 21℃ are 1.33 ± 0.56 and 1.59 ± 0.12 respectively. Between Ω =
0.9~1, the reaction order n is smaller, indicated by a shallower slope in the log (dissolution rate)
vs. log (1-Ω) plot (Figure 2b). The Ω
NOPQPNRS
between the two rate law formulations is around 0.9
(log (1-Ω
NOPQPNRS
) = -1.0).

Figure 3.2 Aragonite dissolution rates measured in the field along the North Pacific transect (2~7℃), and in the lab (5℃ and
21℃). (a) Specific dissolution rates (g cm
-2
day
-1
) vs. (1-Ω); (b) log-log scale of specific dissolution rates (g cm
-2
day
-1
) vs. (1-Ω).
The reaction orders (slopes of the fitting lines) of field, lab 5℃, lab 21℃ rate laws at Ω = 0.45~0.87 are 1.37 ± 0.18, 1.33 ± 0.56,
and 1.59 ± 0.12 respectively. At Ω = 0.9~1 [log(1- Ω)<-1.0], reaction order n is smaller than at lower Ω. The transition between
the two reaction orders (Ω
NOPQPNRS
) occurs at Ω
NOPQPNRS
~ 0.9.

3.3.2 Sinking C fluxes captured by sediment traps
Mass export fluxes in the upper 200 m are higher in the North Pacific subarctic gyre than
in the subtropical gyre by a factor of 2~6 (Figure 3.3a-d). Total mass fluxes at 100 m in the
subtropical gyre are approximately 100~200 mg m
-2
day
-1
, and increase sharply to ~600 mg m
-2

day
-1
at the transition zone. Total C fluxes account for 11~23% of total fluxes in the upper 200 m
of the transect. PIC fluxes at 100 m vary from 0.1~0.7 mmol m
-2
day
-1
in the subtropical gyre, to
1.0~2.5 mmol m
-2
day
-1
in the subarctic gyre; while at 200 m, PIC fluxes in the subarctic gyre are
approximately 0.8 mmol m
-2
day
-1
(Table 3.1). Sinking PIC/POC molar ratios vary from 0.2 to

66
0.6, with no specific trend versus latitude. Total mass and C fluxes are higher at 100 m than 200
m by about a factor of 2 (Figure 3.3e-h) although the attenuation with depth is much greater at
higher latitude. Though all stations are supersaturated with respect to calcite between 100 to 200
m, and only Station 5 attains aragonite under-saturation above 200 m, a factor of 1.5~3 decrease
in sinking PIC fluxes is observed at Station 2, 4 and 5 (Figure 3.3g). Sample amounts were
insufficient for C analysis at 200 m depth for Station 1 and 3.

67

Figure 3.3 Sinking fluxes captured by sediment traps at 100m and 200m along the North Pacific transect in August 2017. Station
1, 2, 3 were in the North Pacific subtropical gyre, while Station 4 and 5 were in the subarctic gyre. The solid bar for the 100 m
0
200
400
600
800
20 30 40 50
Mass Flux (mg m
-2
day
-1
)
Latitude (°N)
Mass Flux
100m
200m
0
30
60
90
120
150
20 30 40 50
Total C Flux (mg m
-2
day
-1
)
Latitude (°N)
Total C Flux
100m
200m
0.0
0.5
1.0
1.5
2.0
2.5
20 30 40 50
PIC Flux (mmol m
-2
day
-1
)
Latitude (°N)
PIC Flux
100m
200m
0
3
6
9
20 30 40 50
POC Flux (mmol m
-2
day
-1
)
Latitude (°N)
POC Flux
100m
200m
(a) (b)
(c) (d)
50
100
150
200
250
0 200 400 600 800
Depth (m)
Mass Flux (mg m
-2
day
-1
)
Station 1
Station 2
Station 3
Station 4
Station 5
50
100
150
200
250
0 30 60 90 120 150
Depth (m)
Total C Flux (mg m
-2
day
-1
)
50
100
150
200
250
0.0 0.7 1.4 2.1 2.8
Depth (m)
PIC Flux (mmol m
-2
day
-1
)
50
100
150
200
250
0 2 4 6 8 10
Depth (m)
POC Flux (mmol m
-2
day
-1
)
(e) (f)
(g) (h)

68
sample at 42°N represent the difference in the duplicate samples at that site (The two ends of the bars are the duplicate values,
and the trend lines are fitted through the average of the duplicates). (a-d) fluxes vs. latitude; (e-h) fluxes vs. depth.

Table 3.1 PIC fluxes, carbonate percentages and PIC/POC in sediment trap samples of CDisK-IV (August, 2017)
Station Latitude Longitude Depth
(m)
Total C Flux
(mg m
-2
day
-1
)
CaCO 3 Flux
(mmol m
-2
day
-1
)
Aragonite/CaCO 3 PIC/POC
Station 1 22°45’N 157°59’W 100 44.8 0.71 44% 0.27
200 N.A. N.A. 67% N.A.
Station 2 27°45’N 155°15’W 100 20.0 0.50 75% 0.46
200 10.5 0.32 74% 0.60
Station 3 35°16’N 150°59’W 100 10.2 0.13 50% 0.19
200 N.A. N.A. 46% N.A.
Station 4 41°45’N 148°16’W 100 110.7±19.7 2.23±0.21 10% 0.34±0.09
200 34.8 0.82 14% 0.42
Station 5 49°50’N 149°39’W 100 51.7 1.03 9% 0.36
200 38.7 0.75 8% 0.33

Because Station 4 had abundant material in the trap tubes, only 3 tubes were combined
from the 100 m trap whereas all other samples included material collected from 6 tubes.
Therefore, a replicate analysis of flux values was determined for the 100 m trap at Station 4,
demonstrating the variability in tube-to-tube collection efficiency and composition. The
reproducibility of mass, total C, PIC and POC flux is approximately 20%, 36%, 18% and 52%
respectively.
In terms of mineral composition in the sediment traps, significantly higher percentages of
calcite are observed in the subarctic gyre (Figure 3.4a, 3.4b, Table 3.1). At both 100 m and 200

69
m in the subtropical gyre, calcite accounts for 20~60% of total CaCO3. However, in the subarctic
gyre, the calcite percentage of total CaCO3 exceeds 80% and constitutes the dominant sinking
CaCO3 phase. As a consequence, the apparent higher sinking flux observed in the subarctic gyre
is mainly due to a significantly higher calcite flux, while the aragonite flux remains similar along
the whole transect (Figure 3.4c).

Figure 3.4 CaCO 3 mineral composition (a, b) and calculated mineral fluxes (c) in the sinking material along the North Pacific
transect at 100m and 200m in August 2017.

3.3.3 Suspended particulate C concentrations measured with in situ pumps
Suspended PIC and POC show quite different distribution patterns from each other along
the North Pacific transect (Figure 3.5a, 3.5b), resulting in a PIC/POC ratio that varies by a factor

70
of 8~10. A high concentration pool of suspended PIC (30 ~ 100 µm) is detected in the upper 800
m centered at Station 3 (~35°N). In contrast, suspended POC at this latitude does not show high
concentration at mid depth but simply decreases with depth as it does at all other stations. The
PIC pool as measured on filter sub-samples agrees with observations of a concentrated
suspended calcite pool detected by Raman scanning (Figure 3.5c). The ‘bullseye’ of high PIC
around Station 3 is supported by the Raman measurements and identified as suspended calcite.
Throughout the water column, suspended calcite concentrations are higher than aragonite
concentrations by a factor of 5~100 (Figure 3.5c, 3.5d). Aragonite particles are detected below
3000 m in the subtropical gyre (Figure 5e). These aragonite particles include relatively large
fragments (up to 190 µm) compared to calcite particles (<80 µm), indicating that large pteropod
fragments are able to sink to the bottom of the ocean in the subtropical gyre. The average error of
calcite or aragonite percentage in total PIC is 26% based on the three replicate filters analyzed.

71

72
Figure 3.5 Suspended PIC and POC captured by in situ pumps along the North Pacific transect in August 2017. (a) Suspended
PIC (µg/L), (b) Suspended POC (µg/L), (c) Suspended calcite concentration (count/L), (d) Suspended aragonite concentration
(count/L), (e) Calcite and aragonite percentages.

3.4 DISCUSSION
3.4.1 Kinetics of aragonite dissolution in the lab versus in the field
In previous studies, reported dissolution rate laws for aragonite as suspended particles in
seawater were nonlinear, but the reaction order n varied (n=2.93 in Morse et al. (1979); n=4.2 in
Keir (1980); n=1.87 in Acker et al. (1987)). In this study, I found n = 1.4 ± 0.2 at 5℃ (lab and
field) and 1.6 ± 0.1 at 21℃ (lab) for synthetic aragonite. A summary of dissolution rate data for
aragonite in seawater (Figure 3.6) calls for distinctions between lab studies vs. field studies, and
experiments with synthetic aragonite vs. pteropods. First, dissolution rates measured in situ in the
ocean water column are orders of magnitude slower than rates measured in the lab. Second,
whether the dissolution rates of synthetic aragonite agree with rates for pteropods appears to be
dependent of the intrinsic crystal properties of the synthetic aragonite. In two lab dissolution
studies, rates for synthetic aragonite were ~30 times faster than pteropod rates when normalized
to surface area (Morse et al., 1979; Keir, 1980). However, in a field dissolution study using a
different synthetic aragonite, dissolution rates of synthetic aragonite and pteropods were in
agreement (Honjo and Erez, 1978). The collection of all rates obtained from the literature and
our work show a convergence of dissolution rate (surface area normalized) when Ω > 0.9.
However, at lower saturation states, rates diverge by as much as 4 orders of magnitude.
This study is, as far as I know, the first to show consistency between lab and field CaCO3
dissolution rates. It is also the first to use the exact same material in both environments. The field
data agree with lab measurements conducted at 5°C, whereas lab measurements at 21°C are ~2
times faster for equivalent saturation state. While this temperature effect is not a central focus of

73
our discussion, I note that it is similar to observations of the influence of temperature on calcite
dissolution (Naviaux et al., 2019). An Ω
NOPQPNRS
at ~0.9 is observed for aragonite dissolution rates,
similar to the value of Ω
NOPQPNRS
(0.87±0.05) between step retreat dissolution and defect-assisted
dissolution for calcite dissolution recently reported in Dong et al. (2018), and Ω
NOPQPNRS
(~0.85)
for coccolith and foraminiferal biogenic calcite reported in Subhas et al. (2018).
The synthetic aragonite in this study dissolves significantly slower than the synthetic
aragonites used by Morse et al. (1979) and Keir (1980). This could be due to lower defect
density resulting from slower crystallization rates during our synthesizing process. The gel-
diffusion methodology in this study allowed long crystal growth time, up to 3-6 months, whereas
the precipitation time for the method of aragonite formation used by Morse et al. (1979) and Keir
(1980) was ~30 min. Furthermore, the specific surface area of the synthetic aragonite produced
in this study (2.46 m
2
g
-1
) is similar to that of pteropods (2.17 m
2
g
-1
, Honjo and Erez, 1978)
(Table 3.2). I thus propose that the synthetic aragonite in this study is more representative of the
pteropod aragonite. Our rates lie between the published dissolution rates of pteropods measured
in the lab and in the field (Milliman, 1975; Honjo and Erez, 1978; Morse et al., 1979; Keir, 1980;
Acker et al., 1987).

74

Figure 3.6 Comparison of the aragonite dissolution rates in this study and previous studies. Closed symbols are dissolution rates
of synthetic aragonite, whereas open symbols are rates of pteropods. Note that dissolution rates of synthetic aragonite are ~30
times larger than pteropods in Morse et al. (1979) and Keir (1980). Additionally, previously published field dissolution rates are
orders of magnitude lower than lab dissolution rates. Dissolution rates reported in this study show a consistency between field
and lab.

Table 3.2 Experimental conditions and materials used in published aragonite dissolution studies
Material Experimental
condition
SSA (m
2
g
-1
) BET flow gas Reference
Aragonite Field N.A. N.A. Milliman (1975)
Synthetic aragonite Field 1.53 He Honjo and Erez (1978)
Pteropds Field 2.17 He Honjo and Erez (1978)
Synthetic aragonite Lab 1.4 Kr Morse et al. (1979)
Pteropods Lab 1.6 Kr Morse et al. (1979)
Synthetic aragonite Lab 1.4 Kr Keir (1980)
Pteropods Lab 2.17 He Keir (1980)

75
Pteropods Lab N.A. N.A. Acker et al. (1987)
Synthetic aragonite Lab and field 2.46 Kr This study
* SSA (specific surface area) not measured in Acker et al. (1987), specific rates (g cm-2
day-1) calculated assuming the same pteropods SSA as in Honjo and Erez (1978) and Keir
(1980). Synthetic aragonite used in Morse et al. (1979) and Keir (1980), pteropod assemblage
used in Honjo and Erez (1978) and Keir (1980) were samples of the same batch.

3.4.2 Total C and PIC export production rates, PIC/POC ratios in sinking fluxes and
suspended materials in the North Pacific
Measured total C fluxes in the upper 200 m in the N. Pacific subtropical and subarctic
gyre are 0.85–3.73 mmol m
-2
day
-1
, 2.90–9.23 mmol m
-2
day
-1
, respectively (Table 1);
comparable to the annual rate of net community production in the N. Pacific mixed layer
reported by Lee (2001) (1.4–4.1 mmol m
-2
day
-1
in the subtropical gyre, 5.5–8.2 mmol m
-2
day
-1

in the subarctic gyre). PIC fluxes in the upper 200 m in the two gyres are 0.13–0.71 mmol m
-2

day
-1
, 0.75–2.23 mmol m
-2
day
-1
; slightly lower than the annual rate of net CaCO3 production in
Lee (2001) (0–1.6 mmol m
-2
day
-1
in the subtropical gyre, 1.6–4.4 mmol m
-2
day
-1
in the
subarctic gyre).
PIC/POC molar ratios in sinking particles vary from 0.19~0.60 at 100 m and 200 m along
the N. Pacific transect (Table 1), with no apparent trend with latitude. This range is inclusive of
values measured with shallow sediment traps in the N. Pacific, including Station P
(PIC/POC=0.54 at 200 m), the Equatorial Pacific (PIC/POC=0.1–0.5 at 125–340 m), and the W.
Pacific Warm Pool (PIC/POC=0.2–0.5 at 105–320m) (Wong et al., 1999; Rodier and Le Borgne,
1997; Honjo et al., 2008 and references therein). It is higher, however, than the estimated global

76
mean rain ratio of 0.056 obtained by applying the vertical gradients of potential alkalinity and
nitrate to an ocean biogeochemical-transport box model (Sarmiento et al., 2002 and references
therein).
PIC/POC ratios in suspended particles cover a wider range than in sinking particles,
varying from 0.06 to 0.81. Suspended particle concentrations measured by in situ pumps are
considered to represent a semi-decomposed phase generated from sinking particles but also from
particles embedded in an advecting water mass. The high concentration of suspended PIC
observed in the upper 800 m between 30~40°N is confirmed by PIC analysis via filter
acidification (Figure 3.5a) and by calcite grain counts via Raman scanning (Figure 3.5c). One
explanation for this high PIC concentration pool is potentially the not-yet-dissolved carbonate
produced during the previous weeks to months when the transition zone resided near that
latitude. The absence of high aragonite concentrations within this pool could be due to a) lack of
production or b) the lower saturation state of aragonite at this depth than calcite, and thus more
rapid dissolution. Further to the north, where there is high PIC export, seawater is under-
saturated at shallower depth and the suspended PIC is much lower. The fact that there is no
bullseye pattern observed in suspended POC indicates that POC associated with this PIC has a
much shorter residence time in the upper ocean. This agrees with the finding of higher PIC/POC
ratios in deeper sediment traps reported in Berelson et al. (2007). One consequence of POC
remineralization is that it can leave PIC unprotected and exposed to seawater for dissolution.
Although this bullseye is distinct, dissolution of the suspended PIC in this pool will generate a
maximum excess alkalinity of only 0.3 µmol/kg, which is trivially small. However, this bullseye
pattern is potentially a tracer of dissolution of a larger sinking PIC flux. It is the transformation

77
from the flux of particulate sinking carbonate to suspended carbonate, and subsequently to
dissolved carbonate that helps define the ocean alkalinity balance.
The presence of suspended aragonite particles below 3000 m in the subtropical gyre
(Figure 3.5e) provides direct evidence that aragonite aggregates can sink towards the bottom of
the ocean without being fully dissolved. Because the sediments at these depths have PIC < 0.1
wt. %, benthic dissolution of aragonite should therefore be considered when constructing carbon
cycle simulations.

3.4.3 PIC dissolution and POC remineralization rates in sinking fluxes and shallow depth
dissolution
The following discussion of trap flux data is predicated on our assumption that trapping
with PIT traps is accurate and equally efficient at 100 m and 200 m. Work with these identical
traps (Haskell et al., 2013) exhibited good agreement between Th-based POC flux and sediment-
trap-based POC flux, although the traps occasionally yielded fluxes smaller than those obtained
via Th methodologies. Another confirmation that our traps are representative of the true export in
this region is a comparison of the fluxes I obtained at Station 1 and work by others at Station
Aloha. The floating PITs trap results of Karl and Church (2014) yielded an average POC flux of
2.5 mmol m
-2
day
-1
at 150 m, comparable to our 100 m trap flux of 2.6 mmol m
-2
day
-1
.
The difference in sediment trap PIC fluxes between 100 m and 200 m at the same
location indicates that a large fraction of the sinking PIC dissolves in this zone (27~63%, Figure
3.3g). This occurs at one site that is under-saturated with respect to aragonite between 100 m and
200 m (e.g., Station 5; Ω=1.3~0.5), but also at stations super-saturated at this depth (Station 2
and 4; Ω=3~2.7 and Ω=1.8~1.5 respectively). This analysis is in direct conflict with

78
determinations of calcium carbonate dissolution rates (Figure 3.2), which show no appreciable
dissolution when Ω >1. Although supported by only a few data, a strong correlation between PIC
dissolution and POC remineralization rates is observed (Figure 3.7), with a molar ratio of 0.29 at
all three stations (Figure 3.7a). Higher percentages of POC loss are associated with higher
percentages of PIC loss, with 20~70% POC loss between 100 m and 200 m, and 30~60% PIC
loss (Figure 3.7b). The similar calcite to aragonite ratios at 100 m and 200 m (Figure 3.4a, 3.4b)
indicates that calcite and aragonite dissolve at similar ratios relative to their abundance.
Therefore, dissolution mechanism for the shallow depth dissolution is not sensitive to carbonate
mineral phase, and has less direct dependence on saturation state.

Figure 3.7 Amount of PIC dissolved and POC remineralized between 100 m and 200 m in the sinking fluxes during CDisK-IV
(August, 2017). (a) PIC weight loss as a function of POC weight loss. PIC dissolves and POC remineralizes at a ratio of 0.29. (b)
Percent of PIC and POC lost from 100 m to 200 m. High PIC dissolution is associated with high POC remineralization.

It has been suggested that organic matter respiration can produce microenvironments
with locally enhanced CO2 concentration, and therefore promote dissolution well above the
saturation horizon (Jansen et al., 2002). Aerobic respiration has been shown to decrease oxygen
concentrations within sinking aggregates (Ploug, 2001). I propose that respiration-driven

79
dissolution is one potential mechanism underlying the observed attenuation of sinking PIC fluxes
observed in this study. Another possible reason is that both POC and PIC are consumed by large
organisms (metazoans) or zooplankton. PIC is dissolved during gut passage (Bishop et al., 1980;
Harris, 1994). The percentage of PIC and POC loss may have temporal variation because
microbial and metazoan/zooplankton abundances may vary seasonally. At Station 5, where
seawater is under-saturated for aragonite below 100 m, thermodynamically-driven dissolution
may also occur and contribute alongside respiration-driven dissolution, and/or
metazoan/zooplankton consumption to the observed attenuation.

