See also: Ocean CDR Center of Excellence — a broader map of the labs, companies, and funders working the marine CDR space.
Ocean carbon removal · the chemistry, drawn
The ocean keeps a two-column ledger. Every carbon scheme is a bet on one number.
Alkalinity in one column, dissolved carbon in the other. Their ratio decides whether the sea
inhales CO₂ or exhales it. Growing plankton, dissolving rock, building calcite shells — each one
is just a different way of nudging that ratio. Here is what actually happens, bond by bond.
The rest of this post is just me re-jogging my own memory and trying to break down the current
state of Ocean CDR. Posting in case anyone else is doing a similar deep dive and it can be useful.
Very subject to mistakes, as it's just me laying out thinking.
Meet the cell — two ledgers at once
A coccolithophore (celebrity species Emiliania huxleyi, “Ehux”) is one cell. Light: photosynthesis = sink into soft tissue (POC). Same hour: calcite plates (coccoliths, PIC) = source at the surface. Climate net is the difference.
Note that: carbon fixed is not carbon removed. A coccolithophore
building a beautiful calcite shell is emitting CO₂ to build it. Weathered rock is the
only route here that adds alkalinity outright — and alkalinity is the column that does the pulling.
01 / THE ONE IDEA
Two columns, and only their ratio matters
Seawater holds carbon in three chemical costumes at once. Which costume dominates —
and therefore whether CO₂ wants to leave the water or enter it — is set by a tug-of-war between
two bookkeeping totals.
Column 1 · DIC
Dissolved Inorganic Carbon. Every carbon atom
in the water, counted once, no matter its costume:
DIC = [CO₂*] + [HCO₃−] + [CO₃2−]
Push DIC up and the water becomes more likely to
exhale CO₂.
Column 2 · Total Alkalinity
Not "how basic the water is." It is a
charge count — the excess of proton-acceptors over proton-donors:
Alkalinity is a supply of negative charge that has to be balanced. When you add it,
the water can only stay electrically neutral by converting bicarbonate into carbonate — which drops the
free CO₂ concentration, which makes the sea suck CO₂ down out of the air. Every drawdown scheme
on this page is, underneath, a way of tipping this beam to the left.
~90%
of surface-ocean DIC is bicarbonate, HCO₃−
Roughly 9% carbonate, and well under 1% actual dissolved CO₂. The ocean is
essentially a bicarbonate solution.
~10×
the Revelle buffer penalty
Add 1% to DIC and surface pCO₂ rises ~10%. The ocean resists being loaded — and
that same stiffness is why alkalinity is such a powerful lever.
~42×
more carbon in the ocean than the atmosphere
37,200 ± 200 Gt C dissolved versus ~891 Gt C in the air. The ratio was ~62 before
industrialisation — the air's share has been growing.
02 / THE CAST
The anions, and which ones the ledger counts
You asked how CO₂ sequestration changes depending on which seawater anion is involved.
It comes down to this: an anion contributes to alkalinity in proportion to how many protons it can
soak up. A doubly-charged carbonate ion is worth twice a bicarbonate. Sulfate and chloride, which
never grab a proton at seawater pH, are worth nothing at all.
TA +0
Carbon dioxide
CO₂(aq)
Neutral. Carries no charge, so contributes nothing to alkalinity — but it is
the species that exchanges with the air.
TA +1
Bicarbonate
HCO₃−
The workhorse. ~90% of ocean carbon. One negative charge, accepts one proton.
TA +2
Carbonate
CO₃2−
Doubly charged — worth two units of alkalinity. Raising this is what
pulls CO₂ from the air.
TA +1
Hydroxide
OH−
Direct alkalinity. This is what dissolving lime or running an electrochemical
cell actually delivers.
TA +1
Borate
B(OH)₄−
The quiet second-place contributor — a few percent of TA, and the reason precise
alkalinity accounting is fiddly.
TA +1*
Nitrate
NO₃−
*Not itself alkalinity — but when plankton consume it, they leave a
negative charge behind and TA rises by 1 per nitrate taken up.
TA +0
Sulfate
SO₄2−
Abundant, doubly charged — and irrelevant. Its conjugate acid is so strong it
never takes a proton at pH 8. Charge without alkalinity.
