Supporting Information for “Green Propylene and Polypropylene Production from

Glycerol: Process Simulation and Economic Evaluation”

Felipe Blanco Guerra1, Raquel Massad Cavalcante1 and André Ferreira Young2*

1
Escola de Química, Universidade Federal do Rio de Janeiro, Av. Athos da Silveira

Ramos, 149, Rio de Janeiro, Brasil.

2
Departamento de Engenharia Química e de Petróleo, Universidade Federal Fluminense,

Rua Passo da Pátria, 156, Niterói, Brasil.

*
[email protected]

CONTENTS

This document presents:

- 34 pages;

- 3 figures;

- 11 tables.

S1
POLYMERIZATION PROCESS SIMULATION

The Aspen Polymers® tool allows the simulator to identify selected components

as polymers so that specific properties and their particle size distribution can be calculated

during simulation.

Polymer characterization

Aspen Polymers® offers recommended attribute sets for different types of

polymerizations. In this case, the Ziegler-Natta type was selected and the attributes

recommended by the tool were used. The attributes have the function of characterizing

the polymer by means of properties such as degree of polymerization, molecular weight

and composition.

There are three different classes of attributes. Class 0 composes the attributes

whose value is calculated from other attributes. Class 1 are structural properties per unit

of mass, however they are not used for polymers. Class 2 are structural properties per unit

of time. Class 2 attributes are zero-order (ZMOM) and two-order (SMOM) moments, and

segment molar flow (SFLOW). Moments are calculated from population balances applied

to the kinetic model and, by definition, can be calculated by Equation S1:

𝜆 = ∑ 𝑛𝑚 𝑄𝑛 (S1)

where 𝜆 is the moment value, m is the moment order, n is the size of chain or

polymerization degree, and Qn is the number of mols of polymer with size n.

The main class 0 attributes to be calculated are the numeral and weight average

degrees of polymerization (DPN and DPW), or the numeral and weight average molecular

masses (MMN and MMW), and the polydispersity index (PDI). They characterize the

S2
average molecular size or mass and the degree of dispersion around it, respectively. They

can be calculated directly from the moment values as described in Equations S2 to S6:

𝜆
𝐷𝑃𝑛 = 𝜆1 (S2)
0

𝜆
𝐷𝑃𝑤 = 𝜆2 (S3)
1

𝑀𝑛 = 𝐷𝑃𝑛 𝑀𝑠𝑒𝑔 (S4)

𝑀𝑤 = 𝐷𝑃𝑤 𝑀𝑠𝑒𝑔 (S5)

𝐷𝑃𝑤
𝑃𝐷𝐼 = (S6)
𝐷𝑃𝑛

where Mseg is the molecular mass of the segment, witch for a homopolymer is the

molecular mass of the monomer. The attributes can be calculated for the total polymers

and for the fraction of “living” polymers. In addition, they can be calculated for the overall

composition and for the composition of each site type.

Within the Aspen Polymers® environment, it must also be specified which

component is the catalyst and its type. In this case, the catalyst used was TiCl₄, of the

Ziegler-Natta type and, although the activator TEAl was not included in the simulation,

it was considered in the economic analysis. This type of catalyst has more than one type

of site by nature, and each of them produces a polymer chain at a different rate, generating

polymers with different sizes and characteristics.1 The number of sites must be selected

according to the kinetic modeling to be used, so only one site was selected, according to

the base reference.2 When choosing the type of catalyst, attributes are automatically

selected to define its properties, such as the fraction and flow rate of potential (CPS),

vacant (CVS), inhibited (CIS) or deactivated (CDS) catalyst sites. In this work, the

S3
attributes related to inhibited sites were removed, since the reaction system does not

involve any inhibitor.

Another important parameter is the concentration of sites defined as kmols of sites

/ kg of catalyst. This parameter directly impacts the catalyst productivity, so it was used

to adjust the polymer production in the simulation based on the reference being followed.3

Zheng et al.1 proposed that the concentration of catalyst sites could vary from 10-3 to 10-
5
kmol/kg, allowing to vary the polymer production without changing its composition and

molecular weight. However, in the work by Kim et al.2 the value of this parameter was

not provided and, to reach the productivity obtained in this reference, the value of 1.85 x

10-2 kmol/kg was calculated, which is greater than the upper limit of the suggested range.