3.4.4 Calcite/aragonite ratios in sinking and suspended materials in the North Pacific water
column
Previous studies of aragonite and calcite sinking fluxes were determined using traps and
defined mineral species by microscopic identification (Betzer et al., 1984; Fabry and Deuser,
1991) and X-ray diffraction analysis (Berner and Honjo, 1981; Fabry and Deuser, 1991). Betzer
et al. (1984) reported a 0.3 mmol m
-2
day
-1
average flux of aragonite at 100 m, compared to
0.057 mmol m
-2
day
-1
at 400 m and 0.009 mmol m
-2
day
-1
at 2200 m along a Northwestern
Pacific section, yielding a 90% loss in the upper 2200 m. Their absolute flux rates showed that
foraminifera were more abundant in the northern areas (aragonite/total CaCO3 = 9~33% north of
42°N) whereas pteropods dominated in the southern areas (aragonite/total CaCO3 = 66~95%
south of 30°N). A rather constant aragonite flux through depth was reported in the Sargasso Sea,
with aragonite accounting for 8~35% of total CaCO3 (Fabry and Deuser, 1991). Their average
pteropod mass flux from 7 trap deployments were 0.026 ± 0.006 mmol m
-2
day
-1
at 500 m, 0.024
± 0.004 mmol m
-2
day
-1
at 1500 m, and 0.024 ± 0.002 mmol m
-2
day
-1
at 3200 m. Similar values

80
were reported at the Panama Basin (~0.06 mmol m
-2
day
-1
at 667 m to ~0.03 mmol m
-2
day
-1
at
3791 m) and the Equatorial Atlantic (0.168 mmol m
-2
day
-1
at 389 m to 0.034 mmol m
-2
day
-1
at
988 m), with aragonite accounting for 4~39% of total CaCO3 (Berner and Honjo, 1981).
In this study, lower aragonite/calcite ratios are observed in sediment trap materials in the
subarctic gyre (aragonite < 20%) than in the subtropical gyre (aragonite 37~75%) (Figure 3.4a,
3.4b). This distribution pattern is similar to the Northwestern Pacific section in Betzer et al.
(1984). The increase in PIC and POC fluxes across the transition zone (from south to north) is
mainly due to an increase in calcite species, and the aragonite fluxes remain similar along the
whole transect (Figure 3.4c). The aragonite fluxes I determined; 0.07 to 0.38 mmol m
-2
day
-1
at
100 m, and 0.06 to 0.24 mmol m
-2
day
-1
at 200 m, are comparable to the average aragonite flux at
100 m (0.3 mmol m
-2
day
-1
) in Betzer et al. (1984). The higher calcite percentage north of the
transition zone also agrees with the conclusion of Juranek et al. (2012), wherein the authors
related the increased Transition Zone Chlorophyll Front productivity to a high amount of
coccolithophorids and diatoms.

3.4.5 Aragonite dissolution fluxes in the North Pacific water column
According to the dissolution experiments in Byrne et al. (1984), abundant small size class
aragonite particulates were suggested to provide substantial contributions to the “excess
alkalinity” of the upper water column of the Pacific Ocean. However, Fabry (1990) estimated the
pteropod and heteropod growth rate to yield 0.029 mmol CaCO3 m
-2
day
-1
, only sufficient to
account for ~8% of the calculated rate of CaCO3 dissolution in the North Pacific, were this to
completely dissolve upon sinking. Similar to aragonite, high-Mg calcite has also been suggested
to account for the shallow dissolution phenomenon due to its high solubility. Carbonate formed

81
within fish, a Mg-rich calcite, was reported to explain up to a quarter of the increase in titratable
alkalinity within 1000 m of the ocean surface (Wilson et al., 2009).
In the North Pacific, maximum in situ dissolution rates of approximately 1.1 √mol kg
-1

yr
-1
were reported at 530 m by plotting excess alkalinity (TA*) versus water parcel ages derived
from chlorofluorocarbon-11 (CFC-11) (Feely et al., 2004). This high dissolution rate at shallow
depths was associated with either dissolution in locally acidic conditions, including guts of
zooplankton (Bishop et al., 1980; Harris, 1994), and where bacterial oxidation of organic matter
takes place (Jansen and Wolf-Gladrow, 2001); or dissolution of the more soluble forms of
CaCO3 in shallow waters, including pteropods and high-Mg calcite (Byrne et al., 1984). Recent
studies also pointed out the significance of the enzyme carbonic anhydrase in enhancing
carbonate dissolution rates in the ocean (Subhas et al., 2017) and the kinetic enhancement of
CaCO3 dissolution rates due to pressure (Dong et al., 2018). The calculation below assesses
whether the inorganic dissolution of sinking aragonite alone can account for the excess alkalinity
observed (Figure 3.8a, 3.8b), and how much of the dissolution can be driven by organic matter
respiration and metazoan/zooplankton grazing (Figure 3.8c, 3.8d).

82

Figure 3.8 (a) Aragonite sinking flux below the saturation horizon (430 m) if dissolution only occurs as abiotic aragonite
dissolution; (b) in situ dissolution rate in the water column (µmol kg
-1
yr
-1
) if dissolution only occurs as abiotic aragonite
dissolution. Different symbols in (a) and (b) represent different sinking rates of aragonite particles. Dissolution rates are
calculated based on the saturation states measured in Station 3 (35° N, 151° W) and the dissolution rate law determined in this
chapter. Aragonite sinking flux of 0.1 mmol m
-2
day
-1
is assumed to reach 430 m, below which seawater starts to become under-
saturated. (c) PIC sinking flux assuming 40% dissolution every 100 m due to organic matter respiration driven dissolution or
metazoan/zooplankton consumption; (d) in situ dissolution rate generated by respiration driven dissolution or
metazoan/zooplankton consumption. For (c), (d), total PIC sinking flux of 1.2 mmol m
-2
day
-1
, calcite flux of 0.8 mmol m
-2
day
-1
,
aragonite flux of 0.4 mmol m
-2
day
-1
are assumed at 100 m.

I construct a box model of the water column to diagnose the sinking and dissolution of
CaCO3 particles in the upper 2000 m (Appendix 2). Particles are considered to be produced in
the upper 100 m within the euphotic zone, and are exposed to dissolution while sinking. PIC,
calcite and aragonite fluxes at the bottom of the euphotic zone, saturation states in the water

83
column, and dissolution rates are based on the measurements in this study. A three orders of
magnitude difference in aragonite sinking rate (0.1 m/day to 100 m/day) is adopted to investigate
the effect of sinking rates on aragonite dissolution fluxes. First, to estimate the dissolution flux
due to inorganic aragonite dissolution, sinking aragonite particles are considered to be exposed to
dissolution when the water column first reaches Ω < 1, and dissolution rates are determined as a
function of Ω according to the relationship shown in Figure 3.2. Details of the model can be
found in supplementary materials.
Model results show that the sinking rate of aragonite aggregates significantly alters the
sinking flux vs. depth and in situ dissolution signals (Figure 3.8a, 3.8b). To generate the
observed in situ CaCO3 dissolution rates (0.8~1.2 √mol kg
-1
yr
-1
between 250~750 m) in the
water column as reported in Feely et al. (2004), the sinking rate of aragonite flux will have to be
much less than 1 m day
-1
. This sinking rate is lower than generally accepted pteropod sinking
rates of 80~1080 m day
-1
(Noji et al., 1998). Assuming a pteropod sinking rate of 100 m day
-1
,
our measured aragonite sinking flux and dissolution rate only account for 0.2% of the excess
alkalinity signal. The implication of this model result is that the thermodynamically-driven
dissolution of sinking aragonite particles alone is inadequate to explain the excess alkalinity
observed in the North Pacific water column, and that other mechanisms, e.g. respiration-driven
dissolution and/or metazoan/zooplankton consumption, are required to explain the shallow depth
dissolution.
The mechanism for both respiration-driven dissolution and metazoan/zooplankton
consumption is that acidic local micro-environment is provided and drives PIC dissolution. As a
result, CaCO3 dissolution can take place above the saturation horizon. To approximate this type
of dissolution, a model is developed in which sinking fluxes at 100 m are taken from

84
measurements using sediment traps, and a constant dissolution rate of 40% every 100 m is
assumed based on the calculated dissolution rate between 100 and 200 m (Figure 3.7b). PIC,
calcite and aragonite sinking fluxes at 100 m (Flux100m) are 1.2 mmol m
-2
day
-1
, 0.8 mmol m
-2

day
-1
, and 0.4 mmol m
-2
day
-1
, respectively.
Model results show that the in situ dissolution signal produced by either respiration-
driven dissolution or metazoan/zooplankton consumption in the upper 600 m is significantly
higher than thermodynamically-driven dissolution fluxes (Figure 3.8c, 3.8d), and is comparable
in magnitude to the observed dissolution signal reported in Feely et al. (2004). However, the
shape of the modeled in situ dissolution rate appears to be different than the observed pattern.
This could be due to vertical mixing in the mixing layer, horizontal transport, and/or the re-
utilization of alkalinity in the surface by shell-making biota. The magnitude of shallow depth
dissolution may be sensitive to seasonality because microbial and metazoan/zooplankton
abundances may vary seasonally.

3.5 CONCLUSIONS
I report dissolution rates of synthetic aragonite in seawater both in the lab (at 5℃ and
21℃) and in the field (2~7℃) along a transect from Hawaii to Alaska, and show consistency
between lab rates at 5℃ and field rates. Aragonite dissolution rates (at 5°C) fit a non-linear rate
equation: Rate (g g
-1
day
-1
) = 0.013*(1- Ω)
1.37
. Rates are ~2 times faster at 21℃ than at 5℃ for
equivalent saturation state. As determined by floating sediment traps deployed on this transect,
sinking carbon fluxes are significantly higher in the subarctic gyre than in the subtropical gyre,
and yet there is no geographical trend in PIC/POC mole ratio, which is 0.2 ~ 0.6 for material
sinking between 100~200 m. The calcite/aragonite ratio is lower in the subtropical gyre than in

85
the subarctic gyre. Calcite fluxes in the two gyres are different by a factor of 5~10, whereas
aragonite fluxes appear relatively constant along the North Pacific transect. Measured aragonite
fluxes are 0.07~0.38 mmol m
-2
day
-1
at 100 m, and 0.06~0.24 mmol m
-2
day
-1
at 200 m, with no
specific trend over latitude. A comparison of fluxes at depths 100 m to 200 m indicates that
30~60% PIC dissolves between these depths with a simultaneous 20~70% attenuation in POC
fluxes. The presence of suspended aragonite below 3000 m in the subtropical gyre indicates that
large pteropod fragments can sink to the bottom of the ocean without being fully dissolved. One
significant implication of this study is that predictions based on a strictly thermodynamic view of
aragonite dissolution underestimate observed alkalinity excess and measured PIC attenuation.
This conclusion is indicated by the unreasonably low sinking rate of pteropod aragonite (< 1 m
day
-1
) required to generate the excess alkalinity described in Feely et al. (2004). Our measured
aragonite flux and inorganic dissolution rate only account for 9% and 0.2% of the excess
alkalinity observed in the North Pacific, assuming a pteropod sinking rate of 1 m day
-1
and 100
m day
-1
, respectively. Respiration-driven dissolution or metazoan/zooplankton grazing, indicated
by the simultaneous attenuation of PIC and POC in the sediment traps versus depth, produces an
amount of dissolution comparable to that suggested by excess alkalinity but not consistent with
the depth distribution.

ACKNOWLEDGEMENTS
I would like to acknowledge editor Derek Vance, reviewer Jack Middelburg and another
anonymous reviewer for their invaluable comments of the original manuscript. I thank the
captain and crews on Kilo Moana for their assistance at sea. I also acknowledge Christopher
Moore, Loraine Martell-Bonet for their help measuring pH and alkalinity during CDisK-IV; Yi

86
Hou for leak-checking the Niskin Incubators by measuring dissolved Si concentrations; Doug
Hammond for providing the in situ pumps; Johnny Stutsman and James Rae for their help
deploying and recovering in situ pumps; as well as Abby Lunstrum and Huanting Hu for their
help picking out swimmers from the sediment trap samples.

REFERENCES
Acker, J. G., Byrne, R. H., Ben-Yaakov, S., Feely, R. A., & Betzer, P. R. (1987). The effect of
pressure on aragonite dissolution rates in seawater. Geochim. Cosmochim. Acta, 51,
2171-2175.
Anderson, L. A. and Sarmiento, J. L. (1994) Redfield ratios of remineralization determined by
nutrient data analysis. Global Biogeochem. Cycles, 8, 65–80.
Balch, W. M., Gordon, H. R., Bowler, B. C., Drapeau, D. T., & Booth, E. S. (2005). Calcium
carbonate measurements in the surface global ocean based on Moderate-Resolution
Imaging Spectroradiometer data. Journal of Geophysical Research: Oceans, 110(C7).
Berelson, W. M. (2001). Particle settling rates increase with depth in the ocean. Deep Sea
Research Part II: Topical Studies in Oceanography, 49(1-3), 237-251.
Berelson, W. M., Balch, W. M., Najjar, R., Feely, R. A., Sabine, C., & Lee, K. (2007). Relating
estimates of CaCO3 production, export, and dissolution in the water column to
measurements of CaCO3 rain into sediment traps and dissolution on the sea floor: A
revised global carbonate budget. Global Biogeochemical Cycles, 21(1).
Berger, W. H. (1978). Deep-sea carbonate: pteropod distribution and the aragonite compensation
depth. Deep Sea Research, 25(5), 447-452.
Berner, R. A., & Honjo, S. (1981). Pelagic sedimentation of aragonite: its geochemical
significance. Science, 211(4485), 940-942.
Betzer, P. R., Byrne, R. H., Acker, J. G., Lewis, C. S., Jolley, R. R., & Feely, R. A. (1984). The
oceanic carbonate system: a reassessment of biogenic controls. Science, 226(4678), 1074-
1077.
Bishop, J. K., & Wood, T. J. (2008). Particulate matter chemistry and dynamics in the twilight
zone at VERTIGO ALOHA and K2 sites. Deep Sea Research Part I: Oceanographic
Research Papers, 55(12), 1684-1706.
Bishop, J. K., Collier, R. W., Kettens, D. R., & Edmond, J. M. (1980). The chemistry, biology,
and vertical flux of particulate matter from the upper 1500 m of the Panama Basin. Deep
Sea Research Part A. Oceanographic Research Papers, 27(8), 615-640.
Boyd, P. W., & Trull, T. W. (2007). Understanding the export of biogenic particles in oceanic
waters: is there consensus? Progress in Oceanography, 72(4), 276-312.
Byrne, R. H., J. G. Acker, P. R. Betzer, R. A. Feely, and M. H. Cates (1984) Water column
dissolution of aragonite in the Pacific Ocean, Nature, 312, 321– 326.
Dong, S., Subhas, A. V., Rollins, N. E., Naviaux, J. D., Adkins, J. F., & Berelson, W. M. (2018).
A kinetic pressure effect on calcite dissolution in seawater. Geochimica et Cosmochimica
Acta, 238, 411-423.

87
Fabry, V. J. (1989). Aragonite production by pteropod molluscs in the subarctic Pacific. Deep
Sea Research Part A. Oceanographic Research Papers, 36(11), 1735-1751.
Fabry, V. J. (1990). Shell growth rates of pteropod and heteropod molluscs and aragonite
production in the open ocean: implications for the marine carbonate system. Journal of
Marine Research, 48(1), 209-222.
Fabry, V. J., & Deuser, W. G. (1991). Aragonite and magnesian calcite fluxes to the deep
Sargasso Sea. Deep Sea Research Part A. Oceanographic Research Papers, 38(6), 713-
728.
Feely, R. A., Sabine, C. L., Lee, K., Berelson, W., Kleypas, J., Fabry, V. J., & Millero, F. J.
(2004). Impact of anthropogenic CO2 on the CaCO3 system in the
oceans. Science, 305(5682), 362-366.
Gangstø, R., Gehlen, M., Schneider, B., Bopp, L., Aumont, O., & Joos, F. (2008). Modeling the
marine aragonite cycle: changes under rising carbon dioxide and its role in shallow water
CaCO3 dissolution. Biogeosciences, 5(4), 1057-1072.
Gehlen, M., Bassinot, F. C., Chou, L., & McCorkle, D. (2005). Reassessing the dissolution of
marine carbonates: II. Reaction kinetics. Deep Sea Research Part I: Oceanographic
Research Papers, 52(8), 1461-1476.
Harris, R. P. (1994). Zooplankton grazing on the coccolithophore Emiliania huxleyi and its role
in inorganic carbon flux. Marine Biology, 119(3), 431-439.
Haskell II, W. Z., Berelson, W. M., Hammond, D. E., & Capone, D. G. (2013). Particle sinking
dynamics and POC fluxes in the Eastern Tropical South Pacific based on
234
Th budgets
and sediment trap deployments. Deep Sea Research Part I: Oceanographic Research
Papers, 81, 1-13.
Hong, G.H., Chen, C.T.A. (2002). Aragonitic pteropod flux to the interior of the East Sea (Sea of
Japan). Terrestrial, Atmospheric and Oceanic Sciences, 13(2), 205-210.
Honjo, S., & Erez, J. (1978). Dissolution rates of calcium carbonate in the deep ocean; an in-situ
experiment in the North Atlantic Ocean. Earth and Planetary Science Letters, 40(2), 287-
300.
Honjo, S., Manganini, S. J., & Cole, J. J. (1982). Sedimentation of biogenic matter in the deep
ocean. Deep Sea Research Part A. Oceanographic Research Papers, 29(5), 609-625.
Honjo, S., Manganini, S. J., Krishfield, R. A., & Francois, R. (2008). Particulate organic carbon
fluxes to the ocean interior and factors controlling the biological pump: A synthesis of
global sediment trap programs since 1983. Progress in Oceanography, 76(3), 217-285.
Iglesias-Rodriguez, M. D., Armstrong, R., Feely, R., Hood, R., Kleypas, J., Milliman, J. D.,
Sabine, C., & Sarmiento, J. (2002). Progress made in study of ocean's calcium carbonate
budget. EOS, Transactions American Geophysical Union, 83(34), 365-375.
Jansen, H., & Wolf-Gladrow, D. A. (2001). Carbonate dissolution in copepod guts: a numerical
model. Marine Ecology Progress Series, 221, 199-207.
Jansen, H., Zeebe, R. E., & Wolf-Gladrow, D. A. (2002). Modeling the dissolution of settling
CaCO3 in the ocean. Global Biogeochemical Cycles, 16(2), 11-1.
Juranek, L. W., Quay, P. D., Feely, R. A., Lockwood, D., Karl, D. M., & Church, M. J. (2012).
Biological production in the NE Pacific and its influence on air-sea CO2 flux: Evidence
from dissolved oxygen isotopes and O2/Ar. Journal of Geophysical Research:
Oceans, 117(C5).
Karl, D. M., & Church, M. J. (2014). Microbial oceanography and the Hawaii Ocean Time-series
programme. Nature Reviews Microbiology, 12(10), 699.

