See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/286627683

Adsorption separation for high purity propane from liquefied petroleum gas
in a fixed bed by removal of alkanes

Article in Separation and Purification Technology · December 2015

DOI: 10.1016/j.seppur.2015.12.003

CITATIONS READS

22 2,364

4 authors, including:

Huan Xiang
Southwest Petroleum University
25 PUBLICATIONS 603 CITATIONS

SEE PROFILE

All content following this page was uploaded by Huan Xiang on 08 February 2019.

The user has requested enhancement of the downloaded file.

 Separation and Purification Technology 158 (2016) 1–8

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur

Adsorption separation for high purity propane from liquefied petroleum

gas in a fixed bed by removal of alkanes
Pengfei Liu, Huiping Zhang, Huan Xiang, Ying Yan ⇑
School of Chemistry and Chemical Engineering, South China University of Technology, Guangzhou 510640, PR China

a r t i c l e i n f o a b s t r a c t

Article history: The process for high purity propane production via adsorption separation of propane from liquefied pet-
Received 10 July 2015 roleum gas (LPG) was investigated at ambient temperature and atmospheric pressure under dynamic
Received in revised form 28 September 2015 conditions in a fixed bed. Breakthrough curves for LPG adsorption on 4A, 5A, 13X and NaY zeolite pellets
Accepted 7 December 2015
and granular activated carbon were measured. The length of unused bed (LUB) values for isobutane and
Available online 7 December 2015
butane adsorption in fixed beds with different adsorbent samples were determined by analyzing break-
through curves. Effects of bed height (5–10 cm) and flow rate (60–100 mL/min) on adsorption character-
Keywords:
istics of propane, isobutane and butane onto activated carbon were also examined. Among all these five
Propane
LPG
adsorbent samples, activated carbon exhibited the highest adsorption capacity and the best separation
Breakthrough curve performance, and the concentration of propane in the outlet gas could reach as high as 99.7% (volume
Fixed bed fraction). Besides, the LUB value of activated carbon fixed bed was the smallest, which indicated that acti-
Adsorption separation vated carbon fixed bed was more suitable for adsorption separation of propane from LPG. Adsorption
dynamics also showed that relatively lower flow rate and relatively longer bed length were favorable
for removal of isobutane and butane from LPG. The Yoon–Nelson model provided a good fit to the exper-
imental data and there was a nice agreement between the numerical simulations and the corresponding
data obtained from experiments. All the correlation coefficients in fitted equations exceeded 0.98, indi-
cating a good confidence level for the fit.
Ó 2015 Elsevier B.V. All rights reserved.

1. Introduction
Global warming and depletion of the ozone layer are the major
environmental issues today [1,2]. The use of R12 (dichlorodifluo-
romethane) and R134a (1,1,1,2-tetrafluoroethane) as refrigerants
in refrigeration industry contribute a lot to greenhouse effect
and/or ozone depletion in spite of their excellent cooling character-
istics. From the environmental point of view, it is essential to find
substitutes for these traditional refrigerants [3]. Hydrocarbons and
in particular, R290 (propane of high purify, normally higher than
99.5% (volume fraction)), are known as excellent refrigerants
which have negligible ozone depletion potential and pretty low
global warming potential [4]. Also, R290 has several positive char-
acteristics such as non-toxicity, good miscibility with mineral oils
and acceptable compatibility with common materials employed
in refrigeration equipment [5].
Liquefied petroleum gas (LPG) are mixtures of hydrocarbon
gases (normally C1–C4) usually used as fuel in heating appliances,
cooking equipment and vehicles. The composition of LPG may dif-
⇑ Corresponding author.
E-mail address: yingyan@scut.edu.cn (Y. Yan).
fer from country to country and from season to season [6], but gen-
erally its main constituent is propane. LPG is inexpensive and
commercially available in local markets, so the process for refining
green refrigerant R290 from LPG by purification is a front subject
with both theoretical significance and engineering value [7].
Various treatment processes such as cryogenic distillation [8],
absorption [9], membrane separation [10] and adsorption [11]
have been proposed for the purification of propane from hydrocar-
bon mixtures. The adsorption process has become a major unit
operation in chemical and petrochemical industries. Furthermore,
fixed bed adsorption for separation and purification has been
widely used in industrial applications due to its advantages of large
capacity, easy operation and low energy consumption [12].
Breakthrough characteristics of the fixed bed play a dominant
role in the evaluation of the efficacy of an adsorbent under flowing
gas conditions [13]. Therefore, it is of great importance to study
breakthrough curves for light hydrocarbons on different adsor-
bents. Kawakami et al. [14,15] provided a method for purification
of propane from LPG at 0.50–0.60 MPa and with molecular sieve
4A adsorbing ethane and propene while with activated carbon
adsorbing isobutane and butane. Approximately 99.9% (volume
fraction) of the propane was obtained from the LPG which had

