Energy Fuels 2010, 24, 1094–1098 : DOI:10.

1021/ef9009586
Published on Web 10/07/2009

Recovery of Bitumen from Oil or Tar Sands Using Ionic Liquids

Paul Painter,* Phillip Williams, and Ehren Mannebach
Department of Materials Science and Engineering, Penn State University, University Park, Pennsylvania 16802

Received September 1, 2009. Revised Manuscript Received September 22, 2009

The extraction and separation of bitumen from oil sands for the purpose of processing fuels is relatively
expensive and poses several environmental challenges. Roughly two tons of oil sands are required to produce
a barrel of oil, and the separation of the bitumen from sand and clay requires significant amounts of energy
and the use of large quantities of water. It is shown here that bitumen in a sample of Canadian tar sands can be
recovered using ionic liquids (ILs) and organic solvents. Essentially, a multiphase system;consisting of a
sand and clay slurry, an ionic liquid layer, and an organic layer containing the bitumen;can be formed by
simply mixing the components at somewhat elevated (∼55 °C) or ambient temperatures (∼25 °C). Essentially
all of the bitumen is released from the sand. Water is not used in this stage of the separation, but relatively
small amounts are used to separate entrained IL from the sand and clays. Because both the IL and water can
be recycled through the system and used repeatedly, this process has the potential for ameliorating many of
the environmental problems associated with current extraction methods.

Introduction and other materials used in processing. It is highly toxic to

aquatic life.1-4 As mentioned above, large quantities of tailing
Oil or tar sands compose a significant proportion of the
pond water are now recycled through the process, but this can
world’s known oil reserves. The largest deposits are found in
lead to scaling and corrosion problems and can adversely
Canada and Venezuela, which have combined oil sand re-
affect bitumen recovery. In addition, very fine mineral parti-
serves estimated to be equal to the world’s total reserves of
cles and emulsified salty water are coextracted with the bitu-
conventional crude oil.1 Significant quantities (estimated to be
men, and these can also lead to problems in subsequent
32 billion barrels of oil) can also be found in Eastern Utah in
processing.5-7
the US.1 Oil or tar sands are complex mixtures of sand, clays,
A number of different extraction processes have been
water, and bitumen, a “heavy” or highly viscous oil. Extrac-
evaluated over the years, but these can also introduce sig-
tion and separation of bitumen from surface-mined oil sands
nificant problems (see refs 7 and 8 and citations therein). For
for the purpose of processing fuels is much more expensive
example, solvents such as benzene or toluene can dissolve and
than extracting conventional oil by drilling and involves
extract bitumen from tar sands, at least those containing more
the use of significant amounts of energy and water.2-4
than 7% bitumen,9 thus reducing water consumption and
Although a large proportion of the water used in the process
energy use in an extraction process. However, such solvents
is now recycled from tailing ponds, the production of each
introduce problems associated with handling volatile, flam-
barrel of oil still requires importing additional barrels of
mable, and toxic materials and are not always easy to remove
“fresh” water.3,4
from the spent sand.8
The most commonly used process for the recovery of
In this communication, we present preliminary results
bitumen involves the addition of hot or warm water and
demonstrating that the bitumen in a medium grade Canadian
processing aids (historically NaOH) to the tar sands to form
tar sand can be readily recovered using ionic liquids and
a slurry, which is then “conditioned” in processes involv-
organic solvents at ambient (∼25 °C) or somewhat elevated
ing shearing the tar sand particles so that the bitumen is
(∼55 °C) temperatures. Ionic liquids (ILs) consist entirely of
detached from the minerals, forming suspended droplets. The
ionic species. Although they are not new, interest in these
density of bitumen and water are closely matched, so sepa-
materials has increased dramatically in the past few years,
ration is achieved by aeration to form an oil-containing
focusing especially on those ILs that are liquids at tempera-
froth that can be skimmed off the surface. Additional proces-
tures below 100 °C. This is a consequence of their unusual
sing of the water is performed to remove residual bitumen.
range of properties and their potential for use in a number of
This process water is ultimately stored in vast tailing ponds. It
industrial processes. ILs have outstanding chemical and
is a complex mixture of water, dissolved salts, minerals,
thermal stability; they are nonflammable (at temperatures
residual bitumen, surfactants released from the bitumen,
below the degradation point) and have a negligible vapor