88
Keir, R. S. (1980). The dissolution kinetics of biogenic calcium carbonates in
seawater. Geochimica et Cosmochimica Acta, 44(2), 241-252.
Klaas, C., & Archer, D. E. (2002). Association of sinking organic matter with various types of
mineral ballast in the deep sea: Implications for the rain ratio. Global Biogeochemical
Cycles, 16(4).
Lee, K. (2001). Global net community production estimated from the annual cycle of surface
water total dissolved inorganic carbon. Limnology and Oceanography, 46(6), 1287-1297.
Lohmann, G. P. (1995) A model for variation in the chemistry of planktonic foraminifera due to
secondary calcification and selective dissolution. Paleoceanography, 10, 445–457.
Milliman, J. D. (1975). Dissolution of aragonite, Mg-calcite, and calcite in the North Atlantic
Ocean. Geology, 3(8), 461-462.
Mohan, R., Verma, K., Mergulhao, L. P., Sinha, D. K., Shanvas, S., & Guptha, M. V. S. (2006).
Seasonal variation of pteropods from the Western Arabian Sea sediment trap. Geo-
Marine Letters, 26(5), 265-273.
Morse, J. W., de Kanel, J. O. H. N., & Harris, K. A. R. E. N. (1979). Dissolution kinetics of
calcium carbonate in seawater; VII, The dissolution kinetics of synthetic aragonite and
pteropod tests. American Journal of Science, 279(5), 488-502.
Mucci, A. (1983). The solubility of calcite and aragonite in seawater at various salinities,
temperatures, and one atmosphere total pressure. American Journal of Science, 283(7),
780-799.
Naviaux, J. D., Subhas, A. V., Rollins, N. E., Dong, S., Berelson, W. M., & Adkins, J. F. (2019).
Temperature dependence of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 246, 363-384.
Nickl, H. J., & Henisch, H. K. (1969). Growth of calcite crystals in gels. Journal of the
Electrochemical Society, 116(9), 1258-1260.
Noji, T. T., Bathmann, U. V., & Von Bodungen, B. (1998). Clearance of picoplankton-sized
particles and formation of rapidly sinking aggregates by the pteropod, Limacina
retroversa. Oceanographic Literature Review, 1(45), 107.
Ploug, H. (2001). Small-scale oxygen fluxes and remineralization in sinking
aggregates. Limnology and Oceanography, 46(7), 1624-1631.
Rodier, M., & Le Borgne, R. (1997). Export flux of particles at the equator in the western and
central Pacific ocean. Deep Sea Research Part II: Topical Studies in
Oceanography, 44(9-10), 2085-2113.
Sarmiento, J. L., Dunne, J., Gnanadesikan, A., Key, R. M., Matsumoto, K., & Slater, R. (2002).
A new estimate of the CaCO3 to organic carbon export ratio. Global Biogeochemical
Cycles, 16(4).
Schiebel, R. (2002) Planktonic foraminiferal sedimentation and the marine calcite budget. Global
Biogeochem. Cycles, 16, 1065, doi:10.1029/2001GB001459.
Subhas, A. V., Adkins, J. F., Rollins, N. E., Naviaux, J., Erez, J., & Berelson, W. M. (2017).
Catalysis and chemical mechanisms of calcite dissolution in seawater. Proceedings of the
National Academy of Sciences, 114(31), 8175-8180.
Subhas, A. V., Rollins, N. E., Berelson, W. M., Dong, S., Erez, J., & Adkins, J. F. (2015). A
novel determination of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 170, 51-68.

89
Subhas, A. V., Rollins, N. E., Berelson, W. M., Erez, J., Ziveri, P., Langer, G., & Adkins, J. F.
(2018). The dissolution behavior of biogenic calcites in seawater and a possible role for
magnesium and organic carbon. Marine Chemistry, 205, 100-112.
Wilson, R. W., Millero, F. J., Taylor, J. R., Walsh, P. J., Christensen, V., Jennings, S., & Grosell,
M. (2009). Contribution of fish to the marine inorganic carbon cycle. Science, 323(5912),
359-362.
Wong, C. S., Whitney, F. A., Crawford, D. W., Iseki, K., Matear, R. J., Johnson, W. K., Page,
J.S., & Timothy, D. (1999). Seasonal and interannual variability in particle fluxes of
carbon, nitrogen and silicon from time series of sediment traps at Ocean Station P, 1982–
1993: relationship to changes in subarctic primary productivity. Deep Sea Research Part
II: Topical Studies in Oceanography, 46(11-12), 2735-2760.

90
Appendix. Model of aragonite dissolution fluxes in the North Pacific water column.
(1) Thermodynamically-driven (abiotic) aragonite dissolution
I construct a box model of the water column to diagnose the sinking and dissolution of
CaCO3 particles in the upper 2000 m. First, to estimate the dissolution flux due to abiotic
aragonite dissolution, sinking aragonite particles are considered to be exposed to dissolution
when the water column first reaches Ω < 1, and dissolution rates are determined as a function of
Ω according to the relationship shown in Fig. 2. The whole water column is separated into
multiple boxes (box height from 20 to 800 m depending on the depth of CTD deployment). The
amount of aragonite dissolution in each box depends on the aragonite flux that reaches the box,
the average saturation state of the box, and the aragonite particles’ residence period in the box.
The sinking flux in the i
th
box Flux Zi (mmol m
-2
day
-1
) and in situ dissolution rate in the i
th
box In
Situ Rate Zi (√mol kg
-1
yr
-1
) are calculated based on the following equations:

ƒÑ≈∆
«P
= ƒÑ≈∆
«P∞ú
−ƒÑ≈∆
«P∞ú
∗»… ÀÑ≈V…ÀÖ TUVW∗V…´W
= ƒÑ≈∆
«P∞ú
−ƒÑ≈∆
«P∞ú
∗[0.013(1−Ω
P
)
ú.öÃ
]∗
(Õ
P
−Õ
P∞ú
)
Œ

Eq. 1
œÖ w…V≈ TUVW
«P
=
ƒÑ≈∆
«P
∗w∗[0.013(1−Ω
P
)
ú.öÃ
]∗
(Õ
P
−Õ
P∞ú
)
Œ
w∗(Õ
P
−Õ
P∞ú
)∗Å
=
ƒÑ≈∆
«P
∗[0.013(1−Ω
P
)
ú.öÃ
]
Å∗Œ

Eq. 2

where Zi (m) is the depth of the i
th
box, ƒÑ≈∆
«P
(mmol m
-2
day
-1
) is the sinking flux of
aragonite at depth Zi, Ω
P
is the saturation state of water column at Zi, Œ (m day
-1
) is the sinking
rate of aragonite flux, œÖ w…V≈ TUVW
«P
(√mol kg
-1
yr
-1
) is the in situ dissolution rate of aragonite at
depth Zi, S (m
2
) is the area of the water-column cross section, and Å (1029 kg m
-3
) is the density
of seawater. In Eq. 1 and 2, aragonite dissolution rate is calculated as established by our
dissolution experiments: Rate (g g
-1
day
-1
) = 0.013*(1- Ω)
1.37
.

91
To estimate the average aragonite dissolution fluxes in the North Pacific, water column
saturation is taken from measurements at Station 3 (35° N, 151° W); and an aragonite sinking flux
of 0.1 mmol m
-2
day
-1
is assumed at the saturation horizon (430 m). Values adopted in the model
are chosen to be generally representative of the N. Pacific transect.

(2) Respiration-driven or metazoan/zooplankton consumption-driven PIC dissolution
For the shallow depth dissolution indicated by PIC losses in the sediment trap samples, a
model is developed to simulate the process in which sinking fluxes at 100 m are taken from
measurements using sediment traps, and a constant dissolution rate of 40% every 100 m is
assumed based on the calculated dissolution rate between 100 and 200 m (Fig. 7b). The sinking
flux (mmol m
-2
day
-1
) and the in situ dissolution rate (√mol kg
-1
yr
-1
) are calculated based on the
following equations:

ƒÑ≈∆
«P
= ƒÑ≈∆
«P∞ú
−ƒÑ≈∆
«P∞ú
∗
40%
100´
∗(Õ
P
−Õ
P∞ú
)
Eq. 3
œÖ w…V≈ TUVW
«P
=
ƒÑ≈∆
«P
∗w∗
40%
100´
∗(Õ
P
−Õ
P∞ú
)
w∗(Õ
P
−Õ
P∞ú
)∗Å
=
ƒÑ≈∆
«P
∗40%
100´∗Å

Eq. 4

PIC, calcite and aragonite sinking fluxes at 100 m (Flux100m) are 1.2 mmol m
-2
day
-1
, 0.8
mmol m
-2
day
-1
, and 0.4 mmol m
-2
day
-1
, respectively.

92
CHAPTER 4: AN ATOMIC FORCE MICROSCOPY STUDY OF
CALCITE DISSOLUTION IN SEAWATER AND THE
DEPENDENCE OF DISSOLUTION RATE ON SURFACE
PROPERTIES

This chapter is a manuscript in preparation as:
Dong, S., Berelson, W. M., Adkins, J.F., Rollins, N. E., Naviaux, J. D., Pirbadian, S., El-Naggar,
M. Y., & Teng, H. H. (in preparation). An Atomic Force Microscopy study of calcite dissolution
in seawater and the dependence of dissolution rate on surface properties.

ABSTRACT
This chapter presents the first examination of calcite dissolution in seawater using
Atomic Force Microscopy (AFM). I quantify step retreat velocity, etch pit density, and introduce
a new parameter, the “edge length density” in order to compare calcite dissolution under AFM
conditions to those conducted in bulk solution experiments (e.g. Subhas et al. 2015) and to
compare dissolution in seawater to low ionic strength water. No significant difference in etch pit
morphology is observed between seawater and previous studies in low ionic strength water. Bulk
dissolution rates and step retreat velocities are slower at high and mid W values and become
comparable to low ionic strength water rates at low W, with a ratio of obtuse to acute step
velocity of 5 to 10. At 21℃, I observe a significant increase in etch pit density (to ~10
8
cm
-2
)
occurring at W = 0.7 in seawater, compared to W < 0.1 in low ionic strength water, suggesting a
transition from defect-assisted dissolution to homogeneous dissolution closer to equilibrium in
seawater. The change in Ωcritical for enhanced etch pit nucleation in seawater is consistent with the
published effect of Mg
2+
and SO4
2-
(Ruiz-Agudo et al., 2009). The kinetic coefficient ü is found
to increase from 1.4×10
-5
m s
-1
to 5.5×10
-4
m s
-1
as W decreases from 0.9 to 0.4. Step retreat

93
velocities of two calcite surfaces with distinctly different topography, but at similar saturation
states, differ by a factor of 4, whereas “bulk” dissolution rates calculated in the image areas
(normalized to surface area) are different by 200 times. The tight correlation between dissolution
rate and the “edge length density” suggests that this new parameter captures an Angstrom-scale
roughness that BET surface area does not capture and is essential in determining bulk dissolution
rates.

4.1 INTRODUCTION
Calcite plays a critical role in regulating geochemical cycles and thus, ocean chemistry
and atmospheric CO2 through dissolution and precipitation in aqueous environment due to the
mineral’s wide occurrence and high reactivity at earth surface (Berner, 1981). For the past two
decades, an increasing number of dissolution studies have focused on direct observation and
quantification of the kinetics of dissolution processes on calcite surfaces using AFM and Vertical
Scanning Interferometry (VSI) (e.g. Stipp et al., 1994; Dove and Platt, 1996; Liang et al., 1996;
Liang and Baer, 1997; McCoy and LaFemina, 1997; Shiraki et al., 2000; Lea et al., 2001; Morse
and Arvidson, 2002; Arvidson et al., 2003; Teng, 2004; Bisschop et al., 2006; Lüttge and
Arvidson, 2010; Smith et al., 2013), and these experiments have greatly complemented
interpretations of results obtained from bulk calcite dissolution studies (Plummer et al., 1978;
Sjoberg and Rickard, 1984; Schott et al., 1989; Gutjahr et al., 1996; Cubillas et al., 2005; Xu et
al., 2012). The linkage between bulk experiments and dissolution occurring on the microscopic
scale is the ultimate goal of this study.
Although the dissolution/precipitation reactions that help regulate atmospheric CO2 on
millennial timescales occur primarily in seawater, all previous AFM and VSI studies have been
conducted in low-ionic strength waters. Even though effects of electrolytes, inorganic ions and

94
organic molecules on calcite dissolution in low ionic strength water have been studied separately
to analogize seawater in contact with rock-forming minerals (e.g. Kowacz and Putnis, 2008;
Arvidson et al., 2006; Teng and Dove, 1997), no previous AFM experiment was ever done in
seawater. Bulk dissolution experiments we have conducted show that calcite reacts at distinct
rates and respond to W much differently in seawater than in freshwater, a problem that has
plagued marine chemists for decades (Subhas et al., 2017). Whether calcite dissolution in
seawater is only a combined effect of individual components based on the relatively fixed
seawater composition or not is still an open question.
Another problem that is widely acknowledged but yet to be fully resolved is the effect of
mineral surface properties on dissolution rates. Intrinsic sources of variation within crystals, such
as step and defect density, have been noted to affect dissolution rates in addition to extrinsic
variability such as temperature and saturation state (Lüttge et al., 2013). Even though differences
in surface topography result in distinct dissolution rates (Arvidson et al., 2003; Arvidson and
Lüttge, 2010), BET surface area has been shown to be an unsatisfactory measure of dissolution
active sites (Gautier et al., 2000) as dissolution rates appear to be more closely proportional to
geometric surface area than BET surface areas (Wolff-Boenisch et al., 2004). As a result, caution
has been suggested in selecting geometric or BET area when comparing rates between different
studies (Cubillas et al., 2005; Subhas et al., 2015; Dong et al., 2019). Lüttge et al. (2013)
proposed that the variability in dissolution rates of samples with distinct surface properties
reflects a fundamentally stochastic interaction between solution and crystal surfaces with a
heterogeneous defect distribution, and applied a kinetic Monte Carlo model to simulate the
evolution of surface area, the total surface-integrated rate, and the distribution of kinks and other
surface sites during dissolution. Their work provides critical insight into understanding the

95
intrinsic variability of crystals, but falls short of establishing a more fundamental rate normalizer.
An experimental approach to quantify the intrinsic variability is still needed to complement the
model results.
This chapter aims to bridge the seawater and low ionic strength water dissolution studies
from an experimental and microscale perspective. Comparisons of dissolution rates and the value
of Ωcritical between dissolution mechanisms are made between seawater and low ionic strength
water. In addition, the effect of surface properties on dissolution is investigated and a new
approach to normalize dissolution rate is proposed.

4.2 METHODS
4.2.1 Sample and solution preparation
Calcite (104) surfaces were obtained by using a razor blade to cleave a large crystal of
optical-quality Iceland spar. An air burst was then applied to the cleaved fragment
(approximately 5×5×0.5 mm) to remove small adhering particles. The fragment was
subsequently adhered to a magnetic plate using double-sided adhesive tape.
The experimental solution was standard reference Dickson seawater, Batch 176
(https://www.nodc.noaa.gov/ocads/oceans/Dickson_CRM/batches.html), acidified to desired
saturation states by adding HCl. The undersaturated seawater was kept in gas-impermeable bags
with no headspace. Dissolved Inorganic Carbon (DIC) and alkalinity were measured to
determine calcite saturation state (W). W was calculated from the CO2SYS program (Van
Heuven et al., 2011) using K’1 K’2 (apparent dissociation constants of carbonic acid in seawater)
reported by Mehrbach et al. (1973) and refit by Dickson and Millero (1987); K SO4 reported by
Dickson et al. (1990); and borate to salinity ratio reported by Uppström (1974). The uncertainty

96
of W was calculated based on the standard errors in DIC and alkalinity as described in Subhas et
al. (2015) and averaged ±0.026.

4.2.2 In situ dissolution experiment set-up and AFM imaging
The AFM in situ dissolution experiment setup is shown as in Figure 4.1. Experiments
were conducted at 21℃ at atmospheric pressure. In situ fluid cell imaging was conducted using
an Asylum Research Cypher ES Environmental Atomic Force Microscope. Images were
obtained in either Tapping or Contact Mode, and no obvious difference was observed between
the two modes. All commercially available AFM probes we tried corroded within 1~2 hours in
seawater, even for Au-coated silicon probes. The two probes we used for experiments in this
chapter are: Arrow UHFAuD from Asylum Research
(https://afmprobes.asylumresearch.com/arrow-uhfaud.html), and SNL-10 from Bruker
(https://www.brukerafmprobes.com/p-3693-snl-10.aspx). Because the Cypher AFM had a gas
headspace in the fluid cell (~3 mL), the headspace was manually adjusted by adding CO2-
controlled gas that was in equilibrium with the solution. Alkalinity and DIC measurements taken
from the inflow and outflow confirmed that Ω remained constant throughout the experiment
(Table 4.1). All experiments were conducted at a flow rate of 15 mL h
-1
; at this flow rate, water
was in contact with the mineral surface for ≤1 minute (residence time).

97

Figure 4.1 AFM in situ dissolution experiment setup.

Table 4.1 DIC, alkalinity and W of the fill bag, solution in inflow syringe and solution in outflow syringe.
DIC
(µmol
kg
-1
)
No. of
DIC
samples
DIC std
error
(µmol kg
-1
)
Alkalinity
(µmol kg
-1
)
No. of
alk.
samples
Alk. std
error
(µmol kg
-1
)
W W
error
FB 2125.2 2 10.0 1989.2 3 0.95 0.398 0.021
In 2139.4 1 N.A. (15) * 1989.6 1 N.A. (2) * 0.373 0.028
Out 2145.8 3 16.8 1993.8 2 0.45 0.370 0.031

* Standard errors not available because there was only enough sample for single DIC and
alkalinity measurement. Errors in the parenthesis were used to calculate the error of W.
Abbreviations: No. = number; std = standard; alk. = alkalinity; FB = fill bag; In = inflow
solution; Out = outflow solution; N.A. = not available.

4.2.3 Dissolution rate analysis and edge length measurement
Because we observed non-negligible drifts between scans in most of our experiments, we
used etch pit widening rates to present step retreat velocities to eliminate the factor of scan area
drifting. The scan rates for all experiments were either 6.5 Hz or 9.8 Hz, equivalent to 0.2~1.5

98
µm s
-1
when taking the scan sizes into consideration. This scan rate was significantly larger than
the step retreat rates measured in all experiments (0.003~0.5 nm s
-1
), and therefore should have a
negligible influence in generating step velocity error. In addition, all step velocity calculations
were made with images of the same scan direction during an experiment (either frame-up or
frame-down). Because opposite parallel edges in an etch pit are different in edge type (acute vs.
obtuse), the measured step velocities (half of etch pit widening rates) were treated as the average
of acute and obtuse step retreat velocities. The separation of acute and obtuse edges was only
made in two dissolution experiments when no obvious drifting was visually observed, to give a
rough estimate of acute and obtuse velocity ratio during calcite dissolution in seawater. Step
velocities were measured perpendicular to the step edges. Uncertainty of step velocity was
determined as the standard error of step velocities at 1 to 8 different etch pits and at 2 to 6
different time periods. Measured step velocity was found to be independent of time and location
(Figure 4.2), and the variation was largely due to the limited precision in width measurement
within the image analysis program Gwyddion. Specifically, the precision of the distance
measurement was ±0.01µm, whereas step velocities (average of acute and obtuse velocities) in
our experiments were 0.003~0.5 nm s
-1
. Due to the fast probe corrosion in seawater, experiments
were generally less than 1 h. To obtain widths at multiple time points, e.g. every 15 min, the
changes in width were 0.005~0.9 µm. Therefore, for determinations of widths within several
minutes, especially in slow dissolution experiments, errors existed because changes in widths is
comparable to precision of measurement.

99

Figure 4.2 Step velocities of individual etch pits during different time periods for two experiments. Measured step velocities are
independent of time and etch pit location. The variation is largely due to the limited precision in width measurement.