http://dx.doi.org/10.1016/j.seppur.2015.12.003
1383-5866/Ó 2015 Elsevier B.V. All rights reserved.
2 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8

an initial propane concentration of 98.4% (volume fraction). Cho adsorbents were calculated via DFT (Density Functional Theory)
et al. [16] discussed adsorption separation, purification apparatus and BJH method (determined from the adsorption branch of the
and the process for obtaining high purity isobutane by removing N2 isotherms).
paraffins (methane, ethane, propane and butane) and olefins (iso-
butene, 1-butene and 2-butene) from light hydrocarbon mixtures
2.3. Adsorption separation measurements
containing isobutane with zeolite 5A and carbon molecular sieve
adsorbents. Ahmed at al. [17,18] studied breakthrough curves of
Adsorption experiments of LPG were performed in a stainless
methane, ethane and propane mixture and the possibility of pro-
steel tube (20 mm i.d.). The breakthrough tests of LPG were carried
pane separation from the gas mixture on 4A molecular sieve zeo-
out at ambient temperature (298 ± 1 K) and atmospheric pressure
lite as well as effects of the feed flow rate, feed concentration
with flow rate of 60, 80 and 100 mL/min and bed height of 5, 8,
and composition of adsorbate on propane purity and productivity.
10 cm, respectively. The inlet gas flow rate was measured by a
Although a number of researchers have reported the adsorption
rotameter. The inlet and outlet concentrations of methane, ethane,
separation of light hydrocarbons, only a few articles have focused
propane, butane and isobutane were monitored by a gas chro-
on the adsorption dynamics of separation of propane from LPG or
matography (Aglient 7890A, USA) equipped with a flame ioniza-
its performance in a fixed bed absorber. In this work, the adsorp-
tion detector (FID) and a GS-GasPro capillary column using N2 as
tion of LPG onto five adsorbents (4A, 5A, 13X, NaY zeolite pellets
carried gas. The data for ethane from effluent stream of adsorption
and granular activated carbon) have been studied at ambient con-
column were measured at 1 min sampling interval while those for
ditions in a fixed bed. The dynamics measurements of propane,
butane and isobutane were analyzed at 2 min sampling interval
isobutane and butane on granular activated carbon have been per-
until breakthrough of all adsorbate species were completed. The
formed to investigate effects of different operation parameters
experimental results (breakthrough carves) were expressed in
such as flow rate and bed height on the separation of propane from
the form of variations of C/C0 verse time (C and C0 refer to the con-
LPG. Adsorption models have been employed to simulate break-
centration of outlet and inlet gas respectively). In this work, when
through curves of propane, isobutane and butane. Moreover, some
the concentration of adsorbate was detected above 0.5% of the cor-
characteristic parameters from models have been analyzed to test
responding C0, it was regarded that the breakthrough point was
their validity against the experimental data. The performance of
reached.
those adsorbents on separation of propane from LPG in lab-scale
enables us to collect necessary information for industrial scale
applications of refining green refrigerants R290 from LPG. 3. Mathematical models

2. Experimental Successful design of a fixed bed adsorption process requires

prediction of the concentration–time profile or breakthrough curve
2.1. Materials for the adsorbate. The maximum adsorption capacity of an adsor-
bent is also needed for rational design. Several models have been
Commercial granular activated carbons (GAC) with average par- used to fulfil those purposes.
ticle diameter of 0.9 mm were supplied by Shanxi Xinhua activated
carbon Co., Ltd. The zeolite 13X and NaY were manufactured by 3.1. The length of unused bed (LUB)
Anhui Mingmei Minchem Co., Ltd and Dalian absorbent Co., Ltd,
respectively. The zeolite 4A and 5A were both purchased from In the design of an adsorption column which can obtain a high
Tianjin Kemeiou Chemical Reagrnt Co., Ltd. All zeolites in this work sorbent productivity as well as a high product purity, a short LUB is
were spherical grain with average diameter of 4 mm and all adsor- desired. This length represents the length of the mass-transfer
bent samples were milled and sieved to size of 250–425 lm. Before zone, which is the part of the column where the adsorbent is not
adsorption breakthrough experiments, zeolite pellets 13X, NaY, 4A completely used and but where adsorption is still occurring [19].
and 5A were degassed at 300 °C for 4 h in air and cooled to ambient The LUB is defined by [20]
temperature while GAC was put in an oven at 110 °C for about


