(193c) Adsorptive Separation of Water-Acetonitrile Mixtures

Acetonitrile is a commonly used
chemical with applications in pharmaceutical, analytical and synthetic domains.
Most applications actually involve a mixture of acetonitrile and water,
including Reversed Phase-HPLC. Acetonitrile, a byproduct of acrylonitrile
production, has seen a large rise in cost due to decreasing supply since 2008. The
lack of a direct production method for acetonitrile resulted in falling supply
and a dramatic increase in price. The problem grew to such proportions that it
became known as "The Great Acetonitrile Shortage" [1]. It is therefore of great
importance to recuperate acetonitrile after its use, from both economic as
ecologic point of view. The separation of both molecules is a challenge since
both are highly polar and have small kinetic diameters. The conventional method
to separate water and acetonitrile requires a pressure swing or azeotropic
distillation due to the presence of an azeotrope. These techniques are quite
expensive in both capital costs as well as operating costs. The use of
pervaporation in this separation problem is quite promising, but suitable
membranes are difficult to synthesize in a defect-free manner and the capital
cost are quite high.

The use of an adsorption column
to preferentially adsorb water out of these water-acetonitrile mixtures could
provide a straight-forward and more cost-effective method to produce high grade
acetonitrile. Such a device is also suited to be used with smaller waste
streams, as encountered in HPLC, where distillation units are simply too large.
Such a separation device has to be used in bulk separation mode as the amount
of water is typically 50 wt.%. This is a significant difference with more
conventional adsorption columns where the concentration of the adsorbents is
often no more than several wt.%.

The present work deals with adsorption and separation of water and acetonitrile
on solid adsorbents. In a first screening round, 29 sorbents of different classes including
zeolites, alumina, Metal-Organic Frameworks and activated carbons were evaluated
for their selectivity in the adsorption of water/ acetonitrile mixtures, by
performing batch adsorption experiments. Composition of the liquid mixtures was
determined using GC-MS and Karl-Fisher titration. These results from these experiments were
implemented in numerical evaluation tool to rank the performance of the
sorbents to separate water and acetonitrile. This screening allowed the
selection of a distinct number of sorbents showing outstanding preferential
adsorption of water. Only 3 of the 29 sorbents showed a preference towards
acetonitrile, while most sorbents have a preference towards water ranging from
very poor to excellent. Surprisingly even aluminum-deficient hydrophobic
zeolites showed a slight preference towards water. A strong correlation between
the Si/Al ratio and water adsorption capacity for a low water content mixture
was found for all zeolites with different structures. The adsorption enthalpy
(kJ/mol) of acetonitrile is in fact quite strong [2], but due to the large
molar volume of the acetonitrile compared to water, it is energetically more
favorable to adsorb water for most sorbents.

The selected sorbents were further investigated by assessing their
competitive water-acetonitrile isotherms at 296K. Two sorbents were identified
showing excellent separation behavior. The first sorbent shows a quasi-rectangular
isotherm and a very large Langmuir constant. Average adsorption enthalpies were
in excess of 100 kJ/mol. The second sorbent shows a lower Langmuir constant,
but a slightly higher water saturation capacity of about 0.25 g/g. These two
sorbents were studied in more detail. The kinetic uptake curves of water out of
water-acetonitrile mixtures were determined for both powders as pellets at room
temperature. Diffusion constants were obtained via fitting with a numerically
solved diffusion model. Additional isotherms at 307K, 313K and 333K allowed to
evaluate the temperature dependency of capacity and selectivity. Separation was
tested in breakthrough mode using 30 cm columns packed with pellets and powder.
Experimental breakthrough curves, obtained at different flow rate, feed
composition and column temperature, were compared to simulated breakthrough
curves. The model accounts for velocity variation due to the large amount
adsorbed. Adsorption parameters obtained in the independent isotherm
experiments were used in the model. Breakthrough of water was accurately
predicted by the model, as shown in Figure 1

Figure  SEQ Figure \* ARABIC 1: Experimental
breakthrough curve (dots) of water in 30wt.%water-70wt.%acetonitrile on a 30 cm
column (diameter ½ inch) packed with 1 mm sorbent pellets at a flow rate
of 0.6 ml/min compared to simulation (dotted line).

In conclusion, it is shown in the present work that adsorption processes
using porous solids are promising for the removal of acetonitrile from waste
stream in small or even larger scale processes.

References

[1] Lowe D., (2009), The Great Acetonitrile Shortage,
Available at http://pipeline.corante.com, Retrieved August 26, 2009

[2] Janchen J., van Wolput J.H.M.C., van Well W.J.M, Stach H., (2001),
Adsorption of water, methanol and acetonitrile in ZK-5 investigated by
temperature desorption, microcalorimetry an FTIR, Thermochimica Acta, 379,
pp. 213-225