Contents lists available atScienceDirect
Applied Energy
journal homepage:www.elsevier.com/locate/apenergy
Two-train elevated-temperature pressure swing adsorption for high-purity hydrogen production
Xuancan Zhu, Yixiang Shi
⁎, Shuang Li, Ningsheng Cai
Key Laboratory for Thermal Science and Power Engineer of Ministry of Education, Department of Thermal Engineering, Tsinghua University, Beijing 100084, People's Republic of China
H I G H L I G H T S
•
Two-train elevated temperature pressure swing adsorption for H2purification.•
Optimal process achieves 99.9994% hydrogen purity and 97.51% hydrogen recovery ratio.•
Total steam consumption is significantly reduced with reflux structures.A R T I C L E I N F O
Keywords:
Elevated-temperature pressure swing adsorption
Potassium-promoted layered double oxide Hydrogen production
Shifted gas
A B S T R A C T
The trade-offbetween hydrogen recovery ratio (HRR) and hydrogen purity (HP) is one of the main drawbacks in normal temperature pressure swing adsorption (NT-PSA) for producing high-purity hydrogen from shifted gas. In this paper, a two-train elevated-temperature pressure swing adsorption (ET-PSA) process that achieved 99.999%
HP and over 95% HRR is proposed, which has wide application potentials in fuel cells and chemical industries.
Potassium-promoted layered double oxide (K-LDO), which shows reasonable working capacity and fast ad- sorption/desorption kinetics at elevated temperatures (200–450 °C), is adopted as the CO2adsorbent. CO in the shifted gas is co-purified by high-temperature water gas shift (WGS) catalysts added to the columns. Thefirst- train ET-PSA adopted an eight-column thirteen-step configuration with shorter step time to remove most of the CO/CO2in the shifted gas, and the second-train ET-PSA adopted a double-column seven-step configuration with longer step time to purify the residual gas impurities. The introduction of co-current high-pressure steam rinse and counter-current low-pressure steam purge is the key to achieve both high HRR and HP. The high-tem- perature steam is the main energy consumption of ET-PSA rather than low HRR in NT-PSA, and the total steam consumption is reduced by adopting the tail gas from second-train ET-PSA as the purge gas forfirst-train ET-PSA.
The optimal results achieved 97.51% HRR and 99.9994% HP with only 0.188 rinse-to-feed ratio and 0.263 purge-to-feed ratio, which are the highest values reported for PSAs producing high-purity hydrogen from carbon- based fuels.
1. Introduction
With the growing demand for energy and increasing environmental pressure around the world, it has become necessary to replace the current main energy source (combustion of fossil fuels) with cleaner and more efficient energy systems. Hydrogen is expected to be a fuel of the future and has the potential to contribute to 90% energy generation in 2080 [1]. Emerging technology is expected to break the present energy systems by combining hydrogen as an energy carrier with fuel cells as an electricity generator[2]. Hydrogen is also an important raw material that is widely used in chemical industries (for instance,
ammonia synthesis)[3]. Hydrogen is conventionally produced by hy- drocarbon reforming [4], biomass processes [5], and water splitting (including photocatalytic processes[6], electrolytic processes[7], and thermal processes[8]). Despite renewable sources attracting attention, steam methane reforming (SMR) and coal gasification (CG) are the most cost-effective processes for producing hydrogen, at present and for the near future[9].
In hydrogen production processes from SMR or CG, a two-stage water gas shift (WGS) reactor is adopted to remove carbon gases from reformed gas, followed by CO2 absorption and methanation for CO purification, or alternatively, a pressure swing adsorption (PSA) unit
https://doi.org/10.1016/j.apenergy.2018.08.093
Received 7 June 2018; Received in revised form 2 August 2018; Accepted 16 August 2018
⁎Corresponding author.
E-mail address:[email protected](Y. Shi).
Applied Energy 229 (2018) 1061–1071
0306-2619/ © 2018 Published by Elsevier Ltd.
T
[4]. From the perspective of exergetic evaluation, PSA is a desirable hydrogen purification technology achieving above 99.9% hydrogen purity (HP) [5]. Common PSA systems work at normal temperature ranges (NT-PSA) by adopting physical adsorbents such as activated carbon, zeolite, and silica gel[10]. NT-PSA is a cyclic process where the gas impurities arefirst adsorbed on the surfaces of adsorbents and then desorbed by reducing the pressure[11]. Thermal energy duty for re- generation is avoided during NT-PSA operation. However, for achieving high HP, NT-PSA faces limitations such as low hydrogen recovery ratio (HRR) and high system complexity[12]. For hydrogen production from typical steam methane reformed gas (70–80% H2, 15–25% CO2, 3–6%
CH4, 1–3% CO, and trace N2), a typical NT-PSA achieves 98–99.999%
HP and 70–90% HRR[10].
When fixing the column number of NT-PSA, there is a trade-off between HRR and HP. Ribeiro et al.[13]demonstrated that shortening the adsorption time from 160 s to 120 s increased HP of a four-column eight-step NT-PSA from 99.8193% to 99.9992% while decreasing HRR from 71.8% to 62.7%. A similar phenomenon was found when counter- current purge rate was increased when hydrogen was commonly adopted as purge gas[14]. When the column number was increased, both HRR and HP increased due to increase in pressure-equalization steps. Moon et al.[15]indicated that a two-column NT-PSA achieved 99.77–99.95% HP with 73.30–77.64% HRR, and a four-column NT-PSA could achieve above 99.97% HP with 79% HRR. Lopes et al. [16]
proposed a ten-step NT-PSA with three pressure-equalizations, which achieved 99.981% HP and 81.6% HRR from afive-component gas (79%
H2, 17% CO2, 1.2% CO, 2.1% CH4, and 0.7% N2). HRR could be further increased to 92.7% with 99.993% HP for a twelve-column thirteen-step NT-PSA to purify raw hydrogen fuel gas in IGCC power plants[17].
However, the large column number decreased productivity and in- creased the operating complexity and capital cost (CAPEX)[18].
Alternatively, the concept of elevated-temperature pressure swing adsorption (ET-PSA) has been proposed for producing hydrogen from shifted gas[19]. ET-PSA works at elevated temperatures (200–450 °C), thus allowing the shifted gas to directly enter the purification unit without pre-cooling. Chemisorbents like potassium-promoted layered double oxides (K-LDOs) [20] and molten salt-promoted MgO [21], which show high CO2working capacity and fast adsorption/desorption kinetics at elevated temperatures, are adopted as CO2adsorbents in ET- PSA system. ET-PSA allows the use of steam as rinse and purge gas, where the co-current steam rinse squeezes residual hydrogen out of the column after the adsorption step and counter-current steam purge is adopted to replace hydrogen purge in the NT-PSA process[22]. The high temperature steam, which could be produced in a heat recovery boiler with waste heats from the gasifier, the exhaust gas from gas turbine or other subsystems, was the main energy consumption of ET- PSA [23]. The ET-PSA process can be applied in sorption-enhanced water gas shift (SEWGS), where WGS catalysts are mixed into CO2
adsorption columns to transform the residual CO in the shifted gas to produce extra H2[24].