Total edge length on the calcite surface and “bulk” dissolution rates were calculated by
analyzing the grayscale of the AFM images using programs we developed in MATLAB. The
purpose of determining total edge length per unit imaging area is to serve as a measure of surface
roughness. Total edge length in images was calculated by binning a convoluted surface to one
monolayer increment and detecting abrupt grayscale changes using MATLAB (Figure 4.3a and
4.4). One way to envision this parameter is to imagine a line drawing of a set of steps, and the
sum of all line segments used to draw the steps would be total edge length. An edge of “n” layers
and “x” µm would have a total length of “nx” µm. During a dissolution experiment, the evolution
of total edge length with time might change by ~20%. This variation is considered insignificant
compared to the orders of magnitude difference between different surfaces. Because image
quality significantly affected edge detection, images with the least noise in each experiment were

100
selected for edge analysis (usually towards the beginning of each experiment because imaging
quality decreased as the AFM probe corroded).

Figure 4.3 Images showing how the MATLAB code determines (a) edges and (b) dissolution rates (AF-5). (a-1) raw image; (a-2)
processed image after binning the convoluted surface to one monolayer increment; (a-3) detected edges overlying the raw image.
(b-1) image at t 1 during AF-5; (b-2) image at t 2 during AF-5; (b-3) number of monolayers dissolved between t 1 and t 2.

101

Figure 4.4 Edge length analysis using MATLAB. (a) to (e) are images of the four experiments in Table 4.3. (b) and (c) are sub-
regions of a larger image. Black curves are edges detected after smoothing the images.

To calculate the “bulk” dissolution rate within the imaging area, images at different time
points were compared, the difference in grayscale was translated to a difference in height, and
integrated with the area to obtain the volume lost (Figure 4.3b). Volumes dissolved were then
plotted versus time, and the slope was normalized to the size of the image to get the “bulk” rate
(Figure 4.5). Because all calcite samples were at the (104) cleavage surface and had less than 30
monolayers (9 nm) relief, the additional topographic relief area was negligible compared to the
2D image size (2×2 µm to 10×10 µm). Thus, rates normalized to real surface area and image
area were different by less than 1%. For the different images in one experiment, grayscale of the
same monolayer (reference height in the images) was manually adjusted to be the same to correct
for the automatic color adjustment of AFM as top monolayers were removed (e.g. Figure 4.3b-1
and 4.3b-2).

102

Figure 4.5 One example of “bulk” rate determination by MATLAB (AF-3, spot 3). Image area = 107.9 µm
2
. “Bulk” rate
calculated = 1.58×10
-6
g cm
-2
day
-1
.

4.3 RESULTS
4.3.1 Effect of flow rate on dissolution and the variation of step velocity on the calcite
surface
Average step velocities of acute and obtuse edges vary by as much as 6 times between
different etch pits during a single dissolution experiment at Ω = 0.37 ± 0.01 (Figure 4.6). Such
variability can only be accounted for by taking many different measurements. The mean values
of average step velocities at fluid flow rates of 15 mL h
-1
, 30 mL h
-1
, 45 mL h
-1
are within the
standard deviation of each other, indicating that dissolution is not limited by diffusion above 15
mL h
-1
. Because Ω = 0.37 ± 0.01 is the lowest saturation state among all dissolution experiments
in this study, and thus the step velocity is the highest in this study, all dissolution experiments
conducted with flow rates = 15 mL h
-1
can be considered surface-controlled instead of diffusion-
controlled. Calculated Ω values for the inflow and the outflow solutions, based on DIC and

103
alkalinity measurements, are similar within error (Table 4.1) and this confirms our ability to
conduct dissolution rate measurements with a steady saturation state.

Figure 4.6 Average step velocity of acute and obtuse edges at three different flow rates (15 mL h
-1
, 30 mL h
-1
, 45 mL h
-1
) at Ω =
0.37 ± 0.01. The grey crosses are velocities at different individual etch pits. The squares are the mean values of the crosses, with
the error bars representing the standard deviation of the population.

4.3.2 Etch pit morphology in seawater and etch pit density vs. undersaturation
Etch pits in seawater are rhombic and no significant rounding of the rhomb is observed to
occur during dissolution (Figure 4.7). Dissolution experiments at Ω = 0.88 ± 0.04 and 0.87 ±
0.04 show no etch pit formation, and dissolution only occurs as step retreat of existing edges
(Figure 4.8a). The highest Ω at which etch pit formation is observed is 0.82 ± 0.04 (1 etch pit in
31.9 µm
2
throughout 60 min). Below Ω = 0.82 ± 0.04, dissolution happens both as step retreat
and etch pit formation (Figure 4.8b). Etch pit density increases significantly below Ω ~ 0.7, and
the rise is used to distinguish the defect-assisted etch pit mechanism and the homogeneous etch
pit mechanism, which will be discussed in Section 4.4.2.

104

Figure 4.7 One example of dissolution on the calcite (104) cleavage surface in seawater (W = 0.46 ± 0.01). 0 min in Figure 4.7a
is actually 37 min after the start of the continuous seawater flow. Field of view is 11×11 µm. Color scale spans 4 nm in surface
height.

Figure 4.8 (a) Ω = 0.88 ± 0.04; (b) Ω = 0.50 ± 0.02 (etch pits formed before t=31 min, but image quality was poor). At Ω = 0.88
± 0.04, dissolution only happens as step retreat (black arrows). No etch pit formation was found for 30 min within the total
scanned area of 17.04 √m × 17.04 √m. At Ω = 0.50 ± 0.02, dissolution happens both at existing step edges (black arrows) and

105
at newly-formed etch pits (yellow arrows). The highest Ω observed for etch pit formation is 0.82 ± 0.04 with a pit density of
3.1×10
6
cm
-2
.

4.3.3 The dependence of step velocity and bulk rate on saturation state
Despite the scatter of data points, average step velocity increases as Ω decreases (Figure
4.9), with a reaction order n=2.7. The scatter results from the variation of step velocities between
different etch pits and time periods (Figure 4.2 and Figure 4.6), and is potentially due to the
limited precision in etch pit width measurement (see Section 4.2.3 for detail). For two
experiments conducted at similar Ω but with distinct surface features of calcite (one is smooth,
and the other has lots of pre-existing holes, yellow diamonds in Figure 4.9), the average of acute
and obtuse step velocities at the top monolayer are different by a factor of 4, whereas “bulk”
rates are different by a factor of 200 (Figure 4.9 and 4.10). Notable differences in surface
morphology are observed for the two calcite samples (Figure 4.10). The step velocity for AF-5
(Figure 4.10b) refers to the top monolayer (0.059 ± 0.007 nm s
-1
), whereas the step velocity for
the deep etch pits is lower (0.028 ± 0.010 nm s
-1
), potentially due to limitation in the diffusion of
dissolved ions into the bulk solution.
The ratio of obtuse and acute step velocities is determined in AF-3 (W = 0.46) and AF-20
(W = 0.37), during which experiments no obvious image drift is visually observed. At two
different etch pits in AF-3 (W = 0.46), v ob/vac = 9 and 5.6, respectively. In AF-20 (W = 0.37), for
the continuous spreading of an etch pit, average step velocities at 4 different time intervals give
vob/vac = 9.6. Therefore, our study suggests that calcite dissolution in natural seawater has a
vob/vac of roughly 5 to 10. A more accurate determination of vob/vac will require more dissolution
experiments to be done, ideally with a manually-added reference point.

106

Figure 4.9 Average step velocity of acute and obtuse edges vs. undersaturation. The two yellow diamonds are two dissolution
experiments at similar Ω but on calcite crystals with distinct surface features (see Figure 4.10).

Figure 4.10 (a) AF-18, Ω = 0.56, dissolution at point defects; (b) AF-5, Ω = 0.58, dissolution at dislocations or pre-existed holes
on calcite surface. (c) Ratios of step velocity, bulk rate and surface area per unit image area between the two surfaces (surface (a)

107
is considered as 1). One monolayer = 0.3 nm. Note that in AF-5, “0 min” is not the real start time of dissolution, but 38 min after
seawater was in contact with the calcite sample. “SA” is short for surface area.

4.4 DISCUSSION
4.4.1 Comparison of etch pit morphology, bulk rate, step velocity and etch pit density
between seawater and freshwater
Various ions and organic matter have been reported to affect calcite dissolution by
altering etch pit morphology, density, spreading and deepening rate. These components include
Mg
2+
(Arvidson et al., 2006; Ruiz-Agudo et al., 2009; Xu and Higgins, 2010), PO4
3-
(Klasa et al.,
2013), Mn
2+
(Lea et al., 2001; Vinson et al., 2007), Sr
2+
(Lea et al., 2001; Vinson and Lüttge,
2005), CO3
2-
(Lea et al., 2001; Arvidson et al., 2006; Vinson et al., 2007), NH4
+
(Klasa et al.,
2013); SO4
2-
(Ruiz-Agudo et al., 2009), Co
2+
(Freij et al., 2004), background major ions (e.g.
Na
+
, K
+
, Cl
-
, F
-
, etc. Ruiz-Agudo et al., 2009, 2010), and organic molecules (Teng and Dove,
1997; Perry et al., 2004; Teng et al., 2006; Oelkers et al., 2011). The combined effect from these
components may be complex and lead to potentially distinct dissolution phenomena in seawater
as opposed to low ionic strength water.
In terms of etch pit morphology, several studies have reported rounded etch pits with the
addition of different ions. Ions that were reported to alter etch pit morphology include: CO3
2-
,
PO4
3-
, Mn
2+
and Mg
2+
(Lea et al., 2001; Vinson et al., 2007; Klasa et al., 2013; Arvidson et al.,
2006; Ruiz-Agudo et al., 2009; Xu and Higgins, 2010). While the concentration of CO3
2-
, PO4
3-
,
and Mn
2+
in seawater is below the reported concentration that has an effect on etch pits
morphology, the concentration of Mg
2+
in seawater ~ 0.05 mol kg
-1
, higher than the reported
[Mg
2+
] that affects etch pit morphology in Arvidson et al. (2006) (0.01×10
-3
mol kg
-1
) and Ruiz-
Agudo et al. (2009) (0.0008 mol kg
-1
). Specifically, in the Dickson seawater used in our AFM

108
experiments, [CO3
2-
] varies between 20 √mol kg
-1
and 60 √mol kg
-1
, lower than the reported
[CO3
2-
] that alters etch pit morphology in Lea et al. (2001) (~900 √mol kg
-1
). [PO 4
3-
] for the
experiments in this study is 0.29 √mol kg
-1
, which is lower than the minimum [PO4
3-
] added in
Klasa et al. (2013) (5 √mol kg
-1
). Seawater [Mn
2+
] is 3 orders of magnitude lower than the
concentration that has an effect on dissolution in Lea et al. (2001). Our results of calcite etch pit
morphology in seawater show no obvious difference than previously published freshwater
morphology (Figure 4.7). Specifically, etch pits are rhombohedral and no obvious change in
morphology is observed, which does not agree with the observations in Arvidson et al. (2006)
and Ruiz-Agudo et al. (2009).
Bulk dissolution rate and step velocity show similar patterns of difference between
seawater and low ionic strength water, with near equilibrium rates being lower in seawater
whereas far from equilibrium rates are more comparable (Figure 4.11a and 4.11b, Table 4.2). For
the comparison of step velocity, however, it is worth noting that most freshwater studies were
conducted at extremely low saturation states, and the only study conducted at high W was aimed
at conditions of geological carbon sequestration, and thus had high temperature (Xu et al., 2010).
Step velocities are shown to be smaller at lower temperatures. However, because there is only
one data point at 50 ºC between W = 0.5 and 1 in Xu et al. (2010), it is difficult to extrapolate the
high temperature velocities to 21 ºC at near equilibrium conditions.

109

Figure 4.11 Comparisons of bulk dissolution rate, step velocity and pit density between dissolution in seawater (SW) and in
freshwater (FW). Subhas et al. (2017) and this study are in seawater, all others are in freshwater. “Other fresh water studies” in
Figure 4.11a and 4.11b include: Shiraki et al., 2000; De Giudici, 2002; Arvidson et al., 2003; Arvidson et al., 2006; Vinson and
Luttge, 2005; Lea et al, 2001; Harstad and Stipp, 2007. Except for Xu et al. (2010) that has experimental temperatures of 50~70
℃, all other studies are between 20 ℃ and 25 ℃. Experimental details are listed in Table 4.3 and 4.4.

110

Etch pit density in seawater dissolution experiments is ~10
8
cm
-2
between W = 0.4 and
0.7, higher than the ~10
6
cm
-2
in equivalent W in fresh water reported by Teng (2004) (Figure
4.11c, Table 4.3). However, the magnitude of etch pit density converges at low saturation states
(~10
8
cm
-2
). The higher etch pit density in seawater agrees with the enhancement of etch pit
nucleation by Mg
2+
and SO4
2-
, as suggested by Ruiz-Agudo et al. (2009), and will be discussed
further in Section 4.4.2.
In summary, the combination of ions in seawater has a dual and opposing effect on
calcite dissolution. Specifically, they significantly increase etch pit density between W = 0.82 and
0.37, while inhibiting etch pit spreading at high and mid-undersaturation conditions. The overall
influence in bulk dissolution rate is an inhibition near equilibrium when etch pits cannot form
and step velocities are suppressed. In mid-undersaturation states, etch pit density is higher
whereas step velocity is lower, and the “product” of these two processes leads to increasing, but
still lower bulk dissolution rates. Far from equilibrium, etch pit density and step velocity are
similar to those at low ionic strength water conditions.

111
Table 4.2 A comparison of bulk rate and step velocity between this study and previous publications.
Log bulk rate
(mol cm
-2
s
-1
)
Step velocity
(nm/s)
(vo+va)/2
(nm/s)
Material Solution Ωcalcite pH T (ºC) Method Reference
-14.1 ~ -11.9 N.A. 0.003~0.5
Iceland Spar
(104) surface
Natural seawater 0.4 ~ 0.9 7.0-7.3 21 AFM This study
-14.5 ~ -9.9 N.A. N.A. Calcite powder Natural seawater 0.02 ~ 0.99 5.9-7.3 21
Bulk
dissolution
Subhas et al. (2017)
-13.0 ~ -10.4 N.A. N.A. Iceland Spar
NaCl-NaHCO3-
CaCl2 solution
0.1 ~ 0.8 8.0-8.1 20
VSI and bulk
dissolution
Smith et al. (2013)
-12.0 ~ -10.2 N.A. N.A.
Fragmental and
powder samples
NaHCO3-CaCl2
solution
10
-4
~ 0.9 >8 25
Bulk
dissolution
Xu et al. (2012)
-9.8 N.A. N.A. Calcite powder HCl solution 10
-3
7.3 25
Bulk
dissolution
Cubillas et al. (2005)
N.A.
vo=0~3;
va=0~0.6
0~1.7
Iceland Spar
(104) surface
NaCl-NaHCO3-
CaCl2 solution
0.09 ~ 1.2 7.8-8.3 50 AFM Xu et al. (2010)
-9.5
vo=3;
va=1
2
Iceland Spar
(104) surface
NaCl solution 10
-7
7.6 21 AFM Shiraki et al. (2000)
N.A.
vo=0.90;
va=0.67
0.79
Iceland Spar
(104) surface
Na2CO3 solution 0.07 8.9 22 AFM Lea et al. (2001)
-10.6
vo=4.2;
va=0.9
2.6
Iceland Spar
(104) surface
HCl solution 10
-2.15
7.5 22 AFM De Giudici (2002)
-11.0 0.00337 0.00337
Iceland Spar
(104) surface
NaHCO3-
Na2CO3 solution
10
-3.41
8.8 25 VSI
Arvidson et al.
(2003)

112
-11.6
vo=0.04;
va=0.74
0.39
Iceland Spar
(104) surface
NaHCO3
solution
10
-3.39
8.8 25 VSI, AFM
Arvidson et al.
(2006)
-11.2
vo=0.27;
va=1.14
0.71
Iceland Spar
(104) surface
Na2CO3 solution 10
-3.59
8.7
22 VSI, AFM
Vinson and Lüttge
(2005)
-10.6
vo=1.98;
va=1.51
1.75 NaCl solution N.A. 8.6
N.A.
vo=0.29~2.10;
va=0.19~0.55
0.24~1.3
Iceland Spar
(104) surface
Milli-Q water N.A. 5.6-8.3 30 AFM
Harstad and Stipp
(2007)

* Except for this study and Subhas et al. (2017), all other studies in Table 4.2 are considered as “freshwater” studies (Milli-Q water
w/wo addition of certain ions).

113
Table 4.3 Measured etch pit density against solution undersaturation.
Experiment
No.
W
Average etch
pit numbers
Error etch
pit numbers
Image area
(µm
2
)
Average etch pit
density (cm
-2
)
Error etch pit
density (cm
-2
)
AF-9 0.82 1 0 31.9 3.1×10
6
0
AF-14 0.72 10 2 31.6 3.2×10
7
6.3×10
6

AF-17 0.68 17 2 13.9 1.2×10
8
1.4×10
7

AF-18 0.56 8 0 6.4 1.2×10
8
0
AF-11 0.50 9 3 9.2 9.8×10
7
3.3×10
7

AF-12 0.50 23 3 14.7 1.5×10
8
1.7×10
7

AF-20 0.37 33 8 25.0 1.3×10
8
3.0×10
7

Teng (2004)
0.012~0.54 N.A. N.A. N.A. <4×10
6
N.A.
0.007 N.A. N.A. N.A. 5×10
8
N.A.

* Error in etch pit numbers is determined by the variation between images at different time.