Particle size distribution

To model the particle size distribution, Aspen Plus® v8.8 uses two approaches

simultaneously. One of them is the method of moments, which can calculate average

properties of the polymer for the streams with a certain simplicity of calculation but does

not produce the complete distribution function. The second approach is the instantaneous

properties method, which, despite demanding greater computational effort and storage of

reaction rate data, manages to produce the complete distribution with greater detail.

The function used by the instantaneous properties method to obtain the particle

size distribution is the Schultz-Flory Most Probable Distribution function (Equations S7

and S8). This function only needs one parameter (p), which is the probability of a “live”

polymer chain propagating (Equation S9). This parameter is obtained in the simulation as

a result of calculating the ratio between local rates of transfer reaction and chain

propagation of a “living” polymer (𝜏) (Equation S10).

S4
𝑊(𝑟) = 𝑝𝑟−1 (1 − 𝑝) (S7)

(numerical distribution)

𝑊(𝑟) = 𝑟𝑝𝑟−1 (1 − 𝑝)2 (S8)

(mass distribution)

1
𝑝 = (𝜏+1) (S9)

(𝑅𝑡𝑑 +𝑅𝑓 )
𝜏= (S10)
𝑅𝑝

In these equations, Rtd is the rate of termination by disproportionation, Rf is the

sum of all rates of chain transfer and Rp is the rate of propagation.

The other properties of the polymer can be calculated as a function of p (Equations

S11, S12 and S13). It is worth noting that the distribution function can be calculated in

an analogous way for both chain size (r) and molecular mass (MW).

1
𝐷𝑃𝑛 = (1−𝑝) (S11)

(1+𝑝)
𝐷𝑃𝑤 = (1−𝑝) (S12)

𝑃𝐷𝐼 = 1 + 𝑝 (S13)

Looking at Equation S13, by the nature of parameter p, the maximum value of

PDI will be 2 for a Schultz-Flory distribution. However, for cases where the

polymerization kinetics involves different catalytic sites, such as for the Ziegler-Natta

type, this value may be higher. This is because the particle size distribution is calculated

for each site and, later, they are all superimposed to form a single curve using Equation

S14:

S5
 𝑚1 𝑚2
𝑊𝑜𝑢𝑡 = 𝑊1 (𝑟) + 𝑊2 (𝑟) (S14)
𝑚 𝑚

where 𝑚 = 𝑚1 + 𝑚2 is the total mass of polymer produced by each kind of site in a

reactor.

Equation S14 also applies in the case where more than one CSTR reactor is used

in series. Aspen Plus® v8.8 calculates the distribution of polymer being formed locally in

each reactor according to the reactor's operating conditions. However, for the process

streams after the reactors, the distributions obtained in each reactor are cumulatively

superimposed. The same occurs in the case of a PFR reactor, which can be approximated

by multiple CSTR reactors.

Validation of the polymerization kinetics

Before starting to simulate the process at the desired scale, it was necessary to

validate the polymerization kinetics presented in the previous sections, with the same

conditions of feed and scale as the reference. In this step, the loop reactor model used was

the one proposed by Kim et al.,2 represented in the flowsheet of Figure S1, which contains

only the reaction system necessary for validation.

Fig. S1. Polymerization loop reactors modeled by PFRs with internal recycle for

validation of the polymerization kinetics

S6
 The results obtained for the mass flow rate of produced polymer (W), number

average molecular weight (MWn), weight average molecular weight (MWw) and

polydispersity index (PDI) for each reactor are compared with the values reported in the

reference in Table S1. These attributes characterize the average molecular size or mass

and the degree of dispersion around it, respectively.