12 h. tb
LPG was obtained from Guangzhou Huakai Oil and Gas Co., Ltd. LUB ¼ 1  L ð1Þ
t
The specific LPG mainly contained propane with small concentra-
tion of methane, ethane, propene, isobutane and butane. The com- where L is the length of the bed (m), tb is the breakthrough time (s)
position for the test LPG by volume fraction was: propane, at which the outlet concentration reaches its maximum permissible
97.05 ± 0.10%; ethane, 0.17 ± 0.01%; isobutane, 2.16 ± 0.01%; level and t* is the stoichiometric time (s) which can be obtained by
butane, 0.54 ± 0.01%; all other constituents, 0.08%. Therefore, the drawing a vertical line on breakthrough curve at a time when the
concentration of methane, propene and corrosive substance (e.g. used adsorbent capacity equals the unused adsorbent capacity
sulphur, mercaptans and water) were very low. [21]. A single dynamic adsorption experiment which generates
the entire breakthrough curve is sufficient to enable tb and t* to be
2.2. Characterization determined since
Z tb

C
The pore structure characteristics of adsorbents were deter- tb ¼ 1 dt ð2Þ
mined by N2 adsorption–desorption at 77 K with a Micromeritics 0 C0
ASAP 2020 automatic adsorption instruments. Prior to the mea-
Z 1

surement, samples were outgassed at 200 °C in a vacuum system C
t ¼ 1 dt ð3Þ
for 4 h. The specific surface areas were calculated from adsorption 0 C0
branches using the BET (Brunauer–Emmett–Teller) equation. The
micropore, mesopore and total volumes were determined via And t* can be also regarded as the adsorption time for reaching the
t-plot, BJH (Barret–Joyner–Halenda) and single point method, half of the equilibrium concentration for a symmetrical break-
respectively. The micropore and mesopore size distribution of all through curve, it is the time at which C/C0 equals 0.5 [22].
 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8 3

3.2. Yoon–Nelson model Table 1

Textural properties of five adsorbents calculated from N2 adsorption data at 77 K.

Yoon–Nelson model is a relatively simple model based on the Adsorbent SBET (m2/g) Vt (cm3/g) Vmeso (cm3/g) Vmicro (cm3/g)
assumption that the rate of decrease in the probability of adsorp- 4A 17.80 0.0998 0.1132 0.0004
tion for each molecule is proportional to the probability for adsorp- 5A 375.12 0.2181 0.0714 0.1528
tion and breakthrough on the adsorbent [23]. One of the 13X 410.43 0.3179 0.1160 0.1993
advantages of this model is that it requires no detailed data con- NaY 542.14 0.3365 0.1268 0.2172
GAC 884.27 0.4485 0.0337 0.4491
cerning the characteristics of adsorbate, the type of adsorbent,
and the physical properties of adsorption bed. The equation for
the 50% breakthrough concentration from a fixed bed is
be explained by the poor adsorption of N2 molecules at the pore