A series of ET-PSA processes including four-column eight-step cycle [25], six-column eight-step cycle [26], seven-column ten-step cycle [27], eight-column eleven-step cycle[28], and nine-column eleven-step cycle[29]has been proposed. One of the optimization objectives for ET-PSA is to reduce the steam amount for rinse and purge, as it is the main energy consumer in ET-PSA[23]. For SEWGS process, Reijer et al.
[26]showed that the rinse mainly affected CO2purity and the purge mainly affected CO2capture ratio. The rinse-to-carbon and purge-to- Nomenclature
Ai pre-exponential factor for reaction rate, s−1or kg/mol s c fitting parameter for pressure effect,–
Ci gas molar concentration, mol/m3 Cfeed,i feed gas molar concentration, mol/m3 Crinse,i rinse gas molar concentration, mol/m3 Cpurge,i purge gas molar concentration, mol/m3
Cpep,i given gas molar concentration for pressure-equalization pressurization, mol/m3
Cpp,i given gas molar concentration for product pressurization, mol/m3
CT total gas molar concentration, mol/m3
Ctransfer,̇ i mass transfer between bulk gas and particles, mol/m3s Dax,i axial dispersion coefficient, m2/s
Dp diameter of adsorbent/catalyst particles, m Ei activation energy, J/mol
Ei0 initial activation energy, J/mol ki reaction rate, s−1or kg/mol s
kped valve rate for pressure, –equalization depressurization, s−1
kdep valve rate for blowdown, s−1
Keq equilibrium constant for WGS reaction,– Lb column length, m
madsorbents,total total adsorbents mass, kg MCO2 CO2molar mass, kg/mol Mg mixed gas molar mass, kg/mol p pressure, Pa
p0 standard atmospheric pressure, Pa pi partial pressure of componenti, Pa pfeed feed pressure, Pa
prinse rinse pressure, Pa
pped given pressure for pressure, –equalization depressuriza- tion, Pa
pdep given pressure for blowdown, Pa ppurge purge pressure, Pa
qi site concentration, mol/kg qi,0 initial site concentration, mol/kg qtotal total CO2adsorption capacity, mol/kg qAS sum ofqAandqO s( ), mol/kg
Qproduct,outoutletflow rate of the product column, m3/s Qfeed feedflow rate of the product column, m3/s ratea CO2adsorption rate, mol/kg s
ratec catalytic reaction rate, mol/kg s R ideal gas constant, J/mol K SA specific surface area, m2/kg
t time, s
ttotal total operating time, s T temperature, K v gas velocity, m/s vfeed feed velocity, m/s vrinse rinse velocity, m/s vpurge purge velocity, m/s
vpep given velocity for pressure, –equalization pressurization, m/s
vpp given velocity for product pressurization, m/s vol ratio_ a/c packing volume ratio of adsorbent and catalyst,– xi molar fraction of componenti,–
xproduct,i molar fraction of componentiin the product column,– xK ratio of K-related sites to non-related sites,–
α fitting parameter forE1f, J/mol β fitting parameter forE1b, J/mol εb void ratio offixed bed,– ρa adsorbent density, kg/m3 ρc catalyst density, kg/m3 ρg gas density, kg/m3 μ gas viscosity, Pa s
carbon ratios were 0.44 and 1.06, respectively, for achieving 95%
carbon capture ratio and 99% CO2purity[29]. Boon et al.[24]con- sidered surface adsorption and competitive adsorption of CO2and H2O in the adsorption model and concluded that the enhancement of steam rinse was underestimated. However, it should be noted that the SEWGS is commonly adopted for pre-combustion CO2 capture in integrated gasification combined cycle (IGCC) or natural gas combined cycle (NGCC) power plants, which has substantial differences with hydrogen production. The aim of SEWGS is in situ capture of high-purity CO2—which can thus be directly utilized or stored—rather than the production of high-purity hydrogen [30]. Typical HP in an eight- column eleven-step SEWGS process for achieving 95% carbon capture ratio and 99% CO2purity is lower than 99%[20].
Previous work indicated that the HP of ET-PSA, after mixing high- temperature WGS catalysts in the columns, mainly depended on the performance of CO2 adsorbents and the steam gas ratio (S/G ratio) [31]. K-MG30 (MgO:Al2O3mass ratio: 30:70, K2CO3impregnation mass ratio: 25%, Sasol, Germany), which is a kind of potassium-promoted MgAl-CO3layered double oxide, was proven to have reversible trace- CO2 purification ability at 400 °C [22]. Recently, a double-column seven-step ET-PSA adopting K-MG30 as adsorbent was proposed for high-purity hydrogen production from hydrogen-rich gas (1% CO, 1%
CO2, 10% H2O, and 88% H2), achieving 99.6% HRR and 99.9991% HP with only 0.24 total steam-to-feed ratio[19].
In this work, a two-train ET-PSA is proposed for producing high- purity hydrogen directly from typical coal-fired shifted gas (29% CO2, 40% H2, 1% CO, and 30% H2O). An eight-column thirteen-step ET-PSA with shorter step time served as thefirst train for removing most CO/
CO2from the shifted gas, and a double-column seven-step ET-PSA with longer step time served as the second train purifying the residual CO/
CO2to ppm level. The necessity by adopting two-train ET-PSA rather than one-train ET-PSA to achieve both high HP and HRR was proved.
To reduce the total steam consumption, the tail gas in the second-train ET-PSA was adopted as the purge gas for thefirst-train ET-PSA. The operating conditions were optimized considering the system efficiency (HP and HRR), purification energy consumption (rinse-to-feed (R/F) ratio, and purge-to-feed (P/F) ratio), and the performance of the pro- posed two-train ET-PSA was compared with conventional PSA cycles for H2production.
2. Mathematical model
A one-dimensional composite column with mixed packing of K- MG30 adsorbent and Fe2O3/Cr2O3catalysts (HT-B113, Liaoning Haitai, China) was considered for ET-PSA simulation. The mathematical model was based on a series of partial differential and algebraic equations (PDAEs)[32]. Detailed model descriptions have been listed in previous work[19]. With assumptions of surface adsorption controlled kinetics, axially dispersed plug flow, no radical mass/velocity gradients, and isothermal operation condition, the mass balance of the column is given by Eqs.(1) and (2).
