An ionic liquid process for mercury removal from natural gas

Abstract

Efficient scrubbing of mercury vapour from natural gas streams has been demonstrated both in the laboratory and on an industrial scale, using chlorocuprate(ii) ionic liquids impregnated on high surface area porous solid supports, resulting in the effective removal of mercury vapour from natural gas streams. This material has been commercialised for use within the petroleum gas production industry, and has currently been running continuously for three years on a natural gas plant in Malaysia. Here we report on the chemistry underlying this process, and demonstrate the transfer of this technology from gram to ton scale.

Full text

Dalton
Transactions
PAPER
Cite this: Dalton Trans., 2015, 44,
8617
Received 23rd October 2014,
Accepted 11th February 2015
DOI: 10.1039/c4dt03273j
www.rsc.org/dalton
An ionic liquid process for mercury removal
from natural gas†
Mahpuzah Abai,
a,b
Martin P. Atkins,*
b
Amiruddin Hassan,
a
John D. Holbrey,*
b
Yongcheun Kuah,
a,b
Peter Nockemann,
b
Alexander A. Oliferenko,
b
Natalia V. Plechkova,
b
Syamzari Rafeen,
a,b
Adam A. Rahman,
a
Rafin Ramli,
a,b
Shahidah M. Shariff,
a
Kenneth R. Seddon,*
b
Geetha Srinivasan*
b
and Yiran Zou
b
Efficient scrubbing of mercury vapour from natural gas streams has been demonstrated both in the lab-
oratory and on an industrial scale, using chlorocuprate(II) ionic liquids impregnated on high surface area
porous solid supports, resulting in the effective removal of mercury vapour from natural gas streams. This
material has been commercialised for use within the petroleum gas production industry, and has currently
been running continuously for three years on a natural gas plant in Malaysia. Here we report on the chem-
istry underlying this process, and demonstrate the transfer of this technology from gram to ton scale.
Introduction
Mercury is a natural component of the earth’s crust, and
through a number of natural and anthropogenic cycles is
released into the environment as a toxic pollutant.
1
All forms
of mercury viz. elemental mercury and oxidised, Hg(I) and Hg(II)
are intrinsically toxic.
2
One of the main sources of environ-
mental pollution by mercury is from fossil fuels, either during
combustion of coal or emitted from natural gas.
3
Many natural
gas fields contain mercury released from mercury-containing
ores through secondary geothermal processes. Natural gas
fields typically contain mercury, in either elemental or
combinations of elemental and organometallic forms, in con-
centrations in the range <0.1–5000 μgm
−3
depending on geo-
graphical location and geology. Although these concentrations
are usually relatively low, for a typical gas processing plant
treating 250 MMSCFD (million standard cubic feet per day =
28 300 m
3
d
−1
) of natural gas, the cumulative quantities can be
significant and can lead to problems through accumulation
via condensation and amalgamation. High mercury concen-
trations in oil and gas production correlate with regions of
high mercury emission to the environment (see Fig. 1).
4
Mercury can be extremely corrosive, causing destructive
damage to process equipment, particularly aluminium heat
exchangers, through liquid metal embrittlement.
5
For
example, an explosion in 1973 at the Skikda liquefied natural
gas plant in Algeria led to 27 fatalities and financial losses of
$1 billion due to catastrophic failure of an aluminium heat
exchanger through reaction with mercury contaminants.
6
Mercury control and treatment within the oil and gas
supply chain is increasingly recognised as important, both to
protect equipment and personnel from deleterious exposure,
and in order to comply with increasingly stringent discharge
regulations. For natural gas processing, mercury control usually
signifiesthe use of fixed-bed scrubbers containing a solid adsor-
bent that can capture mercury vapour either through adsorp-
tion, amalgamation, or oxidation followed by adsorption. The
Fig. 1 Global mercury transport, indicating the major areas of emission
(in red) taken from United Nations Environment Programme, Global
Mercury Assessment.
4
†CCDC 1029551. For crystallographic data in CIF or other electronic format see
DOI: 10.1039/c4dt03273j
a
PETRONAS Research Sdn. Bhd., Lot 3288 & 3289, OffJalan Ayer Itam,
Kawasan Institusi Bangi, 43000 Kajang Selangor, Malaysia
b
QUILL, The Queen’s University of Belfast, Belfast, BT9 5AG, UK.
E-mail: quill@qub.ac.uk
This journal is © The Royal Society of Chemistry 2015 Dalton Trans.,2015,44,8617–8624 | 8617
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online
View Journal
| View Issue

most common approaches employ commercially available
scrubbers, as shown in Table 1, incorporating sulfur, metal
oxides/sulfides, or metals (particularly silver) as active agents on
porous alumina, zeolite, or activated carbon supports.
7
Sulfur-impregnated activated carbons are probably the most
widely used adsorbents for mercury control within the gas
industry. Impregnation of sulfur into activated carbons over-
comes some of the slow kinetics of the reaction of mercury
with elemental sulfur under ambient conditions. Vidićand co-
workers have proposed that this is may be due to the involve-
ment of sulfur allotropes with greater number reactive end
groups.