TA +0
Chloride
Cl−
The most abundant anion in the sea by far, and a complete spectator. Proof that
"negative" and "alkaline" are different things.
⚠
The lesson hidden in this grid: an ion's charge is not its carbon value. Sulfate has two
negative charges and does nothing. Carbonate has two and does everything. What separates them is
proton affinity — the willingness to accept an H⁺ and thereby mop up acidity.
03 / THE LADDER
One carbon atom, three costumes — climbing by shedding protons
Dissolved CO₂ is not stable in seawater. It hydrates, then sheds protons one at a
time, climbing a two-rung ladder. Where the population sits on that ladder is set entirely by pH —
and pH is set by alkalinity.
The Bjerrum ladder · bond by bond
Each step to the right removes an H⁺ and adds a negative charge. Watch the bonds:
CO₂ is a rigid linear molecule with two double bonds. Hydration bends it into carbonic acid — a
tetrahedral carbon centre with two hydroxyl arms. Stripping the first proton gives bicarbonate, where the
remaining negative charge is delocalised across two oxygens (the dashed bonds). Stripping the second
gives the flat, symmetric, doubly-charged carbonate ion — the shape that fits into a calcite crystal.
Where the ocean actually sits
The Bjerrum plot. Modern surface seawater sits at pH ≈ 8.1 — the vertical line — deep in
bicarbonate territory. Note how close the line has drifted toward the left since 1750 (pH ≈ 8.2):
that leftward drift is ocean acidification, and it means less carbonate, which is precisely why
shell-builders are under stress and why the ocean's ability to absorb more CO₂ is weakening.
04 / THE MASTER MAP
Every carbon scheme, on one diagram
Here is the payoff. Plot any chemical process by what it does to the two columns —
how much it changes DIC (across) and how much it changes alkalinity (up). A single diagonal line
splits the whole plane into the ocean inhales and the ocean exhales. Nothing else
is needed.
The ΔTA / ΔDIC plane
The rule: if a process raises alkalinity faster than it raises dissolved
carbon (ΔTA > ΔDIC), it lands above the diagonal and seawater pCO₂ falls —
the sea pulls CO₂ down out of the air. If it lands below, pCO₂ rises and the sea gives CO₂
back. Note where calcification sits: firmly in the exhale half.
Note that silicate weathering products land exactly on the
line — the ocean neither gains nor loses, because the CO₂ was already captured on land, and the
ocean's job is only to hold it. And note that pure alkalinity
addition travels straight up: the single most efficient direction on this map.
●
Read the map once and the rest of this page is commentary. Alkalinity is the vertical axis, and
up is the direction of drawdown. Ocean alkalinity enhancement points straight up. Enhanced weathering
travels up the diagonal. Growing plankton moves sharply left. Building a shell moves down. That is the
whole field.
05 / THE PARADOX
Coccolithophores: the shell is made of carbon and made by emitting carbon
Coccolithophores are single-celled algae roughly five micrometres across that armour
themselves in intricate plates of calcite. Blooms of the best-known species —
Gephyrocapsa huxleyi, long called Emiliania huxleyi — can cover more than 100,000
km² and turn the sea milky turquoise, plainly visible from orbit. Their fossil plates built the White
Cliffs of Dover. And the act of building those plates releases CO₂ into the water.
Watch the atoms · calcification, step by step
Two bicarbonate ions and a calcium ion drift together.
Count the carbons. Two go in, riding on two bicarbonate ions.
One ends up locked in the solid calcite plate. One walks out as CO₂.
The cell has not removed carbon from the water — it has taken two dissolved carbons and converted one of
them into a gas that will equilibrate with the atmosphere.
ΔDIC −1 per CaCO₃
ΔTA −2 per CaCO₃
ΔTA / ΔDIC = 2 → below the line
Result pCO₂ RISES ↑
Alkalinity falls twice as fast as dissolved carbon. That is the entire mechanism. The carbonate
ion the cell consumes was worth two units of alkalinity; the carbon atom it consumes is worth one unit of
DIC. Take away twice as much of the pulling column as the pushing column, and the water's grip on CO₂
loosens.
One honest refinement before we go further: not all of that CO₂ escapes. Seawater
buffering reabsorbs part of it, so the real figure is about 0.6 moles of CO₂ released per mole
of CaCO₃ formed, not 1.0. That number has a name — ψ — and it turns out to be
the hinge on which the whole coccolithophore question swings. We come back to it in section 8. It also gets
worse as the ocean acidifies: ψ climbs toward 1.0 as CO₂ rises and water cools.