∂
∂ = ∂
∂
⎛
⎝
∂
∂
⎞
⎠−∂
∂ − −
ε C
t ε
z D C x
z z(vC) (1 ε C) ̇
i i i
i i
b b ax, T b transfer,
(1)
= + +
C vol ratio +
vol ratio ρ sto rate
vol ratio ρ sto rate
̇ _
1 _
1
1 _
i i i
transfer, a/c
a/c a a, a
a/c c c, c
(2) wherestoais 0 except +1 for CO2,stocis−1 for CO2/H2and +1 for CO2/H2O. The axial dispersion coefficient, Dax,i, in the momentum balance equation was calculated by the Wakao correlation[33]. The assumption of isothermal operation was applied both in this work and previous work[19]. In actual application, the temperaturefluctuation of the column for ET-PSA was small due to a low CO2adsorption heat of K-LDO[34]. The ideal gas law was adopted to calculate the column pressure (Eq.(3)).
=
p C RTT (3)
The pressure drop along the column is described by the Ergun equation (Eq.(4)). This steady-state momentum equation was proven to be valid in PSA simulation[35].
−∂
∂ = −
+ −
p z
μ ε
ε D v ε ρ
ε D v v 150 (1 )
1.75(1 )
b2 | |
b3 p2
b g b3
p (4)
The CO2adsorption/desorption rate,ratea, was calculated based on an elementary kinetic model that wasfirst proposed by Ebner et al.[36]
and then modified by Zheng et al.[25]. The model assumed that there were three types of CO2adsorption sites on the surface of K-LDO. The CO2 in the bulk phase was quickly chemisorbed on the unsaturated oxygen site (O(s)). The adsorbed CO2(A) was then transformed by the potassium-modified adsorption site (E) to carbonate (B) or the un- modified adsorption site (D) to carbonate (C). The mathematical de- scription of the proposed model is shown in Eqs.(5)–(10).
⎜ ⎟
= ⎛
⎝
⎞
⎠
− − + − +
dq
dt k q p
p k q k q q k q k q q k q
c A
1f O(s) CO
0
1b A 2f A E 2b B 3f A D 3b C
2
(5)
= −
dq
dtB k q q2f A E k q2b B
(6)
= −
dq
dtC k q q3f A D k q3b C
(7)
= = + +
+ + +
rate dq
dt 1 (q q q )M
dq dt
dq dt
dq a dt
A,0 B,0 C,0 CO
A B C
2 (8)
where
= + x q +q
q q
K B E
C D (9)
= + + +
qtotal qB qC qD qE (10)
The reaction rates in Eqs. (5)–(7) were fitted by the Arrhenius equation. However, due to the heterogeneous adsorption sites on the surface of K-LDO, the activation energy of the chemisorption step is described by Elovich-type equations (Eqs.(11) and (12))[37].
= +
E1f E1f0 αq q/
A AS (11)
= −
E1b E1b0 βq qA/ AS (12)
The kinetics of the WGS catalysts is described by the power law equation in Eq.(13), where the kinetic parameters are those measured by Hla et al.[38].
= ⎛
⎝
− ⎞
⎠
⎛
⎝⎜ − ⎞
⎠⎟
− −
rate exp
RT p p p p
K p p p p 4.436 88000
1 1
c CO
0.9 H O
0.31 CO
0.156 H
0.05 eq
CO H
CO H O
2 2 2
2 2
2 (13)
Table S1 (Supporting Information, the same below) shows the model parameter values of the adsorption model adopting K-MG30 as the CO2 adsorbent. Note that the thermogravimetric analyzer (TGA) could only evaluate the CO2adsorption/desorption performance below 1 atm partial pressure. The parameters for the low-pressure model with feed CO2partial pressure lower than 1 atm were calibrated with TGA results and verified withfixed bed results[39]; for cases with feed CO2
partial pressure high than 1 atm, a simplified Elovich-type model con- sidering only the chemisorption step was adopted and the parameters werefitted with high-pressure kinetic data tested using a static volu- metric method [40]. The fitted model was successfully adopted to predict ET-PSA performance with a double-column seven-step process [19]and a four-column eight-step process[25].
Table S2 lists the ET-PSA boundary conditions for each step, which were adopted in the previous work[19]and also in other published papers[14]. The mathematical model was built using the gPROMS®
simulation package (Process Systems Enterprise Limited, UK). The
backwardfinite difference method (BFDM) and the forwardfinite dif- ference method (FFDM) withfirst order wereflexibly applied to solve the column equations [15]. Table 1lists the column parameters and operating conditions adopted in the simulation. The column was 2 m in height and 0.5 m in inner diameter, with an adsorbents/catalysts vo- lume ratio of 5. The feed gas was 400 °C, 30 atm, and consisted 1% CO, 30% H2O, 29% CO2, 40% H2, which was the typical composition for syngas after the WGS reactor[23].
The performance of ET-PSA was evaluated based on HP, HRR, and Productivity in Eqs.(14)–(16). CO2and CO were the only gas impurities in the definition ofHP, and theHRRboth considered the H2in the feed gas and the H2produced from the shift reaction. A 100%HRRmeans that all the H2in the feed gas was recovered, and all the CO was con- versed to H2and was then recovered[19]. The R/F ratio and P/F ratio were defined as the ratios of the total moles over one cycle.
∫
=
∫
+ +
x Q dt
x x x Q dt
HP
( )
t t
0 product,H product,out
0 product,H product,CO product,CO product,out total
2 total
2 2 (14)
∫
=
∫
+
x Q dt
x x Q dt
HRR
( )
t t
0 product,H product,out
0 feed,CO feed,H feed
total 2 total
2 (15)
=
∫
x +x +x Q dtm t
Productivity t ( )
0 product,H product,CO product,CO product,out adsorbents,total total
total
2 2
(16)
3. Results and discussion 3.1. First-train ET-PSA
Fig. 1 shows theflow diagram and process sequence of an eight- column thirteen-step ET-PSA as thefirst train to remove most CO/CO2
from the shifted gas. The aim of the first-train ET-PSA was to achieve > 95% HP. The sequence for each column was: during the adsorption step, the CO in the feed gas was shifted to CO2and H2while CO2was adsorbed in situ. The H2-rich gas produced was collected in a product columnfive times larger than the adsorption column before flowing away. Co-current high-pressure steam rinse was adopted to squeeze the remaining H2out of the column. The feedflow in the ET- PSA process was discontinuous when the rinse step was implemented.