8
However, activated carbons are not as mechanically
robust as oxide supports such as alumina or silica, and there-
fore suffer from attrition which shortens their operational life-
times.
7
Mercury capture from gas using copper(II) chloride
impregnated on carbon as a ‘chloride’source has been
investigated,
9
however the resulting mercury(II) chloride was
incompletely captured and leached under high pressure
conditions.
10
Ionic liquids
11
have also been explored for liquid/liquid par-
titioning of mercury, as mercury(II), from water,
12
initially
using hydrophobic ionic liquids containing pendant sulfur
ligands.
13
Subsequently, it was shown that mercury(II) can
partition into unfunctionalised hydrophobic ionic liquids.
14,15
The effective involatility of most ionic liquids means that
gas–liquid contacting scenarios can be considered without
contaminating process gas streams. Pinto and co-workers have
reported the use of ionic liquids to capture elemental and oxi-
dised mercury from combustion flue gases, combining the
ionic liquid with permanganate(VII) as an oxidant.
16
Rogers
and Holbrey
17
have suggested using ionic liquids containing
perhalide anions for the oxidative dissolution of metals, and
Sasson and co-workers
18
have reported using similar ionic
liquid systems as liquid scrubbers to remove mercury from
power-plant combustion gas. However, the corrosive nature of
these highly oxidising ionic liquids (analogous to chlorine or
bromine) to metals such as iron may place restrictions on their
applicability.
Building on previous work developing ionic liquid
approaches to oxidative dissolution,
19,20
and selective extrac-
tion and separations,
21
we considered whether ionic liquids
22
incorporating metal complexes might oxidise elemental
mercury, leading to the formation of stable anionic mercurate(II)
species. Such anions should then become integral com-
ponents of a new, more complex, ionic liquid system,
23
which
would also be non-volatile. Here, we report on the investigation
of chlorocuprate(II)-based ionic liquids for the direct oxidation
of elemental mercury,
24
and the use of these ionic liquids in a
solid-supported ionic liquid phase (SILP)
25
for reactive capture
of mercury from gas streams, which has led to the scale-up
and deployment on industrial plants in Malaysia.
26
Experimental
199
Hg NMR spectra were measured using a Bruker 500DRX
spectrometer with a 1 M solution of HgCl
2
(Sigma Aldrich) in
dmso-d
6
as an internal standard (−1501 ppm relative to Hg-
(CH
3
)
2
and deuterium lock.
27
IR spectroscopy was carried out
using a Perkin Elmer Spectrum 100 FT-IR spectrophotometer
with a Universal diamond Attenuated Total Reflectance (ATR)
top plate.
Mercury solubility in ionic liquids was measured by dissol-
ving an accurately weighed sample of the ionic liquid phase
(ca. 0.05 g) in water (10 cm
3
), removing the small quantity of
precipitate produced by filtration, and then diluting to 50 cm
3
,
followed by subsequent dilution of a 1 cm
3
aliquot to 50 cm
3
.
A sample (<50 μL) calculated to contain less than 1000 ng
mercury was then analysed for mercury using a Milestone
DMA-80 direct mercury analyser. Based on the sample size,
initial mass of ionic liquid used for the contact test, and
mercury content determined, the wt% of water soluble
mercury present in the ionic liquid was calculated.
Mercury capture from the carrier gas by SILPs was tested
using an accelerated breakthrough test rig constructed in-
house, incorporating a mercury generator and Sir Galahad ana-
lyser (PS Analytical). SILPs (and commercial activated carbons
for comparison) were crushed and sieved using a 300–500 μm
mesh. Adsorbents (30–100 mg) were packed into a thermo-
stated glass column of diameter 1 mm. Mercury removal from
test gas streams (nitrogen and methane) was examined using a
high flow rate (600 cm
3
min
−1
) and high mercury content
(2000 ng l
−1
) in the carrier gas; the outlet mercury concen-
tration was monitored continuously.
Crystallography
A suitable crystal of [N
4441
]
2
[Cu
2
Cl
6
] was selected and mea-
sured on a Rigaku Saturn724+ (2 × 2 bin mode) diffracto-
meter. The crystal was kept at 120.0 K during data collection.
Using Olex2,
28
the structure was solved with the ShelXS
29
struc-
ture solution program using Direct Methods and refined with
the ShelXL refinement package using Least Squares minimis-
ation. Crystal Data for C
13
H
30
Cl
3
CuN (M= 370.27 g mol
−1
):
monoclinic, space group P2
1
/n(no. 14), a= 9.2771(4) Å, b=
15.6784(5) Å, c= 12.5521(5) Å, β= 100.016(2)°, V= 1797.88(12)
Å
3
,Z=4,T= 120.0 K, μ(MoK
α
) = 1.647 mm
−1
,D
calc
= 1.368 g
cm
−3
, 21 158 reflections measured (5.98° ≤2Θ≤54.96°), 4122
unique (R
int
= 0.0469, R
σ
= 0.0382) which were used in all cal-
culations. The final R
1
was 0.0377 (>2σ(I)) and wR
2
was 0.0951
(all data).