So why is chalk made of dead plankton if they emit carbon?
Because of a timescale switch, and it is the single most-confused point in this whole
subject.
Over decades to centuries — the timescale a carbon credit is sold on — calcification
is a CO₂ source. The released CO₂ is in surface water, in contact with the air, within
months.
Over tens of thousands to millions of years — the timescale of the White Cliffs of
Dover — calcite that reaches the seafloor and is buried removes calcium and carbonate from the ocean
permanently, and the geological weathering cycle rebalances the books. On that clock, carbonate
burial is a genuine long-term sink.
Both statements are true. They are answers to different questions. A company selling carbon removal is
answering the first question, not the second.
The counter-argument: calcite as ballast
There is a real effect pulling the other way, and honest accounting has to include it. Calcite is dense —
about 2.7 times the density of seawater. A sinking aggregate of soft organic tissue is barely denser than
the water around it and drifts down slowly, giving bacteria weeks to eat it and return its carbon to the
dissolved pool. Stud that aggregate with mineral plates and it drops like a stone, outrunning the bacteria
and delivering more of its organic carbon to depth. This is the ballast hypothesis, and it
is why the relationship between calcifiers and the carbon cycle is genuinely contested rather than simply
negative.
Two pumps, opposite signs
The soft-tissue pump moves carbon down and pCO₂
down. The carbonate counter-pump moves carbon down and
pCO₂ up. They run through the same organisms, in the same water, at the same time. Which one
wins is a question about a ratio — and that ratio is the subject of section 8.
06 / THE ROCK ROUTE
Weathering: the planet's own thermostat, run fast
Over geological time, rain that has absorbed CO₂ dissolves silicate rock, and the
resulting alkalinity washes to the sea where it holds carbon as bicarbonate. This is the slow feedback that
has kept Earth habitable for billions of years. Enhanced weathering is the proposal to run it on a human
schedule by grinding the rock up first.
Route A · Silicate weathering — the four-for-one deal
Olivine dissolution · Mg₂SiO₄
Forsterite olivine is the poster mineral because its stoichiometry is unusually generous:
one formula unit consumes four molecules of CO₂, converting each into a bicarbonate
ion that the ocean will hold for on the order of 100,000 years. The magnesium goes into solution as a
spectator cation whose positive charge is exactly what permits four extra negative bicarbonates to exist.
4 CO₂ consumed per Mg₂SiO₄
ΔTA +4 · ΔDIC +4
Theoretical max ~1.25 t CO₂ per t olivine
Durability ~10⁵ years as bicarbonate
Route B · Carbonate weathering — calcification, run backwards
Limestone dissolution · CaCO₃
Look closely: this is the coccolithophore reaction with the arrow reversed. Dissolving
limestone consumes a CO₂ and produces two bicarbonates — ΔTA +2, ΔDIC +1,
which lands above the diagonal. But half the carbon in the products came out of the rock, not the air,
so the net atmospheric benefit is only one CO₂ per formula unit — roughly a quarter
of what the same mass of olivine delivers.
Route C · Ocean alkalinity enhancement — skip the rock, add the base
The most direct version of all: manufacture alkalinity and put it in the sea. Slaked lime,
Ca(OH)₂, dissolves to give a calcium ion and two hydroxides. Electrochemical approaches split seawater
to make a sodium hydroxide stream and a matching acid stream that is disposed of separately. Either way
ΔDIC starts at zero and ΔTA jumps — a vector pointing straight up on the master
map, the most efficient direction available.
Alkalinity in, CO₂ follows
Adding hydroxide instantly converts bicarbonate into carbonate. That drops the free CO₂
concentration in the water, which opens a gradient across the sea surface, and over the following months
the atmosphere refills it. The removal is not the dissolving — the removal is the slow inhale
that follows. This is why "equilibration efficiency" is the number that matters, and why an
alkalinity plume released in the wrong place, or subducted below the surface before it can breathe,
delivers a fraction of its paper value.
Where it gets hard
Grinding costs energy. Dissolution rate scales with surface area, so finer is faster
— but milling to fine grades can eat a meaningful fraction of the CO₂ you are trying to capture.
The energy budget has to be counted against the removal.
Olivine carries passengers. Natural olivine contains nickel and chromium. Spread at
scale, those metals go into the coastal environment along with the alkalinity.
Push too hard and it snaps back. Add alkalinity too fast or too concentrated and the
water becomes supersaturated enough to precipitate CaCO₃ spontaneously — which is calcification,
which runs the reaction you just paid for in reverse. This "runaway precipitation" threshold is a hard
engineering constraint on how any alkalinity is delivered.
The inhale is invisible. Nothing is captured at the moment of release. The CO₂
crosses the sea surface over months, spread across a moving patch of ocean. There is no meter. There is
only a model.
07 / THE LIVING ROUTE
Growing plankton: the biggest carbon pump on Earth, and the hardest to sell
Every year, microscopic algae in the sunlit surface ocean fix roughly as much carbon as
all land plants combined. A fraction of that sinks out of the surface as dead cells, faecal pellets and
aggregates — "marine snow". This is the biological pump, and it is why the deep ocean holds more carbon
than the atmosphere.
Photosynthesis, in Redfield proportions
The stoichiometry of plankton growth is remarkably consistent — roughly 106 carbons to 16
nitrogens to 1 phosphorus, the Redfield ratio. Two things happen at once here.
Dissolved carbon plunges: ΔDIC −106. And alkalinity rises slightly, ΔTA +17,
because every nitrate ion the cell absorbs takes a negative charge out of solution with it. Both effects
push pCO₂ down — which is why this vector points so steeply into the inhale half of the master map.
⚠
The sentence that decides whether any plankton scheme is real: fixing carbon at the surface
lowers the DIC of that surface water — but nothing has been removed from the atmosphere until
the air refills that deficit across the sea surface. That exchange takes months to a year. If the
water sinks, mixes away, or the carbon is eaten and released before the inhale completes, the paper
removal never happened.
Depth is durability
Carbon that sinks is not gone; it is only away, and for a length of time set almost entirely by
how deep it got before something ate it. Sinking flux falls off sharply with depth — the
Martin curve — because bacteria and zooplankton intercept most of the falling material in the first few
hundred metres and respire its carbon straight back into dissolved form.
How long does it stay down?
The clock is not set by the sinking — it is set by the ventilation age of the water
the carbon lands in: how long until that parcel of water next touches the surface and can breathe. This
is why the depth a project can honestly claim is the single largest lever on the durability it can sell,
and why "below the thermocline" and "permanent" are very different statements.
The cautionary history: iron fertilisation
The idea of stimulating plankton to draw down carbon is not new. Thirteen major open-ocean iron
fertilisation experiments were run between 1993 and 2009. They reliably produced blooms. What they did
not reliably produce was measurable, durable carbon export — and the field's own verdict became
markedly more cautious as a result.
Export was small and wildly variable. Big blooms often produced little sinking flux;
most of the fixed carbon was recycled in the surface layer rather than exported.
Nutrient robbing. Nutrients consumed in one patch are nutrients unavailable
downstream. Some of the apparent local gain is productivity stolen from elsewhere.
Side effects at depth. Respiring a large pulse of organic matter consumes oxygen and
can generate nitrous oxide and methane — greenhouse gases far more potent than CO₂ per molecule.
Verification proved brutally hard. Attributing a change in a moving, mixing, patchy
ocean to your intervention rather than to weather is the central unsolved problem, and it is unsolved for
every ocean-biology approach, not just iron.
08 / THE DECIDING NUMBER
The rain ratio: how much shell per unit of flesh
Sections 5 and 7 described two pumps running in the same water with opposite signs.
Which one wins is settled by a single ratio: how much mineral carbon (PIC) a plankton community exports
per unit of organic carbon (POC). This is the rain ratio, and the threshold is not where
most people guess.
Where the number 1.65 comes from
Not all the CO₂ released by calcification escapes — seawater's
buffering absorbs some of it. Frankignoulle and colleagues (1994) named that fraction
ψ: the moles of CO₂ actually released per mole of CaCO₃ precipitated.
ψ ≈ 0.6 mol CO₂ per mol CaCO₃
threshold = 1 / ψ
≈ 1.65 PIC : POC
One unit of organic carbon fixed removes one
unit of DIC. One unit of calcite formed releases only 0.6. So it takes about 1.65 units
of calcite to cancel one unit of soft tissue — not 1.0, which is the shorthand you will often see
and which is off by about 65%.