In actual application, the feedflow rate was smoothened with a buffer tank[41]. A simple evaluation was shown in Fig. S1 to verify the fea- sibility of the buffer tank to smooth the feedflow rate. The column pressure was then reduced to near 1 atm by four-stage pressure-equal- ization depressurization and blowdown steps. Counter-current low- pressure steam purge replaced the hydrogen purge in NT-PSA to re- generate the adsorbents. Xiu et al. [42] proposed a similar concept called “reactive regeneration” in a sorption enhanced process using steam purge or hydrogen purge. The purge time of the proposed process sequence inFig. 1was longer than that of the process sequence in lit- eratures [24], which was beneficial to produce a product gas with higher than HP. Finally, the column was re-pressurized with four-stage step pressure-equalization pressurization with gases from the pressure- equalization depressurization steps and product pressurization steps.
Note that the PSA process was complex and integrated, only parameters that were directly connected the purification energy consumption (R/F ratio and P/F ratio) with the system efficiency (HP and HRR) were studied. In the proposed ET-PSA process, the operating time for each step wasfixed. The R/F ratio and P/F ratio could be changed by varying the feed, rinse, or purge rate.
Fig. 2shows the pressure profile of the column in a typical cycle.
The pressure change during pressure-equalization depressurization and blow down steps matched valve equations listed in Table S2. The maximum pressure drop rate for pressure-equalization depressurization
was limited to 0.1 atm/s to avoid miscalculations caused by fast column pressure change. It was also meaningful in actual PSA operation be- cause fastflow rate during pressure-equalization might wear the ad- sorbent/catalyst particles. Even after product pressurization, the column pressure did not reach the exact adsorption pressure due to a small pressure drop of the product column[19]. The column was then pressurized by feed gas to 30 atm at the beginning of the adsorption step.
The performance of the base case (case 1) with 400 h−1feed rate, 200 h−1rinse rate, and 30 h−1purge rate is listed inTable 2, and its stable-state gas distributions along the column at the end of each step are shown inFig. 3. Theflow rate in this work was defined as the gas hourly space velocity (GHSV, h−1). The results inFig. 3(a) and (b) show that the outlet CO2and CO did not penetrate the column during ad- sorption and rinse steps, thus leading to a high HP of 94.542%. In ad- dition, the base case also reached a high HRR of 97.92%, as most of the remaining H2after the adsorption step was squeezed out of the column (Fig. 3(c)), either to the product column in the rinse step or to other adsorption columns during pressure-equalization depressurizations steps. The R/F ratio (0.167) was significantly reduced by increasing the pressure-equalization number to four. The improvement in HRR by adopting the rinse step was more obvious for cases with higher CO/CO2
concentrations in the feed gas (shorter adsorption time), as the ratio of the remaining H2in the column to the produced H2during adsorption step was higher.
The performance of the product column was investigated in Fig. S2.
One challenge of ET-PSA is to ensure the continuity of operation. Fig.
S2(a) shows that the inlet gas to the product column was not continuous during a cycle, which was due to the pressurizing in the adsorption step, the gas consumption in product pressurization step, and the different flow rate between the feed gas and the rinse gas. However, the accu- mulatedflow of the inlet gas was in direct proportion to the cycle time (Fig. S2(b)), and the pressure change of the product column could be controlled within ± 0.5 atm if an average value was adopted as the outletflow rate (Fig. S2(c)).
The introduction of steam rinse step into ET-PSA was the key to achieving high HRR, while a large amount of H2that remained in the column in NT-PSA after adsorption was wasted in the subsequent blowdown and purge step. For comparison, the rinse step was canceled in case 2 by lengthening the adsorption time to 120 s. The feed rate was reduced to 300 h−1to maintain the same feed amount as the base case.
As expected, HRR dropped to 82.11%, which was a typical value for NT-PSA[16]. When the feed rate was increased to 400 h−1without changing the purgeflow rate (case 3) the HRR only increased by about 5%, but the HP decreased rapidly from 99.647% to 87.484% due to the breakthrough of CO/CO2during the adsorption step.
Fixing the feed rate at 400 h−1,Fig. 4(a) shows CO2profiles at the
Table 1
Column parameters and operating conditions.
Column parameters
Lb(m) 2
Db(m) 0.5
εb(–) 0.717
ρa(kg/m3) 1943.5 ρc(kg/m3) 5335.7
Dp(m) 0.005
vol ratio_ a c/ (–) 5
T(K) 673.15
μ(Pa s) CO2: 3.05 × 10−5, H2: 1.57 × 10−5, CO: 3.11 × 10−5, H2O:
2.45 × 10−5 Operating conditions
pfeed(atm) 30
xfeed(–) 1% CO, 30% H2O, 29% CO2, 40% H2
prinse(atm) 30
ppurge(atm) 1.2
end of the adsorption and rinse steps with different R/F ratios (cases 1, 4, 5). The increase in rinse rate slightly reduced the HP by pushing the CO2profile fronts forward, making it easier for them to penetrate the column. The results inFig. 3(a) indicate that in the base case, although CO2did not penetrate the column during the adsorption and rinse steps, the outlet CO2concentration increased from 15 mol/m3to 39 mol/m3 during the third and fourth pressure-equalization depressurization steps. The penetrated CO and CO2were then transferred to another column during pressure-equalization pressurization steps, reducing the HP in the next cycle by contaminating the adsorbents at the top of the column[27]. When the R/F ratio was increased to 0.208 (case 4), CO2
still did not penetrate the column after the rinse step and HRR increased to 99.08% as more H2was pushed to the product column in the rinse step. Further increasing the R/F ratio to 0.250 (case 5) did not improve HRR much, but rapidly decreased HP due to the breakthrough of CO2
during the rinse step.
When the P/F ratio was increased to 0.3, the HP of case 6 exceeded the required value of 95%.Fig. 4(b) shows that the significant increase in HP from 94.952% to 97.715% was due to: (1) CO2profiles moving backward and (2) decrease in the balanced-CO2 concentration. The improvement effect on HP was reduced when the P/F ratio was further Fig. 1.Flow diagram of and process sequence of thefirst-train ET-PSA (tED=tBD=tEP=tPP=60 s;tAD=90 s;tR=30 s;tP=240 s,tcycle=960 s).
Fig. 2.Pressure profile of column in the eight-column thirteen-step ET-PSA (feed rate: 400 h−1; rinse rate: 200 h−1; purge rate: 30 h−1).
Table 2
Performance offirst-train ET-PSA.