114

4.4.2 Identification of changes in calcite dissolution mechanisms in seawater
Distinct dissolution mechanisms at different saturation states in freshwater have been
both theoretically proposed and experimentally identified by previous publications (Holdren and
Berner, 1979; Brantley et al., 1986; Gratz et al., 1991; Stipp et al., 1994; Lasaga and Lüttge,
2001; Teng, 2004; Dove et al., 2005; Arvidson and Lüttge, 2010). From near-equilibrium to
farther from equilibrium, dissolution occurs as (1) retreat of pre-existing steps at edges, corners
and dislocations; (2) opening of etch pits at defects; and finally (3) opening of etch pits
homogenously across the mineral surface. After formation, etch pits can either spread as “2D
pancakes” (Dove et al, 2005), or step-waves that contain multi-layers (Lasaga and Lüttge, 2001).
The transitions of different dissolution mechanisms happen at Wcritical’s, and may imply
discontinuous rate to undersaturation correlations below and above Wcritical’s.
Bulk dissolution experiments conducted in bags in seawater have shown different rates
and Wcritical’s than in low ionic strength water (Subhas et al., 2017; Dong et al., 2018; Naviaux et
al., 2019). The determination of Wcritical’s in these publications is by fitting bulk dissolution rates
to a mechanistic model (Dove et al, 2005) and identifying the breaks in slope in a plot of rate vs.
undersaturation. These transitions in mechanism can be verified by AFM observations.
According to Teng (2004), The onset of the defect-assisted etch pit mechanism is marked by the
highest W observed for etch pit formation, whereas the onset of the homogeneous etch pit
mechanism is revealed by a precipitous increase of pit density as W falls below a critical value.
Measured pit densities against solution undersaturation in weak electrolyte solutions show the
sudden rise at W = 0.007, as pit density increases from 4×10
6
cm
-2
to 5×10
8
cm
-2
(Teng, 2004).
Compared to the observations in weak electrolyte solutions, we have previously reported

115
Wcritical’s for the opening of defect-assisted etch pits in seawater is W = 0.9 (versus W = 0.54 in
freshwater), and the Wcritical for homogeneous etch pit formation is W = 0.75 (versus W = 0.007 in
freshwater) in bulk dissolution experiments (Naviaux et al., 2019).
The AFM measurements reported here generally support our earlier reported Wcritical
values in seawater. No etch pits are observed during dissolution experiments at W = 0.87 and
0.88, dissolution only occurs as step retreat (Figure 4.8a). The highest W observed for etch pit
formation is 0.82, with a pit density of 3.1×10
6
cm
-2
(Figure 4.11c, Table 4.3). The transition
between step retreat and defect-assisted etch pit mechanisms is therefore between 0.82 and 0.87,
comparable to Wcritical = 0.87 in Dong et al. (2018) and Wcritical = 0.9 in Naviaux et al. (2019).
Below W = 0.7, pit density increases abruptly to ~10
8
cm
-2
(Figure 4.11c), similar to the pit
density reported for the homogeneous etch pit formation mechanism far from equilibrium in
Teng (2004) (ns = 10
8
sites cm
-2
) and in Ruiz-Agudo et al. (2009) (ns = 10
9
sites cm
-2
). The
significant difference in etch pit density above and below W = 0.7 may indicate the onset of the
homogeneous etch pit formation mechanism, which agrees with the Wcritical = 0.75 in Naviaux et
al. (2019).
The fact that the onsets of both defect-assisted and homogeneous etch pit mechanisms
occur at higher W in seawater than in low ionic strength water is closely related to the effects of
other ions besides Ca
2+
and CO3
2-
(e.g. Mg
2+
, SO4
2-
etc.) on etch pit density and step velocity.
Mg
2+
has been reported to increase the density and depth of etch pits nucleated on calcite
surfaces (Ruiz-Agudo et al., 2009). A molecular dynamics (MD) simulation by Kerisit and
Parker (2004) has shown that Mg
2+
is able to attract water molecules from the calcite surface to
retain a full coordination shell (i.e. 6 water molecules) once it adsorbs as an inner-sphere
complex directly above a surface carbonate group. As a result, water molecules could be

116
transferred from surface calcium of calcite during magnesium adsorption. Such a strong
magnesium-surface interaction and the fact that magnesium can disrupt the surface hydration
layer can lead to surface destabilization, and ultimately favor 2D nucleation of etch pits. A
reduction in the kinetic barrier associated with the magnesium-calcite surface interaction initiates
etch pit nucleation which manifests itself as an increase in etch pit density.
In addition to Mg
2+
, SO 4
2-
has also been reported to increase the etch pit deepening rate
and etch pit density during calcite dissolution, and the effect is related to an increase in Mg-
adsorption on carbonates (Ruiz-Agudo et al., 2009). The rate limiting step for Mg
2+
adsorption
and incorporation into carbonates is its dehydration (Lippmann, 1973). SO4
2-
is known to
enhance cation desolvation through the formation of ion pairs (Piana et al., 2006). Specifically,
Mg
2+
and SO4
2-
hydrated ions in the solution combine to form double solvent separated ion pairs,
and water molecules are lost from such complexes. In this respect, Brady et al. (1996) have also
shown that adsorption of magnesium on carbonates is enhanced in sulfate-rich solutions during
dolomite growth. Therefore, the effect of SO 4
2-
on calcite dissolution is attributed to the
adsorption of Mg
2+
on carbonates.
In addition, growth rates of calcite steps were found to depend on the aqueous Ca/CO3
2-

ratio, and the response of obtuse and acute steps to Ca/CO3
2-
ratio is highly variable (Stack and
Grantham, 2010; Nielsen et al., 2012). The effect of Ca/CO3
2-
ratio on calcite dissolution is still
an open question. However, it is highly possible that the Ca/CO3
2-
ratio may alter the gross
dissolution rate and etch pit morphology, especially near equilibrium, by affecting the back-
precipitation process even though the net dissolution process may not be significantly affected.
Therefore, the distinct Ca/CO3
2-
ratio between seawater and the low ionic strength water in

117
previous laboratory experiments can also potentially result in different dissolution behaviors at
the same saturation states.

4.4.3 The role of calcite surface properties on step velocity and bulk rate
In surface nucleation models, the speed of a moving step, v, is related to the kinetic
coefficient " and the solution saturation state W via (Chernov, 1984; Malkin et al., 1989):
# = ω"&
'
(1 −Ω)
where ω is the molecular volume of a molecule in the crystal (6.12×10
-29
m
3
molecule
-1
), and &
'

is the mineral solubility (2.59×10
22
atoms m
-3
). In this formulation, # (measurable on AFM) will
be a linear function of Ω if " is a constant. We set out to test this premise.
In our previous study (Naviaux et al., 2019), we assumed a constant " across W values
within the homogeneous etch pit spreading mechanism and the defect-assisted etch pit
mechanism (0 < W < 0.9), and that there was a different " for the step retreat mechanism (0.9 <
W < 1). From this assumption, we derived " based on the mechanistic framework in Dove et al.
(2005) using measured bulk dissolution rates. The reason Naviaux et al. assumed constant "
across the two mechanisms far from equilibrium was that although the two mechanisms initiate
differently, once started, they are assumed to procced via the same opening and spreading of 2D
pits. In addition, linear dependence of step velocity on saturation state in low ionic strength
solutions was reported by Xu et al. (2010), indicating a constant " across W. Naviaux et al. found
the " for homogenous 2D dissolution and defect-assisted dissolution (0 < W < 0.9) to be 5×10
-3
m s
-1
and a value of " = 3×10
-7
m s
-1
was reported for step propagation (0.9 < W < 1).
In this study, with our measured mean step velocity of acute and obtuse steps, the mean
kinetic coefficient for the two types of edges is calculated at different saturation states. Note that

118
because we use etch pits widening rates to calculate step velocity, the "s we obtain are only
between 0 < W < 0.9, for homogenous 2D dissolution and defect-assisted dissolution. Contrary to
our previous assumption in Naviaux et al. (2019), our results here show changing " across an W
range of 0.4 to 0.9 in seawater, with significantly higher values at lower W (Figure 4.12a). From
near equilibrium to W = 0.4, " increases from 1.4×10
-5
m s
-1
to 5.5×10
-4
m s
-1
, smaller than " =
5×10
-3
m s
-1
reported in Naviaux et al. (2019). The discrepancy between " measured in this
study and " calculated from bulk dissolution rates in Naviaux et al. (2019) could be due to
inherent differences in the nature of the calcite crystals, i.e. faster bulk dissolution rates for fine-
grained calcite powder used by Naviaux compared to large Iceland spar crystals we used in this
AFM work.
To further test whether " is fundamentally different between different dissolution
mechanisms, or in other words whether newly-opened etch pits and long straight step edges have
the same step velocity, we compare step velocities (and hence ") at long step edges and at
smaller etch pits during the same experiment, and show " as a function of initial etch pit width
(Figure 4.12b). " for step propagation well after the formation of etch pits is clearly not orders of
magnitude lower than when the etch pit initially forms (initial etch pit width = 0). Instead, the
initial step velocity during a pit formation and spreading appears slight lower, even though the
difference (10~20%) is within the error of the step velocity determination (Figure 4.12b and 4.2).
These findings suggest that our previous assumption of constant " across W within the
2D dissolution and defect-assisted mechanisms may not be true for Iceland spar dissolution in
seawater. One potential reason that step velocity vs. W is more linear in low ionic strength water
(Xu et al., 2010) than in seawater (this study) is due to the surface complexation processes
between the ions in seawater and the calcite surface. The complexation of W-dependent ion

119
species (e.g. PO4
3-
/HPO4
2-
/H2PO4
-
, SO4
2-
/HSO4
-
) may have effects on step velocity and therefore
generates the nonlinearity in the step velocity vs. W correlation, which yields changing ". Over
all, our new data shows that the kinetic coefficient " is not fundamentally different between
different mechanisms. Instead of changing abruptly at Ω
./010.23
, " changes continuously across
W. In other words, " is W-dependent instead of mechanism-dependent.

Figure 4.12 (a) Kinetic coefficient β as a function of undersaturation state (β calculated for each different dissolution
experiment). (b) β calculated for different etch pits and at different time points during one dissolution experiment (AF-20, W =
0.37). As noted, β is slightly lower (~30%) when an etch pit first forms.

4.4.4 The role of calcite surface properties on step velocity and bulk rate
Dissolution preferentially occurs at specific surface sites that are characterized as having
“excess surface energy”, including defects, dislocations, and grain boundaries (Holdren and
Berner, 1979; Helgeson et al., 1984; MacInnis and Brantley, 1992). Pit walls can be the source of
steps that emanate from the outskirts of the pits and travel across the crystal surface (Lasaga and
Lüttge, 2001, 2003). Etch pit nucleation at surface defects can be shallow at point defects, or
deep at dislocations (Liang et al., 1996). As a result, the spreading of defect-assisted etch pits can
be one monolayer or multi-layers deep.

120
During two dissolution experiments at similar saturation states (Figure 4.10a and 4.10b,
AF-18 and AF-5), distinct “bulk” dissolution rates were observed with significantly different
dissolution patterns on the calcite surface. Etch pits formed in AF-18 are one monolayer deep,
whereas pit depths in AF-5 are 0.3 to 6 nm (1 to 20 monolayers). Step retreat velocity at the top
monolayer of AF-5 is 4 times faster than AF-18 (Figure 4.10c). The step velocity deeper in the
etch pit wall is half of the rate for the top monolayer, indicating that dissolution within pits may
be diffusion-controlled. The difference in step velocity between the two surfaces, however, is
within the uncertainty of step velocity determination (Figure 4.2, Figure 4.6). Due to higher pit
density and pit depth, the surface area normalized “bulk” dissolution rate of AF-5 is higher than
AF-18 by 200 times, whereas the surface area per unit imaging area are similar due to the
significantly smaller scale of the z axis compared to x and y (Figure 4.10c). The large
discrepancy between the “bulk” rate calculated between AF-18 and AF-5 indicates the effect of
Angstrom-scale surface roughness on dissolution, and indicates the limitation in normalizing
dissolution rate to surface area, which supports previous findings of disproportionality of
dissolution rates to BET surface area (Gautier et al., 2001; Wolff-Boenisch et al., 2004). An
alternative way to quantify the Angstrom-scale surface roughness by edge length density is
suggested below in discussion section 4.4.5.

4.4.5 The correlation between edge length density and bulk dissolution rate
To explain the effect of surface roughness on albite dissolution kinetics observed in
Arvidson and Lüttge (2010), Lüttge et al. (2013) introduced the concept of “intrinsic” sources of
variation in rates in addition to “extrinsic” sources including temperature, saturation state, flow
rate, etc. This variability reflects a fundamentally stochastic interaction between a fluid and

121
crystal surface having heterogeneous defect distributions. Important intrinsic factors are
structural discontinuities of crystals and minerals that result in heterogeneities of surface energy,
which include step-edge sites, kink sites, 2-bonded chain atoms and terrace atoms. By simulating
the evolution of a dissolving particle’s surface (in terms of different sites) with a kinetic Monte
Carlo model, a large variation (2.5 orders of magnitude) in total dissolution rate over time was
observed. This variation reflects a direct influence of the diversity and distribution of surface
sites on dissolution rates. To complement the model result, we present here an alternative
experimental explanation of the phenomenon.
During dissolution on a crystal surface, the relationship between step velocity and bulk
rate can be described as:
5
'67'
(cm)× # :
;<
=
>×ℎ (;<)×@ :
A
;<
B
>
CDEC (;<
F
)
= G
HI3J
(
A
;<
F
∙=
)
where Ledge is the total length of the edge front, v is the step velocity, h is the monolayer
height (3×10
-8
cm), @ is the mineral density (@calcite = 2.71 g cm
-3
), area is the surface area of the
crystal surface, Rbulk is the bulk dissolution rate normalized to surface area. Reorganizing the
equation gives:
G
HI3J
#
:
A
;<
B
>= ℎ (;<)×@ :
A
;<
B
> ×
5
'67'
CDEC
(
1
;<
)
Therefore, plotting
L
MNOP
Q
versus
R
STUS
2/'2
should give a linear correlation with a slope of ℎ@.
R
STUS
2/'2
is termed as “edge length density” in the discussion below and is a measure of surface
roughness (but note it has a unit of length
-1
). Six sub-regions of four experiments at W ~ 0.5 are
selected to investigate the correlation of edge length density and dissolution rate (Figure 4.4).
Total edge lengths in the images are calculated using a MATLAB program and the edges

122
detected are shown as black curves on Fig. 4.3. For the six sub-regions,
L
MNOP
Q
and
R
STUS
2/'2
scale
linearly with R
2
= 0.97, and the correlation (dashed line) shows good agreement with theoretical
calculation based on what we assume to be mono-layer height and mineral density (yellow solid
line) (Figure 4.13, Table 4.4). This result suggests that edge length density is an ideal measure of
surface roughness, and therefore normalizing dissolution rate to edge length density will give a
rate independent of sample surface property.

Figure 4.13 Bulk dissolution rate / step velocity vs. edge length density. Circles are six distinct regions on the calcite surface
during four dissolution experiments at W~0.5. Dashed line is fit to the circles with R
2
=0.97. Yellow line is the theoretical
calculation of the correlation with a slope of ℎ@=8.13×10
-4
g cm
-3
/

µm
-1
. Note that experimental data points are in very good
agreement with theoretical calculation.

123

Table 4.4 Bulk dissolution rates, step velocities, and edge lengths for six sub-regions on the calcite (104) surface during dissolution experiments at W~0.5.
Experiment
No.
Subplot in
Figure 4.4
W Image
area
(µm
2
)
Surface
area /
image area
Total edge
length
(µm)
Edge length
per area (µm
-1
)
Bulk rate
(g cm
-2
day
-1
)
Step
velocity
(nm s
-1
)
Rateedge *
(g cm
-1
day
-1
)
AF-5 (a) 0.58 5.8 1.0055 107 18.45 1.21×10
-5
0.06 1.0×10
-10

AF-3 (b) 0.46 76.3 1.0003 72 0.94 7.25×10
-7
0.11 6.2×10
-11

AF-3 (b)+(c) 0.46 107.9 1.0003 117 1.08 1.58×10
-6
0.11 1.2×10
-10

AF-3 (c) 0.46 19.6 1.0026 173 8.83 8.51×10
-6
0.12 7.8×10
-11

AF-12 (d) 0.50 12.3 1.0005 20 1.62 1.64×10
-6
0.05 9.9×10
-11

AF-18 (e) 0.51 6.5 1.0001 3 0.47 6.28×10
-8
0.02 1.4×10
-11

* Rateedge =
#$%% '(%%)*+,'
-)-$* ,'., *,/.-0 × -(#,

124

A schematic illustration is shown to further interpret the concept of edge length density
normalized rate (Figure 4.14). Within a 2D area of A µm
2
, assuming there is a spreading etch pit
that is one monolayer deep and a µm × b µm large (Figure 4.14a). The total edge length is
{2(a+b)} µm, and the surface area is {A+2(a+b)h} µm
2
. If, within the same area, the etch pit is
two monolayers deep (Figure 4.14b) and the step velocity remains the same, the mass dissolved
per unit time (Rm) will double when dissolution is not limited by diffusion. Total edge length is
twice as large as the one monolayer pit, whereas the surface area becomes {A+4(a+b)h} µm
2

instead of doubling. As a result, normalizing dissolution rate to surface area does not give the
same value for the two scenarios, but normalizing to edge length gives the same result. This
simple illustration may explain the previous conundrum of discrepancies in BET area-
normalized dissolution rates between different samples, and indicates that edge length density is
a notably better measure of the Angstrom-scale surface roughness that BET area cannot capture.
Nevertheless, this interpretation only applies to dissolution that is not diffusion-limited. Deep
etch pits (steep cliffs) and highly porous surfaces may have inhibited dissolution within the
diffusion boundary layer. In addition, the determination of edge length density is difficult to
obtain and can only be acquired at the atomic scale over a limited region of a grain’s surface.

125

Figure 4.14 Schematic illustration of calcite dissolution at etch pit that is 1 monolayer deep vs. 2 monolayers deep. 2D area = A
µm
2
. h = monolayer thickness = 0.3 nm.

4.5 CONCLUSIONS
I report AFM observations of calcite dissolution in seawater for the first time and show
no significant difference of etch pit morphology between dissolution in seawater and low ionic
strength water. Etch pit density is higher in seawater and the rise of etch pit density happens at
higher W than in low ionic strength water, indicating a transition from defect-assisted dissolution
mechanism to homogeneous etch pit spreading mechanism at significantly higher W. Although
etch pit opening is enhanced, step velocities are inhibited at high and mid-undersaturation states
in seawater compared to low ionic strength water, leading to net lower bulk dissolution rates near
equilibrium. The ratio of obtuse to acute step velocities is around 5-10 for calcite dissolution in
seawater. The kinetic coefficient # is not constant across W and is higher at lower W. In addition,
surface roughness is found to affect bulk dissolution rate significantly, but only has a small effect
in step velocity. We further demonstrate that edge length density is a notably better measure of

126
Angstrom-scale surface roughness than surface area, and is a good indicator of dissolution active
sites.

ACKNOWLEDGEMENTS
This work was supported by NSF Ocean Acidification grants (numbers OCE1220600 and
OCE1220302), and USC Dornsife Doctoral Fellowship. The authors would like to thank Mina
Hong, Zibo Li and Liang Zhao for helpful discussions on AFM operations.

REFERENCES
Arvidson, R. S., & Lüttge, A. (2010). Mineral dissolution kinetics as a function of distance from
equilibrium–New experimental results. Chemical Geology, 269(1-2), 79-88.
Arvidson, R. S., Collier, M., Davis, K. J., Vinson, M. D., Amonette, J. E., & Lüttge, A. (2006).
Magnesium inhibition of calcite dissolution kinetics. Geochimica et Cosmochimica
Acta, 70(3), 583-594.
Arvidson, R. S., Ertan, I. E., Amonette, J. E., & Luttge, A. (2003). Variation in calcite
dissolution rates: A fundamental problem?. Geochimica et cosmochimica acta, 67(9),
1623-1634.
Berner, R. A. (1981). Kinetics of weathering and diagenesis. Rev. Mineral.;(United States), 8.
Bisschop, J., Dysthe, D. K., Putnis, C. V., & Jamtveit, B. (2006). In situ AFM study of the
dissolution and recrystallization behaviour of polished and stressed calcite
surfaces. Geochimica et cosmochimica acta, 70(7), 1728-1738.
Brady, P. V., Krumhansl, J. L., & Papenguth, H. W. (1996). Surface complexation clues to
dolomite growth. Geochimica et Cosmochimica Acta, 60(4), 727-731.
Brantley, S. L., Crane, S. R., Crerar, D. A., Hellmann, R., & Stallard, R. (1986). Dissolution at
dislocation etch pits in quartz. Geochimica et Cosmochimica Acta, 50(10), 2349-2361.
Chernov A.A. (1984). Modern Crystallography III: Crystal Growth. Springer Science &
Business Media.
Cubillas, P., Köhler, S., Prieto, M., Chaïrat, C., & Oelkers, E. H. (2005). Experimental
determination of the dissolution rates of calcite, aragonite, and bivalves. Chemical
Geology, 216(1-2), 59-77.
De Giudici, G. (2002). Surface control vs. diffusion control during calcite dissolution:
Dependence of step-edge velocity upon solution pH. American Mineralogist, 87(10),
1279-1285.
Dickson, A. G., & Millero, F. J. (1987). A comparison of the equilibrium constants for the
dissociation of carbonic acid in seawater media. Deep Sea Research Part A.
Oceanographic Research Papers, 34(10), 1733-1743.