 openings [24].
C
ln ¼ K YN t  t 0:5 K YN ð4Þ Micropore and mesopore size distributions of 4A, 5A, 13X and
C0  C
NaY zeolite and activated carbon which were presented in Fig. 2.
It was interesting to notice that there existed numerous microp-
where KYN is the rate constant (min1); t0.5, the time requires for
ores in the AC sample provided by Shanxi Xinhua activated carbon
50% adsorbate breakthrough (min) and t is the breakthrough time
Co., Ltd despite the fact that AC is typical adsorbent with rich
(min). The values of KYN and t0.5 can be obtained from the slope
mesopores. The micropore and mesopore size distributions of all
and the intercept, respectively, of a linear plot of ln[C/(C0  C)] verse
zeolite samples were complicate. All zeolite adsorbents except 4A
the breakthrough time t.
showed corresponding characteristic peaks in the micropore range
(0–2 nm). All zeolite adsorbents, however, had wide miscellaneous
pore distributions in the micropore range (2–50 nm), which helped
4. Results and discussion
explain the hysteresis loops in N2 adsorption isotherm shown in
Fig. 1. One possible reason for the mesoporous structure in zeolite
4.1. Pore structure characterization
adsorbents is that these samples were industrial or commercial
adsorbents, containing impurity such as binder.
N2 adsorption isotherms of zeolite adsorbents shown in Fig. 1
were sort of type IV, indicating significant mesopore behavior.
While the isotherm curve for AC was of type I, indicating the pres- 4.2. Adsorption breakthrough studies on different adsorbents
ence of microporous structure in this sample. Above a relative
pressure of about 0.8, monolayer coverage was completed follow- Fig. 3 showed the adsorption performance of 4A, 5A, 13X, NaY
ing which adsorption occurred in successive layers with a remark- zeolite pellets and activated carbon for LPG at ambient tempera-
able hysteresis loop in N2 adsorption isotherm curves of 4A, 5A, ture and atmospheric pressure. Inlet gas flow rate and bed height
13X and NaY zeolites. Characteristic features of N2 adsorption iso- were maintained at 80 mL/min and 8 cm, respectively.
therm curves of adsorbents which have mesoporous structure are For all adsorbents, except 5A zeolite, ethane penetrated into
its hysteresis loop, which is associated with capillary condensation adsorbents first, because it was a fast-diffusing and less-strongly-
taking place in mesopores. Therefore, there existed open meso- adsorbed species in this system. Then ethane was displaced by
pores in zeolites 4A, 5A, 13X and NaY which are tested in this work. later-coming propane, therefore, a marked roll-up or overshoot
The textural characteristics such as surface area, total pore volume, effect was observed in breakthrough curves of ethane. The momen-
mesopore volume and micropore volume were evaluated from the tary higher concentration of ethane in the outlet gas than in the
low temperature nitrogen adsorption and summarized in Table 1. feed gas was due to desorption of ethane when propane, isobutane
It can be seen that GAC had both the largest BET surface area and butane were competitively adsorbed in the column. The roll-
and pore volume among these five adsorbents, and the pore vol- up effect may also be attributed to a temperature wave caused
ume of 13X and NaY were nearly the same. It was interesting to by the exothermic adsorption of propane [25]. The concentration
notice that the BET surface area and pore volume of 4A zeolite
of propane in the feed was extremely high, so this thermal effect
were much smaller than those of others, this phenomenon might might largely contribute to the classical roll-up and should not

300
0.015
0.010
Incremental Pore Volume (cm3• g-1 )

AC 0.005
NaY
0.000
NaY
Volume adsorbed (cm3/g STP)

0.015
13X
0.010
200 5A 0.005 AC
4A 0.000
0.015
0.010
0.005 4A
0.000
100 0.015
0.010
0.005
0.000 13X
0.015
0.010
0.005
0 0.000
5A
0.0 0.2 0.4 0.6 0.8 1.0 0 100 200 300 400 500
Relative pressure (P/P0) Pore diameter (Å)

Fig. 1. N2 adsorption isotherm of 4A, 5A, 13X and NaY zeolite and activated carbon Fig. 2. Micropore and mesopore size distributions of 4A, 5A, 13X and NaY zeolite
at 77 K. and activated carbon.
4 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8

10 10
ethane ethane
propane propane
isobutane isobutane
8 8
butane
butane

6 6

C/C0
C/C0

4 4

2 2

0 0
0 3 6 9 12 15 0 3 6 9 12 15
time (min) time (min)

(a) (b)
10 10
ethane ethane
propane propane
8 isobutane 8 isobutane
butane butane

6 6
C/C0
C/C0

4 4

2 2

0 0
0 10 20 30 40 50 0 10 20 30 40 50
time (min) time (min)

(c) (d)
10
ethane
propene
8 isobutane
butane

6
C/C0

0
0 20 40 60 80 100
time (min)

(e)
Fig. 3. Breakthrough curves for LPG adsorption in 4A (a), 5A (b), 13X (c), Y (d) zeolite and activated carbon (e) fixed bed at flow rate of 80 mL/min and bed height of 8 cm.