Case Feed rate (h−1)
Rinse rate (h−1)
Purge rate (h−1)
HP (%) HRR (%) Productivity (mol/kg/day)
1a 400 200 30 94.542 97.92 37.23
2b 300 0 30 99.647 82.11 29.62
3b 400 0 30 87.484 87.99 48.20
4 400 250 30 93.005 99.08 38.29
5 400 300 30 91.617 99.61 39.08
6 400 200 45 97.715 97.80 35.97
7 400 250 45 96.413 99.02 36.91
8 400 300 45 95.065 99.59 37.65
9 400 200 60 98.984 97.69 35.47
10 400 250 60 98.198 98.96 36.22
11 400 300 60 97.182 99.57 36.82
12 300 200 30 99.893 95.64 25.81
13 350 200 30 99.233 97.13 30.78
14 450 200 30 88.930 98.31 44.70
a Base case.
b Rinse step was replaced by lengthening the adsorption step to 120 s.
increased to 0.4 (case 9). The desorption kinetics of K-MG30 was slower than the adsorption kinetics, and a large amount of purge steam was required to totally regenerate the adsorbent [22]. The results in Fig. 3(a) also indicate that although the purge gas cleaned the top of the column, a large fraction of CO2remained at the bottom of the column after the steam purge. The increased P/F ratio reduced the CO2partial pressure in the column, thus causing thorough regeneration of the ad- sorbent.
Cases 12 and 13 discuss conditions with reduced feed rates. When reducing the feed rate to lower than 350 h−1by keeping the same rinse and purge rates as that of the base case, HP was increased to above 99%
as less CO/CO2was able to penetrate the column during the adsorption, rinse, and pressure-equalization depressurization steps. However, re- duced feed rate caused problems such as low HRR, higher R/F ratio and P/F ratio, and lower productivity.Fig. 5summarizes the performance of first-train ET-PSA at different operating cases. The total steam con- sumption ratio, which was defined as the sum of R/F ratio and P/F ratio, was shown in brackets. Fixing HP higher than 95%, the total steam consumption ratio lower than 0.5, case 6 shows the highest HRR of 97.80% among all the cases.
Fig. S3 shows the tail gas of case 6 during the blowdown and purge steps. The tail gas mostly consisted of CO2and H2O with a small amount of H2remaining in the columns, carried out during the blowdown step and the beginning of the purge step. The average composition of tail gas forfirst-train ET-PSA was 39.51% CO2, 3.30% H2, 0.0554% CO, and 57.13% H2O for the blowdown and 41.90% CO2, 0.21% H2, 0.00534%
CO, and 57.88% H2O for the purge, and the total composition was 41.12% CO2, 1.22% H2, 0.0217% CO, and 57.64% H2O.
3.2. Second-train ET-PSA
The total inlet CO/CO2 concentration for first-train ET-PSA was 30%. Therefore, a short step time was required to avoid the break- through of CO/CO2during the adsorption and rinse steps, which caused incomplete desorption of K-MG30. The simulation results offirst-train ET-PSA also proved that it was not energy efficient to achieve high HP in single train ET-PSA. Even if adopting cases with large P/F ratios (case 9, 10, 11), the HP of the first-train ET-PSA didn’t exceed 99%.
Therefore, a second-train ET-PSA was proposed to produce hydrogen with HP higher than 99.999%. By adopting case 6 as the operating condition forfirst-train ET-PSA, the product gas consisted of 1.137%
CO2, 52.44% H2, 0.0893% CO, and 46.33% H2O, withflow rate 90.07 Nm3/h (average value). The low feed CO/CO2 concentration made second-train ET-PSA possible and economical to choose a sequence with longer step time. Second-train ET-PSA could also adopt less pressure- equalization number as the ratio of wasted H2during the blowdown step to the H2produced during the adsorption/rinse step was lower.
Herein, a double-column seven-step process with only one pressure- equalization step was selected for the second-train ET-PSA (Fig. 6).
Unlike the sequence offirst-train ET-PSA, the feed and rinse times were changeable by keeping the same feed rate as that of product gas for first-train ET-PSA. The P/F ratio was varied by changing the purgeflow Fig. 3.Stable-state gas distributions along the column of the base case (feed rate: 400 h−1; rinse rate: 200 h−1; purge rate: 30 h−1).
rate.
Fig. S4 shows the pressure profile of the base case with 2400 s ad- sorption time, 60 s rinse time, and 0.05P/F ratio. Table 3 lists the performance of the base case (case 15). Due to reduced pressure- equalization number, there was a higher pressure drop during the blowdown step and larger gas requirement for product pressurization.
However, the seven-step process achieved a high purity of hydrogen with extremely simple configuration[19]. The R/F ratio (0.025) was
much lower than that offirst-train ET-PSA, and the purge time was lengthened in second-train ET-PSA leading to high HP of the base case (99.967%).
Our previous work indicated that HRR of double-column ET-PSA mainly depended on the rinse step[19]; On the other hand, the results in Table 2indicate that there was a trade-offbetween HRR and HP when changing the R/F ratio, as the rinse steam caused desorption of CO2adsorbents due to reduced CO2partial pressure. Therefore, the R/F ratio for second-train ET-PSA was optimized to achieve over 99% HRR and the highest HP.Fig. 7(a) shows HRR and HP with varying rinse time from 0 to 60 s. Owing to longer adsorption time (2400 s), the HRR of second-train ET-PSA reached a high value of 97.12%, even without the rinse step. The value was higher than that of NT-PSA for high-purity H2production from raw H2fuel gas due to the replacement of H2purge with steam purge[17]. When the rinse time was increased to 50 s, HRR increased to 99.24% with HP 99.973%.Fig. 7(b) shows the profiles of H2O, H2, and CO2at the end of the rinse step. By adopting 50 s rinse time, about half of the H2remaining in the column was pushed out, and the other half was further squeezed out to another column during the next pressure-equalization step. The desorbed CO2 was re-adsorbed during the rinse step at the top of the column[24]. However, the rinse steam penetrated the column if the rinse time was more than 60 s, which reduced HP by carrying CO/CO2out to the product gas. There- fore, the rinse time of 50 s was an optimal value for second-train ET- PSA.
To achieve extremely high HP, gas impurities were not allowed to penetrate the column even during the pressure-equalization step as the penetrated CO/CO2would contaminate the adsorbents at the tops of the columns[19]. The data inFig. 7(b) indicate that with P/F ratio 0.05, the CO2front profile at the end of the rinse step had already reached the top of the column. Therefore, a larger P/F ratio was needed to push back the CO2profile.Fig. 8shows the effect of P/F ratio on HP with adsorption time. Fixing the adsorption time at 2400 s, the HP reached 99.9993% when the P/F ratio was increased from 0.05 to 0.10. Further increase in P/F ratio had little effect on improving HP but could in- crease the adsorption time to 3600 s (P/F ratio: 0.15, HP: 99.9996%) and 4800 s (P/F ratio: 0.20, HP: 99.9992%). The optimized cases (16, 17, and 18) with different adsorption times for achieving over 99.999%
HP are listed inTable 3.