Table 1 Various mercury removal systems for natural gas streams
7
Active compound Support/
medium Fate of mercury
Sulfur Carbon/
alumina HgS
Metal sulfide Carbon/
alumina HgS
Silver Zeolite Ag–Hg amalgam
Thiol/oxidising agent/
chelating agent Scavenger
solution Soluble Hg(II)
compound
Metal oxide/sulfide Metal oxide HgO/HgS
Paper Dalton Transactions
8618 |Dalton Trans.,2015,44,8617–8624 This journal is © The Royal Society of Chemistry 2015
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online

Results and discussion
Synthesis and characterisation of copper(II) ionic liquids
Chlorocuprate(II) ionic liquids were prepared by combining
1-butyl-3-methylimidazolium chloride ([C
4
mim]Cl), 1-ethyl-3-
methylimidazolium chloride ([C
2
mim]Cl), or tributylmethyl-
ammonium chloride ([N
4441
]Cl, (ex. Sigma Aldrich) with copper(II)
chloride (either anhydrous or dihydrate, ex. Sigma Aldrich) in
methanol in either a 1 : 1 or 2 : 1 molar ratio, followed by
solvent removal under reduced pressure, and then in vacuo at
80 °C. It is convenient to define the ionic liquid systems by the
mole fraction χ
CuCl
2
, which is defined by χ
CuCl
2
=n
i
/(n
i
+n
c
),
where n
i
is the number of moles of CuCl
2
and n
c
is the
number of moles of cation. The resulting ionic liquids formed
as dark yellow-brown liquids which, in all cases, slowly solidi-
fied on standing at room temperature. Additionally, a 1 : 2 : 1
cholinium chloride–diethylene glycol–CuCl
2
·2H
2
O deep eutec-
tic
30
was prepared as a dark green liquid. The ionic liquids
were characterised by DSC, TGA, microanalysis, electronic
absorption and vibrational spectroscopy, Karl Fischer titration
and surface analysis by SEM/EDAX.
Small single crystals were isolated from the bulk solidified
[N
4441
]
2
[Cu
2
Cl
6
] ionic liquid (1 : 1 CuCl
2
:[N
4441
]Cl), and analysed
by X-ray crystallography. The structure confirmed the presence
of a dimeric copper(II) dianion with four-coordinate copper
centres separated by 3.33 Å (Fig. 2). Li et al.
31
have reported
that [C
2
mim]Cl/copper(II) chloride mixtures form [CuCl
4
]
2−
chlorocuprate(II) anions even in the presence of substantial
concentrations of water. It was thus not anticipated that
copper speciation would differ in ionic liquids formed with
anhydrous or hydrated copper(II) chloride. Crystal structures of
many chlorocuprate(II) salts with a variety of organic cations
have been previously reported,
32
and are dominated by salts of
the discrete tetrachlorocuprate(II) anion, which can vary in
structure from tetrahedral to square planar, with many inter-
mediate geometries.
Reaction of copper(II) ionic liquids with mercury
Because the dominant form of mercury in natural gas is
elemental, tests were conducted to determine the solubility of
Hg(0) in the ionic liquids by heating a droplet of mercury
(Sigma Aldrich) with the ionic liquids at 60 °C overnight in a
sealed vial. No precautions to exclude air were made during
sample loading. In each case, a colour change was observed in
the ionic liquid, with the characteristic brown colour of the
chlorocuprate(II)-based systems changing to (i) a clear pale
green ionic liquid at χ
CuCl
2
= 0.33, or to (ii) a pale orange-yellow
ionic liquid with precipitation of a pale grey-white solid at
χ
CuCl
2
= 0.50, as shown in Fig. 3. After heating, the size of the
mercury drops remaining in the vials had visibly reduced and
the mercury content in ionic liquid phase was determined by
mercury analysis. The results are shown in Table 2.
Over 15 wt% mercury dissolution was observed (Table 2) for
each of the chlorocuprate(II) ionic liquids tested, as deter-
mined from the water soluble mercury content of diluted ali-
quots of the ionic liquids. Reactive dissolution with similar
end results was observed for the ionic liquids prepared from
anhydrous and hydrated copper(II) chloride. A similar ability to
dissolve mercury was observed for both χ
CuCl
2
= 0.33 and 0.50,
although the apparent outcomes of the dissolution process
appear to be slightly different. The amount of mercury that
was solubilised by the ionic liquids can be expressed by the
molar concentration ratio [Cu(II)
IL
] : [Hg(II)
IL
] (in Table 2) from
the initial concentration of copper in the ionic liquids and
mass of mercury extracted. In all cases, a small excess of
unreacted copper(II) in solution was apparent due to the
residual paler brown colouration. This would imply that one
Fig. 2 Cation and anion within the unit cell of [N
4441
]
2
[Cu
2
Cl
6
]; d¯(CuCl)
t
=
2.19, d¯(CuCl)
b
=2.32,d¯(Cu⋯Cu) = 3.33 Å.