be neglected. The roll-up effect in 4A zeolite was less marked,
partly because of the lower adsorption capacity for LPG. The resi-
dence time of isobutane in the fixed bed was longer than the figure
for propane and butane which were last detected in the eluting gas.
This was in agreement with Song’s article [26], which showed that
the diffusivity of the adsorbed normal hydrocarbon molecules in
silicalite-1 decreases as the chain-length increases.
Breakthrough curves for 5A zeolite in Fig. 3(b) were totally dif-
ferent from others in Fig. 3. When the mixture gases were passing
through the fixed bed, isobutane came out first and was followed
by propane, ethane and butane in a row. According to Mori and
Tomita [27], kinetic diameters of butane and isobutane are 0.43
and 0.50 nm, respectively, so isobutane cannot easily penetrate
the pore openings of zeolite 5A at atmospheric pressure because
 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8
of the steric hindrance effect and was purged out quickly. While
the kinetic diameters of other components are smaller than the
average pore size of zeolite 5A, they were adsorbed and the con-
centration of ethane, propane and butane in the gas phrase were
little, which explained the reason for the roll-up effect in the
breakthrough curve of isobutane. However, it was interesting to
find that zeolite 4A can adsorb isobutane and butane despite of
the fact that their kinetic diameters are bigger than 0.40 nm. The
possible reason was that the zeolite 4A in this experiment has
mesoporous and macroporous structure which was formed during
manufacture so that large molecules such as isobutane and butane
can penetrate these pores and thus the possibility of adsorption.
Fig. 3(c) and (d) also showed that breakthrough curves for 13X
and NaY zeolite were nearly the same, even though the Si/Al ratio
between 13X and NaY zeolites, the range 1.0–1.5 for 13X and 1.5–
3.0 for NaY, are different. This is mainly because the ionic nature of
the 13X and NaY zeolite makes them have a high affinity for polar
molecules. But this affinity may have little influence on non-polar
molecules such as alkanes.
It was evident from Fig. 3 that all these adsorbents, except 5A
zeolite, were able to separate isobutane and butane from feed
gas, while ethane cannot be removed effectively under the operat-
ing conditions. Among these adsorbents, activated carbon showed
the highest adsorption capacity and excellent separation perfor-
mance. This can be attributed to large surface area and wide pore
distribution. For activated carbon, the electrical charges are too
weak to exert any significant electric field or field gradient on
the surface. So the adsorption strength between activated carbon
and those non-polar alkane molecules are stronger. The break-
through time of isobutane and butane in activated carbon fixed
bed were 31 and 59 min, respectively. The time required for com-
plete column saturation was 21 min for propane. At this point the
concentration of propane in the outlet gas reached the highest
(about 99.7% (volume fraction)). It indicated that high purity pro-
pane could be obtained. With experimental results that AC showed
the highest adsorption capacity and desirable separation perfor-
mance among these adsorbents, as well as the explanation for pore
size distribution of these adsorbents above, it was reasonable to
say that the micropores played more important role than meso-
pores here in terms of the adsorption separation process for high
purity propane from LPG.
Although the separation of ethane from propane in LPG was not
successful under ambient conditions, these adsorption break-
through studies of ethane on various adsorbents may allow us to
compare the adsorbents’ performance at atmospheric pressure
and provide us with useful guidance to process the separation
experiments at different operation pressures in the near future.
4.3. Determination of length of unused bed
The length of unused bed (LUB) is one of the important factors
for the design of the separation process. The LUB values for isobu-
tane and butane in different fixed beds were determined by anal-
ysis of breakthrough curves data according to Eq. (1). The results
were given in Table 2.
Table 2
Length of unused bed of isobutane and butane adsorption in different fixed beds.
Fixed bed Bed height (cm) Flow rate Isobutane (cm)
(mL/min)
LUB
4A 8 80 2.00
5A – 4.60
13X 1.52
Y 1.52
GAC 1.47
5
In the case of fixed bed adsorption of isobutane and butane, the
LUB values onto activated carbon were smaller than that of other
adsorbent samples, which indicated that the utilization of bed
capacity of activated carbon was the largest. A small LUB, or a nar-
row mass-transfer zone, was desirable to make efficient use of the
adsorbent and to reduce the energy costs in regeneration [28].
Therefore, it can be found that the activated carbon fixed bed
may be more suitable for the separation of propane from LPG via
adsorption.
4.4. Adsorption dynamics on activated carbon
4.4.1. Effect of bed height
To study the effect of bed height on the separation performance
of propane, isobutane and butane, the fixed bed with different bed
heights (5, 8, 10 cm) at a constant flow rate of 80 mL/min were
examined. It can be observed from Fig. 4(a), (c), and (e) that break-
through curves with different bed height presented nearly the
same characteristic S shape, but the time required for outlet gas
to reach the concentration of the inlet gas increased with increas-
ing bed height. For the given inlet concentration and flow rate, the
shape of concentration profile maintained stable for the same
adsorbent and adsorbate in a fixed bed, which indicated that vari-
ation in bed height has no or slight effect on the shape of break-
through curve. As the adsorbent doses increased, the adsorption
capacity for LPG increased and the length of mass transfer zone
in per unit bed height became smaller, indicating a greater utiliza-
tion of fixed bed. Therefore, it is advisable to use relatively longer
bed for the separation of LPG because the treated gas volume will
increase. However, much higher bed height values are not accepted
when flow resistance and pressure drop of the fixed bed are taken
into consideration.
4.4.2. Effect of flow rate
Fig. 4(b), (d), and (f) showed breakthrough curves for the
adsorption of propane, isobutane and butane in a continuous-
flow fixed bed with activated carbon at ambient temperature and
atmospheric pressure. The bed height maintained 8 cm. The effect
of flow rate was studied by varying its value between 60 and
100 mL/min. It was found that a decrease in flow rate at constant
bed height increased the breakthrough time of adsorbate. At higher
flow rate, the adsorbate molecular will not have sufficient time to
diffuse within the pores of activated carbon and propane, isobu-
tane and butane leave the fixed bed before reaching the equilib-
rium, resulting in low adsorbent utilization. For a smaller value
of flow rate, the adsorption process is controlled by interparticle
diffusion, the front of the adsorption zone reaches the top of col-
umn later and the residence time of each component is more, this
is especially obvious for isobutane and butane, indicating that the
relatively lower flow rate is in favor of separating isobutane and
butane from LPG in activated carbon column. Moreover, an
increase in flow rate is likely to increase the sharpness of the
breakthrough curve. A Possible reason for the tailing of a break-
through curve (a slow approach of C/C0 toward 1) at lower flow
rate is that the interparticle diffusion is the rate-limiting transport
process and external mass transfer rate within the activated carbon
fixed bed is slow. The higher external diffusion resistance means
the slower mass transfer rate, resulting in slower adsorption rate.
Butane (cm) In other words, the length of mass transfer zone will be longer.
Therefore, a less sharpness breakthrough curve is obtained at rela-
tively slower flow rate.
2.66
1.65 4.5. Application of the Yoon–Nelson model
1.65
1.44 Breakthrough curves of propane, isobutane and butane onto
activated carbon fixed bed were studies through a simple theoret-
6 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8