Fig. S5 investigates the tail gas during the blowdown and purge steps for case 16, which shows the possibility for use as the purge gas forfirst-train ET-PSA. According to the results, the tail gas consisted of CO2/H2/H2O = 5.64%/14.33%/80.00% for the blowdown gas and 9.88%/0.64%/89.48% for the purge gas. The P/F ratio in case 19 was increased to 0.2, which further decreased the CO2concentration in the tail gas. As expected, the total concentration of CO2/H2/H2O changed from the original 9.19%/2.87%/87.94% to 5.24%/1.57%/93.20%.
3.3. Two-train ET-PSA
The two proposed ET-PSAs were combined to produce high-purity hydrogen directly from the shifted gas. One combination method was to simply chain the two systems together, where the product gas of the first-train ET-PSA served as the feed gas for the second-train ET-PSA and all the rinse and purged gases came from the steam. Cases 20, 21, and 22 inTable 4list the performances of double train ET-PSA, where case 6 was adopted as the operating condition forfirst-train ET-PSA and cases 16, 17, and 18 were adopted as operating conditions for second- train ET-PSA.
Another method reduced the total purge steam amount by reusing the tail gas of the second-train ET-PSA. As shown inFig. 9, the tail gas from the blowdown and purge steps for second-train ET-PSA was col- lected in the second tail gas column and used as purge gas forfirst-train ET-PSA. The extra steam was mixed with the tail gas to meet the re- quired P/F ratio. As shown in Fig. S6, the adoption of the blowdown gas as the tail gas might cause a temporary pressure rise of the tail gas Fig. 4.Effect of (a) R/F ratio and (b) P/F ratio on CO2profiles at the end of
adsorption and rinse steps.
Fig. 5.Performance offirst-train ET-PSA with different operating conditions.
column. Therefore, the pressure of the purge gas for the second-train ET-PSA should be increased to guarantee the purgeflow.
One problem with this configuration was that CO2in the tail gas of second-train ET-PSA reduced the HP of first-train ET-PSA, thus in- creasing the workload of second-train ET-PSA. As shown in Table S3, the tail gas of second-train ET-PSA with 2400 s adsorption time and 0.2 R/F ratio contained 9.44% CO2, which reduced to 5.86% after mixing with steam. The HP of first-train ET-PSA decreased from the original 97.715% to 95.638%. As a result, a small amount of CO2/CO penetrated the second-train ET-PSA. Thefinal HP of two-train ET-PSA (99.780%) thus could not reach the required value of 99.999%. To avoid the breakthrough of CO/CO2, the adsorption time of second-train ET-PSA was further reduced to 1800 s (Table S4). HP increased to 99.9994%
during this time with only 5.3 ppm CO2and 0.4 ppm CO (dry basis).
The R/F ratio (0.188) was slightly higher than those with series struc- tures due to reduced adsorption time in second-train ET-PSA. However, the P/F ratio (0.263) was reduced due to partial replacement of purge gas with tail gas from second-train ET-PSA, which was important for reducing the operating energy consumption of two-train ET-PSA.
Fig. S5 indicates that the CO2concentration of the tail gas forfirst- train ET-PSA decreased remarkably with purge time. Therefore, a third method considered the pure tail gas for second-train ET-PSA as the purge gas forfirst-train ET-PSA followed by steam purge to reduce the residual CO2in the column. As shown in Table S5, the successive purge configuration increased the purification efficiency of thefirst-train ET- PSA, thus lengthening the adsorption time for second-train ET-PSA and
Fig. 6.Flow diagram of and process sequence of second-train ET-PSA (tED1=tBD=tEP1=tPP=90 s).
Table 3
Performance of second-train ET-PSA.
Case Adsorption time (s)
R/F ratio (–)
P/F ratio (–)
HP (%) HRR (%) Productivity (mol/kg/day)
15a 2400 0.025 0.05 99.9671 99.57 140.26
16 2400 0.021 0.10 99.9993 99.28 139.80
17 3600 0.014 0.15 99.9996 99.48 140.08
18 4800 0.010 0.20 99.9992 99.59 140.25
19 2400 0.021 0.20 99.9995 99.29 139.82
a Base case.
Fig. 7.Effect of rinse time on (a) HP and HRR; (b) gas distribution along the column at the end of rinse step (adsorption time: 2400 s; rinse time: 0–60 s; P/F ratio: 0.05).
reducing the total rinse steam amount.
Fig. 10and Table S6 compares the performances (HP and HRR) of two-train ET-PSA in this work with those of the optimal cases in pub- lished papers of NT-PSA for pre-combustion CO2capture[11,43], NT- PSA for H2production[10,13–17,44–48], ET-PSA for pre-combustion CO2 capture [25,49], and SEWGS [20,24,26–28,50,51]. Most of the proposed NT-PSA systems for H2production achieved above 99.99%
HP and 50–90% HRR. The only case that achieved HRR above 90%
adopted twelve-column thirteen-step PSA with low inlet CO/CO2con- centration (2.12% CO2 and 2.66% CO)[17]. The PSAs for pre-com- bustion CO2capture typically achieved about 95% HP. However, a high HP was not a requirement for pre-combustion CO2capture cases, rather a high carbon capture ratio (95%) and high CO2purity (99%). The ET- PSA proposed by Liu et al.[49]had the highest HRR of 96.93% because Fig. 8.Effect of P/F ratio on HP with different adsorption times (adsorption
time: 2400, 3600, 4800 s; rinse time: 50 s; P/F ratio: 0.05–0.20).
Table 4
Performance of two-train ET-PSA.
Case Coupling method
R/F ratio (–)
P/F ratio (–)
HP (%) HRR (%) Productivity (mol/kg/day)
20 Series 0.183 0.376 99.9993 97.26 27.96
21 Series 0.177 0.415 99.9996 97.45 28.02
22 Series 0.174 0.453 99.9992 97.57 28.05
23 Reflux1 0.183 0.269 99.7801 97.60 28.13
24 Reflux1 0.188 0.263 99.9994 97.51 28.09
25 Reflux2 0.182 0.272 99.9995 97.48 28.08
Fig. 9.Flow diagram of two-train ET-PSA with tail gas from second-train ET-PSA as purge gas forfirst-train ET-PSA.
Fig. 10.Comparison of the performances of PSAs for H2production or pre- combustion CO2capture from carbon-based fuels.
it utilized steam purge step. The SEWGS reached about 95% HP and above 97% HRR due to the introduction of steam rinse step. The per- formances of one-train ET-PSA with (case 6) and without (case 3) steam rinse in this work were similar to those of PSAs for CO2capture and SEWGS, respectively. The two-train ET-PSA, however, achieved HP si- milar to that of NT-PSA for H2production, with higher HRR. Case 24 reached 99.9994% HP and 97.51% HRR, which were the highest values among all PSAs investigated. In addition, the steam consumption is compared inTable 5, which indicates that the R/F ratio (0.188) and P/F ratio (0.263) of two-train ET-PSA were similar to that of SEWGS.