127
Dickson, A. G., Wesolowski, D. J., Palmer, D. A., & Mesmer, R. E. (1990). Dissociation
constant of bisulfate ion in aqueous sodium chloride solutions to 250. degree. C. Journal
of Physical Chemistry, 94(20), 7978-7985.
Dong, S., Berelson, W. M., Rollins, N. E., Subhas, A. V., Naviaux, J. D., Celestian, A. J., ... &
Adkins, J. F. (2019). Aragonite dissolution kinetics and calcite/aragonite ratios in sinking
and suspended particles in the North Pacific. Earth and Planetary Science Letters, 515, 1-
12.
Dong, S., Subhas, A. V., Rollins, N. E., Naviaux, J. D., Adkins, J. F., & Berelson, W. M. (2018).
A kinetic pressure effect on calcite dissolution in seawater. Geochimica et Cosmochimica
Acta, 238, 411-423.
Dove, P. M., & Platt, F. M. (1996). Compatible real-time rates of mineral dissolution by atomic
force microscopy (AFM). Chemical Geology, 127(4), 331-338.
Dove, P. M., Han, N., & De Yoreo, J. J. (2005). Mechanisms of classical crystal growth theory
explain quartz and silicate dissolution behavior. Proceedings of the National Academy of
Sciences, 102(43), 15357-15362.
Freij, S. J., Putnis, A., & Astilleros, J. M. (2004). Nanoscale observations of the effect of cobalt
on calcite growth and dissolution. Journal of Crystal Growth, 267(1-2), 288-300.
Gautier, J. M., Oelkers, E. H., & Schott, J. (2001). Are quartz dissolution rates proportional to
BET surface areas?. Geochimica et Cosmochimica Acta, 65(7), 1059-1070.
Gratz, A. J., Manne, S., & Hansma, P. K. (1991). Atomic force microscopy of atomic-scale
ledges and etch pits formed during dissolution of quartz. Science, 251(4999), 1343-1346.
Gutjahr, A., Dabringhaus, H., & Lacmann, R. (1996). Studies of the growth and dissolution
kinetics of the CaCO3 polymorphs calcite and aragonite I. Growth and dissolution rates
in water. Journal of crystal growth, 158(3), 296-309.
Harstad, A. O., & Stipp, S. L. S. (2007). Calcite dissolution: Effects of trace cations naturally
present in Iceland spar calcites. Geochimica et Cosmochimica Acta, 71(1), 56-70.
Helgeson, H. C., Murphy, W. M., & Aagaard, P. (1984). Thermodynamic and kinetic constraints
on reaction rates among minerals and aqueous solutions. II. Rate constants, effective
surface area, and the hydrolysis of feldspar. Geochimica et Cosmochimica Acta, 48(12),
2405-2432.
Holdren Jr, G. R., & Berner, R. A. (1979). Mechanism of feldspar weathering—I. Experimental
studies. Geochimica et Cosmochimica Acta, 43(8), 1161-1171.
Kerisit, S., & Parker, S. C. (2004). Free energy of adsorption of water and metal ions on the
{1014} calcite surface. Journal of the American Chemical Society, 126(32), 10152-
10161.
Klasa, J., Ruiz-Agudo, E., Wang, L. J., Putnis, C. V., Valsami-Jones, E., Menneken, M., &
Putnis, A. (2013). An atomic force microscopy study of the dissolution of calcite in the
presence of phosphate ions. Geochimica et Cosmochimica Acta, 117, 115-128.
Kowacz, M., & Putnis, A. (2008). The effect of specific background electrolytes on water
structure and solute hydration: Consequences for crystal dissolution and
growth. Geochimica et Cosmochimica Acta, 72(18), 4476-4487.
Lasaga, A. C., & Lüttge, A. (2001). Variation of crystal dissolution rate based on a dissolution
stepwave model. Science, 291(5512), 2400-2404.
Lasaga, A. C., & Lüttge, A. (2003). A model for crystal dissolution. European Journal of
Mineralogy, 15(4), 603-615.

128
Lea, A. S., Amonette, J. E., Baer, D. R., Liang, Y., & Colton, N. G. (2001). Microscopic effects
of carbonate, manganese, and strontium ions on calcite dissolution. Geochimica et
Cosmochimica Acta, 65(3), 369-379.
Liang, Y., & Baer, D. R. (1997). Anisotropic dissolution at the CaCO3 (101̄4)—water
interface. Surface Science, 373(2-3), 275-287.
Liang, Y., Baer, D. R., McCoy, J. M., Amonette, J. E., & Lafemina, J. P. (1996). Dissolution
kinetics at the calcite-water interface. Geochimica et Cosmochimica Acta, 60(23), 4883-
4887.
Lippmann F. (1973) Sedimentary Carbonate Minerals. Springer-Verlag, Berlin.
Lüttge, A., & Arvidson, R. S. (2010). Reactions at surfaces: A new approach integrating
interferometry and kinetic simulations. Journal of the American Ceramic Society, 93(11),
3519-3530.
Lüttge, A., Arvidson, R. S., & Fischer, C. (2013). A stochastic treatment of crystal dissolution
kinetics. Elements, 9(3), 183-188.
MacInnis, I. N., & Brantley, S. L. (1992). The role of dislocations and surface morphology in
calcite dissolution. Geochimica et Cosmochimica Acta, 56(3), 1113-1126.
Malkin, A. I., Chernov, A. A., & Alexeev, I. V. (1989). Growth of dipyramidal face of
dislocation-free ADP crystals; free energy of steps. Journal of Crystal Growth, 97(3-4),
765-769.
McCoy, J. M., & LaFemina, J. P. (1997). Kinetic Monte Carlo investigation of pit formation at
the CaCO3 (1014) surface-water interface. Surface Science, 373(2-3), 288-299.
Mehrbach, C., Culberson, C. H., Hawley, J. E., & Pytkowicx, R. M. (1973). Measurement of the
apparent dissociation constants of carbonic acid in seawater at atmospheric
pressure1. Limnology and Oceanography, 18(6), 897-907.
Morse, J. W., & Arvidson, R. S. (2002). The dissolution kinetics of major sedimentary carbonate
minerals. Earth-Science Reviews, 58(1-2), 51-84.
Naviaux, J. D., Subhas, A. V., Rollins, N. E., Dong, S., Berelson, W. M., & Adkins, J. F. (2019).
Temperature dependence of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 246, 363-384.
Nielsen, L. C., DePaolo, D. J., & De Yoreo, J. J. (2012). Self-consistent ion-by-ion growth
model for kinetic isotopic fractionation during calcite precipitation. Geochimica et
Cosmochimica Acta, 86, 166-181.
Oelkers, E. H., Golubev, S. V., Pokrovsky, O. S., & Bénézeth, P. (2011). Do organic ligands
affect calcite dissolution rates?. Geochimica et Cosmochimica Acta, 75(7), 1799-1813.
Perry IV, T. D., Duckworth, O. W., McNamara, C. J., Martin, S. T., & Mitchell, R. (2004).
Effects of the biologically produced polymer alginic acid on macroscopic and
microscopic calcite dissolution rates. Environmental science & technology, 38(11), 3040-
3046.
Piana, S., Jones, F., & Gale, J. D. (2006). Assisted desolvation as a key kinetic step for crystal
growth. Journal of the American Chemical Society, 128(41), 13568-13574.
Plummer, L. N., Wigley, T. M. L., & Parkhurst, D. L. (1978). The kinetics of calcite dissolution
in CO2-water systems at 5 degrees to 60 degrees C and 0.0 to 1.0 atm CO2. American
journal of science, 278(2), 179-216.
Ruiz-Agudo, E., & Putnis, C. V. (2012). Direct observations of mineral-fluid reactions using
atomic force microscopy: the specific example of calcite. Mineralogical Magazine, 76(1),
227-253.

129
Ruiz-Agudo, E., Kowacz, M., Putnis, C. V., & Putnis, A. (2010). The role of background
electrolytes on the kinetics and mechanism of calcite dissolution. Geochimica et
cosmochimica acta, 74(4), 1256-1267.
Ruiz-Agudo, E., Putnis, C. V., Jiménez-López, C., & Rodriguez-Navarro, C. (2009). An atomic
force microscopy study of calcite dissolution in saline solutions: The role of magnesium
ions. Geochimica et Cosmochimica Acta, 73(11), 3201-3217.
Schott, J., Brantley, S., Crerar, D., Guy, C., Borcsik, M., & Willaime, C. (1989). Dissolution
kinetics of strained calcite. Geochimica et Cosmochimica Acta, 53(2), 373-382.
Shiraki, R., Rock, P. A., & Casey, W. H. (2000). Dissolution kinetics of calcite in 0.1 M NaCl
solution at room temperature: an atomic force microscopic (AFM) study. Aquatic
Geochemistry, 6(1), 87-108.
Sjoeberg, E. L., & Rickard, D. T. (1984). Calcite dissolution kinetics: surface speciation and the
origin of the variable pH dependence. Chemical Geology, 42(1-4), 119-136.
Smith, M. E., Knauss, K. G., & Higgins, S. R. (2013). Effects of crystal orientation on the
dissolution of calcite by chemical and microscopic analysis. Chemical Geology, 360, 10-
21.
Stack, A. G., & Grantham, M. C. (2010). Growth rate of calcite steps as a function of aqueous
calcium-to-carbonate ratio: independent attachment and detachment of calcium and
carbonate ions. Crystal Growth & Design, 10(3), 1409-1413.
Stipp, S. L. S., Eggleston, C. M., & Nielsen, B. S. (1994). Calcite surface structure observed at
microtopographic and molecular scales with atomic force microscopy
(AFM). Geochimica et Cosmochimica Acta, 58(14), 3023-3033.
Subhas, A. V., Adkins, J. F., Rollins, N. E., Naviaux, J., Erez, J., & Berelson, W. M. (2017).
Catalysis and chemical mechanisms of calcite dissolution in seawater. Proceedings of the
National Academy of Sciences, 114(31), 8175-8180.
Subhas, A. V., Rollins, N. E., Berelson, W. M., Dong, S., Erez, J., & Adkins, J. F. (2015). A
novel determination of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 170, 51-68.
Teng, H. H. (2004). Controls by saturation state on etch pit formation during calcite
dissolution. Geochimica et Cosmochimica Acta, 68(2), 253-262.
Teng, H. H., & Dove, P. M. (1997). Surface site-specific interactions of aspartate with calcite
during dissolution: Implications for biomineralization. American Mineralogist, 82(9-10),
878-887.
Teng, H. H., Chen, Y., & Pauli, E. (2006). Direction specific interactions of 1, 4-dicarboxylic
acid with calcite surfaces. Journal of the American Chemical Society, 128(45), 14482-
14484.
Uppström, L. R. (1974). The boron/chlorinity ratio of deep-sea water from the Pacific Ocean.
In Deep Sea Research and Oceanographic Abstracts (Vol. 21, pp. 161-162).
Van Heuven, S. M. A. C., Pierrot, D., Rae, J. W. B., Lewis, E., & Wallace, D. W. R. (2011).
MATLAB program developed for CO2 system calculations. ORNL/CDIAC-105b.
Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, US
Department of Energy, Oak Ridge, Tennessee, 530.
Vinson, M. D., & Lüttge, A. (2005). Multiple length-scale kinetics: an integrated study of calcite
dissolution rates and strontium inhibition. American Journal of Science, 305(2), 119-146.
Vinson, M. D., Arvidson, R. S., & Lüttge, A. (2007). Kinetic inhibition of calcite (1 0 4)
dissolution by aqueous manganese (II). Journal of Crystal Growth, 307(1), 116-125.

130
Wolff-Boenisch, D., Gislason, S. R., Oelkers, E. H., & Putnis, C. V. (2004). The dissolution rates
of natural glasses as a function of their composition at pH 4 and 10.6, and temperatures
from 25 to 74 C. Geochimica et Cosmochimica Acta, 68(23), 4843-4858.
Xu, J., Fan, C., & Teng, H. H. (2012). Calcite dissolution kinetics in view of Gibbs free energy,
dislocation density, and pCO2. Chemical Geology, 322, 11-18.
Xu, M., & Higgins, S. R. (2011). Effects of magnesium ions on near-equilibrium calcite
dissolution: Step kinetics and morphology. Geochimica et Cosmochimica Acta, 75(3),
719-733.
Xu, M., Hu, X., Knauss, K. G., & Higgins, S. R. (2010). Dissolution kinetics of calcite at 50–70°
C: An atomic force microscopic study under near-equilibrium conditions. Geochimica et
Cosmochimica Acta, 74(15), 4285-4297.

131

CHAPTER 5: A MECHAMISTIC STUDY OF THE EFFECT OF
CARBONIC ANHYDRASE ON CALCITE DISSOLUTION AND ITS
IMPLICATIONS FOR THE OCEAN CARBON CYCLE

This chapter is a manuscript in preparation as:
Dong, S., Berelson, W. M., Teng, H. H., Rollins, N. E., Pirbadian, S., El-Naggar, M. Y., &
Adkins, J.F. (in preparation). A mechanistic study of the effect of carbonic anhydrase on calcite
dissolution and its implications for the ocean carbon cycle.

ABSTRACT
Carbonic anhydrase (CA) is ubiquitous in aqueous and terrestrial life, primarily acting to
regulate pH and CO2 balance inside, or surrounding, organisms by catalyzing the equilibrium
interconversion between CO2, water and the dissociated ions of carbonic acid. Our work shows
that CA also promotes calcite dissolution in seawater (Subhas et al., 2017); i.e. for the same
degree of undersaturation, calcite dissolves 2-150 times faster in waters containing 0.01-0.04 mg
mL
-1
CA. Understanding the catalytic mechanism of enhanced dissolution will facilitate the
proper estimate of dissolution fluxes in natural environments and the marine alkalinity cycle.
This study is the first to use atomic force microscopy (AFM) to directly observe calcite
dissolution in CA-bearing solution. No significant difference in calcite step retreat velocity is
observed with CA compared to water of the same saturation state without CA. However, CA is
found to enhance dissolution when in contact with, or is very close (~nm) to, the calcite surface.
The “etch pits” that CA generates by randomly touching the surface enhance dissolution rate
most pronouncedly near equilibrium when etch pits cannot form in the absence of CA. The

132
possible catalytic mechanism is through the adsorption of CA on the calcite surface, followed by
proton transfer from the CA catalytic center to the calcite surface during CO2 hydration.

5.1 INTRODUCTION
The enzyme carbonic anhydrase catalyzes the equilibration of CO2 and H2CO3. It is
ubiquitous through prokaryotes and eukaryotes (Moroney et al., 2001) and crucial to many
physiological processes involving CO2. Biological processes which require either rapid
production or consumption of CO2 require enzymatic catalysis of the CO2 hydration reaction,
and CA has been located at the site of biomineralization in many marine calcifiers (Miyamoto et
al., 1996; Tambutte et al., 2007; Rahman et al., 2007). In addition to its effects in vivo, CA has
also been experimentally shown to catalyze the dissolution kinetics of calcium carbonate, in karst
environments far from equilibrium (Liu, 2001; Liu et al., 2005), and in seawater near equilibrium
(Subhas et al., 2017). The elevated carbonate dissolution rates in the presence of CA may have
critical, yet underestimated, impact on the rate of chemical weathering (including carbonate rock
dissolution and silicate weathering) (Liu et al., 2005; Thorley et al., 2015; Xie and Wu, 2013)
and the rate at which alkalinity is recycled in the ocean (Subhas et al. (in review)). In fact, the
substantial amount of calcium carbonate dissolution in mid-depths (600 to 900 m) of the water
column in the North Pacific (Feely et al., 2002; Berelson et al., 2007; Dong et al., 2019) is
suggested to occur within confined acidic environments such as diffusively limited marine snow
aggregates or the guts of zooplankton (Dong et al., 2019), and this leaves open the possibility
that CA produced by marine organisms can catalyze the dissolution of calcium carbonate in these
confined settings in situ (Subhas et al., in review).
Isoforms of CA can be found intracellularly and extracellularly, bound to outer cell walls
or membranes, or untethered in organisms (Elzenga et al., 2000; Martin and Tortell, 2008;

133
Mustaffa et al., 2017). The quantification of CA catalysis on carbonate dissolution in laboratory
experiments was measured as a function of the equivalent bulk CA (bovine) concentration (Liu
et al., 2005; Subhas et al., 2017) in solution. Whether the catalysis requires direct interaction of
CA and the carbonate surface or occurs by altering local water chemistry, and if the latter, how
“local” is the effect, has not been determined.
This study uses AFM to investigate how CA catalyzes calcite dissolution in seawater.
The catalytic mechanism of CA is discussed based on the observed dissolution patterns in CA-
free and CA-bearing seawater, and these results both deepen our understanding of CA catalysis
mechanisms and shed light on the role of calcite dissolution in CA in the context of the global
carbon and alkalinity cycles.

5.2 METHODS
Calcite (104) surfaces were obtained by using a razor blade to cleave a large crystal of
optical-quality Iceland spar. The experimental solution was standard reference Dickson seawater
(https://www.nodc.noaa.gov/ocads/oceans/Dickson_CRM/batches.html), acidified to a desired
saturation state by adding HCl. The undersaturated seawater was prepared in and stored in gas-
impermeable bags with no headspace. Dissolved Inorganic Carbon (DIC) and alkalinity were
measured to determine W (using carbonate equilibria parameters and formulations described in
Dong et al. (2018) as prescribed in the worksheet CO2SYS). In CA-bearing seawater,
lyophilized carbonic anhydrase from bovine erythrocites purchased from Sigma Aldrich (C2624)
was added to seawater with the total [CA] = 0.04 mg mL
-1
. This concentration is one of the
values used by Subhas et al. (2017) who demonstrated a large effect at this level.

134
Experiments were conducted at 21℃ and 1 atm. In situ fluid cell imaging was conducted
using an Asylum Research Cypher ES Environmental Atomic Force Microscope in Contact
Mode. Two commercially available AFM probes were used; Arrow UHFAuD from Asylum
Research (https://afmprobes.asylumresearch.com/arrow-uhfaud.html), and SNL-10 from Bruker
(https://www.brukerafmprobes.com/p-3693-snl-10.aspx). Because the Cypher AFM has a gas
headspace in the fluid cell (~3 mL), the headspace was manually adjusted by adding CO2-
adjusted gas that was in equilibrium with the solution. Measured DIC and alkalinity of the inflow
and outflow solution showed that W was within the error of each other (Dong et al., in
preparation), thus demonstrating that degassing of CO2 during the measurements did not alter the
saturation state during dissolution. All experiments were conducted at a flow rate of 15 mL h
-1
.
At this flow rate, the water residence time in contact with the mineral surface is ≤1 minute.
Step retreat velocities were measured as half of the etch pit widening rates to eliminate
the potential artifact of drifting between images, and this pit-widening velocity equals the
average of acute and obtuse step velocities. Uncertainty in step velocity was determined as the
standard error of step velocities at 1-8 different etch pits and at 2-6 different time periods. In
experiments that had no etch pits, step retreat velocity was determined as the movement of edge
fronts when no obvious drifting was detectable between images. More details of the experimental
materials and methods can be found in Dong et al. (in preparation).

5.3 RESULTS
5.3.1 Effect of dissolved CA on calcite step retreat velocity
The average step velocities of acute and obtuse edges in dissolution etch pits are similar
in CA-bearing (dissolved-phase CA) seawater (0.04 mg mL
-1
) compared to CA-free seawater

135
(Figure 5.1a). To eliminate the potential effect of calcite surface properties on dissolution rate,
three experiments at W ~ 0.86, 0.73, 0.51 were conducted by using the same Iceland spar sample
and flowing CA-bearing seawater immediately after CA-free seawater. The slight difference in
W between the CA-free and CA-bearing seawater is corrected for by normalizing the rates to the
same W using the rate-undersaturation correlation in Figure 5.1a. The ratios of step velocity with
dissolved-phase CA to without CA are not significantly higher than 1 for all three saturation
states (Figure 5.1b). In the AFM images of these experiments, calcite surfaces show no
particulate CA, indicating that CA is present in the dissolved phase. However, according to
Subhas et al. (2017), CA enhances dissolution rate of calcite across all saturation states, and the
effect is most pronounced close to equilibrium. In Subhas et al. (2017), a rate increase of ~2.5
orders of magnitude at [CA] = 0.04 mg mL
-1
is observed at W = 0.85; whereas far from
equilibrium, the enhancement is a factor of 2~3. Because I find no significant difference in
calcite dissolution step velocity in seawater with and without dissolved CA across a wide range
in W, I argue that the enhancement of calcite dissolution rate by the addition of CA is not due to
a change in bulk seawater chemistry.