1.2 1.2

0.8 0.8
C/C0

C/C0
0.4 0.4
5 cm
60 mL/min
8 cm
80 mL/min
10 cm
100 mL/min

0.0 0.0
0 10 20 30 40 0 10 20 30 40
time time

(a) (b)
1.0 1.0
60 mL/min
5 cm
8 cm 80 mL/min
10 cm 100 mL/min
C/C0

C/C0

0.5 0.5

0.0 0.0
0 20 40 60 80 100 0 20 40 60 80 100
time time

(c) (d)
1.0 1.0

5 cm 60 mL/min
8 cm 80 mL/min
100 mL/min
10 cm
C/C0
C/C0

0.5 0.5

0.0 0.0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
time time

(e) (f)
Fig. 4. Breakthrough curves for propane (a), isobutane (c) and butane (e) adsorption at different bed heights (flow rate = 80 mL/min) and breakthrough curves for propane (b),
isobutane (d) and butane (f) adsorption at different flow rate (bed height = 8 cm) in activated carbon fixed bed. Symbols: experimental data, solid lines: predictions of the
Yoon–Nelson model.

ical model developed by Yoon–Nelson. The values of KYN (a rate
constant) and t0.5 (the time required for 50% adsorbate break-
through) were determined from plots of ln[C/(C0  C)] versus time
(t) at different values of bed height and flow rate. The values of
these two parameters along with the correlation coefficient (R2)
related to propane, isobutane and butane adsorption were listed
in Tables 3–5 respectively. Fig. 4 showed theoretical breakthrough
curves of propane, isobutane and butane at different values of
height bed, flow rate for the Yoon–Nelson model.
The variation of these parameters for each adsorbate showed
nearly the same trend. The values of t0.5, which were similar to
the experimental values of t0.5, increased with decrease in flow rate
and with increase in bed height, while the values of KYN were found
to decrease with decrease in flow rate and maintained stable with
increase in bed height, which was in agreement with the experi-
mental results. According to Yoon and Nelson [23], KYN character-
ized the shape of the breakthrough curve and a sharp breakthrough
curve always had a large KYN. At higher flow rate, the external
 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8 7

Table 3
Parameters of Yoon–Nelson model for propane adsorption under different conditions.

Concentration (L/L) Bed height (cm) Flow rate (mL/min) KYN (min1) T0.5 (min) T0.5(experimental) (min) R2
97.05 5 80 1.69445 7.2 7.5 0.9956
8 80 1.66252 11.3 11.5 0.9905
10 80 1.59882 13.6 14.0 0.9893
8 60 1.19557 13.2 14.5 0.9930
8 100 2.99493 8.3 8.0 0.9825

Table 4
Parameters of Yoon–Nelson model for isobutane adsorption under different conditions.