4. Conclusions
A two-train ET-PSA is proposed to produce high-purity hydrogen directly from shifted gas, where thefirst train with eight-column thir- teen-step configuration removes most of the CO/CO2from the shifted gas and the second train with double-column seven-step configuration purifies the residual trace impurities. In the simulation, thefirst-train ET-PSA achieved 97.715% HP and 97.80 HRR, with 0.167 R/F ratio and 0.300P/F ratio. Increase in the R/F ratio improved HRR but decreased HP by pushing the CO/CO2 concentration profile forward. The re- placement of H2purge with steam purge increased the HP, but the purge time in thefirst-train ET-PSA was not enough to regenerate the adsorbents completely. The second-train ET-PSA with longer step time purified the H2-rich gas from 95% HP to higher than 99.999%. Two- train ET-PSA—by adopting the tail gas from second-train ET-PSA as the purge gas forfirst-train ET-PSA—reached 99.9994% HP and 97.51%
HRR, which are the highest values among proposed PSAs for pre- combustion CO2capture and H2production. In addition, the total steam ratio of two-train ET-PSA was comparable to that of the SEWGS process, which further proves its technical feasibility.
Acknowledgement
This research is financed by Shanxi Province Science and Technology Major Projects of (MH2015-06), National Key R&D Program of China (2017YFB0601900), and China Postdoctoral Science Foundation (2017M610890).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in the online version, athttps://doi.org/10.1016/j.apenergy.2018.08.093.
References
[1] Veras TD, Mozer TS, dos Santos D, Cesar AD. Hydrogen: trends, production and characterization of the main process worldwide. Int J Hydrogen Energy 2017;42(4):2018–33.
[2] McDowall W, Eames M. Forecasts, scenarios, visions, backcasts and roadmaps to the hydrogen economy: a review of the hydrogen futures literature. Energy Policy 2006;34(11):1236–50.
[3] Andersson J, Lundgren J. Techno-economic analysis of ammonia production via integrated biomass gasification. Appl Energy 2014;130:484–90.
[4] Steinberg M, Cheng HC. Modern and propective technologies for hydrogen-pro- duction from fossil-fuels. Int J Hydrogen Energ 1989;14(11):797–820.
[5] Iribarren D, Susmozas A, Petrakopoulou F, Dufour J. Environmental and exergetic evaluation of hydrogen production via lignocellulosic biomass gasification. J Clean Prod 2014;69:165–75.
[6] Ahmad H, Kamarudin SK, Minggu LJ, Kassim M. Hydrogen from photo-catalytic water splitting process: a review. Renew Sustain Energy Rev 2015;43:599–610.
[7] Anantharaj S, Ede SR, Sakthikumar K, Karthick K, Mishra S, Kundu S. Recent trends and perspectives in electrochemical water splitting with an emphasis on sulfide, selenide, and phosphide catalysts of Fe Co, and Ni: a review. Acs Catal 2016;6(12):8069–97.
[8] Muhich CL, Ehrhart BD, Al-Shankiti I, Ward BJ, Musgrave CB, Weimer AW. A re- view and perspective of efficient hydrogen generation via solar thermal water splitting. Wires Energy Environ 2016;5(3):261–87.
[9] Nikolaidis P, Poullikkas A. A comparative overview of hydrogen production pro- cesses. Renew Sust Energ Rev 2017;67:597–611.
[10] Sircar S, Golden TC. Purification of hydrogen by pressure swing adsorption. Sep Sci Technol 2000;35(5):667–87.
[11] Riboldi L, Bolland O. Evaluating Pressure Swing Adsorption as a CO2separation technique in coal-fired power plants. Int J Greenh Gas Con 2015;39:1–16.
[12] Riboldi L, Bolland O. Overview on pressure swing adsorption (PSA) as CO2capture technology: state-of-the-art, limits and potentials. Energ Procedia
2017;114:2390–400.
[13] Ribeiro AM, Grande CA, Lopes FVS, Loureiro JM, Rodrigues AE. Four beds pressure swing adsorption for hydrogen purification: case of humid feed and activated carbon beds. Aiche J 2009;55(9):2292–302.
[14] Ribeiro AM, Grande CA, Lopes FVS, Loureiro JM, Rodrigues AE. A parametric study of layered bed PSA for hydrogen purification. Chem Eng Sci 2008;63(21):5258–73.
[15] Moon DK, Lee DG, Lee CH. H2pressure swing adsorption for high pressure syngas from an integrated gasification combined cycle with a carbon capture process. Appl Energy 2016;183:760–74.
[16] Lopes FVS, Grande CA, Rodrigues AE. Activated carbon for hydrogen purification by pressure swing adsorption: multicomponent breakthrough curves and PSA per- formance. Chem Eng Sci 2011;66(3):303–17.
[17] Luberti M, Friedrich D, Brandani S, Ahn H. Design of a H2PSA for cogeneration of ultrapure hydrogen and power at an advanced integrated gasification combined cycle with pre-combustion capture. Adsorption 2014;20(2–3):511–24.
[18] Nikolic D, Giovanoglou A, Georgiadis MC, Kikkinides ES. Generic modeling fra- mework for gas separations using multibed pressure swing adsorption processes.
Ind Eng Chem Res 2008;47(9):3156–69.
[19] Zhu X, Shi Y, Cai N. Elevated temperature pressure swing adsorption process for reactive separation of CO/CO2in H2-rich gas. Int J Hydrogen Energy 2018;43(29):13305–17.
[20] Najmi B, Bolland O, Colombo KE. A systematic approach to the modeling and si- mulation of a Sorption Enhanced Water Gas Shift (SEWGS) process for CO2capture.
Sep Purif Technol 2016;157:80–92.
[21] Lee CH, Lee KB. Sorption-enhanced water gas shift reaction for high-purity hy- drogen production: application of a Na-Mg double salt-based sorbent and the di- vided section packing concept. Appl Energy 2017;205:316–22.
[22] Zhu X, Shi Y, Cai N. CO2residual concentration of potassium-promoted hydrotalcite for deep CO/CO2purification in H2-rich gas. J Energy Chem 2017;26(5):956–64.
[23] Zhu X, Shi Y, Cai N. Integrated gasification combined cycle with carbon dioxide capture by elevated temperature pressure swing adsorption. Appl Energy 2016;176:196–208.
[24] Boon J, Cobden PD, van Dijk HAJ, Annaland MVS. High-temperature pressure swing adsorption cycle design for sorption-enhanced water-gas shift. Chem Eng Sci 2015;122:219–31.
[25] Zheng Y, Shi Y, Li S, Yang Y, Cai N. Elevated temperature hydrogen/carbon dioxide separation process simulation by integrating elementary reaction model of hydro- talcite adsorbent. Int J Hydrogen Energ 2014;39(8):3771–9.