One axis decides the sign
Measured PIC:POC across coccolithophore species spans roughly 0.17 to 2.08, and
Gephyrocapsa huxleyi — the species long known as Emiliania huxleyi — usually
sits between about 0.16 and 1.2. Which means most real coccolithophores land below the
threshold: taken as whole organisms, at the moment of growth, they are typically a modest net
CO₂ sink.
⚠
Both popular claims are wrong. "Coccolithophores lock carbon away in their shells" is wrong
— the shell-building step is a CO₂ source. But "coccolithophores are CO₂ emitters" is an
over-correction — the same cell photosynthesises, and for most species the organic half wins. The
honest statement is narrower and more useful: calcification is a carbon cost that the organism's
photosynthesis has to pay for, and how much is left over depends on the species, the season and the
strain.
Three further complications keep this from being a clean calculation:
The threshold itself is moving. ψ climbs toward 1.0 as water cools and as
CO₂ rises — so the tipping point falls as the ocean acidifies. The penalty on
calcification gets worse over time, not better.
Ballast cuts the other way. Klaas and Archer (2002), standardising sediment-trap data
from 52 sites below 1,000 m, found mineral ballast explains 85–90% of the variance
in deep organic carbon flux, with calcite the dominant carrier — about 0.072 g of organic carbon
delivered per gram of CaCO₃. A calcifier may cost you carbon at the surface and still deliver more
of it to depth.
Acidification is reshaping the organisms. Riebesell et al. (2000) found reduced
calcification under high CO₂; Iglesias-Rodriguez et al. (2008) found the opposite. Meyer and
Riebesell's 2015 meta-analysis of 23 studies and 48 experiments came down on the side of reduced
calcification in the two dominant species — but with four qualitatively different response shapes
across 14 strains. No single sentence about "how coccolithophores respond" survives contact with the data.
09 / SIDE BY SIDE
The routes, compared honestly
The same seven columns for every approach. The one to read first is
"what it does to the ledger" — it predicts the rest.
Approach
Mechanism
Effect on the ledger
CO₂ accounting
Durability
Maturity
The main doubt
10 / THE HARD PART
None of this is the hard part. Proving it happened is.
Everything above is textbook chemistry. It is not in dispute. The reason marine carbon
removal is difficult has almost nothing to do with the reactions and almost everything to do with the fact
that you cannot put a meter on the ocean. The removal happens as a diffuse, months-long
inhale across a moving patch of sea. There is no flue to measure. There is only a model with a measured
input.
⚠
Sit with the first two numbers together. Roughly 600,000 tonnes contracted against roughly
730 tonnes independently verified. That is not an accusation against anyone in particular — it is
simply what a market of forward promises looks like before the verification catches up.
Four things worth knowing about how this market actually behaves
11 / A COMPANY IN THIS SPACE
Gigablue: where their idea sits on this map
Gigablue is a marine carbon-removal company operating in New Zealand waters. Their approach
belongs to the biological pump family from section 7 — not the alkalinity family — which
means the questions that decide whether it works are the export-and-durability questions, not the
chemistry ones.
The concept, as described
An engineered particle is released at the surface. Algae grow on it, fixing dissolved carbon
by photosynthesis. The loaded particle sinks below the thermocline, carrying that carbon out of contact
with the atmosphere. It is the biological pump with a manufactured substrate — and note where the
atmospheric removal actually lives on this diagram: not in any of the four steps, but in the air-to-sea
inhale that has to refill the dissolved carbon the algae consumed.
What the company says
What is publicly documented
What independent scientists have said
This is contested territory and it would be dishonest to present it otherwise. The criticism below is
attributed; it is reporting on what named researchers and journalists have said, not a verdict of ours.
The questions that would settle it
⚠
Read this section as a case study in where the hard part lives. Nothing about growing algae on
a particle is chemically doubtful. The doubt is entirely about the two things this whole page has been
building toward: did the atmosphere actually inhale, and did the carbon actually stay down? Those
are the questions to ask of any marine CDR company, including this one.
WATCH
If you would rather be told than read
Four videos worth the time, chosen for accuracy rather than production values.