Table S1. Validation of the polymerization kinetics

Parameters Simulation Reference2

R1 R2 R3 R1 R2 R3
W (kg/h) 403.86 16789.93 27071.95 335.77 16860.42 27181.42
PDI 1.91 2.00 2.02 1.98 1.93 1.92
MWn 24,182 131,891 136,024 27,000 205,746 215,053
MWw 46,303 263,856 274,489 51,000 398,085 412,558

It is observed that the simulation results, despite not being identical to the

reference, are close enough to demonstrate that the model is able to express well the

behavior of the polymerization kinetics. The main differences are in the properties of the

polymer, since the polypropylene obtained in the simulation had a slightly smaller chain

size than expected. These disagreements were probably generated by divergences in

relation to the thermodynamic model used, which was not specified in the work by Kim

et al.,2 and the calculation method used in the Aspen Plus® v8.8 simulator compared to

gPROMS, used in the reference.

Because the PFR reactor model with recycle requires a high computational effort

to converge, the same approximation performed by Zheng et al.1 was assumed in the

complete simulation. The authors showed that the loop reactor can be represented, for

simulation purposes, by a CSTR when using a high recycle ratio. The approach was tested

in the simulator, replacing each PFR system by a CSTR and the results were quite similar,

S7
given the high recycle ratios in the reactors in this work (75 for the pre-reactor and 153

for the main reactors).

Design equations and parameters of the steamer

The mathematical relationship used to calculate the steam flow rate necessary for

the catalyst deactivation to be successful in the steamer is showed in Equation S15.4 It is

a function of the polymer inlet flow, the polymer temperature difference between the inlet

and outlet streams of the equipment, and a constant that depends on the physicochemical

properties of the polymer and of the steam under operating conditions (Equation S16).

𝐹𝑠𝑡𝑒𝑎𝑚 = 𝐹𝑝𝑜𝑙 𝐾∆𝑇𝑖𝑛/𝑜𝑢𝑡 (S15)

𝐾 = 𝐾 ∗ 𝐾𝑒 (S16)

In these equation, K* is the specific heat capacity of the polymer ((Cp)pol) divided

by the heat of condensation of the steam (Hl(steam)), and Ke is a factor that considers the

theoretical surplus of steam needed to heat the polymer from Tin to Tout. In the patent,4 it

is recommended to use saturated steam and a temperature range of 60 to 95 °C for the

inlet stream, and 90 to 115 °C for the polymer outlet stream. The parameters used in the

calculation of the steam flow rate in the simulation are found in Table S2.

Table S2. Parameters for calculating the steam inlet flow rate in the steamer

Parameter Value Unit

Fpol 479.3233 kg/h
Tin 85 °C
Tout 105 °C
K* 0.001082 1/°C
Ke 1.5 -
(Cp)pol 0.58 kcal/kg.°C
Hl(steam) 536.2572 kcal/kg
Tsteam 105 °C
Psteam 1.2 Bar
Fsteam 15.55266 kg/h
S8
 DETAILED RESULTS FOR PROCESS STREAMS AND EQUIPMENT
OPERATING CONDITIONS

Table S3. Operating conditions of some glycerol hydrogenolysis process equipment

Equipment Flow rate (kg/h) Te (°C) Ts (°C) Pe (bar) Ps (bar) Utility

HT1 4288.39 265 300 1 1 Fuel oil

HT2 26712.79 10 35 50 50 Low pressure steam

HT3 542.54 8.7 50 21 21 Low pressure steam

COOLER1 4288.34 420 35 10 10 Water

COOLER 2 27255.33 113 10 50 50 Ammonia

COOLER 3 3037.73 35 25 1 1 Ammonia

COMP 1 4288.34 35 420 1 10 -

COMP 2 3584.37 35 280 10 50 -

VALV 542.54 10 8.6 50 21 -

S9
 Table S4. Glycerol hydrogenolysis process streams

Streams C3H8 C3H10 F1 F2 F3 F4 F5

Temperature (°C) 50.0 50.0 25.0 264.4 300.0 300.0 35.0
Pressure (bar) 25.0 21.0 1.0 1.0 1.0 1.0 1.0
Vapor fraction 0.00 1.00 0.99 1.00 1.00 1.00 1.00
Enthalpy (cal/g) 41.31 -542.69 -482.54 203.88 297.63 192.45 -493.98
Density (g/cm³) 0.460 0.028 0.000 0.000 0.000 0.000 0.000
Molar mass (g/mol) 42.07 31.38 2.77 2.77 2.77 2.75 2.75
Mass flow (kg/h) 483.73 58.81 4288.39 4288.39 4288.39 4288.34 4288.34
Mass fraction
GLYCEROL 0.00000 0.00000 0.27876 0.27876 0.27876 0.00000 0.00000
PROPYLENE 0.99606 0.05443 0.00000 0.00000 0.00000 0.11464 0.11464
PROPANE 0.00353 0.83407 0.00000 0.00000 0.00000 0.01201 0.01201
H2O 0.00000 0.00000 0.00000 0.00000 0.00000 0.16360 0.16360
H2 0.00000 0.01513 0.72124 0.72124 0.72124 0.70836 0.70836
METHANE 0.00014 0.03267 0.00000 0.00000 0.00000 0.00049 0.00049
ETHANE 0.00027 0.06370 0.00000 0.00000 0.00000 0.00091 0.00091
Streams F6 F7 F8 F9 F10 F11 F12
Temperature (°C) 419.5 35.0 35.0 35.0 279.1 113.4 10.0
Pressure (bar) 10.0 10.0 10.0 10.0 50.0 50.0 50.0
Vapor fraction 1.00 0.98 1.00 1.00 1.00 1.00 0.99
Enthalpy (cal/g) 506.61 -567.30 -103.30 35.32 776.31 51.48 -40.47
Density (g/cm³) 0.000 0.001 0.001 0.001 0.003 0.017 0.024
Molar mass (g/mol) 2.75 2.75 2.45 2.36 2.36 10.93 10.93
Mass flow (kg/h) 4288.34 4288.34 3744.58 3584.37 3584.37 27255.33 27255.33
Mass fraction
GLYCEROL 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
PROPYLENE 0.11464 0.11464 0.13073 0.13657 0.13657 0.72797 0.72797
PROPANE 0.01201 0.01201 0.01369 0.01430 0.01430 0.06830 0.06830
H2O 0.16360 0.16360 0.04278 0.00000 0.00000 0.00000 0.00000
H2 0.70836 0.70836 0.81120 0.84746 0.84746 0.13935 0.13935
METHANE 0.00049 0.00049 0.00056 0.00058 0.00058 0.04734 0.04734
ETHANE 0.00091 0.00091 0.00104 0.00109 0.00109 0.01704 0.01704

S10
 Table S4. Glycerol hydrogenolysis process streams (continuation)

Streams F13 F14 F15 F16 F17 FEED H2

Temperature (°C) 10.0 35.0 35.0 8.7 50.0 25.0 35.0
Pressure (bar) 50.0 50.0 50.0 21.0 21.0 1.0 1.0
Vapor fraction 1.00 1.00 0.76 0.04 1.00 0.68 1.00
Enthalpy (cal/g) -40.17 -18.62 -58.27 -55.13 37.64 -1654.49 34.20
Density (g/cm³) 0.023 0.021 0.066 0.333 0.044 0.002 0.000
Molar mass (g/mol) 10.77 10.77 24.32 40.57 40.57 30.98 2.02
Mass flow (kg/h) 26712.79 26712.79 23675.06 542.54 542.54 1250.67 3037.73
Mass fraction
GLYCEROL 0.00000 0.00000 0.00000 0.00000 0.00000 0.95585 0.00000
PROPYLENE 0.72460 0.72460 0.81758 0.89399 0.89399 0.00000 0.00000
PROPANE 0.06778 0.06778 0.07648 0.09356 0.09356 0.00000 0.00000
H2O 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
H2 0.14215 0.14215 0.03208 0.00164 0.00164 0.04415 1.00000
METHANE 0.04823 0.04823 0.05441 0.00366 0.00366 0.00000 0.00000
ETHANE 0.01724 0.01724 0.01945 0.00715 0.00715 0.00000 0.00000
Streams H2OA H20B HC REC
Temperature (°C) 35.0 35.0 10.0 25.0
Pressure (bar) 10.0 10.0 50.0 1.0
Vapor fraction 0.00 0.00 0.00 1.00
Enthalpy (cal/g) -3762.66 -3779.24 -55.13 -0.03
Density (g/cm³) 0.980 0.984 0.511 0.000
Molar mass (g/mol) 18.05 18.02 40.57 2.02
Mass flow (kg/h) 543.76 160.20 542.54 3037.73
Mass fraction
GLYCEROL 0.00000 0.00000 0.00000 0.00000
PROPYLENE 0.00383 0.00000 0.89399 0.00000
PROPANE 0.00049 0.00000 0.09356 0.00000
H2O 0.99557 1.00000 0.00000 0.00000
H2 0.00010 0.00000 0.00164 1.00000
METHANE 0.00000 0.00000 0.00366 0.00000
ETHANE 0.00001 0.00000 0.00715 0.00000

S11
 Table S5. Operating conditions of some propylene polymerization process

equipment

Equipment Flow rate (kg/h) Te (°C) Ts (°C) Pe (bar) Ps (bar) Utility

HT 877.64 54 70 35 35 Low pressure steam

FP 912.90 70 85 35 35 Low pressure steam

COOLER1 422.45 305 50 25 25 Water

COOLER 2 35.04 54 20 35 35
Ammonia
COMP 422.45 85 305 1 25 -

PUMP 912.68 51 54 25 35 -

S12
 Table S6. Propylene polymerization process streams

Streams C3H8 CAT F1 F2 F3 F4 F5

Temperature (°C) 50.0 10.0 20.3 70.0 70.0 85.0 304.3
Pressure (bar) 25.0 40.0 35.0 35.0 35.0 35.0 25.0
Vapor fraction 0.00 0.00 0.00 0.00 0.00 0.33 1.00
Enthalpy (cal/g) 40.31 -826.55 -159.45 -285.72 -298.07 -269.72 94.04
Density (g/cm³) 0.461 0.339 0.578 0.623 0.631 0.188 0.023
Molar mass (g/mol) 42.07 117.15 41.99 41.84 41.83 41.83 41.54
Mass flow (kg/h) 490.25 0.22 35.26 593.38 912.90 912.90 422.45
Mass fraction
TEA 0.00000 0.93610 0.00578 0.00034 0.00022 0.00022 0.00000
TICL4 0.00000 0.06390 0.00039 0.00002 0.00002 0.00002 0.00000
PP 0.00000 0.00000 0.20016 0.50362 0.52757 0.52757 0.00000
H2 0.00000 0.00000 0.00003 0.00003 0.00003 0.00003 0.00006
N2 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
H2O 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
PROPY 0.99617 0.00000 0.69080 0.39255 0.36871 0.36871 0.78082
PROPANE 0.00343 0.00000 0.09124 0.09177 0.09178 0.09178 0.19437
ETHANE 0.00026 0.00000 0.00750 0.00755 0.00755 0.00755 0.01601
METHANE 0.00014 0.00000 0.00410 0.00412 0.00412 0.00412 0.00874
TICL4 CPSFRAC 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
TICL4 CDSFRAC 0.00000 0.00001 0.00075 0.00124 0.00124 0.00124
TICL4 CVSFRAC 1.00000 0.00162 0.00014 0.00015 0.00015 0.00015
PP DPN 575 3.373 3.495 3.495 3.495
PP DPW 1.148 6.705 6.939 6.939 6.939
PP PDI 2.00 1.99 1.99 1.99 1.99
PP MWN 24,177 141,929 147,077 147,077 147,077
PP MWW 48,312 282,133 292,016 292,016 292,016
Streams F6 F7 F8 F9 F10 FEED1 FEED2
Temperature (°C) 50.0 53.7 53.7 53.7 70.0 20.0 70.0
Pressure (bar) 25.0 35.0 35.0 35.0 35.0 35.0 35.0
Vapor fraction 0.03 0.00 0.00 0.00 0.00 0.00 0.00
Enthalpy (cal/g) -109.76 -27.36 -27.36 -27.36 -13.42 -50.48 -13.42
Density (g/cm³) 0.374 0.451 0.451 0.451 0.400 0.518 0.400
Molar mass (g/mol) 41.54 41.83 41.83 41.83 41.83 41.83 41.83
Mass flow (kg/h) 422.43 912.68 35.04 877.64 877.64 35.04 558.12
Mass fraction
TEA 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
TICL4 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
PP 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
H2 0.00006 0.00003 0.00003 0.00003 0.00003 0.00003 0.00003
N2 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
H2O 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
PROPY 0.78082 0.89650 0.89650 0.89650 0.89650 0.89650 0.89650
PROPANE 0.19437 0.09181 0.09181 0.09181 0.09181 0.09181 0.09181
ETHANE 0.01601 0.00755 0.00755 0.00755 0.00755 0.00755 0.00755
METHANE 0.00874 0.00412 0.00412 0.00412 0.00412 0.00412 0.00412
TICL4 CPSFRAC 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
TICL4 CDSFRAC 0.00124 0.00124 0.00124 0.00124 0.00124 0.00124 0.00124
TICL4 CVSFRAC 0.00015 0.00015 0.00015 0.00015 0.00015 0.00015 0.00015
PP DPN 3.495 3.495 3.495 3.495 3.495 3.495 3.495
PP DPW 6.939 6.939 6.939 6.939 6.939 6.939 6.939
PP PDI 1.99 1.99 1.99 1.99 1.99 1.99 1.99
PP MWN 147,088 147,075 147,075 147,075 147,075 147,075 147,075
PP MWW 292,016 292,011 292,011 292,011 292,011 292,011 292,011