136

Figure 5.1 (a) Average step velocity of acute and obtuse edges vs. undersaturation with and without CA. Trend line is fit to all
“no CA” rates. (b) Ratios of step velocity with dissolved-phase CA to without CA. The absolute step velocities at W ~ 0.86 and
0.74 are not included in Figure 5.1a, because the step velocities are not the average of acute and obtuse step retreat velocity. The
arrows point to comparison points obtained at W ~ 0.53.

5.3.2 Effect of CA aggregate on dissolution through CA-calcite contact
The conclusion formulated above runs directly in contrast with bulk dissolution
experiments previously conducted (Subhas et al., 2017). However, it has been noted that sea
water that has CA added becomes very surfactant-like and even after filtering, coagulated
particles can be seen in this water. During the dissolution experiment in CA-bearing seawater at
W = 0.83, a CA aggregate appeared in the observation window of AFM and landed on the calcite
surface (Figure 5.2). Within 3 minutes, a massive amount of dissolution was observed
underneath (and around) the CA aggregate whereas no obvious change was found at an etch pit
not in contact with CA (black arrow). Because the shape of the CA aggregate changed slightly
and small aggregates were free-floating as seawater flowed through the AFM fluid cell, it was
difficult to identify whether the dissolution sites around the main CA aggregate were covered by
CA at some point within the 3 minutes of observation. Nevertheless, it is clear that CA-promoted

137
dissolution can be an extremely local process where CA touches or is close enough (~ nm
according to the color scale of the images) to the calcite surface. The rate at which the “etching”
happens at CA-calcite interface is greater than 1 monolayer per minute.

Figure 5.2 (a) CA aggregates in contact with the calcite surface and their effect on dissolution at Ω = 0.83. The straight horizontal
dark stripes near the CA aggregate are artifacts in the AFM imaging process that attempt to balance the total grayscale in
individual rows, and are not trenches on the surface. The irregular-shaped dark areas, however, are real etched features (yellow
arrows). (b) Schematic illustration of CA-induced dissolution on the calcite surface. Note that some CA blobs did not induce
dissolution (e.g. the blob that appeared at 2 min next to the black arrow in Figure 5.2a), likely because they did not touch the
surface.

138
This phenomenon we describe as contact-associated dissolution was further confirmed by
the following experiment: A droplet of CA-free seawater at W = 0.56 was added on top of a
calcite crystal (the droplet sat still on the mineral), and a large coagulated CA aggregate was
manually picked out from CA-bearing seawater with tweezers and placed on the dry part of the
calcite surface without touching the seawater droplet. I first scanned the calcite surface in CA-
free seawater to obtain the initial morphology of the area. Then I moved the AFM probe to the
CA aggregate (soft) and submerged it into the CA for ~10 seconds. I subsequently moved the
AFM probe to the area that was scanned previously and scanned again. CA attached to the probe
immediately created non-rhombic etching patterns (Figure 5.3a and 5.3b, “After (0 min)”,
indicated by yellow arrows), which subsequently reshaped to rhombic pits when CA floated
away from the imaging tip (Figure 5.3b).

139

Figure 5.3 Dissolution patterns observed by dipping the AFM probe into a CA aggregate and moving to a CA-free area to scan.
“Before” means before dipping the probe into CA; “After (0 min)” means the image obtained immediately after the probe is
coated with CA; “After (x min)” means x minutes after continuous scanning. Note that Figure 5.3a (before) has a large scan area,
and Figure 5.3a (after) is only a sub-area of the initial surface.

140
5.4 DISCUSSION
5.4.1 Interpretations of bulk dissolution rate vs. undersaturation correlation w/wo CA
From the experiments above, I conclude that the catalysis effect of CA on carbonate
dissolution is attributed to the creation of “etch pits” when CA touches the carbonate surface.
The removal of one calcite monolayer that is in contact with CA happens within seconds (<1
minute). These observations shed interesting light on the interpretations for the bulk rate –
undersaturation correlation w/wo CA reported in Subhas et al. (2017) and also discussed by
Dong et al. (2018) and Naviaux et al. (2019) (Figure 5.4). Far from equilibrium (orange area),
etch pits can form homogeneously on calcite surfaces in the absence of CA; whereas near
equilibrium (green area), the chemical potential of the solution is smaller than the free energy
change upon nucleation of an etch pit, and dissolution only occurs at existing edge fronts as step
retreats. Between these two mechanisms, dissolution is dominated by the formation of defect-
assisted etch pits. The transition between dominate dissolution mechanisms is indicated by the
change of slope in uncatalyzed seawater dissolution rates, and dissolution near equilibrium falls
below the trend line of the rates far from equilibrium because the generation of new edges
(dissolution-active sites) is missing close to equilibrium. Our key new understanding is that CA
in contact with calcite surfaces immediately opens new etch pits for enhanced dissolution.
The etching patterns CA creates when touching the calcite surface can be viewed as the
formation of “etch pits” regardless of solution chemistry. The effect of this “etch pit” generation
on bulk dissolution rate at W values far from equilibrium is to enhance dissolution rate by a
factor of 2~3 at [CA] = 0.04 mg mL
-1
. The addition of CA near equilibrium, however, increases
dissolution rate substantially by creating new edge fronts for subsequent dissolution. Therefore,
the fact that CA elevates bulk dissolution rates more pronouncedly near equilibrium is because

141
CA opens up “etch pits” when dissolution rate is limited by the formation of etch pit due to low
chemical potential of the solution.
Our observation generally agrees with, but slightly modifies the theoretical analysis of
the physical-chemical parameters reported in Subhas et al. (2017), based on the crystal growth
and dissolution theory in Dove et al. (2005). Subhas et al. proposed that the addition of CA
increases either the density of nucleation sites (ns) or the rate of step retreat (β), and meanwhile
decreases the free energy barrier to etch pit nucleation (α) by increasing the concentration of
carbonic acid at defects on the calcite surface near equilibrium. Our new results show that CA
increases the density of nucleation sites ns, but does not change step retreat velocity #. (
decreases because CA produces etch pits and lowers the energy barrier for pit nucleation in
defect-assisted and step retreat dominated regions.

Figure 5.4 The relationship between saturation state, CA concentration, and calcite dissolution rate in bulk dissolution
experiments in seawater reported in Subhas et al. (2017). The background colors of green, white, and orange indicate the
transition of dominating dissolution mechanisms from step retreat, to defect-assisted, to homogeneous etch pit spreading as W
decreases.

142

5.4.2 Possible mechanism for the CA catalysis on calcite dissolution
Based on the observation that calcite dissolution is only enhanced where CA sits on the
mineral surface, a mechanistic analysis of this local process is provided below. First, the
attachment of CA on the calcite surface requires the hydrophilic sites on CA (electrically
charged) to interact with the calcite surface. Semi-empirical quantum-mechanical calculations
show the net charges of different side-chain atoms in bovine carbonic anhydese II (BCA II) at
pH 7.5 (Saito et al., 2004), and the interactions between these partially-charged side-chain atoms
and the Ca
2+
/CO3
2-
on the calcite surface likely drive the adsorption of CA to calcite.
Once CA sits on the calcite surface, protons produced during CO2 hydration will etch the
calcite and generates local dissolution patterns that follow the shape of CA. More specifically,
the hydration of CO2 generates HCO3
-
and H
+
, and the proton transferred from the CA catalytic
center to the bulk solution is potentially the key to enhanced calcite dissolution. Therefore, to
investigate the possible mechanism for the CA catalysis on calcite dissolution, it is first essential
to understand the CA catalysis of CO2 hydration. While understanding of the catalysis
mechanism of CO2 hydration by CA is most detailed for human carbonic anhydrase II (HCA II),
the available evidence suggests that all members of the (-CA family (including BCAs) share the
same mechanism (Krishnamurthy et al., 2008). The active site of CA contains a Zn
II
ion with a
bound hydroxyl group (Zn
II
-OH) surrounded by three histidine residues held in a distorted
tetrahedral geometry (Figure 5.5a). Computational studies suggest that CO2 is not coordinated to
the Zn
II
but instead binds weakly in a hydrophobic region 3-4 Å away from the Zn
II
complex
(Liang and Lipscomb, 1990; Krebs and Fierke, 1993). Evidence suggests that the Zn
II
-bound
hydroxy group attacks CO2 to initiate hydrolysis and produce bicarbonate, which is displaced

143
from the Zn
II
ion by a molecule of water (Silverman and Lindskog, 1988). The Zn
II
-bound water
loses a proton to generate a new Zn
II
-OH for another round catalysis (Krebs and Fierke, 1993;
Krebs et al., 1993). It is generally accepted that this proton is shuttled to bulk solution by a series
of intramolecular and intermolecular proton-transfer steps (An et al., 2002; Ren et al., 1995). The
transfer of a proton from Zn
II
-bound water to the bulk solution is the rate-limiting step in
catalysis (Silverman and Lindskog, 1988; Becker et al., 2011). Proton transfer in HCA II
involves His64, which acts as a proton shuttle. The overall refined structure of BCA II is very
similar to that of HCA II, but it reveals a probable secondary proton-wire pathway that differs
from that of HCA II, which involves the dipole donor group of Gln91 (Saito et al., 2004).
Switching may occur between the different proton wires during the proton transfer process,
making use of the extensive water network leading away from Zn
II
(Figure 5.5b). The catalytic
turnover (kcat) for BCA III is 6400 s
-1
, and the second order rate constant approach the limit of
diffusional control (Ren et al., 1988; Krishnamurthy et al., 2008). The upper limit of the
enhanced dissolution of calcite when the mineral surface is in contact with CA is set by this
proton supply rate.
The etching patterns that follows the shape of CA aggregates indicate that protons
transferred from the CA catalytic center have a very limited effective range before buffered by
seawater. Another evidence is that, the etch pit indicated by the black arrow in Figure 5.2 was
close (within 0.1 µm) to a small CA aggregate at t = 2 min, but barely expanded throughout the
observation duration. The implication of the proximity required for CA to promote calcite
dissolution will be discussed in Section 5.4.3.

144

Figure 5.5 A mechanistic illustration of CA-promoted dissolution on calcite (104) cleavage surface. (a) the proton generated
during the catalysis of CO 2 hydration attacks calcite (104) surface and promotes dissolution; (b) the adsorption of CA on the
calcite (104) surface, and the intramolecular proton transfer from the catalytic center to the mineral surface.

5.4.3 Geological implications of the catalytic effect
CA activity in carbonate-rich environments such as coral reefs or sinking marine particles
would significantly enhance the rate at which alkalinity is cycled between solids and seawater
(Subhas et al., 2017; Eyre et al., 2018). The observation in this study that extremely close
proximity (<1 nm) is necessary for CA to promote CaCO3 dissolution, however, indicates that
estimates for the CA effect in natural environments should be based on the amount of CA readily
accessible to carbonate surfaces in extremely local scales.

145
CA exists in forms as both intracellular CA (iCA), and extracellular CA (eCA) that is
associated with the cell wall, plasma membrane, or periplasmic space (Badger, 2003; Moroney et
al., 1985; Nimer et al., 1999; Elzenga et al., 2000; Burkhardt et al., 2001; Hopkinson et al., 2013;
etc.). In natural seawater, eCA extracted from cell membranes is measured to be 0.10-0.67 nM
(10
-9
M) in the Baltic Sea (Mustaffa et al., 2017) and 0-0.17 pM (10
-12
M) in the North Pacific
(Subhas et al., in review). Subhas et al. (in review) also shows that ~90% of the CA activity is
externally bound in natural oceanic environments by showing similar CA activities between
sonicated and vortexed-only samples; and shows tight correlations of effective CA
concentrations and POC (particulate organic carbon) concentrations both in sinking fluxes and in
suspended materials. From this work by Subhas et al. (in review), it is clear that CA in the ocean
is associated with POC, but I have elsewhere noted that POC and PIC (particulate inorganic
carbon) are not strictly correlated along the same North Pacific transect (Dong et al., 2019).
Scaling measured CA activities by volumes of marine snow particles and pteropod shells gives
localized CA concentration of 0.06~0.32 mg CA g
-1
seawater, comparable to measured CA
activities in extracellular diatom boundary layer and their internal compartments (0.09~6.8 mg
CA g
-1
seawater, Hopkinson et al., 2011 and 2013), and are higher than those documented to
catalyze calcite dissolution in natural seawater (Subhas et al., 2017). Despite the new progress
made in quantifying eCA in natural seawater, whether the eCA associated with the cell
membranes can contact and interact with the carbonate mineral within the organisms is still an
open question and requires further investigation. Overall, to what extent can CA enhance PIC
dissolution and thus the marine alkalinity cycle highly depends on local availability of CA to
PIC, and should be considered as an important yet still unconstrained aspect of oceanic carbon
cycling.

146

5.5 CONCLUSIONS
From our novel use of AFM, no significant difference in step retreat velocity is observed
between calcite dissolution in CA-free and CA-bearing seawater. Enhanced dissolution is only
detected on a carbonate surface that is either in contact with, or in extreme proximity with (~ nm)
CA aggregates. The dissolution pattern closely resembles the outline of CA aggregates that are
observed sitting on the mineral surface. The interaction of CA and calcite generates irregular
etching patterns that do not follow the rhombohedral morphology of normal calcite etch pits. The
generation of etch pits by CA enhances calcite dissolution rate by a modest factor of 2~3 far
from equilibrium when etch pits can form homogeneously at the saturation state without CA.
Near equilibrium, when etch pits cannot form at the chemical potential of the solution, the etch
pits (edge fronts) CA introduces can elevate dissolution rates up to 2.5 orders of magnitude. The
possible catalytic mechanism is through the adsorption of CA on the calcite surface, followed by
proton transfer from the zinc ion to the calcite surface during CO2 hydration. These results point
out caution in estimating carbonate rock weathering and oceanic PIC dissolution with bulk CA
concentrations in natural environments because the key to enhanced dissolution is the contact
that occurs between CA and the mineral surface.

ACKNOWLEGEMENTS
This work was supported by NSF Ocean Acidification grants (numbers OCE1220600 and
OCE1220302), and USC Dornsife Doctoral Fellowship. I thank Adam V. Subhas for helpful
discussions in preparing experiments and the manuscript.

147
REFERENCES

An, H., Tu, C., Ren, K., Laipis, P. J., & Silverman, D. N. (2002). Proton transfer within the
active-site cavity of carbonic anhydrase III. Biochimica et Biophysica Acta (BBA)-
Proteins and Proteomics, 1599(1-2), 21-27.
Badger, M. (2003). The roles of carbonic anhydrases in photosynthetic CO 2 concentrating
mechanisms. Photosynthesis Research, 77(2-3), 83.
Becker, H. M., Klier, M., Schüler, C., McKenna, R., & Deitmer, J. W. (2011). Intramolecular
proton shuttle supports not only catalytic but also noncatalytic function of carbonic
anhydrase II. Proceedings of the National Academy of Sciences, 108(7), 3071-3076.
Berelson, W. M., Balch, W. M., Najjar, R., Feely, R. A., Sabine, C., & Lee, K. (2007). Relating
estimates of CaCO3 production, export, and dissolution in the water column to
measurements of CaCO3 rain into sediment traps and dissolution on the sea floor: A
revised global carbonate budget. Global Biogeochemical Cycles, 21(1).
Burkhardt, S., Amoroso, G., Riebesell, U., & Sültemeyer, D. (2001). CO2 and HCO3
-
uptake in
marine diatoms acclimated to different CO2 concentrations. Limnology and
Oceanography, 46(6), 1378-1391.
Dong, S., Berelson, W. M., Rollins, N. E., Subhas, A. V., Naviaux, J. D., Celestian, A. J., Liu,
X., Turaga, N., Kemnitz, N.J., Byrne, R.H., & Adkins, J. F. (2019). Aragonite dissolution
kinetics and calcite/aragonite ratios in sinking and suspended particles in the North
Pacific. Earth and Planetary Science Letters, 515, 1-12.
Dong, S., Berelson, W.M., Adkins, J.F., Rollins, R.E., Naviaux, J.D., Pirbadian, S., El-Naggar,
M.Y., Teng, H.H. (in preparation). An atomic force microscopy study of calcite
dissolution in seawater and the dependence of dissolution rate on surface properties.
Dong, S., Subhas, A. V., Rollins, N. E., Naviaux, J. D., Adkins, J. F., & Berelson, W. M. (2018).
A kinetic pressure effect on calcite dissolution in seawater. Geochimica et Cosmochimica
Acta, 238, 411-423.
Dove, P. M., Han, N., & De Yoreo, J. J. (2005). Mechanisms of classical crystal growth theory
explain quartz and silicate dissolution behavior. Proceedings of the National Academy of
Sciences, 102(43), 15357-15362.
Elzenga, J. T. M., Prins, H. B., & Stefels, J. (2000). The role of extracellular carbonic anhydrase
activity in inorganic carbon utilization of Phaeocystis globosa (Prymnesiophyceae): a
comparison with other marine algae using the isotopic disequilibrium
technique. Limnology and Oceanography, 45(2), 372-380.
Elzenga, J. T. M., Prins, H. B., & Stefels, J. (2000). The role of extracellular carbonic anhydrase
activity in inorganic carbon utilization of Phaeocystis globosa (Prymnesiophyceae): a
comparison with other marine algae using the isotopic disequilibrium
technique. Limnology and Oceanography, 45(2), 372-380.
Eyre, B. D., Cyronak, T., Drupp, P., De Carlo, E. H., Sachs, J. P., & Andersson, A. J. (2018).
Coral reefs will transition to net dissolving before end of century. Science, 359(6378),
908-911.
Feely, R. A., Sabine, C. L., Lee, K., Millero, F. J., Lamb, M. F., Greeley, D., ... & Ono, T.
(2002). In situ calcium carbonate dissolution in the Pacific Ocean. Global
Biogeochemical Cycles, 16(4), 91-1.

