Design and Proof of Concept of a Continuous Pressurized Multistage Fluidized Bed Unit for Deep Sour Gas Removal Using Adsorption

Authors: 
Driessen, R. T., University of Twente
Knaken, B., University of Twente
Buzink, T., University of Twente
Brilman, D. W. F., University of Twente
Design and proof of concept of a continuous pressurized multistage
fluidized bed unit for deep sour gas removal using adsorption

Rick T. Driessen, Benno Knaken, Tim Buzink,
Derk W.F. Brilman

Sustainable Process Technology, Faculty of Science and
Technology, University of Twente, PO Box 217, 7500 AE Enschede, The Netherlands
Introduction

Adsorption is a widely used separation technology in the
(petro)chemical industry. The conventional process for an adsorption process
are multiple fixed beds, operated in a certain sequence to enable continuous
processing while using adsorption and desorption. Although pressure swing
adsorption (PSA) using fixed beds is an industrially established process, it has
some disadvantages:

1.       Ineffective
use of sorbent: In fixed bed adsorption only the sorbent located within the
mass transfer zone is active. The sorbent upstream and downstream of the mass
transfer zone are either saturated with adsorbates or don’t see any adsorbate
yet. As a consequence, the effective use of the sorbent, during adsorption and
at breakthrough, is relatively low resulting in large adsorption beds.

2.       Slow
heat transfer: When using temperature swing adsorption (TSA), heat has to be
supplied to the fixed bed. However, fixed beds show slow heat transfer. This
implies that (a) heating the fixed bed to the desired temperature takes a
considerable amount of time or (b) additional investment is needed to provide
sufficient heat transfer area. Both lead to higher costs.

This work proposes a process which overcomes these two
disadvantages. First, to overcome the ineffective use of the sorbent, a continuous
adsorption process could be designed in which the adsorber length matches the
length of the mass transfer zone, and where the solid circulates between the adsorption
zone and a desorption zone. This implies that almost every sorbent particle is
either adsorbing or desorbing, minimizing the amount of sorbent in an inactive
zone. Second, with respect to heat transfer, gas-solid fluidization can be
applied, since fluidized beds are known for their excellent heat transfer
because of the vigorous movement of the particles. This work combines the two
mentioned characteristics by using multistage fluidized beds (MSFBs) to adsorb
sour gas compounds (carbon dioxide and hydrogen sulfide) on supported amine
sorbents (SAS) from natural gas. For this application, deep removal is needed
because the pipeline specifications of natural gas are tight: <3 ppm H2S
and 1-2% CO2 (50-150 ppm CO2 for liquefied natural gas).

Recently, continuous adsorption processes employing MSFBs
received renewed attention in view of post-combustion carbon dioxide (CO2)
capture. For a full review of these papers we refer to our earlier work,
nevertheless we want to highlight some [1].
Veneman et al. use a gas-solid trickle flow reactor to adsorb CO2
and use a multistage fluidized bed for desorption [2].
The group of Meikap investigated multistage fluidized beds experimentally with
various sorbents [3,4].
The group of Hofbauer built a bench scale, ambient pressure, multistage
fluidized bed and presented their first experimental results [5].
All these researchers use continuous MSFB processing with TSA to adsorb CO2
from gases. However, to our knowledge, the use of PSA alone, let be
pressure-temperature adsorption (PTSA) in a continuous sorbent process has not
been investigated.

This paper presents the design of our experimental continuous,
pressurized multistage fluidized bed setup for PTSA operation. Furthermore, we
present some first results to prove that this process can be used for the
removal of sour gases from natural gas. Experimental setup

The lay-out of the setup is too extensive to be described
here in detail, but a summary will be given. The core components of the experimental
setup are, in order of circulation: a MSFB adsorber (at elevated pressure), a
high pressure/low pressure (HP/LP) lock to depressurize the sorbent, a MSFB
desorber (at atmospheric pressure), a riser to transport the solids (at
atmospheric pressure), and a low pressure/high pressure (LP/HP) lock to
pressurize the sorbent. The MSFB adsorber (three stages, 50 mm inner diameter
and a shallow bed height of 13 cm) is typically operated at 20-60 °C and at
elevated pressures up to 10 bara. The HP/LP lock is a vessel operated in a
sequential manner: (1) pressurization with nitrogen (N2) to adsorber
pressure; (2) filling with sorbent from adsorber; (3) depressurization to
desorber pressure; (4) emptying the sorbent to the desorber. The MSFB desorber
(seven stages, 200 mm inner diameter) is operated at approximately 100 °C and at
atmospheric pressure. The riser transports the sorbent pneumatically with N2
to a cyclone where the riser gas is separated from the sorbent. The sorbent is
fed to the LP/HP lock, which is operated in a sequential manner similar to the
HP/LP lock. Each main unit operation (adsorber, HP/LP lock, desorber and LP/HP
lock) is equipped with a buffer vessel to facilitate a full continuous flow of
sorbent.

