BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

ABSORPTION OF CARBONYL SULPHIDE IN AQUEOUS

PIPERAZINE

P.J.G. Huttenhuis1,3, A. Mohan1, S. van Loo1, G.F. Versteeg2

1
Procede Group B.V. P.O. Box 328, 7500 AH, Enschede, The Netherlands
(E-mail: Patrick.huttenhuis@procede.nl)
2
University of Twente, Department of Chemical Engineering P.O. Box 217, 7500 AE,
Enschede, The Netherlands
3
The corresponding author is ir. P.J.G. Huttenhuis, Tel: (þ31) 53 489 5340; Fax: (þ31) 53 489
5399, E-mail: patrick.huttenhuis@procede.nl

KEYWORDS: piperazine, carbonyl sulphide, kinetics, gas treating

INTRODUCTION
Acid gas treating has gained a wide importance over the last few decades. With the
implementation of more stringent environmental regulations, the need for efficient
removal of acid gases like CO2 (carbon dioxide), H2S (hydrogen sulphide) has increased
significantly. Therefore additional effort for research in this field is inevitable. Since most
of the current research activities are focused on CO2 and H2S removal, less attention has
been paid to the removal of the substantially lesser reactive sulphur species like e.g. COS
(carbonyl sulphide), CS2 (carbon disulphide) and RSH (mercaptans), which are also
present in (industrial) gas streams. As the pipeline and environmental specifications are
normally given as total amount of sulphur instead of H2S, the removal of the latter sulphuric
species becomes more and more important.
One of the most used processes in the gas treating industry is the absorption of acid
(gaseous) components with aqueous alkanolamine solutions. An extensive overview of
these different treating processes is given by Kohl and Nielsen (1997). For the design
of gas treating equipment the gas absorption rate should be detemined. This absorption
rate is determined by:

. Mass transfer coefficient kg, kl and interfacial area a

. The physical and chemical solubility of the gaseous component
. The enhancement factor, which is determined by the reaction rate of the gaseous com-
ponent with the absorption liquid

The current work is focused on the absorption kinetics of COS in aqueous pipera-
zine solutions. This study was carried out as part of a research program which focuses
on the removal of acid gases using activated aqueous amine solutions. An activated
solvent basically consists of a tertiary amine in combination with a small amount of a
secondary or primary amine (activator). A blend of MDEA (methyl diethanol amine)

581
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

and piperazine (a cyclic secondary amine) was found to be effective for efficient removal
of CO2 (Bishnoi 2000, Derks 2005). With this blend the advantages of a tertiary amine
(high capacity, cheap, stable) are combined with the advantages of primary/secondary
amines (high reaction rates). Because of the similarities between CO2 and COS, it is
expected that these activated amine solutions would be effective for COS removal. It
must be stated, however that the reaction rate of COS with alkanolamines is about two
orders of magnitude lower compared to CO2. The reaction mechanism between COS
and an aqueous solution of MDEA and piperazine is very complex. Before this mechanism
can be studied in detail the reaction rate expressions for the reaction between COS and the
individual amines (MDEA and PZ) must be determined independently. The reaction
between MDEA and COS has been studied before (ref Littel; Al Ghawas). In this work
the reaction of COS in a aqueous solution of piperazine has been studied quantitatively.
Data used in this work can be used for improving the quality of rate based absorption
models used in the gas treating industry.

THEORETICAL
The chemical structure of MDEA and piperazine are given in Figure 1.
From this structure it can be seen that piperazine contains two secondary nitrogen
atoms and MDEA contains one tertiary nitrogen atom. For CO2 it is known that tertiary
amines do not react directly with an amine, but a catalyzed acid-base reaction occurs.
The reaction of COS with primary and secondary amines proceeds similarly to CO2,
via a zwitterion mechanism:

COS þ R2 NH ! R2 NHþ COS
(1)

R2 NHþ COS
þ B ! R2 NCOS
þ BHþ (2)

To determine the reaction rate of the COS-PZ system directly from absorption
experiments, the reaction should be carried out in the so-called pseudo-first order
regime. In this regime the following condition should be satisfied:

2 , Ha ,, ECOS,1
pffiffiffiffiffiffiffiffiffiffiffiffi
k1,1 DCOS DPZ ½PZ
with, Ha ¼ Kl and ECOS,1 ¼¼ 1 þ g
PZ DCOS ½COSi

The above relation for ECOS,1 is only applicable for irreversible reactions. It is
known that the reaction between amines and acid gases are reversible reactions, which
has a decreasing effect on ECOS,1 (Versteeg et al., 1988). This can be taken into
account by applying a more severe condition on Ha , , ECOS,1 (usually a factor 10).

582
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

Figure 1. Chemical structure of piperazine and MDEA

EXPERIMENTAL SETUP AND PROCEDURES

Both the vapour liquid equilibrium and the absorption rate experiments have been carried
out in a 1 liter stirred cell. A schematic sketch of the experimental set-up is given in
Figure 2. The set-up contains the following parts:
. Büchi reactor with water bath heater and stirrer
. gas (N2O and COS) supply vessels

To scrubber

To vacuum
TI pump
PI
N2O from gas N2
cylinder
TI PI Caustic vessel for
neutralizing

TI
COS from PI
lecture bottle

Heating fluid Heating fluid

Figure 2. Sketch of the experimental set-up

583
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

. Vacuum system to remove inert gases

. Gas disposal system (Scrubber with caustic solution)
Before each experiment approximately 0.5 liter aqueous solution is fed to the
reactor. The reactor is kept at the desired temperature with a water bath.
The absorption experiments of COS in aqueous piperazine solutions were carried
out in the batch stirred cell reactor using a flat interface. The experiments were conducted
in a temperature range of 2932313 K and piperazine concentrations between 0.1 and
1.5 M.
The gases used in this work were N2O (purity 99þ%) and COS (99.9%) and were
delivered by Hoekloos and Scott respectively. Piperazine (anhydrous; purity 99þ%) was
delivered by Across and an aqueous alkanolamine solution was prepared by using de-
mineralised water. A more detailed description of experimental setup and procedure can
be found in Blauwhoff et al. (1984).
Knowing the total gas and liquid volume in the reactor the dimensionless gas solu-
bility can be derived from Henry’s relation.

H E (Pint
Peq ) Vg
m¼ ¼ 
RT Peq
Pvap Vl

Where:
m ¼ dimensionless solubility, defined as the ratio of the liquid-phase concentration
to the gas-phase concentration of the solute at equilibrium conditions.
The liquid mass transfer coefficient, kl, was calculated using the decrease in pressure
as a function of time using the following formula
   
P(t)
Peq mV1 þ Vg
ln ¼ kl :a: t
Pint
Peq Vl Vg

The data pressure decrease experiments of COS for both the solubility and the
kinetics respectively were monitored and recorded. The physical solubility of COS in
aqueous piperazine solutions sere derived via the so-called COS-N2O analogy. The solu-
bility of COS and N2O in pure water were taken from Al Ghawas (1988) and Versteeg
(1988) respectively.
For the absorption experiments, the flux equation was used as presented below:
PCOS
JCOS ¼ mCOS kl E
RT
The enhancement factor is equal to Hatta number if pseudo first order conditions are
fulfilled.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Kapp DCOS
Ha ¼
kl

584
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

For pseudo first order reaction 2 , MH  ECOS,1

Da min e RT
ECOS,1 ¼ 1 þ
DCOS gmCOS PCOS

The diffusivity of COS in aqueous piperazine was estimated from the modified
Stokes Einstein relationship given by Versteeg (1988). The diffusivity of COS in water
was found using the diffusivity value in 298 K and using the same activation energy as
used for diffusivity of CO2 in water. Similarly the diffusivity of piperazine in aqueous
piperazine solution was found using another modified stokes Einstein’s relationship devel-
oped by Versteeg et al. (1988). The diffusivity of amine in water was estimated using the
method given by Wilke-Chang (1955). The viscosities of aqueous piperazine were
obtained from the experimental measurements of Derks et al. (2005).

