Membrane Separations, Extractions and Precipitations in Aqueous Solutions for CO2 Mineralisation

G5 Doctoral dissertation (article)

Internal Authors/Editors

Publication Details

List of Authors: Evelina Koivisto
Publisher: Åbo Akademi
Publication year: 2019
ISBN: 978-952-12-3857-4
eISBN: 978-952-12-3858-1


The global carbon dioxide (CO2) emissions have been growing rapidly
since the industrial revolution starting in the 18th century. This has
led to increasing CO2 concentrations in the atmosphere, which in turn
heats up the atmosphere as a result of CO2 being able to absorb heat.
Heating of the atmosphere has naturally taken place many times, so it is
not a new feature on Earth. The problem with the global heating today,
however, is that it apparently goes much faster. The annual increase of
CO2 in the atmosphere is about 100 times faster today than it was the
last time when natural heating occurred, around 300 000 years ago.

Different commitments, like the Kyoto Protocol, have the aim to
decrease the greenhouse gases to certain levels in different commitment
periods. The target for EU by 2020 is to reduce the emissions by 20%,
compared to the CO2-eq. emissions measured in 1990. On a longer-term
basis, Finland has the target to reach a carbon neutral society. Finland
is on its way, but this target will be challenging to reach. However,
finding options to decrease emissions further or develop options for CO2
storage is of highest importance in this context.

The only option for carbon capture and storage, CCS, in Finland is
through CO2 mineralisation. An option has been developed at Åbo Akademi
University during the last years, often referred to as the ÅA route. The
process extracts magnesium from magnesium silicates through a thermal
reaction at ≈ 400 oC, in the presence of ammonium sulphate serving as
flux salt. The extracted magnesium is thereafter precipitated as
magnesium hydroxide by the addition of ammonia, which is released in the
thermal reaction. Magnesium hydroxide will finally react with flue gas
in a pressurized fluidized bed reactor at 500 oC and a pressure of 20
bar to obtain magnesium carbonate, a stable compound that enables
storing CO2 safely. As an alternative ÅA route, the process could take
place in aqueous solutions with an initial extraction step where
ammonium bisulphate and/or ammonium sulphate are present. The final
carbonation would also take place in aqueous solution inserting CO2 gas
and ammonia.

The concept for the ÅA route is extensively studied but the
streamlining and the generation of input chemicals must still be
established. This thesis focused on several different methods, i.e. the
removal of excess water after dissolution (reverse osmosis),
precipitation and separation of precipitated products in one unit
(inclined settlers), extraction of magnesium from overburden rock from
mining activities (thermal solid/solid or aqueous extraction),
separation of two salts from each other (electrodialysis with monovalent
ion selective membranes), or producing acid and bases from input
chemicals used, which could be used in the process again (bipolar
membrane electrodialysis).

Two of the studied rocks, both containing significant amounts of
serpentinite, resulted in CO2 binding capacities of 292 and 260 kg
CO2/ton rock, respectively, in aqueous solution extraction. The
capacities for the same rock types were somewhat lower when applying the
thermal solid/solid extraction step; 240 and 207 kg CO2/ton rock, for
Hitura serpentinite and Vammala serpentinite, respectively. This
indicates that the aqueous extraction route seems to have a better
potential than the thermal solid/solid extraction step, at least under
the studied conditions.

Attempts to separate magnesium and iron precipitates in an inclined
tube settler were made in order to make the ÅA route more continuous.
Magnesium and iron must be stepwise precipitated and separated from the
process stream in the conventional ÅA route. The particle size of both
precipitates was found to be very small according to SEM analyses, which
affected the settling rate negatively. The settling rate for magnesium
hydrocarbonates, however, was more rapid. This was confirmed by SEM
images showing that the magnesium hydrocarbonates had a larger particle
size, which also affected the settling rate positively. Continuous
sedimentation and separation was still hard to achieve with the chosen
set-up, even at low pump speed as to avoid too turbulent conditions.

A reverse osmosis membrane was used in order to try to separate
water from the process stream. Some excess water must be used in order
get all reacted rock washed out from the kiln as the thermal solid/solid
route is applied. Reverse osmosis might therefore be an option to
control the water volumes. The osmotic pressure of typically 23 bar in a
solution after the magnesium extraction step must be overcome before
any separation can take place. No significant separations were achieved,
even with adjustments of the permeate stream composition replaced by
downstream aqueous solutions. At least with the available membrane, RO
will not be an option for controlling the amount of water used in the ÅA

An in-house built set-up of monovalent ion selective membrane
electrodialysis was used to see if the separation of ammonium sulphate
from ammonium bisulphate could be achieved. Initially, the compartment
width was 40 mm. Ion selective electrode analyses (ISE) were done to
measure the change in ammonium ions during the tests. These measurements
show that almost 100% of the ammonia was transported from the feed
compartment to the compartment where ammonium sulphate was collected. By
comparing the exergy of mixing with the input exergy, however, it was
shown that the input exergy (i.e. electricity) needed compared to what
extent the ions were separated was in a magnitude of several thousand
larger. One change in the set-up was therefore to narrow the compartment
distance to 18 mm to overcome part of the electrical resistance in the
solutions. This distance was used for four different set-ups using
bipolar membranes. The bipolar membranes were arranged in different ways
together with monovalent membranes, with the aim to produce acidic and
alkaline solutions that may be used for further extraction and
carbonation steps, respectively.

The energy demands for different type of tests were ranging from 1.7
to 350 MJ/kg separated NH4+. The two best cases, either a set-up placed
after the extraction step or after the carbonation step, respectively,
should be further studied. The need for running experiments in a
commercial lab-scale set-up with sufficiently narrow compartment width
and more repeating units than was used here is now substantial. No
funding could do far be made available for purchase of a commercial
set-up at the time of writing this thesis. Using a commercial labscale
set-up will give better indications of the suitability of the individual
set-ups and could also give more accurate numbers of the exergy
(electricity) input needed for the electrodialysis step. The separation
in more and more narrow repeating units will still use two electrodes so
the energy used for the water electrolysis will not increase. A
significantly lower exergy input will most likely be the case where more
repeating units are used.

The results from all methods used in this thesis indicate that the
all-aqueous variety of the ÅA route seems to be more favorable also when
it comes to the streamlining of the process and the regeneration of the
flux salt used.

Last updated on 2019-09-12 at 02:41