A commercial ion exchange resin functionalized with
benzylamine (Lewatit VP OC 1065, Lanxess) is used as amine sorbent. This amine
sorbent is capable of adsorbing both H2S and CO2. Both
methane, the main compound in natural gas, and N2 do not adsorb on
this amine sorbent. Therefore, N2 is used to mimic methane because
of safety reasons. A LI-COR LI840A CO2 analyzer (0-20 000 mol ppm)
is used to measure the inlet, outlet and concentrations of CO2 per
stage. [6,7] First results

The first round of experiments was conducted with mixtures
of CO2 and N2. CO2 concentrations of a few parts
per million were achieved, while having an inlet concentration of approx. 10 000
mol ppm. Deep removal is therefore shown to be feasible possible, proving the
concept of this process. In the first stage already the majority of CO2
is adsorbed: the concentration is lowered from 9500 mol ppm to 1500 mol ppm.
The second and third stages enable the deep removal of CO2. The
residence time of the gas and solid phase in the adsorber are about 6 s and 45
s respectively, underlining the fast rate of the process.

Gas-based Murphree tray efficiencies, representing the
degree to which adsorption equilibrium is reached, were calculated and show
high values. The measured tray efficiencies are high: generally the tray
efficiency is larger than 0.8 and frequently they are higher than 0.9. This
again underpins that the adsorption process is fast: the gas and the sorbent
are approaching equilibrium. Conclusion

A new process for continuous adsorption is presented employing
MSFBs to (a) maximize the effective use of sorbent and (b) speed up heat
transfer. To our knowledge, we are the first researchers who apply a pressure
swing in a continuous sorbent flow adsorption process. The first results show
that the adsorption process is fast: gas and solid residence times are low, in
the order of seconds, while deep removal down to a few mol ppm (over 99.8%
removal) is possible. This is a prove that this process could be used for the
deep removal of H2S and CO2 from natural gas. In
addition, tray efficiencies indicate that the gas and the sorbent are near their
equilibrium.  

In view of the targeted application, sour gas removal from
natural gas, activities are ongoing to investigate the influence of various
process parameters (such as gas velocity, solid flow and inlet concentrations).
In the next step, hydrogen sulfide will be introduced to mimic natural gas sweeting
for sour gas mixtures. Acknowledgements

This research was carried out in the context of the Compact
Advanced Sour gas Processing (CASPer) project, coordinated by the Institute for
Sustainable Process Technology and co-financed by the Ministry of Economic
Affairs of the Netherlands (RVO.nl project number: TEEI115008). References

[1]         R.T. Driessen, M.J. Bos, D.W.F.
Brilman, A Multistage Fluidized Bed for the Deep Removal of Sour Gases: Proof
of Concept and Tray Efficiencies, Ind. Eng. Chem. Res. 57 (2018) 3866–3875.
doi:10.1021/acs.iecr.7b04891.

[2]         R. Veneman, T. Hilbers, D.W.F. Brilman,
S.R.A. Kersten, CO2 capture in a continuous gas-solid trickle flow
reactor, Chem. Eng. J. 289 (2016) 191–202. doi:10.1016/j.cej.2015.12.066.

[3]         D. Das, D.P. Samal, B.C. Meikap,
Removal of CO2 in a multistage fluidized bed reactor by diethanol
amine impregnated activated carbon, J. Environ. Sci. Heal. Part A. 51 (2016)
769–775. doi:10.1080/10934529.2016.1170462.

[4]         S. Roy, C.R. Mohanty, B.C. Meikap,
Multistage Fluidized Bed Reactor Performance Characterization for Adsorption of
Carbon Dioxide, Ind. Eng. Chem. Res. 48 (2009) 10718–10727.
doi:10.1021/ie901133r.

[5]         F. Dietrich, G. Schöny, J. Fuchs, H.
Hofbauer, Experimental study of the adsorber performance in a multi-stage
fluidized bed system for continuous CO2 capture by means of
temperature swing adsorption, Fuel Process. Technol. 173 (2018) 103–111.
doi:10.1016/J.FUPROC.2018.01.013.

[6]         S. Sutanto, J.W. Dijkstra, J.A.Z.
Pieterse, J. Boon, P. Hauwert, D.W.F. Brilman, CO2 removal from
biogas with supported amine sorbents: First technical evaluation based on
experimental data, Sep. Purif. Technol. 184 (2017) 12–25.
doi:10.1016/j.seppur.2017.04.030.

[7]         E. Sonnleitner, G. Schöny, H. Hofbauer,
Assessment of zeolite 13X and Lewatit VP OC 1065 for application in a
continuous temperature swing adsorption process for biogas upgrading, Biomass
Convers. Biorefinery. (2017) 1–17. doi:10.1007/s13399-017-0293-3.

Abstract: