Chemical Engineering Science 144 (2016) 326–335

Contents lists available at ScienceDirect

Chemical Engineering Science

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

Process design and optimization for etherification of glycerol

with isobutene
Jingjun Liu a,b, Prodromos Daoutidis b,n, Bolun Yang a,n
a
Department of Chemical Engineering, Xi’an Jiaotong University, Xi’an 710049, PR China
b
Department of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455, USA

H I G H L I G H T S

A new process for the production of glycerol ethers is proposed.

Dimerization of isobutene is considered in the new process.

Latest kinetic and liquid–liquid equilibrium data are used.

Optimization based on economic analyses is implemented.

Isobutene cost greatly dominates the production cost of glycerol ethers.

art ic l e i nf o a b s t r a c t

Article history: Surplus glycerol from biodiesel production can be transformed to valuable fuel additives through
Received 29 August 2015 etherification with isobutene. Our previous work has studied the kinetics and thermodynamics of this
Received in revised form reaction. In the present work, a novel process design is proposed. An optimization problem is also for-
9 December 2015
mulated to minimize the cost of production of the di-tert-butyl ethers and tri-tert-butyl ether of glycerol.
Accepted 21 January 2016
Available online 5 February 2016
The resulting mixed-integer nonlinear program enables simultaneous determination of values of key
decision variables including number of column trays, feed tray location of column, volume of reactor,
Keywords: process flowrates etc. The optimal design leads to a minimal production cost of glycerol ethers of 1.201
Process design $/kg, which is dominated by the cost of isobutene.
Optimization
& 2016 Elsevier Ltd. All rights reserved.
Glycerol ether
Etherification
Isobutene

1. Introduction
The development and commercialization of biodiesel has
rapidly expanded worldwide in recent years, driven by sustain-
ability and environmental considerations. Glycerol is the main
byproduct in biodiesel production that cannot be used as a fuel
additive, because of its low solubility and poor thermostability. On
the other hand it can be converted to a variety of high value
chemicals (Zhou et al., 2008; Zheng et al., 2008; Izquierdo et al.,
2012; Behr et al., 2008; Ayoub and Abdullah, 2012; Tan et al., 2013;
Quispe et al., 2013). tert-Butyl glycerol ethers are among the most
promising ones. Glycerol (G) reacting with isobutene (IB) or tert-
butyl alcohol forms mono-tert-butyl ethers of glycerol (ME), di-
tert-butyl ethers of glycerol (DE) and tri-tert-butyl ether of glycerol
n
Corresponding authors.
E-mail addresses: daout001@umn.edu (P. Daoutidis),
blunyang@mail.xjtu.edu.cn (B. Yang).
(TE). DE and TE, the so-called high-ethers, are potential fuel
additives. When they are added into biodiesel or diesel (5–
15 wt%), not only the particulate matter and greenhouse gas
emissions in combustion are reduced (Kesling et al., 1994, 1995;
Frusteri et al., 2013; Beatrice et al., 2013; Asdrubali et al., 2015), but
also low-temperature properties (pour point and cold filter plug-
ging point) and the viscosity of the biodiesel are improved (Melero
et al., 2010). Also, their high octane number makes them an
alternative oxygenate additive to gasoline (Karinen and Krause,
2006), as a substitute for methyl-tert-butyl ether (MTBE) and
ethyl-tert-butyl ether (ETBE). As such their market potential is
significant.
Although tert-butyl alcohol is soluble with glycerol and cheaper
than isobutene in the current market (Chang et al., 2014; Ozbay
et al., 2012; Celdeira et al., 2014), tert-butylation of glycerol using
isobutene has drawn more consideration because of higher gly-
cerol conversion and product selectivity. Most of the research in
this area has focused on finding active and selective catalysts and
optimizing the reaction conditions (Frusteri et al., 2013; González

http://dx.doi.org/10.1016/j.ces.2016.01.055
0009-2509/& 2016 Elsevier Ltd. All rights reserved.
 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335
et al., 2013; González et al., 2014; González et al., 2014; Zhao et al.,
2013; González et al., 2013; González et al., 2012; Frusteri et al.,
2012; Zhou et al., 2014; Singh et al., 2015). Three kinds of pro-
cesses have been proposed, the ARCO process (Vijai and Berwyn,
1995), the Behr and Obendorf process (Behr and Obendorf, 2002)
and the Di Serio et al. process (Di Serio et al., 2010). They differ in
the methods for the recycle of unreacted glycerol and isolation of
ME from the other products. The flowsheets of the three processes
are illustrated in Supplementary material. In the ARCO process
(Vijai and Berwyn, 1995), partially reacted glycerol, isobutene and
products enter into a decanter. The heavier phase containing gly-
cerol, ME and acid catalyst is recycled back to the reactor. The light
phase is fed to a stripping column to recover isobutene, followed
by a water wash column to obtain high-ethers. A large amount of
glycerol and ME in the wastewater is not recovered, resulting in
waste of materials.
Behr and Obendorf (2002) developed a process consisting of a
three-step reactor cascade. The fresh glycerol is used to extract ME
in the reaction product, instead of being fed to the reactor directly.
The extract stream containing glycerol and ME is recycled to the
reactor, and the raffinate stream is passed to a flash tank to recover
unreacted isobutene. The bottom product of the flash tank enters
into a vacuum distillation column to isolate DE and TE from other
impurities including ME and Glycerol. Cheng et al. (2011) modified
this process by changing the three-step reactor cascade to one
reactor to reduce capital costs, replacing the flash tank with a
multistage stripping column to ensure a complete recovery of the
unreacted isobutene, and recycling the bottom product of the
vacuum distillation column to the extraction column instead of to
the reactor to avoid reverse reaction. In this configuration, the
vacuum distillation column used for purification of DE and TE
operates under extreme conditions and steams out a large amount
of distillates with high boiling points.
Di Serio et al. (2010) proposed a process using biodiesel to
extract ME, DE and TE from the etherification products. The extract
is washed by water to remove ME and obtain biodiesel containing
DE and TE. Combining the properties of high-ethers and their
application, this process produces a mixture of biodiesel high-
ethers instead of “pure” high-ethers.
The concept of the Behr and Obendorf process has been
employed in subsequent studies. Vlad et al. (Vlad et al., 2013)
optimized the process with a low-purity isobutene feeding. Martín
and Grossmann (2014) integrated the production of biodiesel
(FAME or FAEE) with glycerol ethers. The integration increases the
production of biofuel by 20%, with a production cost of almost
3 times that of producing biodiesel alone, due to the high price of
isobutene. They concluded that a key factor for improving eco-
nomics is to reduce the cost and improve the availability of iso-
butene. Also they found that it is economically promising to pro-
duce isobutene from biomass containing glucose (Martín and
Fig. 1. Reaction pathway of the etherification process.
327
Grossmann, 2014). Thus, they designed an integrated process for
the production of biodiesel, bioethanol, isobutene and glycerol
ethers from algae (de la Cruz et al., 2014). An affordable production
cost for biofuel was obtained. In summary, although optimization
and economic analyses of specific process designs have been
performed using simplified process models (Vlad et al., 2013;
Martín and Grossmann, 2014) or based on sensitivity analysis
(Cheng et al., 2011), there is no definitive evaluation of the
economic feasibility of this etherification route for glycerol
conversion.
Our previous work has investigated the reaction kinetics (Liu
et al., 2013) and liquid–liquid equilibrium (Liu et al., 2014) of
etherification of glycerol with isobutene. This paper aims at the
design and optimization of a process for this etherification reac-
tion. We initially propose a process flow sheet inspired by the
existing processes, but also modified to avoid waste of raw
materials or extreme operating conditions. The dimerization of
isobutene detected in experiment (Liu et al., 2013) is also con-
sidered in the reaction and separation sections of the process.
Detailed mathematical models are formulated for each unit
operation in the process. Kinetic and liquid–liquid equilibrium
data obtained from our previous experimental study are used,
which is very important for the accuracy of the models. An opti-
mization problem is formulated to seek an optimal process con-
figuration minimizing the production cost of glycerol ethers.
Finally, the effect of raw material prices on the process economics
is studied.
2. Process design
The reaction system contains three consecutive etherification
reactions and one dimerization, shown in Fig. 1. Mono-tert-butyl
ethers of glycerol (ME), di-tert-butyl ethers of glycerol (DE) and
tri-tert-butyl ether of glycerol (TE) are produced continuously. A
side reaction forms dimers of isobutene (DIB). TE and DE are the
desired products. Thus the production process mainly considers
the reaction, the recycle of unreacted glycerol (G), isobutene (IB)
and intermediate ME, and the isolation of DIB from the products.
The proposed process is schematically illustrated in Fig. 2. The
reactions take place in a CSTR under pressure, catalyzed by NKC-9.
ME should be recovered and recycled back to the reactor to be
further converted to high-ethers. Extracting ME from the reaction
products using glycerol proved to be effective in the study of Behr
and Obendorf (Behr and Obendorf, 2002). Thus an extraction
column (E1) is placed after the CSTR, and fresh glycerol is used as
the solvent. Glycerol in the raffinate of E1 (stream 9) is washed by
water in another extraction column (E2). Isobutene in stream 15
can be recovered relatively easily due to its low boiling point
compared with other components. Both the Behr and Obendorf
328 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335

Fig. 2. Process scheme (R – CSTR, E – Extraction column, D – Distillation column, P – Pump, F – Flash tank, H – Heat exchanger, M – Mixer).

process (Behr and Obendorf, 2002) and Di Serio et al. process (Di heat exchanger and pump. The symbols used are listed in the
Serio et al., 2010) use a flash tank with only one stage to separate nomenclature section.
isobutene through depressurization or heating. The ARCO process
(Vijai and Berwyn, 1995) and the modified Behr and Obendorf 3.1.1. CSTR
process (Cheng et al., 2011) replace the flash tank with a stripping Equations for the CSTR include mass and energy balances:
column, because a simple flash could not completely recover the
F in xi;in  F out xi;out þ V 0 ρcat Q E r i ¼ 0; i ¼ 1; 2; …; NC ð1Þ
isobutene. These are reasonable choices as these processes do not
consider the dimerization of isobutene. In our process unreacted XNR
F in H in  F out H out þ V 0 ρcat Q E r
j¼1 j
ΔHj þQ ¼ 0 ð2Þ
isobutene is separated in a distillation column (D1) with a side
draw stream to withdraw DIB simultaneously. The bottom product XNC
of D1 (stream 18) is a mixture of DE and TE. The distillate of D1 i¼1
xi;out ¼ 1 ð3Þ
(stream 16) containing isobutene and small amount of water is
withdrawn in liquid form and recycled back to the reactor. The
3.1.2. Extraction column
extracted ME and glycerol from E1 and E2 are recycled after
It is assumed that a reasonable estimate of the upper bound on
steaming out water from the top of distillation column D2. The
the number of total trays (NTT) of the extraction column is avail-
flash tank F is used to recover isobutene in stream 8 to avoid
able. To determine and optimize the number of trays (NT), the
steaming it out in D2.
withdrawing location of the extract stream (E) has to be deter-
Compared with the existing processes, the proposed process
mined. Thus a binary variable ej ðj ¼ 1; 2; …; NTTÞ denoting the
considers the dimerization of isobutene in the reactor and
position of the extract stream is defined. This is a vector with all
separation sections. The new process avoids steaming out a large
elements being zero except for the one corresponding to the tray
amount of high boiling point materials appearing in the distilla-
on which the extract stream is withdrawn (Fig. 3).
tion column in the Behr and Obendorf process (Behr and (
Obendorf, 2002). All fed glycerol is expected to be converted into 1; j ¼ NT
ej ¼ ð4Þ
the desired product. Also “pure” glycerol ethers are obtained, 0; j a NT
which can not only be mixed with biodiesel as presented in the Di
Serio et al. process (Di Serio et al., 2010), but also be added to Equations for the first tray (j ¼ 1) are as follows:
gasoline and diesel. Sxi;S þ LR2 xRi;2  LR1 xRi;1  LE1 xEi;1  e1 ExEi;1 ¼ 0; i ¼ 1; 2; …; NC ð5Þ

SH S þ LR2 H R2  LR1 H R1  LE1 H E1  e1 EH E1 ¼ 0 ð6Þ

3. Model description
xEi;1 γ Ei;1 ¼ xRi;1 γ Ri;1 ; i ¼ 1; 2; …; NC ð7Þ
3.1. Process model
XNC
i¼1
xEi;1 ¼ 1 ð8Þ
In order to develop a model for the entire etherification pro-
cess, steady state models are considered for all unit operations, i.e. XNC
the CSTR, extraction column, distillation column, flash tank, mixer, i¼1
xRi;1 ¼ 1 ð9Þ
 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335 329

Fig. 3. Scheme of extraction column. Fig. 4. Scheme of distillation column.

xi;R ¼ xRi;1 ; i ¼ 1; 2; …; NC ð10Þ 3.1.3. Distillation column

Binary vector variables f j ðj ¼ 2; 3; …; NTT  1Þ, wj ðj ¼ 2; 3; …; NT
R ¼ LR1 ð11Þ T  1Þ are defined to indicate the positions of feed tray and side
stream tray. Similar to the extraction column model, an upper
TR ¼ T1 ð12Þ bound on the number of total trays (NTT, including condenser and
reboiler) is assumed. A binary variable bj ðj ¼ 2; 3; …; NTT  1Þ is
Equations for tray j (j ¼ 2; 3; …; NTT  1) are of the form:
used to capture the entry position of the boil-up vapor stream
LRjþ 1 xRi;j þ 1 þ LEj 1 xEi;j  1 LRj xRi;j LEj xEi;j  ej ExEi;j ¼ 0; i ¼ 1; 2; …; NC ð13Þ from the reboiler. The entry position also determines the number
of trays (NT) of the distillation column (Fig. 4).
LRjþ 1 H Rjþ 1 þ LEj 1 H Ej 1  LRj H Rj LEj H Ej ej EH Ej ¼ 0 ð14Þ (
1; j ¼ NF
fj ¼ ð26Þ
0; ja NF
xEi;j γ Ei;j ¼ xRi;j γ Ri;j ; i ¼ 1; 2; …; NC ð15Þ
(
XNC 1; j ¼ NW
xE ¼ 1 ð16Þ wj ¼ ð27Þ
i ¼ 1 i;j 0; j aNT
XNC (
i¼1
xRi;j ¼ 1 ð17Þ 1; j ¼ NT  1
bj ¼ ð28Þ
Equations for the last tray (j ¼ NTT) are: 0; j a NT  1

Fxi;F þ LENTT  1 xEi;NTT  1  LRNTT xRNTT  ej ExEi;NTT ¼ 0; i ¼ 1; 2; …; NC ð18Þ Note that there is no vapor flowing up, and no heat or mass
transfer occurs on trays below the tray on which the boil-up vapor
FH F þ LENTT  1 H ENTT  1  LRj H RNTT  ej EH ENTT ¼ 0 ð19Þ stream from the reboiler enters (j ¼ NT; NT þ1; …; NTT  1). To
describe this in our model, we define an indicator aj ðj ¼ 2; 3; …; N
xEi;NTT γ Ei;NTT ¼ xRi;NTT γ Ri;NTT ; i ¼ 1; 2; …; NC ð20Þ TT  1Þ as function of bj , which equals to 1 only for trays that have
vapor flowing up (j ¼ 2; 3; …; NT  1). Specifically, we set aNTT ¼ 0.
XNC (
i¼1
xEi;NTT ¼ 1 ð21Þ Xj 1 for j ¼ 2; 3; …; NT 1
aj ¼ 1 þ bj  b
k¼2 k
¼ ð29Þ
XNC 0 for j ¼ NT; NT þ 1; …; NTT 1
i¼1
xRi;NTT ¼ 1 ð22Þ
The distillation column D1 recovering liquid isobutene uses a
The total mass balance is: total condenser, while D2 removing water in vapor form needs a
F þSRE ¼ 0 ð23Þ partial condenser. Equations for the condensers (j ¼ 1) take the
form:
The temperature and composition of the extract stream are
equal to those of the corresponding tray, i.e.: total condenser : V 2 yi;2  L1 xi;1  Dxi;1 ¼ 0; i ¼ 1; 2; …; NC
OR partial condenser : V 2 yi;2 L1 xi;1  Dyi;1 ¼ 0; i ¼ 1; 2; …; NC
ð30Þ
xi;E ¼ xEi;NT ; i ¼ 1; 2; …; NC ð24Þ
yi;1 P ¼ γ i;1 xi;1 Psi ; i ¼ 1; 2; …; NC ð31Þ
XNC
T E ¼ T NT ð25Þ yi;1 ¼ 1 ð32Þ
i¼1
330 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335

XNC
xi;1 ¼ 1 ð33Þ 3.1.7. Pump
i¼1
Equations for the pump are:
total condenser : V 2 H V2  L1 H L1  DH L1 þ Q C ¼ 0 F in ¼ F out ð57Þ
OR partial condenser : V 2 H V2  L1 H L1 DH V1 þQC ¼ 0 ð34Þ
xi;in ¼ xi;out ; i ¼ 1; 2; …; NC ð58Þ
RR ¼ L1 =D ð35Þ
T in ¼ T out ð59Þ
Equations for tray j (j ¼ 2; 3; …; NTT  1) have the form:
f j Fzi þ aj þ 1 V j þ 1 yi;j þ 1 þ Lj  1 xi;j  1 þ bj V NTT yi;NTT vin ¼ vout ð60Þ
 Lj xi;j  aj V j yi;j  wj Wxi;j ¼ 0; i ¼ 1; 2; …; NC ð36Þ
P out v2 P v2
H0 ¼ þ hout þ out  in  hin  in ð61Þ
yi;j P ¼ γ i;j xi;j Psi ; i ¼ 1; 2; …; NC ð37Þ
ρg 2g ρg 2g

XNC
yi;j ¼ 1 ð38Þ 3.1.8. Stream and unit connection
i¼1
Pressure drop is not considered in streams. A stream flowing
XNC from unit A to unit B can be described as follows:
i¼1
xi;j ¼ 1 ð39Þ
F Aout ¼ F Bin ð62Þ
f j FH F þ aj þ 1 V j þ 1 H Vjþ 1 þ Lj  1 H Lj 1 þ bj V NTT H VNTT
 Lj H Lj  aj V j H Vj  wj WH Lj ¼ 0; ð40Þ xAi;out ¼ xBi;in ; i ¼ 1; 2; …; NC ð63Þ

Equations for the reboiler (j ¼ NTT) are: T Aout ¼ T Bin ð64Þ

LNTT  1 xi;NTT  1  Bxi;NTT  V NTT yi;NTT ¼ 0; i ¼ 1; 2; …; NC ð41Þ

yi;NTT P ¼ γ i;NTT xi;NTT Psi ; i ¼ 1; 2; …; NC ð42Þ 3.2. Cost calculation

XNC Cost calculation includes equipment costs, utility costs, opera-

i¼1
yi;NTT ¼ 1 ð43Þ
tion and maintenance (OM) costs and raw material costs. The
XNC calculation of these costs follows the procedure of Seider et al.
i¼1
xi;NTT ¼ 1 ð44Þ (2009), details of which are shown in the Supplementary material.
The price of raw materials is shown in Table 1.
LNTT  1 H LNTT  1  BH LNTT  V NTT H VNTT þ Q R ¼ 0 ð45Þ
3.3. Parameters, properties and decision variables
3.1.4. Flash tank
Equations for the flash tank are: The pressure of the units P (outlet), H2, R, H1, E1 and E2 is set
to 20 bar to ensure liquid phase reaction and extraction, while all
F in xi;in  V out yi;out  Lout xi;out ¼ 0; i ¼ 1; 2; …; NC ð46Þ
other units (F, D1, D2, M1 and M2) operate under 1 bar for sim-
plicity of operation. No pressure change is considered in any unit
F in H in  V out H Vout  Lout H Lout þ Q ¼ 0 ð47Þ
except the pump. The catalyst loading density in the reactor ρcat is
yi;out P ¼ xi;out γ i;out Psi ; i ¼ 1; 2; …; NC ð48Þ 60 kg/m3 according to the experimental conditions in our previous
work (Liu et al., 2013). Feed temperature of raw materials and
XNC solvent is 298 K. The outlet temperature of the heat exchanger H3
i¼1
xi;out ¼ 1 ð49Þ
is set to 266 K, the boiling temperature of isobutene, to condense
XNC the recycled vapor isobutene.
i¼1
yi;out ¼ 1 ð50Þ Reaction kinetics and reaction enthalpies were obtained from
our previous study (Liu et al., 2013) and summarized in the Sup-
3.1.5. Mixer plementary material. Activity coefficients are calculated by the
All streams to be mixed in the proposed process are in liquid NRTL model. Binary parameters for the NRTL model obtained from
form. Thus equations for a mixer with NS inlet streams and one the experimental study (Behr and Obendorf, 2002; Liu et al., 2014)
outlet stream can be written as: were used, and those not available were estimated (Di Serio et al.,
XNS 2010). These NRTL parameters are presented in Table 2. Enthalpy
F x ¼ F out xi;out ; i ¼ 1; 2; …; NC
k ¼ 1 k;in i;k
ð51Þ and vapor pressure were calculated in Aspen Properties and cor-
related with temperature using polynomials.
XNS
F k;in ¼ F out ð52Þ The degrees of freedom in this model, which are also the
k¼1
decision variables, are listed in Table 3. To simulate the process,
XNS those decision variables were fixed at some reasonable values.
F k;in H k ¼ F out H out ð53Þ
k¼1 Then an optimization problem was formulated and solved using
gPROMS to optimize the design.
3.1.6. Heat exchanger
Equations for the heat exchanger are: Table 1
Price of raw materials.
F in ¼ F out ð54Þ
Material Price, $/kg Reference
xi;in ¼ xi;out ; i ¼ 1; 2; …; NC ð55Þ
Glycerol 0.38 Argus Biofuels (2014)
F in H k  F out H out þ Q ¼ 0 ð56Þ Isobutene 1.60 Argus DeWitt Butadiene (2014)
 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335 331

Table 2 IB utilization factor

NRTL model parameters. moles of IB converted to DE þ TE in the product stream ð18Þ
¼
Component Parameter, K Ref.
moles of IB fed
ð68Þ
i j Aij Aji αij

G ME 612.7 -31.4 0.2 (Liu et al., 2014) 4.2. Constraints

G DE 1851.6 417.4 0.2 (Liu et al., 2014)
G TE 2279.0 1172.5 0.2 (Liu et al., 2014) In addition to the process model equations, the following
ME DE -554.0 1032.9 0.2 (Liu et al., 2014)
constraints are also imposed. The reaction temperature is bounded
ME TE 930.7 -159.7 0.2 (Liu et al., 2014)
DE TE 1195.3 -448.5 0.2 (Liu et al., 2014) to ensure the validity of the kinetic model (Liu et al., 2013),
IB G 937.02 721.75 0.2 (Behr and Obendorf, 2002) 343K r T r 373K. The outlet temperature of the heat exchangers
IB ME 1229.90 -310.24 0.2 (Behr and Obendorf, 2002) using cooling water as cooling medium should be larger than
IB DE -465.14 -742.47 0.2 (Behr and Obendorf, 2002)
310 K, to ensure a driving force of 5 K at least (the inlet tem-
IB TE -465.14 -742.47 0.2 (Behr and Obendorf, 2002)
IB DIB -88.42 56.87 0.329 (Di Serio et al., 2010)
perature of cooling water is 305 K). The glycerol content in the
DIB G 2011.19 1215.89 0.2 (Di Serio et al., 2010) product must be less than 0.2 wt% (Melero et al., 2010; 14214 ESE,
DIB ME 1248.24 260.27 0.275 (Di Serio et al., 2010) 2008.). The water content in the product should be less than
DIB DE 1000.56 -224.58 0.286 (Di Serio et al., 2010) 500 mg/kg (0.05 wt%) (14214 ESE, 2008.). IB, ME and DIB are
DIB TE 746.9 -377.8 0.398 (Di Serio et al., 2010)
soluble in the desired product and acceptable in fuels. We specify
Water G 258.11 -274.35 1.011 (Di Serio et al., 2010)
Water IB 1419.19 1401.79 0.255 (Di Serio et al., 2010) the lower boundary of the product purity as 97 wt%, as there is no
Water ME 1613.06 -415.84 0.392 (Di Serio et al., 2010) restriction for the content of IB, ME and DIB in the product in open
Water DE 2499.9 -20.32 0.345 (Di Serio et al., 2010) literature. The byproduct DIB could be added to gasoline as an
Water TE 2499.99 -46.44 0.202 (Di Serio et al., 2010)
octane enhancer, directly or after subsequent hydrogenation
Water DIB 2499.99 1200.53 0.2 (Di Serio et al., 2010)
(Honkela and Krause, 2004; Goortani et al., 2015; Sarkar et al.,
2014; Fehér et al., 2012). The concentration of DIB in the side draw
Table 3 stream of D1 is set to be larger than 95 wt%.
Values of decision variables used in the base case simulation.

Unit Decision variable Value

5. Results and discussion
R Volume, m3 25.00
Temperature, K 353 5.1. Base case simulation
H1 Outlet temperature, K 313
E1 Number of trays 8
The basis for the design and optimization is the conversion of
E2 Number of trays 4
D1 Number of trays 8 30 kmol/h glycerol to glycerol ethers. Before optimization, the
Feed tray 6 process was simulated with the values of decision variables pre-
Side stream tray 4 sented in Table 3.
Molar reflux ratio 0.50 The composition of streams in the base case is reported in
Flowrate of side stream, kmol/h 11.00
Reboiler duty, MJ/h 8500.00
Table 4. The molar ratio of isobutene to glycerol fed to the process
F1 Temperature, K 323 is 2.83 (85/30), stoichiometrically determined by the composition
D2 Number of trays 6 of the product. However this ratio reaches 8.36 in the feed stream
Feed tray 5 of the reactor (stream 5). The enrichment of IB before the reactor is
Molar reflux ratio 0.40
accomplished by the recycles, and allows achieving high single-
Reboiler duty, MJ/h 2700.00
Global Feed rate of IB, kmol/h 85.00 pass conversion of glycerol and selectivity of high-ethers, which
Feed rate of water, kmol/h 10.00 are 92.17% and 95.69% respectively. The extraction column E1
extracts 84.49% of ME in the reaction product. All residual glycerol
is recovered in E2.
4. Optimization formulation As can be seen in Table 5, the concentrations of the product and
by-product are close to the corresponding specifications. The gly-
4.1. Objective function cerol conversion of the process is 100%, and 99.89% of the pro-
duced glycerol ethers is obtained as the product (stream 18). The
The objective function to be minimized is defined as the cost yield of high-ethers is 95.57%. The IB utilization factor is 71.84%,
per kg of the desired product produced through the process, as which means almost one third of the fed isobutene is converted to
shown below. A payback period of 8 years is considered. DIB or lost in the waste. The production cost of high-ethers is 1.477
Material ½$=yr þ OM ½$=yr þ Utility ½$=yr þ Equipment ½$ $/kg in the base case simulation. As shown in Fig. 5, the cost is
payback period ½yr
min cost ½$=kg ¼ greatly dominated by raw material cost, especially isobutene.
DE þ TE production ½kg=yr
ð65Þ
5.2. Optimization
Other indices including conversion and yield are defined below
to characterize the performance of the process: The mixed-integer nonlinear programming (MINLP) optimiza-
tion problem described in Section 4 was solved using the OAERAP
moles of glycerol converted solver in gPROMS with initial values of decision variables taken
glycerol conversion ¼ ð66Þ
moles of glycerol fed from the base case simulation (Table 3). Table 6 reports the opti-
mal values of the decision variables, leading to a minimum cost of
moles of DE þ TE in the product stream ð18Þ 1.201 $/kg, of which the distribution is shown in Fig. 6. The flow-
DE þ TE yield ¼
moles of glycerol fed rates and compositions of streams under this optimal case are
ð67Þ summarized in Table 7.
332 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335

Table 4
Composition of streams in the base case.

Stream no. 1 2 3 4 5 6 7 8 9 10 11

Temperature, K 298 298 324 324 353 353 313 313 311 323 323
Flowrate, kmol/h 85 30 313.84 313.84 313.84 241.13 241.13 43.42 227.71 41.67 1.75
Mole fraction % G 0.00 100.00 9.83 9.83 9.83 0.35 0.35 64.68 1.21 67.40 0.00
IB 100.00 0.00 82.12 82.12 82.12 72.46 72.46 4.94 75.79 1.07 97.21
ME 0.00 0.00 2.93 2.93 2.93 4.35 4.35 20.41 0.71 21.27 0.00
DE 0.00 0.00 0.00 0.00 0.00 10.36 10.36 0.07 10.96 0.08 0.00
TE 0.00 0.00 0.00 0.00 0.00 1.55 1.55 0.00 1.64 0.00 0.00
DIB 0.00 0.00 3.79 3.79 3.79 9.19 9.19 0.21 9.70 0.13 2.01
Water 0.00 0.00 1.34 1.34 1.34 1.74 1.74 9.68 0.00 10.06 0.77

Stream no. 12 13 14 15 16 17 18 19 20 21
Temperature, K 266 298 311 311 268 337 491 322 371 534
Flowrate, kmol/h 1.75 10.00 9.09 228.63 186.93 11.00 30.69 50.75 10.60 40.15
Mole fraction % G 0.00 0.00 30.32 0.00 0.00 0.00 0.00 60.76 0.00 76.80
IB 97.21 0.00 0.81 75.45 91.48 9.35 1.50 1.02 4.89 0.00
ME 0.00 0.00 3.64 0.57 0.00 0.00 4.23 18.11 0.02 22.89
DE 0.00 0.00 0.00 10.91 0.00 0.02 81.28 0.06 0.25 0.01
TE 0.00 0.00 0.00 1.63 0.00 0.00 12.14 0.00 0.00 0.00
DIB 2.01 0.00 0.01 9.66 6.34 90.59 0.86 0.11 0.52 0.00
Water 0.77 100.00 65.22 1.78 2.18 0.04 0.00 19.93 94.31 0.30

Table 6
Table 5 Optimized values of decision variables.
Summary of main results in the base case.
Unit Decision variable Value
Item Value
3
R Volume, m 28.03
Glycerol conversion 1.0000 Temperature, K 343
DE þ TE yield 0.9557 H1 Outlet temperature, K 310
IB utilization factor 0.7184 E1 Number of trays 8
Production cost, $/kg 1.477 E2 Number of trays 4
Production capacity, 1E6 kg/yr 51.45 D1 Number of trays 9
Product DE þ TE content, wt% 96.08 Feed tray 6
Glycerol content, wt% 1.40E  03 Side stream tray 4
Water content, wt% 2.44E  05 Molar reflux ratio 0.06
By-product DIB content, wt% 95.05 Flowrate of side stream, kmol/h 0.53
Reboiler duty, MJ/h 9795.20
F1 Temperature, K 391
D2 Number of trays 6
Feed tray 5
Molar reflux ratio 0.83
Reboiler duty, MJ/h 3386.20
global Feed rate of IB, kmol/h 64.80
Feed rate of water, kmol/h 11.52

Fig. 5. Production cost [$/kg] break down in the base case.

Regarding the values of decision variables, here are some points

worth pointing out when compared with the base case. The
reaction temperature falls to the lower bound, which could help in
inhibiting the dimerization of isobutene and increasing the
selectivity of high-ethers according to the reaction kinetics (Liu Fig. 6. Production cost [$/kg] break down in the optimal case.

et al., 2013). To compensate for the decrease of the reaction rate

with temperature, the volume of the reactor increases. The outlet solubility decreases with temperature. The tray numbers, feed
temperature of the heat exchanger H1 determines the operating positions and side stream position do not change much in the
temperature of the extraction columns E1 and E2. Low tempera- optimization. The operating temperature of the flash tank for
ture is preferred to improve the efficiency of extraction, for mutual isobutene recovery is much higher than the boiling temperature of
 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335 333

Table 7
Composition of streams in the optimal case.

Stream no. 1 2 3 4 5 6 7 8 9 10 11

Temperature, K 298 298 335 335 343 343 310 310 309 391 391
Flowrate, kmol/h 64.80 30.00 398.33 398.33 398.33 335.06 335.06 82.35 282.72 72.39 9.96
Mole fraction % G 0.00 100.00 9.11 9.11 9.11 1.88 1.88 41.41 0.77 47.09 0.11
IB 100.00 0.00 62.02 62.02 62.02 54.58 54.58 10.00 61.78 0.55 78.66
ME 0.00 0.00 8.50 8.50 8.50 10.43 10.43 40.31 0.62 45.81 0.30
DE 0.00 0.00 0.09 0.09 0.09 7.71 7.71 0.48 9.00 0.54 0.07
TE 0.00 0.00 0.00 0.00 0.00 1.03 1.03 0.01 1.22 0.00 0.01
DIB 0.00 0.00 19.02 19.02 19.02 22.88 22.88 1.75 26.61 0.32 12.13
Water 0.00 0.00 1.25 1.25 1.25 1.49 1.49 6.06 0.00 5.69 8.72

Stream no. 12 13 14 15 16 17 18 19 20 21
Temperature, K 266 298 309 309 276 342 498 385 371 522
Flowrate, kmol/h 9.96 11.52 10.76 283.48 252.68 0.53 30.27 83.14 12.26 70.88
Mole fraction % G 0.11 0.00 20.35 0.00 0.00 0.00 0.00 43.63 0.00 51.18
IB 78.66 0.00 0.68 61.59 69.03 7.94 0.41 0.56 3.82 0.00
ME 0.30 0.00 6.18 0.38 0.00 0.01 3.58 40.68 0.00 47.72
DE 0.07 0.00 0.02 8.97 0.00 0.39 84.02 0.47 0.30 0.50
TE 0.01 0.00 0.00 1.21 0.00 0.03 11.38 0.00 0.03 0.00
DIB 12.13 0.00 0.03 26.54 29.51 91.60 0.61 0.28 1.89 0.00
Water 8.72 100.00 72.74 1.30 1.46 0.03 0.00 14.37 93.98 0.60

Table 8
Summary of main results in the optimal case.

Item Value

Glycerol conversion 1.0000

DE þ TE yield 0.9625
IB utilization factor 0.9443
Production cost, $/kg 1.201
Production capacity, 1E6 kg/yr 51.17
Product DE þTE content, wt% 97.00
Glycerol content, wt% 1.39E  04
Water content, wt% 2.37E  07
By-product DIB content, wt% 95.05

isobutene (266 K). This avoids the waste of isobutene. 95.20% of

the isobutene entering the flash tank is recovered, while the per-
centage is only 79.24% in the base case. The heat duty of the two
distillation columns increases. Those changes ensure that the
proposed process operates under a low feed rate of isobutene. As a
result, the molar ratio of isobutene to glycerol in the feed of the
whole process decreases to 2.16 (64.80/30). We can conclude that
the optimizing direction is to decrease the consumption of the
expensive isobutene. It should be noted that although this is
consistent with intuition, it is virtually impossible to find these
optimal decision variables manually without the systematic plant-
wide optimization described here.
The same conclusion that the optimal design tends to utilize
isobutene efficiently can also be drawn by viewing the stream
results presented in Table 7. The isobutene/glycerol molar ratio in
the feed of the reactor decreases to 6.81, helping in reducing the
amount of isobutene dimers. More intermediate, ME, accumulates
in the reactor, which promotes the formation of high-ethers. As a
result, the single-pass selectivity of high-ethers in the reactor
increases to 97.60%, with a glycerol conversion of 90.71%. More DE
and less TE are obtained than in the base case. That means less
isobutene is consumed, as more isobutene is required to produce
TE than DE. The main impurity of the product is ME, which is
originally the residual of the extraction column E1. It is difficult to
further decrease the concentration of ME in the product with the
setup of the process units. But the specified product purity,
97 wt%, is high enough to provide excellent blending properties
with biodiesel and gasoline (Melero et al., 2010).
Fig. 7. Minimum production cost of high-ethers of glycerol as a function of the
glycerol and isobutene price.
Table 8 summarizes the main results of the optimal case. The IB
utilization factor increases to 94.43%. The concentration of the
product and byproduct satisfies the corresponding specifications.
The production cost of high-ethers is 1.201 $/kg, and still domi-
nated by the cost of isobutene as shown in Fig. 6.
The optimal production cost of high-ethers, 1.201 $/kg is a little
higher than the price of biodiesel (1.00 $/kg, Argus Biofuels, 2014)
and MTBE (1.15 $/kg, Argus DeWitt Fuels and Octane, 2014). As can
be seen in both the base and optimal case, the cost of equipment,
utility and operation and maintenance is about 0.100 $/kg. These
parts of the cost are very small and have little room for further
optimization. The key factor determining the cost of glycerol
ethers is the price of isobutene and glycerol.
The sensitivity of the optimal solution to different initial values
of decision variables was investigated by using random initial
guesses around the ones used for the solution described above. All
integer decision variables like number of trays and feed tray were
fixed to the values shown in Table 6, to simplify the problem. This
simplification is reasonable as those integer variables only influ-
ence the costs of the equipment and utility, which are small
compared to the cost of materials. The resulting optimal values of
decision variables were very close to the ones reported previously
(Table 6), with the production cost being practically the same.
334 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335

5.3. Effect of price of raw materials on the production cost Ps vapor pressure, Pa
1
QE ion exchange capacity of the catalyst, mol H þ kg
Prices of isobutene and glycerol were varied to further explore Q input energy
1
the economic feasibility of the process. Values of integer decision r reaction rate, mol h ðmol H þ Þ1
1
variables were kept constant (the same as those in Table 6) in the R molar flowrate of raffinate, mol h
optimization with different raw material prices. Fig. 7 shows the RR reflux ratio of distillation column in molar
1
changes of minimum production costs with the prices of raw S molar flowrate of solvent, mol h
materials, which drop almost linearly. The reason is that the T absolute temperature, K
material cost still dominates the production cost in the investi- v average velocity of liquid in the pump, m s1
1
gated range of material prices. The values of decision variables V molar flowrate of vapor, mol h
obtained from the optimization do not change much compared to V0 volume, m 3

those reported before. w position indicator for side draw stream of distil-
If we assume that a competitive production cost of glycerol lation column
ethers is 1.00 $/kg, close to the price of biodiesel, the proposed W molar flowrate of side draw stream of distillation col-
1
process would be economically promising when the price of iso- umn, mol h
butene is less than 1.2 $/kg. x molar fraction of liquid phase
y molar fraction of vapor phase
z molar fraction of feed
6. Conclusion
Greek letters
In this paper, we designed a process producing glycerol ethers
from glycerol and isobutene. To obtain high-ethers of glycerol, ME α non-randomness parameter of the NRTL model
is extracted by fresh glycerol, residual glycerol is washed by water, γ activity coefficient
and water, DIB and isobutene are separated by distillation. A ρ density, kg m  3
mathematical model is built in gPROMS to simulate the process.
The production cost including equipment, utilities, operating and Subscripts
maintenance, and raw materials is calculated. Decision variables of
the process are optimized to find the minimum production cost of C condenser
high-ethers, which is 1.201 $/kg. The price of raw materials, E extract
especially that of isobutene largely determines the economic fea- F feed
sibility of this etherification process. The optimal process design i component index
achieves high efficiency utilization of the expensive isobutene. j tray index; alternative component index; reaction index
Thus future efforts should be directed towards improving in inlet stream
the selectivity of DE and inhibiting the side reaction of k alternative tray index; stream index
isobutene. out outlet stream
R reboiler; raffinate