*To whom correspondence should be addressed. E-mail: painter@

matse.psu.edu. (5) Masliyah, J.; Zhou, Z.; Xu, Z.; Czarnecki, J.; Hamza, H. Can. J.
(1) Canada0 s Oil Sands - Opportunities and Challenges to 2015: An Chem. Eng. 2004, 82, 628–654.
Update - June 2006; National Energy Board: 2006. http://www.neb.gc.ca/ (6) Trong, D-V; Jha, R.; Wu, S.-Y.; Tannant, D. D.; Masliyah, J.; Xu,
clf-nsi/rnrgynfmtn/nrgyrprt/lsnd/lsnd-eng.html Z. Energy Fuels 2009, 23, 2628–2636.
(2) Heinberg, R. The Party0 s Over: Oil, War and the Fate of Industrial (7) Budziak, C. J.; Vargha-Butler, E. I.; Hancock, R. G. V.; Neumann,
Societies; New Society Publishers: Gabriola Island, BC, Canada, 2005. A. W. Fuel 1988, 67, 1633–1638.
(3) Allen, E. W. J. Environ. Eng. Sci. 2008, 7, 123–138. (8) Houlihan, R.; Williams, K. H. J. Can. Pet. Technol. 1987, 26, 91–96.
(4) Allen, E. W. J. Environ. Eng. Sci. 2008, 7, 499–524. (9) Jacobs, F. S.; Filby, R. H. Anal. Chem. 1983, 55, 74–77.

r 2009 American Chemical Society 1094 pubs.acs.org/EF

Energy Fuels 2010, 24, 1094–1098 : DOI:10.1021/ef9009586 Painter et al.
10,11
pressure (i.e., they are nonvolatile). We will show that the
bitumen obtained in this process is essentially free from mineral
fines and that residual sand and clays can be obtained in what
appears to be a relatively clean and uncontaminated form.

Materials and Methods

The sample used in this study was a medium grade
Canadian tar sand, with a bitumen content of approximately
Figure 1. The structures of the ionic liquids used in this study; from
10%. It was obtained from the Alberta Research Council. The left-to right 1-butyl-3-methyl-imidazolium trifluoromethanesulfo-
ionic liquids were all purchased from Sigma-Aldrich. nate, [bmim][CF3SO3]; 1-butyl-2,3-dimethyl-imidazolium tetra-
Separations were performed by placing a sample of tar fluoroborate, [bmmim][BF4]; and 1-butyl-2,3-dimethyl-imidazo-
sands, toluene, and an IL in the proportions 1:2:3 by weight lium trifluoromethanesulfonate, [bmmim][CF3SO3].
either in glass vials or centrifuge tubes. Weighed quantities of
tar sand of close to 1 g were used in all separations. The IL was
added to the tar sand first, followed by the toluene. Initial
experiments were conducted by stirring this mixture at 55 °C
overnight using a laboratory magnetic stirrer bar. The tem-
perature control of the oil bath in which the vials were placed
was somewhat erratic and oscillated between 50 and 55 °C.
We chose this temperature on the basis of a review of the
literature, where it has been reported that bitumen yields are
greatly reduced in the water extraction process at tempera-
tures less than 50 °C.5 In subsequent work, however, we found
that simply shaking a vial containing the tar sand/toluene/IL
mixture at room temperature also achieved a degree of
immediate but incomplete separation, as we will describe in
the following section. Stirring for two hours increased the
yield to values comparable to those obtained at higher
temperatures. Figure 2. The agglomerates formed by mixing tar sands with
FTIR spectra were obtained using a Thermo Scientific [bmim][CF3SO3] after removal of the IL with water.
Nicolet 6700 Fourier transform infrared (FTIR) spectro-
brought into the same phase or multiphase processes can be
meter. A wavenumber resolution of 2 cm-1 was used, and
designed. However, it is not obvious that ILs would be good
200 scans were signal averaged. Spectra of the tar sands and
solvents for removing bitumen from tar sands, in the sense
minerals were obtained using a diffuse reflectance accessory.
that their highly polar nature makes most of them immiscible
Samples were prepared by grinding with KBr in a Wig-L-Bug,
with nonpolar hydrocarbon solvents such as toluene and
and spectra were referenced against pure KBr. Spectra of
naphtha, not to mention bitumen. It is their ability to engage
bitumen were obtained by casting toluene solutions onto a
in electrostatic interactions with the surface of sand and clay
KBr window and evaporating the solvent in a vacuum oven at
particles, thus releasing bitumen, that appears to be crucial.
100 °C, forming a thin film for analysis.
Nevertheless, one of our first attempts to determine if ILs
could be used in the separation process ended in dismal failure.
Results
Figure 2 shows the results of mixing the ionic liquid 1-butyl-3-
The apparent complexity of tar sand separation processes methyl imidazolium trifluoromethanesulfonate, [bmim][CF3-
and the various types of processing aids that have been used SO3], with a sample of Canadian tar sands. This IL did not
are a consequence of the range of interactions that occur separate bitumen from the sand, but instead resulted in the
between the various organic and inorganic components of formation of agglomerated, spherical, black balls of bitumen-
these materials. There is a wealth of knowledge that has been encrusted minerals.
obtained over the years on the nature of these interactions Other ILs appeared to at least disperse the tar sands. Upon
(see refs 12-14 and citations therein), and the work of Liu mixing tar sands with the IL 1-butyl-2,3-dimethyl-imidazo-
et al.12 strongly indicates that electrostatic forces play a lium tetrafluoroborate, [bmmim][BF4], at 55 °C, three phases
dominant role. This suggested to us that ILs might prove were formed. Because the layers observed in this multiphase
useful in separating bitumen from sand and clay. A particu- system are not easy to distinguish in photographs, we only
larly versatile group of ILs is based on imidazolium cations, describe our observations here. There is an IL/sand slurry at
and the structures of the ILs used in this study are illustrated in the bottom, a middle dark layer, and a layer of viscous
Figure 1. The properties of these solvents can be “tuned” by bitumen at the top.
varying the imidazolium substituent groups and the nature of We found that a much cleaner and easier way to visualize
the anion, so that unusual combinations of reagents can be separation was achieved by mixing tar sands with an IL and
toluene at temperatures somewhat above ambient, about
55 °C in some separations, and at room temperature (∼25 °C)
(10) Weingartner, W. Angew. Chem., Int. Ed. 2008, 47, 654–670.
(11) Plechkova, N. V.; Seddon, K. R. Chem. Soc. Rev. 2008, 37, in others. A three-phase system is formed and can be clearly
123–150. observed. We tested three different ILs, [bmmim][BF4],
(12) Liu, J.; Xu, Z.; Masliyah, J. Langmuir 2003, 19, 3911–3920. [bmmim][CF3SO3], and [bmim][CF3SO3] (see Figure 1). The
(13) Wang, S.; Liu, J.; Zhang, L.; Xu, Z.; Masliyah, J. Energy Fuels
2009, 23, 863–869. proportions used in these particular separations were 3 parts
(14) Dai, Q.; Chung, K. H. Fuel 1995, 74, 1858–1864. IL to 2 parts toluene to 1 part oil sand (by weight). Other
1095
Energy Fuels 2010, 24, 1094–1098 : DOI:10.1021/ef9009586 Painter et al.