Fig. 3 The reaction of an excess of liquid mercury with 1-butyl-3-
methylimidazolium chlorocuprate(II) ionic liquids at 60 °C. The top three
photographs correspond to χ
CuCl
2
= 0.33 and the bottom three to χ
CuCl
2
= 0.50. At χ
CuCl
2
= 0.33, (a) the initial ionic liquid mixture transforms to
(b), a partially decolourised solution after mixing with Hg(0) for 5 min
and then to (c) after extended mixing. At χ
CuCl
2
= 0.50, (d) the initial ionic
liquid mixture transforms through (e) after mixing with Hg(0) for 5 min
to (f), a clear, largely decolourised ionic liquid phase containing a fine,
grey-white precipitate after extended mixing.
Dalton Transactions Paper
This journal is © The Royal Society of Chemistry 2015 Dalton Trans.,2015,44,8617–8624 | 8619
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online

mole of mercury was required to reduce two moles of copper:
Hgð0Þþ2CuðiiÞ!HgðiiÞþ2CuðiÞ
The simplest mechanisms that can be proposed for this
reaction, depending on the stoichiometry of the chlorocuprate(II)
ionic liquids, are shown in eqn (1)–(3).
2½CuCl42þHgð0Þ!½HgCl42þ2½CuCl2ð1Þ
2½Cu2Cl62þ2Hgð0Þ!½Hg2Cl62þ2½CuCl2þ2CuClðsÞ
ð2Þ
½Cu2Cl62þHgð0Þ!HgCl2ðsÞþ2½CuCl2ð3Þ
The changes in colour of the copper(II)-containing ionic
liquids after reaction with mercury are consistent with
reduction to copper(I). Bolkan and Yoke,
33
described copper(I)
systems ([C
2
mim]Cl–CuCl; 0.33 < χ
CuCl
< 0.67) as having
colours that varied from pale yellow to light green with increas-
ing χ
CuCl
, and that darken rapidly in air and become paramag-
netic as copper(II) is formed.
In eqn (1), all the products are anionic, and will incorporate
into the ionic liquid. eqn (2) and (3) represent two contrasting
interpretations of the observed reaction. In eqn (2), copper(I)
chloride precipitates and in eqn (3), mercury(II) chloride pre-
cipitates. In the former case, the remaining solution contains
both mercury(II) and copper(I) species; in the latter case, there
will be no mercury in solution.
In eqn (2), it is speculated that the speciation of mercury is
[Hg
2
Cl
6
]
2−
. There is no evidence that this would be the
speciation as it may be a monomeric species
34
or a polymeric
species,
35
but the balance of probability is that it is likely to be
dimeric.
36
Whatever the speciation, it does not affect the
validity of the following arguments.
Examples of
199
Hg NMR spectra taken from {[C
4
mim]Cl–
CuCl
2
(χ
CuCl
2
= 0.50) + Hg(0)} and {[C
4
mim]Cl–CuCl
2
(χ
CuCl
2
=
0.33) + Hg(0)} systems are shown in Fig. 4. Well defined
199
Hg
NMR signals were obtained in each case, but with different
chemical shifts. For χ
CuCl
2
= 0.33, this peak was found at
−1106 ppm (relative to 1 M HgCl
2
in dmso); for χ
CuCl
2
= 0.50,
the peak appeared at −1140 ppm. These positions are consist-
ent with the formation of chloromercurate(II) anions,
37
where
the mercury is in the environment of four chlorine atoms. The
lower field peak is assigned to [HgCl
4
]
2−
and the higher peak
to [Hg
2
Cl
6
]
2−
. As it is known that [HgCl
3
]
−
has a peak at
−1197 ppm,
40
this suggests that the mercury complex anion
formed is dimeric rather than monomeric. This provides
strong support for the reaction schemes outlined in eqn (1)
and (2). Eqn (3) has been eliminated as it would imply the
absence of mercury in solution. The formation of a precipitate
from the reaction of liquid mercury with the [C
4
mim]Cl–CuCl
2
(χ
CuCl
2
= 0.50) ionic liquid combined with the presence of
chloromercurate(II) anions in solution as shown from
199
Hg
NMR spectroscopy supports the reaction mechanism in eqn
(2) at this composition and the precipitation of CuCl as
the reaction shifts the ionic liquid towards chloride deficient,
i.e. acidic, compositions. While little is known about the phase
behaviour of chloromercurate(II) ionic liquids with χ
HgCl
2
,
38,39
the limiting point for homogeneity in chlorocuprate(I) systems
is χ
CuCl
= 0.67,
34
that is, with the copper : chloride ratio of 2 : 3.
Copper(II) supported ionic liquids in the laboratory
The above results from the reactions of the copper(II) ionic
liquids with elemental mercury give a good indication that
they might be the basis of an absorption scrubber. This would
combine the thermodynamically favourable oxidation of
mercury(0) to mercury(II) by the copper(II), followed by com-
plexation (and hence stabilisation) by the liberated chloride
ions. Although in the laboratory, it is clearly possible to
develop a mercury scrubber based on an ionic liquid system in
the fluid state, the efficiency of passing natural gas through a
viscous liquid system would prevent its scale-up for an indus-
trial process. The logical way forward was to coat a dispersed
solid with a selected and optimised ionic liquid to form a class
of materials now recognised as a supported ionic liquid phase
(SILP).