Superscripts
Nomenclature
E extract phase
a indicator of distillation column tray L liquid phase
A binary molecular energy interaction parameter of NRTL R raffinate phase
model, K V vapor phase
A0 heat-transfer area
b indicator for the bottom of distillation column Abbreviations
B molar flowrate of bottom product of distillation column,
1
mol h DE di-tert-butyl ethers of glycerol
D molar flowrate of distillate of distillation column, DIB dimer of isobutene
1
mol h ETBE ethyl-tert-butyl ether
e location indicator for extract stream of extraction column FAME fatty acid methyl ester
1
E molar flowrate of extracts, mol h FAEE fatty acid ethyl ester
f location indicator for feed tray of distillation column G glycerol
1
F molar flowrate, mol h IB isobutene
2
g gravity constant, m s ME mono-tert-butyl ethers of glycerol
1
H enthalpy, J mol MTBE methyl-tert-butyl ether
H0 pump head TE tri-tert-butyl ether of glycerol
1
L molar flowrate of liquid, mol h yr year
NC number of components
NF tray index of feed stream
NR number of reactions
NS number of steams
NT number of trays Acknowledgments
NTT number of total trays
NW tray index of side draw steam Financial support from the China Scholarship Council (CSC)
P pressure, Pa (Grant No. 201406280135) is gratefully acknowledged.
 J. Liu et al. / Chemical Engineering Science 144 (2016) 326–335
Appendix A. Supplementary material
Supplementary data associated with this article can be found in
the online version at http://dx.doi.org/10.1016/j.ces.2016.01.055.
References
Zhou, C.H., Beltramini, J.N., Fan, Y.X., Lu, G.Q., 2008. Chemoselective catalytic con-
version of glycerol as a biorenewable source to valuable commodity chemicals.
Chem. Soc. Rev. 37, 527–549.
Zheng, Y., Chen, X., Shen, Y., 2008. Commodity chemicals derived from glycerol, an
important biorefinery feedstock. Chem. Rev. 108 (12), 5253–5277.
Izquierdo, J.F., Montiel, M., Palés, I., et al., 2012. Fuel additives from glycerol
etherification with light olefins: state of the art. Renew. Sustain. Energy Rev. 16
(9), 6717–6724.
Behr, A., Eilting, J., Irawadi, K., Leschinski, J., Lindner, F., 2008. Improved utilisation
of renewable resources: New important derivatives of glycerol. Green Chem. 10
(1), 13–30.
Ayoub, M., Abdullah, A.Z., 2012. Critical review on the current scenario and sig-
nificance of crude glycerol resulting from biodiesel industry towards more
sustainable renewable energy industry. Renew. Sustain. Energy Rev. 16 (5),
2671–2686.
Tan, H.W., Abdul Aziz, A.R., Aroua, M.K., 2013. Glycerol production and its appli-
cations as a raw material: A review. Renew. Sustain. Energy Rev. 27, 118–127.
Quispe, C.A.G., Coronado, C.J.R., Carvalho Jr, J.A., 2013. Glycerol: production, con-
sumption, prices, characterization and new trends in combustion. Renew.
Sustain. Energy Rev. 27, 475–493.
Kesling Jr. H.S., Karas L.J., Liotta Jr. FJ., 1994. Diesel fuel. US patent 5308365.
Kesling Jr. HS, Karas L.J., Liotta Jr. F.J., 1995. Diesel fuel. EP 0641854A1.
Frusteri, F., Cannilla, C., Bonura, G., et al., 2013. Glycerol ethers production and
engine performance with diesel/ethers blend. Top Catal. 56 (1-8), 378–383.
Beatrice, C., Di Blasio, G., Lazzaro, M., et al., 2013. Technologies for energetic
exploitation of biodiesel chain derived glycerol: Oxy-fuels production by cat-
alytic conversion. Appl. Energy 102, 63–71.
Asdrubali, F., Cotana, F., Rossi, F., Presciutti, A., Rotili, A., Guattari, C., 2015. Life cycle
assessment of new oxy-fuels from biodiesel-derived glycerol. Energies. 8 (3),
1628–1643.
Melero, J.A., Vicente, G., Morales, G., Paniagua, M., Bustamante, J., 2010. Oxygenated
compounds derived from glycerol for biodiesel formulation: Influence on EN
14214 quality parameters. Fuel 89 (8), 2011–2018.
Karinen, R.S., Krause, A.O.I., 2006. New biocomponents from glycerol. Appl. Catal. A
306, 128–133.
Chang, J.S., Zhang, Y.C., Chen, C.C., et al., 2014. One-step synthesis of gasoline octane
booster and diesel fuel from glycerol and tert-butyl alcohol. Ind. Eng. Chem. Res.
53 (13), 5398–5405.
Ozbay, N., Oktar, N., Dogu, G., Dogu, T., 2012. Effects of sorption enhancement and
isobutene formation on etherification of glycerol with tert-butyl alcohol in a
flow reactor. Ind. Eng. Chem. Res. 51 (26), 8788–8795.
Celdeira, P.A., Gonçalves, M., Figueiredo, F.C.A., Bosco, S.M.D., Mandelli, D., Carvalho,
W.A., 2014. Sulfonated niobia and pillared clay as catalysts in etherification
reaction of glycerol. Appl. Catal. A 478, 98–106.
González, M.D., Salagre, P., Taboada, E., Llorca, J., Molins, E., Cesteros, Y., 2013.
Sulfonic acid-functionalized aerogels as high resistant to deactivation catalysts
for the etherification of glycerol with isobutene. Appl. Catal. B 136-137,
287–293.
González, M.D., Salagre, P., Mokaya, R., Cesteros, Y., 2014. Tuning the acidic and
textural properties of ordered mesoporous silicas for their application as cat-
alysts in the etherification of glycerol with isobutene. Catal. Today 227,
171–178.
González, M.D., Salagre, P., Linares, M., García, R., Serrano, D., Cesteros, Y., 2014.
Effect of hierarchical porosity and fluorination on the catalytic properties of
zeolite beta for glycerol etherification. Appl. Catal. A 473, 75–82.
335
Zhao, W., Yi, C., Yang, B., Hu, J., Huang, X., 2013. Etherification of glycerol and iso-
butylene catalyzed over rare earth modified Hβ-zeolite. Fuel Process Technol.
112, 70–75.
González, M.D., Salagre, P., Taboada, E., Llorca, J., Cesteros, Y., 2013. Microwave-
assisted synthesis of sulfonic acid-functionalized microporous materials for the
catalytic etherification of glycerol with isobutene. Green Chem. 15 (8),
2230–2239.
González, M.D., Cesteros, Y., Llorca, J., Salagre, P., 2012. Boosted selectivity toward
high glycerol tertiary butyl ethers by microwave-assisted sulfonic acid-
functionalization of SBA-15 and beta zeolite. J. Catal. 290, 202–209.
Frusteri, F., Frusteri, L., Cannilla, C., Bonura, G., 2012. Catalytic etherification of
glycerol to produce biofuels over novel spherical silica supported Hyflon
(R) catalysts. Bioresour. Technol. 118, 350–358.
Zhou, J., Wang, Y., Guo, X., Mao, J., Zhang, S., 2014. Etherification of glycerol with
isobutene on sulfonated graphene: reaction and separation. Green Chem. 16
(11), 4669–4679.
Singh, J., Kumar, J., Negi, M.S., Bangwal, D., Kaul, S., Garg, M.O., 2015. Kinetics and
modeling study on etherification of glycerol using isobutylene by in situ pro-
duction from tert-butyl alcohol. Ind. Eng. Chem. Res. 54 (19), 5213–5219.
Vijai P.G., Berwyn P., 1995. Glycerine ditertiary butyl ether preparation. US patent
5476971.
Behr, A., Obendorf, L., 2002. Development of a process for the acid-catalyzed
etherification of glycerine and isobutene forming glycerine tertiary butyl
ethers. Eng. Life Sci. 2 (7), 185–189.
Di Serio, M., Casale, L., Tesser, R., Santacesaria, E., 2010. New process for the pro-
duction of glycerol tert-butyl ethers. Energy Fuels 24 (9), 4668–4672.
Cheng, J.K., Lee, C.-L., Jhuang, Y.-T., Ward, J.D., Chien, I.-L., 2011. Design and control
of the glycerol tertiary butyl ethers process for the utilization of a renewable
resource. Ind. Eng. Chem. Res. 50, 12706–12716.
Vlad, E., Bildea, C.S., Bozga, G., 2013. Robust optimal design of an glycerol ether-
ification process. Chem. Eng. Technol. 36 (2), 251–258.
Martín, M., Grossmann, I.E., 2014. Simultaneous optimization and heat integration
for the coproduction of diesel substitutes: biodiesel (FAME and FAEE) and
glycerol ethers from algae oil. Ind. Eng. Chem. Res. 53 (28), 11371–11383.
Martín, M., Grossmann, I.E., 2014. Optimal simultaneous production of i-butene and
ethanol from switchgrass. Biomass Bioenergy. 61, 93–103.
de la Cruz, V., Hernández, S., Martín, M., Grossmann, I.E., 2014. Integrated synthesis
of biodiesel, bioethanol, isobutene, and glycerol ethers from algae. Ind. Eng.
Chem. Res. 53 (37), 14397–14407.
Liu, J., Yang, B., Yi, C., 2013. Kinetic study of glycerol etherification with isobutene.
Ind. Eng. Chem. Res. 52, 3742–3751.
Liu, J., Yuan, Y., Pan, Y., Huang, Z., Yang, B., 2014. Liquid–liquid equilibrium for
systems of glycerol and glycerol tert-butyl ethers. Fluid Phase Equilibr. 365,
50–57.
Seider, W.D., Seader, J.D., Lewin, D.R., 2009. Product & process design principles:
synthesis, analysis and evaluation, 3rd edition. John Wiley & Sons.
Argus Biofuels, 2014. 〈〈https://www.argusmedia.com/  /media/Files/PDFs/Samples/
Argus-Biofuels.pdf?la ¼ en〉〉 (accessed 11.06.15.).
Argus DeWitt Butadiene, 2014. 〈〈http://www.argusmedia.com/  /media/Files/PDFs/
PetChems/Argus-DeWitt-Butadiene.pdf?la ¼ en〉〉 (accessed 11.06.15.).
14214 ESE, 2008. Automotive fuels. Fatty acid methyl esters (FAME) for diesel
engines requirements and test methods.
Honkela, M.L., Krause, A.O.I., 2004. Kinetic modeling of the dimerization of iso-
butene. Ind. Eng. Chem. Res. 43 (13), 3251–3260.
Goortani, B.M., Gaurav, A., Deshpande, A., Ng, F.T.T., Rempel, G.L., 2015. Production
of isooctane from isobutene: energy integration and carbon dioxide abatement
via catalytic distillation. Ind. Eng. Chem. Res. 54 (14), 3570–3581.
Sarkar, A., Seth, D., Ng, F.T.T., Rempel, G.L., 2014. Selective oligomerization of iso-
butene on lewis acid catalyst: kinetic modeling. Ind. Eng. Chem. Res. 53 (49),
18982–18992.
Fehér, C., Kriván, E., Hancsók, J., Skoda-Földes, R., 2012. Oligomerisation of iso-
butene with silica supported ionic liquid catalysts. Green Chem. 14 (2),
403–409.
Argus DeWitt Fuels and Octane, 2014. 〈〈http://www.argusmedia.com/Petrochemicals/
 /media/91FBBFB96B3543DEAC2F145E299A6598.ashx〉〉 (accessed 11.06.15.).