Figure 3. The phases formed by mixing tar sands with

Figure 4. The infrared spectra of the original tar sand, the bitumen
[bmmim][BF4] and toluene in the proportions 1:3:2 at room tem-
obtained by extraction, and the clay/sand mixture obtained after
perature.
removing residual IL ([bmmim][BF4]) with water.

proportions can be used, but these were the most useful for Table 1. Yields of Bitumen Obtained by Extraction of Oil Sands
visualizing phase separation. The cleanest separation between Using Recycled Ionic Liquid
the phases was given by [bmmim][BF4] and this is shown in
extraction number extraction yield
Figure 3. This particular separation was obtained by simply
stirring the components at room temperature, about 25 °C. 1 13.5%
2 12.7%
The bottom layer consists of sand and clays suspended in the
3 14.5%
IL, the middle layer contains the ionic liquid with a small 4 15.1%
amount of dissolved or suspended bitumen particles and some 5 13.4%
mineral fines, and the top layer is a toluene/bitumen mixture.
This was easily removed from the other layers using a pipet. assumed there would be some variation in the bitumen
Upon evaporation of the toluene from this mixture, we noticed content of such small samples.
some residual ionic liquid in the bottom of the vial, beneath an We now address the questions concerning the presence of
oil/bitumen layer, presumably because we had entrained some any mineral fines in the bitumen extracts and the degree to
IL in our crude laboratory separation. We added a small which the IL can be separated from the residual minerals. In
amount of fresh toluene to the vial, swirled it around, let the the simple experiments described here, the bitumen/toluene
mixture settle, and then decanted the toluene/bitumen phase. layer was simply removed from the phase-separated mixture
The ionic liquid stayed in the bottom of the vial due to its high using a pipet. Any small amounts of entrained IL were easily
viscosity-it did not move when we poured the toluene/ separated by a second treatment of the bitumen with toluene,
bitumen mixture into a new vial, giving us a clean separation. as described above. The IL middle layer and the IL/mineral
In our initial experiments, involving separation at about slurry at the bottom of the tube were also separated by pipet
55 °C, we obtained bitumen yields of 12.6% in our first and each was poured into cold (i.e., room temperature) water,
extraction and 14.7% in a repeat experiment. Shaking a vial the insoluble residues washed with a second batch of water
containing the oil sand, IL, and toluene at room temperature and then dried under vacuum. The IL layer contained small
gave a yield of 7.5% after 10 min. Stirring at room temperature amounts (∼0.4% by weight of the original oil sand sample) of
for 2 h gave a yield of 12.7%. This bitumen appears to be free of predominantly minerals. However, there were no mineral
mineral fines, as we will show below. Toluene extractions of the fines in the toluene/bitumen layer that could be detected using
tar sand gave yields of 14.3 and 14.5%, although the extracts infrared spectroscopy. This is shown in Figure 4, which
also contained some clay fines, as we will also show below. compares infrared spectra (in the 2000-500 cm-1 region) of
ILs are expensive materials and as such will need to be the original tar sands sample. The bitumen obtained after
reused numerous times in any viable process. Accordingly, we evaporation of the toluene and the sand/clay layer after the
also conducted a set of experiments where we reused the ionic liquid was removed using cold water. These samples
middle IL layer (see Figure 3) five times. IL is also part of were from a 55 °C extraction.
the slurry at the bottom of the vial and can readily be Bands due to methylene and methyl groups near 1450 and
separated from the sand using water. However, in these initial, 1370 cm-1 are prominent in the spectrum of the bitumen and
fairly crude experiments we simply made up the missing IL appear with very weak intensity in the spectrum of the tar
with fresh liquid to restore the initial 1:2:3 proportions by sand. They can be more clearly seen in scale-expanded plots
weight of tar sand, toluene, and IL. The yield of bitumen (not shown), but the mineral bands (predominantly quartz
obtained in each extraction (at 55 °C) is listed in Table 1 and and clay) near 1100, 800, and 500 cm-1 absorb very strongly in
varies between 12.7 and 15.1%. This variation is not surpris- the infrared and tend to mask bands due to organic groups.
ing given the crude nature of these experiments and the fact However, these hydrocarbon absorption modes are essen-
that we were using only 1 g of tar sand in each experiment. We tially undetectable in the spectrum of the sand/clay mixture
1096
Energy Fuels 2010, 24, 1094–1098 : DOI:10.1021/ef9009586 Painter et al.