25
This would lead to good contact between the gas and
the ionic liquid, high throughput flow, and mechanical
strength. A typical SILP is shown in Fig. 5.
Further, most mercury removal units used in the oil and
gas industry are fixed-bed scrubbers, and so a SILP approach
Table 2 Solubility of bulk elemental mercury in metal-containing ionic
liquid system at 60 °C (averages of three measurements)
Ionic liquid composition wt% mercury
dissolved in IL [Cu(II)
IL
]:
[Hg(II)
IL
]
[C
4
mim]Cl–CuCl
2
·2H
2
O (1 : 1) 22.1 ± 1.4 2.6
[C
2
mim]Cl–CuCl
2
(2 : 1) 21.3 ± 2.3 2.2
[N
4441
]Cl–CuCl
2
(2 : 1) 15.9 ± 0.3 2.4
[N
4441
]Cl–CuCl
2
·2H
2
O (2 : 1) 15.2 ± 0.1 2.3
[Chol]Cl–EG–CuCl
2
·2H
2
O(1:2:1)
a
16.9 ± 0.1 3.0
a
Chol = cholinium; EG = diethylene glycol.
Fig. 4
199
Hg NMR (23 °C, neat, 89.57 MHz) spectra of the ionic liquids
obtained from the reaction mixtures of (a) [C
4
mim]
2
[CuCl
4
] + Hg(0) (red
trace) and (b) [C
4
mim]
2
[Cu
2
Cl
6
] + Hg(0) (blue trace). The peak at
−1501 ppm is the reference peak.
Paper Dalton Transactions
8620 |Dalton Trans.,2015,44,8617–8624 This journal is © The Royal Society of Chemistry 2015
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online

to heterogenise the ionic liquids would be a natural fit with
standard industrial practice. SILPs, which incorporate high
surface area thin-films of ionic liquids within a porous solid
scaffold, have many advantages for efficient gas–liquid contact-
ing when a static liquid phase (i.e. as a catalyst or adsorbent) is
required.
25
SILPs containing the chlorocuprate(II) ionic liquids at 10 wt%
loading were prepared by the incipient wetness method,
adding solutions of the respective ionic liquids in methanol to
porous silica beads, followed by drying overnight at 80 °C.
After drying, the SILPs were orange-yellow (see Fig. 5), in con-
trast to the characteristic green of the methanolic solutions of
the copper salts. ATIR spectra of SILPs showed no vibrational
bands that could be assigned to methanol or to water, other
than that residual on the silica surface (Fig. 6).
Mercury adsorption from gas streams was tested by passing
a mercury-containing gas (either N
2
or CH
4
) through a fixed
bed of adsorbent, and comparing the inlet and outlet mercury
concentrations. The more efficient the adsorbent is at
removing mercury from the gas, the longer the time it takes
for mercury to be detected at the outlet. The time from injec-
tion to detection is defined as the breakthrough time. Over the
course of the measurements, the inlet mercury concentration
was held at the high level of 2000 ng l
−1
in order to accelerate
the breakthrough tests. In packed adsorbent beds with no
channelling, the outlet mercury concentration follows the
ideal (and observed) profile illustrated in Fig. 7. The outlet
concentration remained constant at less than 0.2 ng l
−1
prior
to breakthrough, which was characterised by a rapid increase
in its concentration. The breakthrough time was defined as
the time at which mercury capture from the gas stream
decreased to below 99.5% capture efficiency ([Hg]
out
>10ngl
−1
).
Characteristic mercury outlet profiles obtained under test
conditions for a sulfur-impregnated activated carbon (ca.10wt%
sulfur) and for a chlorocuprate(II) containing SILP (10 wt%
[N
4441
]Cl–CuCl
2
,χ
CuCl
2
=0.50onporoussilica,135m
2
g
−1
surface
area and 0.83 cm
3
g
−1
pore volume) are shown in Fig. 7.
Under the conditions defined in Fig. 7, the selected SILP
remained active for 35 h until breakthrough, which corres-
ponded to a mercury uptake of ca. 2.5 wt% (2.52 mg, 1.27 ×
10
−5
mol). This 100 mg SILP sample contained 2.71 × 10
−5
moles of copper(II), and so the breakthrough point represents
93% mercury capture efficiency based on the proposed capture
mechanism requiring oxidation by two copper(II) centres. The
activated carbon, in contrast, showed poorer performance,
with a breakthrough of 12.7 h, which is equivalent to 0.9 wt%
mercury capture on the support. This was lower than antici-
pated, although this may be due to the extremely short resi-
dence times in these accelerated screening tests. Thus, in a
laboratory test rig, the SILP outperformed the activated carbon
by a factor of 3, which was very encouraging for moving from
laboratory to pilot plant.