RESULTS
In order to evaluate the experimental setup and procedures used in this work some cali-
bration experiments were carried out. Therefore the kinetics of the reaction between
COS and aqueous DEA were determined and compared with the data measured by
Littlel et al. (1992). These experiments were also carried out in the pseudo first order
regime (2 , Ha  ECOS,1). It appeared that pseudo first order constant determined by
Littel (1992) could be reproduced in this work within 10%. So it was concluded that
the experimental set-up and procedures are suitable to measure absorption of COS in
aqueous piperazine system.
After the validation experiments of COS with aqueous DEA, absorption studies for
COS in aqueous piperazine were conducted at various concentrations (0.1, 0.3, 0.6 &
1.5 M) at temperatures of 293, 303 and 313 K. The experiments were conducted at a
COS initial partial pressures of approximately 0.5 bar. Figure 3 shows the experimental
measurements of apparent pseudo first order rate constants. The reaction order was
found to be unaffected by temperature within the experimental accuracy. The reaction
order in amine was about 1.6 which confirms that the reaction undergoes a proceeds via
a zwitterion type of mechanism. This supports the findings of Little (1991) who observed
that the reaction order of COS with secondary alkanolamines with respect to amine con-
centration was also between 1–2 and the reaction order with respect to COS was 1.

DISCUSSION AND CONCLUSION

During the kinetic experiments of COS in aqueous piperazine it appeared that the order of
the reaction was about 1.6 in piperazine concentration. This strongly suggests that the
reaction between COS and piperazine can be described by the zwitterions mechanism
as given in reactions 1 and 2. If pseudo steady state conditions are assumed for the
COS-amine zwitterion concentration, the following expression can be derived for the

585
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

10000

1000
Kapp [s-1]p

100
293 K
303 K
313 K
10

1
10 100 1000 10000
[PZ] [mol/m3]

Figure 3. Pseudo first order constant for the reaction of COS in aqueous piperazine

overall forward reaction rate:

½Pz½COS
RCOS ¼ 1 1
¼ Kapp ½COS
þ
K2 KH2 O ½H2 OþKR2 NH ½Pz

with

K2 Kb,H2 O K2 Kb,Pz
KH2 O ¼ ; KPz ¼
K
1 K
1

All the bases in the solution contribute to the zwitterion de-protonation rate. The
contribution of bases other than H2O and the amine itself were found to be negligible.
The observed reaction rate equations were fitted using a levenberg-marquardt fitting
procedure. It appeared that the overall reaction rate was a function of zwitterion
de-protonation rate. This means that the zwitterion formation rate K2 can be neglected.
Figure 2 was fitted simultaneously using 2 independent variables (piperazine and water
concentration). The best fit was obtained and Arrhenius plots for the de-protonation rate
constants were obtained.

586
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

From these best fits the following Arrhenius equations were developed:
 
3
5304
Piperazine: KPZ ¼ 9:33  10 exp ;
T
 

8989
KH2 O, PZ ¼ 2:0  107 exp this work
T
 

5714
DEA: KDEA ¼ 1:16  103 exp ;
T
 
5
9199
KH2 O, DEA ¼ 2:45  10 exp Littel (1992)
T
 

7346
DGA: KDEA ¼ 5:34  105 exp ;
T
 

9409
KH2 O, DGA ¼ 4:9  106 exp Littel (1992)
T

From the expressions described above, it can be observed that the measured acti-
vation energies for piperazine are of the same order of magnitude as DEA and DGA,
being slightly higher for DGA. For KH2 O , it can be seen that the activation energies are
almost identical for all the three amines. Table 1 represents the values of the pseudo
first order constants for 1 M amine solutions piperazine (this work), DEA and DGA
(Little 1992) at 293 K.
From the table above, it can be concluded that the pseudo first order constant of
piperazine is significantly higher, than the ones measured for DEA and DGA. So it
seems that piperazine is an interesting accelerator for the removal of COS from industrial
gas streams as currently used for CO2 removal. Further kinetic research is required to study
the reaction mechanism of blended amine systems like MDEA-piperazine. Data used in
this work can be used for improving the quality of rate based absorption models used in
the gas treating industry.