148
Hopkinson, B. M., Dupont, C. L., Allen, A. E., & Morel, F. M. (2011). Efficiency of the CO2-
concentrating mechanism of diatoms. Proceedings of the National Academy of
Sciences, 108(10), 3830-3837.
Hopkinson, B. M., Meile, C., & Shen, C. (2013). Quantiication of Extracellular Carbonic
Anhydrase Activity in Two Marine Diatoms and Investigation of Its Role. Plant
Physiology 162 (2), 1142-1152.
Krebs, J. F., & Fierke, C. A. (1993). Determinants of catalytic activity and stability of carbonic
anhydrase II as revealed by random mutagenesis. Journal of biological chemistry, 268(2),
948-954.
Krebs, J. F., Rana, F., Dluhy, R. A., & Fierke, C. A. (1993). Kinetic and spectroscopic studies of
hydrophilic amino acid substitutions in the hydrophobic pocket of human carbonic
anhydrase II. Biochemistry, 32(17), 4496-4505.
Krishnamurthy, V. M., Kaufman, G. K., Urbach, A. R., Gitlin, I., Gudiksen, K. L., Weibel, D.
B., & Whitesides, G. M. (2008). Carbonic anhydrase as a model for biophysical and
physical-organic studies of proteins and protein− ligand binding. Chemical
reviews, 108(3), 946-1051.
Liang, J. Y., & Lipscomb, W. N. (1990). Binding of substrate CO2 to the active site of human
carbonic anhydrase II: a molecular dynamics study. Proceedings of the National
Academy of Sciences, 87(10), 3675-3679.
Liu, Z. (2001). Role of carbonic anhydrase as an activator in carbonate rock dissolution and its
implication for atmospheric CO2 sink. Acta Geologica Sinica-English Edition, 75(3),
275-278.
Liu, Z., Yuan, D., & Dreybrodt, W. (2005). Comparative study of dissolution rate-determining
mechanisms of limestone and dolomite. Environmental Geology, 49(2), 274-279.
Martin, C. L., & Tortell, P. D. (2008). Bicarbonate transport and extracellular carbonic
anhydrase in marine diatoms. Physiologia Plantarum, 133(1), 106-116.
Miyamoto, H., Miyashita, T., Okushima, M., Nakano, S., Morita, T., & Matsushiro, A. (1996). A
carbonic anhydrase from the nacreous layer in oyster pearls. Proceedings of the National
Academy of Sciences, 93(18), 9657-9660.
Moroney, J. V., Bartlett, S. G., & Samuelsson, G. (2001). Carbonic anhydrases in plants and
algae. Plant, Cell & Environment, 24(2), 141-153.
Moroney, J. V., Husic, H. D., & Tolbert, N. E. (1985). Effect of carbonic anhydrase inhibitors on
inorganic carbon accumulation by Chlamydomonas reinhardtii. Plant Physiology, 79(1),
177-183.
Mustaffa, N. I. H., Striebel, M., & Wurl, O. (2017). Extracellular carbonic anhydrase: Method
development and its application to natural seawater. Limnology and Oceanography:
Methods, 15(5), 503-517.
Naviaux, J. D., Subhas, A. V., Rollins, N. E., Dong, S., Berelson, W. M., & Adkins, J. F. (2019).
Temperature dependence of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 246, 363-384.
Nimer, N. A., Brownlee, C., & Merrett, M. J. (1999). Extracellular carbonic anhydrase facilitates
carbon dioxide availability for photosynthesis in the marine dinoflagellateprorocentrum
micans. Plant Physiology, 120(1), 105-112.
Rahman, M. A., Oomori, T., & Uehara, T. (2008). Carbonic anhydrase in calcified endoskeleton:
novel activity in biocalcification in alcyonarian. Marine Biotechnology, 10(1), 31-38.

149
Ren, X., Jonsson, B. H., Millqvist, E., & Lindskog, S. (1988). A comparison of the kinetic
properties of native bovine muscle carbonic anhydrase and an activated derivative with
modified thiol groups. Biochimica et Biophysica Acta (BBA)-Protein Structure and
Molecular Enzymology, 953, 79-85.
Ren, X., Tu, C., Laipis, P. J., & Silverman, D. N. (1995). Proton transfer by histidine 67 in site-
directed mutants of human carbonic anhydrase III. Biochemistry, 34(26), 8492-8498.
Saito, R., Sato, T., Ikai, A., & Tanaka, N. (2004). Structure of bovine carbonic anhydrase II at
1.95 Å resolution. Acta Crystallographica Section D: Biological Crystallography, 60(4),
792-795.
Silverman, D. N., & Lindskog, S. (1988). The catalytic mechanism of carbonic anhydrase:
implications of a rate-limiting protolysis of water. Accounts of Chemical Research, 21(1),
30-36.
Subhas, A. V., Adkins, J. F., Dong, S., Rollins, N. E., Berelson, W. M. (in review). The carbonic
anhydrase activity of sinking and suspended particles in the North Pacific
Ocean. Limnology and Oceanography.
Subhas, A. V., Adkins, J. F., Rollins, N. E., Naviaux, J., Erez, J., & Berelson, W. M. (2017).
Catalysis and chemical mechanisms of calcite dissolution in seawater. Proceedings of the
National Academy of Sciences, 114(31), 8175-8180.
Tambutté, S., Tambutté, E., Zoccola, D., Caminiti, N., Lotto, S., Moya, A., ... & Adkins, J.
(2007). Characterization and role of carbonic anhydrase in the calcification process of the
azooxanthellate coral Tubastrea aurea. Marine Biology, 151(1), 71-83.
Thorley, R. M., Taylor, L. L., Banwart, S. A., Leake, J. R., & Beerling, D. J. (2015). The role of
forest trees and their mycorrhizal fungi in carbonate rock weathering and its significance
for global carbon cycling. Plant, cell & environment, 38(9), 1947-1961.
Xie, T., & Wu, Y. (2014). The role of microalgae and their carbonic anhydrase on the biological
dissolution of limestone. Environmental earth sciences, 71(12), 5231-5239.

150
CHAPTER 6: CONCLUSIONS
Accurate values and an understanding of controls on calcite and aragonite dissolution
kinetics in seawater are necessary to construct an accurate oceanic alkalinity and carbon cycle.
This information is also key to predict how fast the dissolution of CaCO3 in the marine
environment neutralizes anthropogenic CO2. Experiments and simulations in this dissertation
promote the understanding of both the kinetics and the mechanism of calcium carbonate
dissolution in seawater. Specifically, this thesis includes laboratory determinations of calcite and
aragonite dissolution kinetics in seawater, in situ measurements of aragonite dissolution and
particulate carbon fluxes in the North Pacific, microscale studies of calcite dissolution
mechanism in seawater, and investigation of the catalytic mechanism of carbonic anhydrase on
calcite dissolution.
Chapter 2 presents experiments demonstrating elevated dissolution rates of calcite in
seawater under high pressure compared to equivalent undersaturation states under ambient
pressure. The discrepancy of dissolution behavior under different pressures is greater than the
uncertainty of Ω based on uncertainties in DV, and is therefore attributed to a pressure effect on
the dissolution rate itself. The calcite dissolution rate in seawater at 700 dbar is 2-4 times faster
than at ambient pressure. The effect of pressure is attributed to a decrease in the surface energy
barrier and is likely due to changes in solid-liquid interfacial bonding under high pressure. The
exact underlying mechanism will require further experimental examinations of surface bonding.
The enhanced pressure-related dissolution of carbonate minerals may explain excess alkalinity
distributions in the ocean as this would effectively decrease the depth at which rapid dissolution
occurs.

151
Chapter 3 reports dissolution rates of synthetic aragonite in seawater both in the lab (at
5℃ and 21℃) and in the field (2~7℃) along a transect from Hawaii to Alaska, and shows
consistency between lab rates at 5℃ and field rates. As determined by floating sediment traps
deployed on this transect, sinking carbon fluxes are significantly higher in the subarctic gyre than
in the subtropical gyre, and yet there is no geographical trend in PIC/POC mole ratio, which is
0.2 ~ 0.6 for material sinking between 100~200 m. A comparison of fluxes at depths 100 m to
200 m indicates that 30~60% PIC dissolves between these depths with a simultaneous 20~70%
attenuation in POC fluxes. One significant implication of this study is that predictions based on a
strictly thermodynamic view of aragonite dissolution underestimate observed alkalinity excess
(Feely et al., 2004) and measured PIC attenuation. This conclusion is indicated by the
unreasonably low sinking rate of pteropod aragonite (< 1 m day
-1
) required to generate the excess
alkalinity observed in the North Pacific. Respiration-driven dissolution or metazoan/zooplankton
grazing, indicated by the simultaneous attenuation of PIC and POC in the sediment traps versus
depth, could produce an amount of dissolution comparable to that suggested by excess alkalinity
but not consistent with the depth distribution.
Chapter 4 is the first study that investigates calcite dissolution in seawater with AFM. We
find that a key dissolution feature, etch pit density, is higher in seawater and the abrupt increase
in etch pit density happens at higher W than in low ionic strength water, indicating a transition
from defect-assisted dissolution mechanism to homogeneous etch pit spreading mechanism at a
significantly higher W. Although etch pit opening is enhanced, step velocities are inhibited at
high and mid W in seawater compared to freshwater, leading to net lower bulk dissolution rates
near equilibrium. Surface roughness is found to affect bulk dissolution rate significantly but only
has a small effect in step velocity. I further demonstrate that edge length density is a notably

152
better measure of Angstrom-scale surface roughness than surface area, and is a good indicator of
dissolution active sites.
Chapter 5 presents interpretations of CA catalysis on calcite dissolution based on direct
observation. No significant difference in step retreat velocity is observed between calcite
dissolution in CA-free and CA-bearing seawater. Enhanced dissolution is only detected on a
carbonate surface that is either in contact with, or in extreme proximity (~ nm) with CA
aggregates. The generation of etch pits by CA enhances calcite dissolution rate by a modest
factor of 2~3 far from equilibrium when etch pits can form homogeneously at the saturation state
without CA. However, near equilibrium, when etch pits cannot form at the chemical potential of
the solution, the etch pits (edge fronts) CA introduces can elevate dissolution rates up to 2.5
orders of magnitude. Therefore, to what extent can CA enhance PIC dissolution and thus the
marine alkalinity cycle highly depends on local availability of CA to PIC surfaces, and should be
considered as an important yet still unconstrained aspect of oceanic carbon cycling.
This dissertation has provided better constraints on calcite and aragonite dissolution rates
in the ocean water column based on laboratory and in situ dissolution experiments. More future
work can be foreseen to testify and complement the results in this dissertation, and to further
deepen the understanding of oceanic carbonate dissolution budgets. First, we did not obtain
enough in situ dissolution rates of calcite during the 2017 cruise in the North Pacific to prove the
pressure effect observed in laboratory experiments. The internal spread of in situ calcite rates
was around a factor of 2, comparable to the pressure effect of 2~4 times. The variability in rates
likely resulted from varying temperature, chlorophyll (indicative of organic matter
concentration), and concentrations of assorted ions. The way to minimize the effects of different

153
physical and chemical parameters on dissolution rates, and to revisit the pressure effect in natural
oceanic conditions is to conduct more in situ experiments.
Seafloor dissolution of CaCO3 minerals has been reported to respond differently to
overlying seawater undersaturation due to the effect of diffusive boundary layer in sediments
(Boudreau, 2001 and 2013; Sulpis et al., 2017 and 2019) and porewater metabolic processes
(Berelson et al., 1990 and 1994; Hales and Emerson, 1996; Eyre et al., 2018). Meanwhile,
dissolution at the seafloor contributes to 85% of carbonate dissolution in the deep sea (Berelson
et al., 1994) and may constitute a primary feedback to ocean acidification over timescales of
centuries to tens of millennia (Sulpis et al., 2018). The application of our isotope labeling
technique to seafloor dissolution in future work will provide more accurate estimate of seafloor
dissolution rates and therefore a better constraint on dissolution fluxes in the water column
versus in the seafloor.
In addition, AFM has opened a new possibility to study the separate effects of different
ions (e.g. Mg
2+
, SO4
2-
), enzymes (e.g. CA), and physical parameters (e.g. pressure, temperature)
in seawater from a mechanistic way. Direct observations will provide solid proof and new
insights for the mechanistic analysis based on theoretical models in our previous work (Subhas et
al., 2017; Dong et al., 2018; Naviaux et al., 2019), and ongoing work on the effects of Mg
2+
,
SO4
2-
; and will be part of our future work.
The results of our kinetic and mechanistic studies will have critical geological and
oceanographic implications, which will be evaluated systematically in future work. First, one
major purpose of the series of our studies is to address the question of excess alkalinity in the
intermediate water column (600~900 m) in the North Pacific. While the role of aragonite
dissolution has been estimated to be negligible (Dong et al., 2019), the role of CA-enhanced

154
dissolution and respiration-driven dissolution in natural environments requires better
quantification. Second, the dissolution of calcium carbonate in the ocean is a process that buffers
ocean acidification and sequesters carbon dioxide. Therefore, the rate of PIC dissolution will
help evaluate the response of water chemistry and ecosystem changes to ocean acidification.
Additionally, the rate of natural carbonate weathering can also be more accurately determined.
Finally, our studies will guide the construction of global carbon and alkalinity budgets
considering the variability of these components in natural environments, both spatially in modern
ocean and temporally in glacial-interglacial timescales. With a more systematic evaluation of the
effects of seawater components on calcium carbonate dissolution, a more precise global carbon
cycle will be constructed in future work.

REFERENCES
Berelson, W. M., Hammond, D. E., & Cutter, G. A. (1990). In situ measurements of calcium
carbonate dissolution rates in deep-sea sediments. Geochimica et Cosmochimica
Acta, 54(11), 3013-3020.
Berelson, W. M., Hammond, D. E., McManus, J., & Kilgore, T. E. (1994). Dissolution kinetics
of calcium carbonate in equatorial Pacific sediments. Global Biogeochemical
Cycles, 8(2), 219-235.
Boudreau, B. P. (2001). Solute transport above the sediment-water interface. The Benthic
boundary layer: Transport processes and biogeochemistry, 104-126.
Boudreau, B. P. (2013). Carbonate dissolution rates at the deep ocean floor. Geophysical
Research Letters, 40(4), 744-748.
Dong, S., Berelson, W. M., Rollins, N. E., Subhas, A. V., Naviaux, J. D., Celestian, A. J., Liu,
X., Turaga, N., Kemnitz, N. J., Byrne, R. H., & Adkins, J. F. (2019). Aragonite
dissolution kinetics and calcite/aragonite ratios in sinking and suspended particles in the
North Pacific. Earth and Planetary Science Letters, 515, 1-12.
Dong, S., Subhas, A. V., Rollins, N. E., Naviaux, J. D., Adkins, J. F., & Berelson, W. M. (2018).
A kinetic pressure effect on calcite dissolution in seawater. Geochimica et Cosmochimica
Acta, 238, 411-423.
Eyre, B. D., Cyronak, T., Drupp, P., De Carlo, E. H., Sachs, J. P., & Andersson, A. J. (2018).
Coral reefs will transition to net dissolving before end of century. Science, 359(6378),
908-911.

155
Feely, R. A., Sabine, C. L., Lee, K., Berelson, W., Kleypas, J., Fabry, V. J., & Millero, F. J.
(2004). Impact of anthropogenic CO2 on the CaCO3 system in the
oceans. Science, 305(5682), 362-366.
Hales, B., & Emerson, S. (1996). Calcite dissolution in sediments of the Ontong-Java Plateau: In
situ measurements of pore water O2 and pH. Global Biogeochemical Cycles, 10(3), 527-
541.
Naviaux, J. D., Subhas, A. V., Rollins, N. E., Dong, S., Berelson, W. M., & Adkins, J. F. (2019).
Temperature dependence of calcite dissolution kinetics in seawater. Geochimica et
Cosmochimica Acta, 246, 363-384.
Subhas, A. V., Adkins, J. F., Rollins, N. E., Naviaux, J., Erez, J., & Berelson, W. M. (2017).
Catalysis and chemical mechanisms of calcite dissolution in seawater. Proceedings of the
National Academy of Sciences, 114(31), 8175-8180.
Sulpis, O., Boudreau, B. P., Mucci, A., Jenkins, C., Trossman, D. S., Arbic, B. K., & Key, R. M.
(2018). Current CaCO3 dissolution at the seafloor caused by anthropogenic
CO2. Proceedings of the National Academy of Sciences, 115(46), 11700-11705.
Sulpis, O., Lix, C., Mucci, A., & Boudreau, B. P. (2017). Calcite dissolution kinetics at the
sediment-water interface in natural seawater. Marine Chemistry, 195, 70-83.
Sulpis, O., Mucci, A., Boudreau, B. P., Barry, M. A., & Johnson, B. D. (2019). Controlling the
diffusive boundary layer thickness above the sediment–water interface in a thermostated
rotating-disk reactor. Limnology and Oceanography: Methods.

Asset Metadata

Creator Dong, Sijia (author)

Core Title Calcite and aragonite dissolution in seawater: kinetics, mechanisms, and fluxes in the North Pacific

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Geological Sciences

Publication Date 07/26/2019

Defense Date 06/04/2019

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag aragonite,CaCO₃ dissolution,calcite,carbonic anhydrase,excess alkalinity,kinetics,micro-scale mechanism,North Pacific,OAI-PMH Harvest,PIC,seawater,shallow dissolution

Format application/pdf (imt)

Language English

Advisor Berelson, William (committee chair), Adkins, Jess (committee member), El-Naggar, Moh (committee member), West, Joshua (committee member)

Creator Email dongsijia.nju@gmail.com,sijiadon@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-194470

Unique identifier UC11662248

Identifier etd-DongSijia-7644.pdf (filename),usctheses-c89-194470 (legacy record id)

Legacy Identifier etd-DongSijia-7644.pdf

Dmrecord 194470

Document Type Dissertation

Format application/pdf (imt)

Rights Dong, Sijia

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Abstract (if available)

Abstract Calcium carbonate minerals play a critical role in regulating geochemical cycles through dissolution and precipitation in aqueous environments due to the mineral’s wide occurrence and high reactivity on the earth surface. The formation of carbonate rocks on long timescales is canonically driven by the interaction of aqueous CO₂ and the cations from silicate rock weathering. Rivers deliver dissolved weathering products to the ocean in the form of alkalinity, which at steady state is removed via the production and burial of calcite and aragonite minerals. Estimates of open ocean calcification vary between 0.4∼1.8 Gt (1 Gt = 10¹⁵ g) PIC yr⁻¹, whereas only 0.1 Gt PIC yr⁻¹ is buried in deep-sea sediments. As a result, the majority of CaCO₃ produced in the surface open ocean must be dissolved either in the water column or in deep sea sediments. The location and rate of CaCO₃ dissolution, and thus where and how fast alkalinity returns to the ocean system, are crucial in determining the response of oceanic system to perturbations in either alkalinity or CO₂ input to the ocean-atmosphere system. ❧ This dissertation makes new measurements of calcite and aragonite dissolution in seawater, in an attempt to constrain dissolution kinetics and to understand the dissolution mechanisms in natural ocean environments with strictly controlled variables in the laboratory. ❧ I first identify a kinetic pressure effect on calcite dissolution that cannot be explained by the influence of pressure on calcite stoichiometric solubility product (K ⃰ₛₚ). My work aims to address a simple question: is dissolution rate under the influence of a saturation state controlled by ion activity product (IAP) the same as controlled by K ⃰ₛₚ? The answer is, surprisingly, no. The enhancement in dissolution rate is a factor of 2-4 at 700 dbar compared to dissolution at the same Ω under ambient pressure (10 dbar). This kinetic pressure effect points out caution in applying lab-determined rate laws to ocean dissolution processes that are under high pressure, and implies that sinking particles would dissolve at shallower depth than previously thought. ❧ Next, I present laboratory and in situ aragonite dissolution rate measurements, and the particulate inorganic carbon (PIC) and particulate organic carbon (POC) fluxes and concentrations in sinking and suspended materials along a North Pacific transect. I show that the measured aragonite flux combined with the inorganic dissolution rate only account for a small fraction of the excess alkalinity observed in the North Pacific, and respiration-driven dissolution of PIC or metazoan/zooplankton driven dissolution is more likely the source of excess alkalinity. ❧ For my third PhD project, I utilize Atomic Force Microscopy (AFM) to investigate calcite dissolution mechanisms in seawater. One key finding is that there is significantly higher etch pit density in seawater than in freshwater, whereas step velocity in seawater is inhibited at high and mid Ω, leading to net lower bulk dissolution rate near equilibrium. I further demonstrate that a newly defined parameter, edge length density, is a notably better measure of Angstrom-scale surface roughness than surface area, and is a good indicator of dissolution active sites. ❧ Finally, the catalysis mechanism of carbonic anhydrase (CA) on calcite dissolution is investigated with AFM. I find that dissolution can be attributed to the adsorption of CA on the calcite surface, and the transfer of protons from CA catalytic center to the mineral. CA does not affect step retreat velocity but creates etch pits by contacting the calcite surface. Therefore, to what extent can CA enhance PIC dissolution and thus the marine alkalinity cycle highly depends on local availability of CA in contact with PIC, and should be considered as an important yet still unconstrained aspect of oceanic carbon cycling.