S13
 Table S6. Propylene polymerization process streams (continuation)

Streams FEED3 GAS1 GAS2 GAS3 H2A H2B H2C

Temperature (°C) 70.0 85.0 105.0 100.0 10.0 10.0 10.0
Pressure (bar) 35.0 0.5 1.2 1.0 35.0 35.0 35.0
Vapor fraction 0.00 1.00 1.00 1.00 1.00 1.00 1.00
Enthalpy (cal/g) -13.42 -4.43 -2988.37 -444.36 -45.50 -45.50 -45.50
Density (g/cm³) 0.400 0.001 0.001 0.001 0.003 0.003 0.003
Molar mass (g/mol) 41.83 42.24 18.65 25.95 2.02 2.02 2.02
Mass flow (kg/h) 319.52 2.59 13.93 17.54 0.00 0.00 0.00
Mass fraction
TEA 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
TICL4 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
PP 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
H2 0.00003 0.00000 0.00000 0.00000 1.00000 1.00000 1.00000
N2 0.00000 0.00000 0.00000 0.85246 0.00000 0.00000 0.00000
H2O 0.00000 0.00000 0.94119 0.14502 0.00000 0.00000 0.00000
PROPY 0.89650 0.79215 0.04678 0.00201 0.00000 0.00000 0.00000
PROPANE 0.09181 0.19831 0.01181 0.00051 0.00000 0.00000 0.00000
ETHANE 0.00755 0.00835 0.00022 0.00000 0.00000 0.00000 0.00000
METHANE 0.00412 0.00119 0.00001 0.00000 0.00000 0.00000 0.00000
TICL4 CPSFRAC 0.00000 0.00000 0.00000 0.00000
TICL4 CDSFRAC 0.00124 0.00124 0.00124 0.00124
TICL4 CVSFRAC 0.00015 0.00015 0.00015 0.00015
PP DPN 3.495 3.495 3.495 3.495
PP DPW 6.939 6.939 6.939 6.939
PP PDI 1.99 1.99 1.99 1.99
PP MWN 147,075 147,077 147,077 147,077
PP MWW 292,011 292,016 292,016 292,016
Streams H2FEED MIXFEED N2 POL1 POL2 POL3 POL4
Temperature (°C) 10.0 50.9 100.0 85.0 85.0 105.0 100.0
Pressure (bar) 35.0 25.0 1.1 2.0 0.5 1.2 1.0
Vapor fraction 1.00 0.00 1.00 0.00 0.00 0.00 0.00
Enthalpy (cal/g) -45.50 -29.15 18.61 -461.55 -463.68 -470.69 -457.98
Density (g/cm³) 0.003 0.452 0.001 0.805 0.807 0.795 0.799
Molar mass (g/mol) 2.02 41.83 28.01 42.09 42.09 41.76 42.05
Mass flow (kg/h) 0.00 912.68 15.00 485.29 482.70 484.77 482.23
Mass fraction
TEA 0.00000 0.00000 0.00000 0.00042 0.00042 0.00042 0.00042
TICL4 0.00000 0.00000 0.00000 0.00003 0.00003 0.00003 0.00003
PP 0.00000 0.00000 0.00000 0.99243 0.99775 0.99350 0.99873
H2 1.00000 0.00003 0.00000 0.00000 0.00000 0.00000 0.00000
N2 0.00000 0.00000 1.00000 0.00000 0.00000 0.00000 0.00010
H2O 0.00000 0.00000 0.00000 0.00000 0.00000 0.00595 0.00071
PROPY 0.00000 0.89650 0.00000 0.00564 0.00143 0.00008 0.00000
PROPANE 0.00000 0.09181 0.00000 0.00142 0.00036 0.00002 0.00000
ETHANE 0.00000 0.00755 0.00000 0.00005 0.00001 0.00000 0.00000
METHANE 0.00000 0.00412 0.00000 0.00001 0.00000 0.00000 0.00000
TICL4 CPSFRAC 0.00000 0.00000 0.00000 0.00000 0.00000
TICL4 CDSFRAC 0.00124 0.00124 0.00124 0.00124 0.00124
TICL4 CVSFRAC 0.00015 0.00015 0.00015 0.00015 0.00015
PP DPN 3.495 3.495 3.495 3.495 3.495
PP DPW 6.939 6.939 6.939 6.939 6.939
PP PDI 1.99 1.99 1.99 1.99 1.99
PP MWN 147,075 147,077 147,077 147,077 147,077
PP MWW 292,011 292,016 292,016 292,016 292,016