Table 1. Pseudo first order reaction rate for 1 M amine

at 293 K

1M amine Kapp [s21] ref.

Piperazine 1.8 E-4 This work

DEA 4.3 E-6 Littel (1992)
DGA 1.0 E-5 Littel (1992)

587
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

NOTATIONS
C Concentration [mole/m3]
P Pressure [bara]
D Diffusion coefficient [m2/s]
E Enhancement factor [2]
Ea,1 Infinity Enhancement factor of component a [2]
Ha Hatta Number [2]
HE Henry’s constant [bar.m3/mol.Kelvin]
JA Molar flux of component A from gas to liquid phase [mole/m2/s]
kapp Apparent reaction constant for a Pseudo- first order reaction. [s21]
k Mass transfer coefficient [m/s]
m Distribution coefficient (Cl/Cg) [2]
A Area of interface [m2]
N Stirrer speed [rpm]
V Volume [m3]
T Temperature [Kelvin]
R Gas constant [8.314e-05 bar.m3/mol.K]
B Base [2]

SUBSCRIPTS
Int Initial
L Liquid
G Gas
Eq At equilibrium conditions
App Apparent

REFERENCES
1. Al Ghawas, “Simultaneous absorption of H2S, CO2 and COS in aqueous MDEA”, Thesis,
(1988).
2. Bishnoi, S., Rochelle, G.T., (2000) Absorption of Carbon Dioxide in Aqueous Piperazine:
Reaction Kinetics, Mass Transfer and Solubility, Chem. Eng. Science, 55, pp. 5531.
3. Blauwhoff, P.M.M., Versteeg, G.F., van Swaaij, W.P.M., A study on the reaction between
CO2 and alkanolamines in aqueous solutions, 1983, Chem. Eng. Sc. 38, (9), pp. 1411.
4. Chang, P., Wilke, C.R., (1955) Correlation of diffusion coefficients in dilute solutions,
AiChE Journal, 1 (2) pp. 264 J., 38, pp. 244.
5. Derks, P.W., Hogendoorn, K.J., Versteeg, G.F., (2005) Solubility of N2O in and Density,
Viscosity, and Surface Tension of Aqueous Piperazine Solutions, J. Chem. Eng. Data, 50
(6), pp. 1947.
6. Derks, P.W.J., Dijkstra, H.B.S., Hogendoorn, J.A., Versteeg, G.F., (2005) Solubility of
Carbon Dioxide in Aqueous Piperazine Solutions, AiChE Journal, 51 (8), pp. 2311.

588
BK1064-ch55_R2_250706

SYMPOSIUM SERIES NO. 152 # 2006 IChemE

7. Kohl, A.L., Nielsen, R.B., (1997) Gas Purification 5th Ed., Gulf Publishing Company,
Houston.
8. Littel, R.J. Versteeg, G.F., van Swaaij, W.P.M., (1992) Kinetics of CO2 with primary and
secondary amines in aqueous solutions., Chemical Engeineering Science, 47 (8), pp. 2027.
9. Littel, R.J., Versteeg, G.F., van Swaaij, W.P.M., Kinetic study of COS with tertiary alka-
nolamine solutions. II. Modeling and experiments in a stirred cell reactor, Ind. Eng. Chem.
Res., 31, pp. 1269.
10. Sada, E., Kumazawa, H., Butt, M (1978) Solubility and Diffusivity of Gases in Aqueous
Solutions of Amines, J. Chem. Eng. Data, 23 (2), pp. 61.
11. Versteeg, G.F., and van Swaaij, W.P.M. (1988) On the kinetics between CO2 and alkano-
lamines both in aqeous and non-aqeous solutions. II Tertiary amines, Chem. Eng. Sci., 43
(3), pp. 587.
12. Versteeg, G.F., van Swaaij, W.P.M., (1988) Solubility and Difussity of acid gases (CO2,
N2O) in aqueous alkanolamine solutions, J. Chem. Eng. Data, 33 (1) pp. 29.

589