Fig. 5 An example of a SILP containing chlorocuprate(II) ionic liquid
impregnated porous 4 mm diameter silica spheres.
Fig. 6 ATR IR spectra comparing a silica support (red) and SILPs pre-
pared with overnight drying at 80 °C (blue) or drying in vacuo (orange),
illustrating the absence of additional –OH stretching frequencies in the
SILPS that could be attributable to methanol or water. Peaks at 2965
and 2882 cm
−1
derive from the ionic liquid cation.
Fig. 7 Mercury breakthrough curves for the capture of mercury vapour
from dinitrogen gas ([Hg]
in
= 2000 ng l
−1
,flow rate = 600 cm
3
min
−1
,
T= 25 °C) using 0.10 g samples of (a) a sulfur-impregnated activated
carbon and of (b) a 10 wt% [N
4441
]Cl–CuCl
2
(χ
CuCl
2
= 0.50) SILP on
porous silica (135 m
2
g
−1
surface area and 0.83 cm
3
g
−1
pore volume).
Similar results were obtained using methane as the carrier gas.
Dalton Transactions Paper
This journal is © The Royal Society of Chemistry 2015 Dalton Trans.,2015,44,8617–8624 | 8621
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online

Ionic liquids containing different cations ([N
4441
]
+
,[C
2
mim]
+
,
and [C
4
mim]
+
) were tested, resulting in similar breakthrough
performance which appears to correlate with the total amount
copper(II) within the tested samples (Fig. 8). At higher load-
ings, a reduction in mercury capture performance was
observed, presumably through plugging of the support
through pore filling. This could be partially alleviated by chan-
ging to supports with large pore geometries.
After exhaustive extraction in contact with mercury-contain-
ing gas streams, the SILPs change colour from orange to pale
green. This transformation is comparable to the colour
changes observed after reaction of the neat ionic liquids with
liquid mercury (Fig. 5). As a useful laboratory application,
Fig. 9 shows a bed of chlorocuprate(II)-containing SILP used as
a mercury trap on the off-gas line from a PSA Sir Galahad
mercury analyser in our laboratory. The mercury breakthrough
front can be clearly seen at the sharp transition from spent
(pale green) absorbent to active (yellow-orange) material.
Interestingly, Haumann and co-workers
40
have also recently
reported the use of chlorocuprate(II)-containing SILPs for gas
absorption applications. They investigated ammonia capture
from air taking advantage of the strong coordination of
nucleophilic amines to copper, and suggested their potential
uses in broadband filters for breathing apparatus.
Copper(II) supported ionic liquids at scale
The laboratory results demonstrated that efficient and compre-
hensive scrubbing of mercury vapour from gas streams could
be achieved with excellent capture kinetics. The accelerated
breakthrough screening used residence times in the order of
50 ms in the test rig, and breakthrough times correlated with
theoretical maximum mercury capacities based on the mecha-
nisms detailed in eqn (1) and (2). Following on from these
small-scale laboratory tests, larger batches (ca. 100 g) of chloro-
cuprate(II) SILPs were prepared and tested in a 200 cm
3
pilot-
scale slipstream scrubber on-site at a Gas Processing Plant in
Malaysia. The pilot trials (over 60 days) yielded a gas stream
with mercury outlet concentrations of <0.01 μgm
−3
(Fig. 10),
at least one order of magnitude below the sales specification
for the natural gas.
This allowed a smooth transition to full plant implemen-
tation using 30 m
3
of a SILP (see Fig. 11). After three years of
continuous operation, the outlet mercury concentration from
the plant remained low, meeting the plant outlet specifica-
tions. This represents a remarkable success of transferring lab-
oratory chemistry to full scale plant operation in a remarkably
fast implementation (less than four years). The SILP is now
commercialised via a marketing licensing agreement between
PETRONAS and Clariant.
41
Fig. 9 Example of a chlorocuprate(II)-containing SILP used as a
mercury trap on the off-gas line from a PSA Sir Galahad mercury analy-
ser in our laboratory. The mercury breakthrough front can be clearly
seen at the sharp transition from spent (pale green) to active (yellow-
orange) absorbent. The packed bed contains ionic liquid impregnated
on two different silica supports; spherical beads and extrudates.
Fig. 10 Results from a pilot trial (over 60 days) of a copper(II) containing
SILP in a slipstream scrubber (200 cm
3
) for mercury removal from
natural gas on-site at a Gas Processing Plant in Malaysia, showing the
mercury outlet concentrations after passing through commercial
material (blue diamonds) and SILPs (red squares). The average extraction
efficiency was 99.998%. The green line indicates specification for sales
quality gas (0.1 μgm
−3
mercury).
27
Fig. 8 The change in mercury breakthrough time for SILPs containing
different cations ([N
4441
]
+
,[C
2
mim]
+
,and[C
4
mim]
+
) as a function of ionic
liquid loading in the SILP, expressed as wt% copper. Compared to the
results shown in Fig. 7, the experimental sample sizes here are reduced to
30 mg, with all other conditions identical ([Hg]
in
=2000ngl
−1
,flow rate =
600 cm
3
min
−1
,T= 25 °C), which results in a corresponding decrease in
the breakthrough time.