Concentration (L/L) Bed height (cm) Flow rate (mL/min) KYN (min1) T0.5 (min) T0.5(experimental) (min) R2
0.0216 5 80 0.38680 30.1 31.0 0.9981
8 80 0.36187 39.5 41.5 0.9995
10 80 0.37089 48.0 47.5 0.9991
8 60 0.20860 55.8 58.0 0.9991
8 100 0.43848 35.5 33.0 0.9994

Table 5
Parameters of Yoon–Nelson model for butane adsorption under different conditions.

Concentration (L/L) Bed height (cm) Flow rate (mL/min) KYN (min1) T0.5 (min) T0.5(experimental) (min) R2
0.0054 5 80 0.21872 55.3 57.0 0.9964
8 80 0.21096 74.7 76.0 0.9956
10 80 0.21090 88.0 89.0 0.9977
8 60 0.18401 96.6 95.0 0.9943
8 100 0.22821 63.7 63.0 0.9983

diffusion resistance is lower. Therefore the adsorption rate
increases and a short mass transfer zone length is obtained, hence
resulting in a faster and sharper breakthrough. At the same time,
all the correlation coefficients exceeded 0.98, indicating an excel-
lent confidence level for all the fitting lines of ln[C/(C0  C)] versus
time (t).
5. Conclusions
The adsorption separation process for high purity propane from
liquefied petroleum gas (LPG) in a fixed bed was studies. The
results obtained are as follows.
All these five adsorbents, except 5A zeolite pellet, were able to
separate isobutane and butane from LPG successfully, while ethane
cannot be removed effectively under ambient temperature and
atmospheric pressure. Among these adsorbents, activated carbon
showed the highest adsorption capacity and desirable separation
performance, and the concentration of propane in the outlet gas
could reach as high as 99.7% (volume fraction). The LUB value of
activated carbon fixed bed was the smallest, which confirmed that
activated carbon fixed bed was more suitable for the purification of
propane from LPG via adsorption.
The adsorption of propane, isobutane and butane in activated
carbon fixed bed were strongly dependent on the flow rate and
bed height. The breakthrough time increased with increase in
bed height while it decreased with increase in flow rate.
The Yoon–Nelson model was found to be suitable for describing
the dynamic behaviors of the whole column and estimating the
characteristic parameters such as rate constant (KYN) and 50%
breakthrough time (t0.5). The value of t0.5 was similar to the exper-
imental time. The KYN value decreased with decrease in flow rate
and maintained stable with increase in bed height.
Acknowledgements
We gratefully acknowledge the financial supports of the
National Natural Science Foundation of China (Grant Nos.
21176086 and 21376101) and Science and Technology Planning
Project of Guangdong Province, China (No. 2013B010403002) for
this work.
References
[1] W.J.M. Martens, Health impacts of climate change and ozone depletion: an
ecoepidemiologic modeling approach, Environ. Health Perspect. 106 (1998)
241–251.
[2] D. de Jager, M. Manning, L. Kuijpers, Safeguarding the ozone layer and the
global climate system: issues related to hydrofluorocarbons and
perfluorocarbons, IPCC/TEAP Special Report. Technical Summary, 2005.
[3] M. Mohanraj, S. Jayaraj, C. Muraleedharan, Comparative assessment of
environment-friendly alternatives to R134a in domestic refrigerators, Energy
Effi. 1 (2008) 189–198.
[4] E. Navarro, J. Urchueguia, J. Gonzalvez, J. Corberán, Test results of performance
and oil circulation rate of commercial reciprocating compressors of different
capacities working with propane (R290) as refrigerant, Int. J. Refrig. 28 (2005)
881–888.
[5] K. Mani, V. Selladurai, Experimental analysis of a refrigerant mixture as drop-in
replacement for CFC 12 and HFC 134a, Int. J. Ther. Sci. 34 (2007) 246–252.
[6] M. El-Morsi, Energy and exergy analysis of LPG (liquefied petroleum gas) as a
drop in replacement for R134a in domestic refrigerators, Energy 86 (2015)
344–353.