Figure 5. The infrared spectrum of the toluene extract of the oil

sand.
Figure 6. A comparison of the infrared spectra of the residual sand
and clays and the IL [bmmim][BF4].
recovered from the bottom of the tube, even in scale-expanded
spectra. Similarly, the mineral bands are absent from the
processes presently in use, naphtha or paraffinic solvents are
spectrum of the bitumen. This is most easily seen through
added to the separated water/bitumen froth (and some resi-
an examination of the right-hand end of the plots in Figure 4,
dual solvent ends up in the tailings). The product is centri-
near 500 cm-1. These results show that we have separated the
fuged and the solvent removed before the bitumen is refined.
bitumen from the sands without carrying over detectable
Naphtha or other nonpolar hydrocarbon solvents could read-
amounts of fine particles, unlike the hot or warm water
ily be used in a process involving ILs. Accordingly, we have
processes presently used. Conversely, the spectrum of the
not introduced the use of additional solvents, but simply
toluene extract of the tar sands clearly displays bands due to
removed the huge problems associated with the use of water
clays, principally kaolinite, as shown in Figure 5.
in the bitumen/sand separation process and lowered the
The spectrum of the residual sands and clays that remained
temperature at which separation can be achieved. It will still
in the form of a slurry showed no evidence for the presence of
be necessary to use some water to remove the IL from the
residual IL after they had been washed with water. This is
separated sand and clay layer (see below), but far less than is
more clearly seen in Figure 6, which compares the infrared
now required. Furthermore, essentially all of this water can be
spectra of these minerals to that of [bmmim][BF4]. Both the
recycled through the process. Because ILs have essentially
washed minerals and this polar IL absorb very strongly in the
zero vapor pressure at temperatures below 300 °C, water can
infrared, and although the most intense bands of each occur in
easily be separated from the IL by heating to 100 °C.
the same region of the spectrum, they do not coincide. Also
There are an enormous number of ionic liquids that are now
shown in this figure as an insert is the 1750-1250 cm-1 region
becoming available. Some, like the ones used here, are based
of the spectrum on an absorbance expanded scale. Here there
on imidazolium ions. Others are based on cations such as
are several bands that do not overlap, and it can be seen that in
tetraalkyl ammonia and pyrrolidinium. At this stage we
terms of limits of infrared spectroscopy there is no detectable
cannot predict which are capable of giving good separations
IL left in the washed minerals.
in a bitumen extraction process. This is a consequence of two
factors. First, in tar sands there are a range of interactions that
Discussion
occur between the various organic and inorganic components,
Although it is scientifically interesting that certain ILs will and we have yet to identify exactly how ILs interact with the
promote the separation of bitumen from sand in such a facile surfaces of sand and clay particles and their associated ions.
manner, the results presented here do not provide the basis for Second, conventional theories of solubility usually deal with
a commercial process, but they do suggest an approach that nonelectrolytes. Ionic liquids are very different and conven-
may be worth exploring. There are a number of issues that tional measurements, such as heat of vaporization to deter-
need to be addressed, such as the choice of IL, the kinetics of mine solubility parameters cannot be made, because these
the separation process, costs, and environmental concerns. materials have extraordinary low vapor pressures at room
There are a range of ILs and organic solvents that could be temperature. These liquids obviously engage in electrostatic
used in a process of this type. Any organic solvent that interactions, but that does not necessarily mean that all ILs
dissolves bitumen but is immiscible with the IL will accom- will facilitate a separation of bitumen from sand, as shown
plish a separation. Toluene happened to be readily available here. We presently have a working hypothesis involving the
and we have determined that most nonpolar hydrocarbon acidity of ILs. Bitumen becomes strongly attached to sand at
solvents (such as hexane) do not mix with the ILs we have used pH values less than 6.14 Although ILs are salts, they have a
so far. Note that in the hot or warm water separation surprising range of pH values in aqueous solutions, between
1097
Energy Fuels 2010, 24, 1094–1098 : DOI:10.1021/ef9009586 Painter et al.