Paper Dalton Transactions
8622 |Dalton Trans.,2015,44,8617–8624 This journal is © The Royal Society of Chemistry 2015
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online

Conclusions
We have demonstrated that it is possible to develop in the lab-
oratory, processes with industrial applicability. Starting at the
gram scale, the translation of the ionic liquid chemistry into a
SILP material that was scaled to full industrial implementation
within four years has been exemplified. The fundamental
chemistry (namely oxidative dissolution and complexation of
mercury into the ionic liquid as a stable anionic mercurate(II)
species) does not change with the scale of the process and the
ability to digest large quantities of mercury (up to 20 wt%),
coupled with efficient gas-contacting, made possible by for-
mation of a SILP, underpins the successful development of
these materials as commercially competitive solid phase
mercury adsorbents in the natural gas industry.
Acknowledgements
We are indebted to our many colleagues at PETRONAS for
their funding and discussions about this work. We would also
like thank Chemviron Carbon, Ltd. for supplying activated
carbon samples for testing, and thank the EPSRC UK National
Crystallography Service at the University of Southampton for
the collection of the crystallographic data.
42
Notes and references
1 C. T. Driscoll, R. P. Mason, H. M. Chan, D. J. Jacob and
N. Pirrone, Environ. Sci. Technol., 2013, 47, 4967–4983.
2 G. Liu, Y. Cai, N. O’Driscoll, X. Feng and G. Jiang, Overview
of Mercury in the Environment, in Environmental Chemistry
and Toxicology of Mercury, ed. G. Liu, Y. Cai and
N. O’Driscoll, John Wiley & Sons, Hoboken, New Jersey,
2011, pp. 1–12.
3 S. M. Wilhelm and N. Bloom, Fuel Process. Technol., 2000,
63,1–27.
4 United Nations Environment Programme, Global Mercury
Assessment and Environmental Transport, UNEP Chemicals
Branch, Geneva, Switzerland, 2013, ISBN: 978-92-807-3310-5.
5 S. M. Wilhelm, Process Saf. Prog., 2009, 28, 259–266.
6 G. T. Kinney, Oil Gas J., 1975, 192.
7 E. J. Granite, H. W. Pennline and R. A. Hargis, Ind. Eng.
Chem. Res., 2000, 39, 1020–1029; N. Eckersley, Hydrocarb.
Process, 2012, 29–35.
8 W. Liu, R. D. Vidićand T. D. Brown, Environ. Sci. Technol.,
1998, 32, 531–538.
9 R. D. Vidićand D. P. Siler, Carbon, 2001, 39,3–14.
10 W. Du, L. Yin, Y. Zhuo, Q. Xu, L. Zhang and C. Chen, Ind.
Eng. Chem. Res., 2014, 53, 582–591.
11 See for example: M. Freemantle, An Introduction to Ionic
Liquids, Royal Society of Chemistry, Cambridge, 2010;
N. V. Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008,
37, 123–150; J. P. Hallett and T. Welton, Chem. Rev., 2011,
111, 3508–3576.
12 M. V. Mancini, N. Spreti, P. Di Profio and R. Germani, Sep.
Purif. Technol., 2013, 116, 294–299.
13 A. E. Visser, R. P. Swatloski, W. M. Reichert, R. Mayton,
S. Sheff, A. Wierzbicki, J. H. Davis Jr. and R. D. Rogers,
Chem. Commun., 2001, 135–136.
14 A. E. Visser, R. P. Swatloski, S. T. Griffin, D. H. Hartman
and R. D. Rogers, Sep. Sci. Technol., 2001, 36, 785–804.
15 N. Papaiconomou, J. M. Lee, J. Salminen, M. von Stosch
and J. M. Prausnitz, Ind. Eng. Chem. Res., 2008, 47, 5080–
5086.
Fig. 11 Transformation from (a) lab-scale preparation of chlorocuprate(II) SILPs through to (b) pilot-scale using 100 cm
3
of SILP on production plant
gas feeds to (c) full-scale mercury removal units with 20 m
3
of SILP at a gas processing plant site.
27
Dalton Transactions Paper
This journal is © The Royal Society of Chemistry 2015 Dalton Trans.,2015,44,8617–8624 | 8623
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online

16 L. Ji, S. W. Thiel and N. G. Pinto, Ind. Eng. Chem. Res.,
2008, 47, 8396–8400; L. Ji, M. Abu-Daabes and N. G. Pinto,
Chem. Eng. Sci., 2009, 64, 486–491.
17 R. D. Rogers and J. D. Holbrey, World Pat. Appl,
WO2010116167, 2010.
18 Y. Sasson, M. Chidambaram and Z. Barnea, US Pat,
US8101144, 2012; Z. Barnea, T. Sachs, M. Chidambaram
and Y. Sasson, J. Hazard. Mater., 2013, 244, 495–500.