[7] B. Purkayastha, P.K. Bansal, An experimental study on HC290 and a commercial
liquefied petroleum gas (LPG) mix as suitable replacements for HCFC22, Int. J.
Refrig.-Rev. Int. Froid 21 (1998) 3–17.
[8] J.R. Alcantara-Avila, F.I. Gomez-Castro, J.G. Segovia-Hernandez, K. Sotowa, T.
Horikawa, Optimal design of cryogenic distillation columns with side heat
pumps for the propylene/propane separation, Chem. Eng. Process. 82 (2014)
112–122.
[9] A. Ortiz, A. Ruiz, D. Gorri, I. Ortiz, Room temperature ionic liquid with silver
salt as efficient reaction media for propylene/propane separation: absorption
equilibrium, Sep. Purif. Technol. 63 (2008) 311–318.
[10] M. Takht Ravanchi, T. Kaghazchi, A. Kargari, Application of membrane
separation processes in petrochemical industry: a review, Desalination 235
(2009) 199–244.
[11] N. Lamia, M. Jorge, M.A. Granato, F.A. Almeida Paz, H. Chevreau, A.E. Rodrigues,
Adsorption of propane, propylene and isobutane on a metal–organic
framework: molecular simulation and experiment, Chem. Eng. Sci. 64 (2009)
3246–3259.
[12] R.T. Yang, Gas Separation by Adsorption Processes, Butterworth-Heinemann,
2013. pp. 1–8.
[13] H. Ma, H. Ren, S. Meng, F. Sun, G. Zhu, Novel porphyrinic porous organic
frameworks for high performance separation of small hydrocarbons, Sci. Rep. 3
(2013).
[14] J. Kawakami, S. Tai, S. Morimoto, M. Miyake, Method and device for purifying
propane, JP Patent, 2011063508, 2011.
8 P. Liu et al. / Separation and Purification Technology 158 (2016) 1–8
[15] J. Kawakami, S. Tai, S. Tanaka, Purification Method and Purification System for
Propane, US Patent, 20140343341, 2014.
[16] S.H. Cho, S.S. Han, J.N. Kim, K.T. Chue, J.I. Yang, H.T. Beum, Adsorption
Separation and Purification Apparatus and Process for High Purity Isobutane
Production, US Patents, 6022398, 2000.
[17] M.J. Ahmed, A.H.A.K. Mohammed, A.A.H. Kadhum, Prediction of breakthrough
curves for light hydrocarbons adsorption on 4A molecular sieve zeolite, Kor. J.
Chem. Eng. 27 (2010) 752–758.
[18] M.J. Ahmed, S.K. Theydan, Modeling of propane separation from light
hydrocarbons by adsorption on 4A molecular sieve zeolite, J. Nat. Gas Sci.
Eng. 18 (2014) 1–6.
[19] I. Mantovaneli, E. Ferretti, M. Simões, C. Silva, The effect of temperature and
flow rate on the clarification of the aqueous stevia-extract in a
fixed-bed column with zeolites, Braz. J. Chem. Eng. 21 (2004)
449–458.
[20] D.M. Ruthven, Principles of Adsorption and Adsorption Processes, John Wiley &
Sons, 1984. pp. 270–271.
[21] W.J. Thomas, B.D. Crittenden, Adsorption Technology and Design,
Butterworth-Heinemann, 1998. pp. 165–166.
View publication stats
[22] W.L. McCabe, J.C. Smith, P. Harriott, Unit Operations of Chemical Engineering,
McGraw-Hill, Singapore, 2001. pp. 821–834.
[23] Y.H. Yoon, J.H. Nelson, Application of gas adsorption kinetics I. A theoretical
model for respirator cartridge service life, Am. Ind. Hyg. Assoc. J. 45 (1984)
509–516.
[24] S.N. Azizi, A.R. Dehnavi, A. Joorabdoozha, Synthesis and characterization of LTA
nanozeolite using barley husk silica: mercury removal from standard and real
solutions, Mater. Res. Bull. 48 (2013) 1753–1759.
[25] L. Hamon, E. Jolimaître, G.D. Pirngruber, CO2 and CH4 separation by adsorption
using Cu-BTC metalorganic framework, Ind. Eng. Chem. Res. 49 (2010) 7497–
7503.
[26] L. Song, Z. Sun, L. Duan, J. Gui, G.S. McDougall, Adsorption and diffusion
properties of hydrocarbons in zeolites, Micropor. Mesopor. Mater. 104 (2007)
115–128.
[27] N. Mori, T. Tomita, Separation of n-butane/iso-butane by self-supporting MFI
membranes with various SiO2/Al2O3, Micropor. Mesopor. Mater. 112 (2008)
88–96.
[28] J. Liu, Y. Yan, H. Zhang, Adsorption dynamics of toluene in composite bed with
microfibrous entrapped activated carbon, Chem. Eng. J. 173 (2011) 456–462.