Figure 7. Schematic plot of a possible separation process.

about 1.6 and 8.3 for the materials studied by Toma et al.15 recycled through the system a large number of times. The ILs
Accordingly, we found that [bmim][CF3SO3], estimated to used here apparently can be, because extraction conditions are
have a pH of about 2.6 in aqueous solutions, did not separate very mild and unlikely to lead to degradation. Also, the
bitumen from sand and clay in the absence of an organic products appear to be cleanly separable with no contamina-
solvent such as toluene. Furthermore, the H2 hydrogen of the tion of the bitumen and residual minerals, either with each
imidazolium cation is weakly acidic, with pKa values of the other or with IL, at least in terms of the limits of detection of
order of 21-23.16 Accordingly, we found that a methyl infrared spectroscopy. The bitumen/solvent mixture that
substituted IL gave the best results in terms of separation of forms on the surface can be separated by decantation and
the ILs we have studied so far. More work clearly remains to any entrained IL removed in a further processing step. The IL
be done to elucidate the mechanism of separation, modes of phase appears to contain a small amount of bitumen and some
interaction and the phase behavior of these mixtures. mineral fines, but these can be simply recycled through the
Although ILs have often been called “green solvents”, this system, as the amount of bitumen would be expected to reach
appears to be largely a reflection of their extremely low vapor equilibrium and not accumulate. In experiments performed so
pressure and imidazolium ILs, particularly those with halogen- far we have reused the middle layer five times in successive
containing anions, are toxic to many forms of aquatic life.17 extractions with no decrease in yield. The sand/clay layer has
Although biodegradable ILs have now been synthesized,18 the entrained IL, but this can be removed by using relatively small
ILs used in this study would have to be cleanly removed from amounts of water. Because the IL has essentially zero vapor
the minerals and recycled through a closed system for a pressure, water is readily distilled from the IL and both can
process to be acceptable. The preliminary results presented then be recycled. A schematic diagram illustrating one possi-
here indicate that this is feasible. ble process is shown in Figure 7.
Another major barrier to a commercial process is economic.
At the present time, ILs are very expensive materials. Even Conclusions
though they are now being widely used in a large number of
Bitumen can be separated from oil or tar sands using an
research projects, getting even ballpark figures for the cost of
ionic liquid. This separation is enhanced by the presence of a
large quantities of ILs is not straightforward, as Plechkova
nonpolar solvent such as toluene. To a first approximation,
and Seddon recently pointed out in a comprehensive review of
given the simple nature of the experiments reported here, it
the use of ILs in the chemical industry.11 However, these
appears that yields in excess of 90% are obtained. Additional
authors estimated that in general ionic liquids are about 5-10
work is required to provide a precise figure. No water at all is
times more expensive than conventional solvents (depending
used in this stage of the separation. Water is used to remove IL
on the specific IL), so that 10-20 recycles would give the same
from the residual sand and clays, but this is easily removed
cost, while a larger number of recycles would make them
from the IL by distillation, because the latter has a; negligible
cheaper. Accordingly, even at today’s prices an economically
vapor pressure under these conditions. There was no detect-
viable industrial scale process would be feasible if the IL can be
able IL contamination of the residual sand and clays and the
(15) Toma, S.; Meciarova, M.; Sebesta, R. Eur. J. Org. Chem. 2009, 3,
bitumen produced in this process was free of both clay fines
321–327. and residual IL, in terms of the detection limits of the
(16) Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth, K. analytical technique used here, infrared spectroscopy.
J. Am. Chem. Soc. 2004, 126, 4366–4374.
(17) Latala, A.; Nedzi, M.; Stepnowski, P. Green Chem. 2009, 11,
1371–1376. Acknowledgment. The authors gratefully acknowledge the
(18) Harjani, J. R.; Farrel, J.; Garcia, M. T.; Singer, R. D.; Scammells, support of the National Science Foundation, Polymers Program,
P. J. Green Chem. 2009, 11, 821–829. under grant DMR-0901180 and DMR-0851897.
1098