19 M. Fields, G. V. Hutson, K. R. Seddon and C. M. Gordon,
World Pat. Appl, WO9806106, 1998.
20 P. Nockemann, B. Thijs, S. Pittois, J. Thoen, C. Glorieux,
K. Van Hecke, L. Van Meervelt, B. Kirchner and
K. Binnemans, J. Phys. Chem. B, 2006, 110, 20978–20992.
21 J. D. Holbrey, I. Lopez-Martin, G. Rothenberg,
K. R. Seddon, G. Silvero and X. Zheng, Green Chem., 2008,
10,87–92.
22 J. Estager, J. D. Holbrey and M. Swadźba-Kwaśny, Chem.
Soc. Rev., 2014, 43, 847–886.
23 M. Y. Lui, L. Crowhurst, J. P. Hallett, P. A. Hunt,
H. Niedermeyer and T. Welton, Chem. Sci., 2011, 2, 1491–
1496.
24 M. Abai, M. Atkins, K. Y. Cheun, J. D. Holbrey,
P. Nockemann, K. R. Seddon, G. Srinivasan and Y. Zou,
World Pat. Appl, WO2012046057, 2012.
25 Supported Ionic Liquids: Fundamentals and Applications, ed.
R. Fehrmann, A. Riisager and M. Haumann, Wiley,
Weinheim, 2014.
26 M. Abai, M. P. Atkins, A. Hassan, J. D. Holbrey, Y. Kuah,
P. Nockemann, A. A. Oliferenko, N. V. Plechkova, S. Rafeen,
A. A. Rahman, K. R. Seddon, S. M. Shariff, G. Srinivasan
and Y. Zou, International Gas Union Research Conference
2011, Seoul, Korea, 2011, http://members.igu.org/old/IGU%
20Events/igrc/igrc2011/igrc-2011-proceedings-and-presenta-
tions/poster-papers-session-4/P4-26_Martin%20Atkins.pdf.
27 R. E. Taylor and F. P. Gabbaï, J. Mol. Struct., 2007, 839,
28–32.
28 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard
and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339–341.
29 G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystal-
logr., 2008, 64, 112–122.
30 A. P. Abbott, K. El Ttaib, G. Frisch, K. J. McKenzie and
K. S. Ryder, Phys. Chem. Chem. Phys., 2009, 11, 4269–
4277.
31 G. Li, D. M. Camaioni, J. E. Amonette, Z. C. Zhang,
T. J. Johnson and J. L. Fulton, J. Phys. Chem. B, 2010, 114,
12614–12622.
32 D. W. Smith, Coord. Chem. Rev., 1976, 21,93–158;
R. D. Willett and U. Geiser, Inorg. Chem., 1986, 25, 4558–
4561; P. De Vreese, N. R. Brooks, K. Van Hecke, L. Van
Meervelt, E. Matthijs, K. Binnemans and R. Van Deun,
Inorg. Chem., 2012, 51, 4972–4981.
33 S. A. Bolkan and J. T. Yoke, Inorg. Chem., 1986, 25, 3587–
3590.
34 P. Nockemann and G. Meyer, Z. Anorg. Allg. Chem., 2003,
629, 123–128.
35 M. Rademeyer, D. G. Billing and A. Lemmerer, Acta Crystal-
logr., Sect. E: Struct. Rep. Online, 2006, 62, m1716–m1718;
A. Linden, B. D. James, J. Liesegang and N. Gonis, Acta
Crystallogr., Sect. B: Struct. Sci., 1999, 55, 396–409.
36 P. Nockemann and G. Meyer, Acta Crystallogr., Sect. E:
Struct. Rep. Online, 2002, 58, m534–m536.
37 R. E. Taylor, S. Bai and C. Dybowski, J. Mol. Struct., 2011,
987, 193–198; G. Klose, F. Volke, G. Peinel and
G. Knobloch, Magn. Reson. Chem., 1993, 31, 548–551.
38 A. Metlen, B. Mallick, R. W. Murphy, A.-V. Mudring and
R. D. Rogers, Inorg. Chem., 2013, 52, 13997–14009.
39 B. Mallick, A. Metlen, M. Nieuwenhuyzen, R. D. Rogers and
A. V. Mudring, Inorg. Chem., 2012, 51, 193–200.
40 F. T. U. Kohler, S. Popp, H. Klefer, I. Eckle, C. Schrage,
B. Böhringer, D. Roth, M. Haumann and P. Wasserscheid,
Green Chem., 2014, 16, 3560–3568.
41 Clariant Newsroom, ‘Clariant And Petronas Sign Licensing
Collaboration’, http://newsroom.clariant.com/clariant-and-
petronas-sign-licensing-collaboration/, 2014.
42 S. J. Coles and P. A. Gale, Chem. Sci., 2012, 3, 683–
689.
Paper Dalton Transactions
8624 |Dalton Trans.,2015,44,8617–8624 This journal is © The Royal Society of Chemistry 2015
Published on 27 February 2015. Downloaded by Queen's Library on 6/1/2022 11:04:26 AM.
View Article Online