[26] Reijers R, van Selow E, Cobden P, Boon J, van den Brink R. SEWGS process cycle optimization. Energy Proc 2011;4:1155–61.
[27] Allam RJ, Chiang R, Hufton JR, Middleton P, Weist EL, White V. Chapter 13—de- velopment of the sorption enhanced water gas shift process. In: Thomas DC, editor.
Carbon dioxide capture for storage in deep geologic formations. Amsterdam:
Elsevier Science; 2005. p. 227–56.
[28] Wright A, White V, Hufton J, van Selow E, Hinderink P. Reduction in the cost of pre- combustion CO2capture through advancements in sorption-enhanced water-gas- shift. Energy Proc 2009;1(1):707–14.
Table 5
Comparison of the steam consumptions of ET-PSA and SEWGS.
HP (%) HRR (%) Process R/F ratio P/F ratio Reference
96.433 98.49 7-column 10-step Unknown Unknown Allam et al.[29]
94.364 98.00 8-column 11-step 0.240 Unknown Van Selow et al.[31]
95.897 98.94 8-column 11-step 0.127 0.190 Wright et al.[30]
96.342 96.34 8-column 11-step 0.120 0.375 Wright et al. [55]
93.153 98.67 6-column 8-step 0.157 0.371 Reijers et al.[28]
76.801 94.52 9-column 11-step 0.018 0.049 Boon et al.[27]
96.410 99.29 9-column 11-step 0.158 0.475 Najmi et al.[23]
97.715 97.80 8-column 13-step 0.167 0.300 This work
99.999 97.51 Two-train ET-PSA 0.188 0.263 This work
[29] Gazzani M, Macchi E, Manzolini G. CO2capture in integrated gasification combined cycle with SEWGS–Part A: thermodynamic performances. Fuel 2013;105:206–19.
[30] Theo WL, Lim JS, Hashim H, Mustaffa AA, Ho WS. Review of pre-combustion capture and ionic liquid in carbon capture and storage. Appl Energy 2016;183:1633–63.
[31] Zhu X, Shi Y, Cai N. Characterization on trace carbon monoxide leakage in high purity hydrogen in sorption enhanced water gas shifting process. Int J Hydrogen Energy 2016;41(40):18050–61.
[32] Riboldi L, Bolland O. Pressure swing adsorption for coproduction of power and ultrapure H2in an IGCC plant with CO2capture. Int J Hydrogen Energy 2016;41(25):10646–60.
[33] Wakao N, Funazkri T. Effect offluid dispersion coefficients on particle-to-fluid mass-transfer coefficients in packed beds: correlation of Sherwood numbers. Chem Eng Sci 1978;33(10):1375–84.
[34] van Selow ER, Cobden PD, Verbraeken PA, Hufton JR, van den Brink RW. Carbon capture by sorption-enhanced water-gas shift reaction process using hydrotalcite- based material. Ind Eng Chem Res 2009;48(9):4184–93.
[35] Sereno C, Rodrigues A. Can steady-state momentum equations be used in modelling pressurization of adsorption beds? Gas Sep Purif 1993;7(3):167–74.
[36] Ebner AD, Reynolds SP, Ritter JA. Nonequilibrium kinetic model that describes the reversible adsorption and desorption behavior of CO2in a K-promoted hydrotalcite- like compound. Ind Eng Chem Res 2007;46(6):1737–44.
[37] Leon M, Diaz E, Bennici S, Vega A, Ordonez S, Auroux A. Adsorption of CO2on hydrotalcite-derived mixed oxides: sorption mechanisms and consequences for adsorption irreversibility. Ind Eng Chem Res 2010;49(8):3663–71.
[38] Hla SS, Park D, Duffy GJ, Edwards JH, Roberts DG, Ilyushechkin A, et al. Kinetics of high-temperature water-gas shift reaction over two iron-based commercial catalysts using simulated coal-derived syngases. Chem Eng J 2009;146(1):148–54.
[39] Zhu X, Shi Y, Cai N. Investigation on the trace amount of released CO in sorption enhanced water gas shift reaction applied in pre-combustion CO2capture and high
purity H2production. Energy Proc 2017;114:2525–36.
[40] Zhu X, Shi Y, Cai N. High-pressure carbon dioxide adsorption kinetics of potassium- modified hydrotalcite at elevated temperature. Fuel 2017;207:579–90.
[41] Zhou L, Lu CZ, Bian SJ, Zhou YP. Pure hydrogen from the dry gas of refineries via a novel pressure swing adsorption process. Ind Eng Chem Res 2002;41(21):5290–7.
[42] Xiu GH, Li P, Rodrigues AE. Sorption-enhanced reaction process with reactive re- generation. Chem Eng Sci 2002;57(18):3893–908.
[43] Casas N, Schell J, Joss L, Mazzotti M. A parametric study of a PSA process for pre- combustion CO2capture. Sep Purif Technol 2013;104:183–92.
[44] Yang SI, Choi DY, Jang SC, Kim SH, Choi DK. Hydrogen separation by multi-bed pressure swing adsorption of synthesis gas. Adsorption 2008;14(4–5):583–90.
[45] Ahn S, You YW, Lee DG, Kim KH, Oh M, Lee CH. Layered two- and four-bed PSA processes for H2recovery from coal gas. Chem Eng Sci 2012;68(1):413–23.
[46] You YW, Lee DG, Yoon KY, Moon DK, Kim SM, Lee CH. H2PSA purifier for CO removal from hydrogen mixtures. Int J Hydrogen Energ 2012;37(23):18175–86.
[47] Rahimpour MR, Ghaemi M, Jokar SM, Dehghani O, Jafari M, Amiri S, et al. The enhancement of hydrogen recovery in PSA unit of domestic petrochemical plant.
Chem Eng J 2013;226:444–59.
[48] Moon DK, Park Y, Oh HT, Kim SH, Oh M, Lee CH. Performance analysis of an eight- layered bed PSA process for H2recovery from IGCC with pre-combustion carbon capture. Energy Convers Manage 2018;156:202–14.
[49] Liu Z, Green WH. Analysis of adsorbent-based warm CO2capture technology for integrated gasification combined cycle (IGCC) power plants. Ind Eng Chem Res 2014;53(27):11145–58.
[50] van Selow ER, Cobden PD, van den Brink RW, Wright A, White V, Hinderink P, et al.
Carbon dioxide capture for storage in deep geologic formations. Pilot-scale devel- opment of the sorption enhaned water gas shift process. Berks, CPL Press. Chapter;
2009. p. 157–80.
[51] Wright AD, White V, Hufton JR, Quinn R, Cobden PD, van Selow ER. CAESAR:
development of a SEWGS model for IGCC. Energy Proc 2011;4:1147–54.