S14
 Table S6. Propylene polymerization process streams (continuation)

Streams PURGA RC1 RC2 S1 VAPOR

Temperature (°C) 83.6 85.0 83.6 83.6 120.0
Pressure (bar) 0.5 2.0 0.5 0.5 2.0
Vapor fraction 1.00 1.00 1.00 1.00 0.00
Enthalpy (cal/g) -17.46 -17.54 -17.46 -17.46 -3697.22
Density (g/cm³) 0.001 0.003 0.001 0.001 0.866
Molar mass (g/mol) 41.54 41.54 41.54 41.54 18.02
Mass flow (kg/h) 7.74 427.61 422.45 430.19 16.00
Mass fraction
TEA 0.00000 0.00000 0.00000 0.00000 0.00000
TICL4 0.00000 0.00000 0.00000 0.00000 0.00000
PP 0.00000 0.00000 0.00000 0.00000 0.00000
H2 0.00006 0.00006 0.00006 0.00006 0.00000
N2 0.00000 0.00000 0.00000 0.00000 0.00000
H2O 0.00000 0.00000 0.00000 0.00000 1.00000
PROPY 0.78082 0.78075 0.78082 0.78082 0.00000
PROPANE 0.19437 0.19434 0.19437 0.19437 0.00000
ETHANE 0.01601 0.01605 0.01601 0.01601 0.00000
METHANE 0.00874 0.00879 0.00874 0.00874 0.00000
TICL4 CPSFRAC 0.00000 0.00000 0.00000 0.00000
TICL4 CDSFRAC 0.00124 0.00124 0.00124 0.00124
TICL4 CVSFRAC 0.00015 0.00015 0.00015 0.00015
PP DPN 3.495 3.495 3.495 3.495
PP DPW 6.939 6.939 6.939 6.939
PP PDI 1.99 1.99 1.99 1.99
PP MWN 147,077 147,077 147,077 147,077
PP MWW 292,016 292,016 292,016 292,016

S15
MEMBRANE SEPARATOR

Modeling and validation

For the membrane separator, the modeling strategy followed by Souza5 was taken

as a basis, on which some modifications were made to adapt it to the system of interest.

The model is based on the analogy with a shell and tube heat exchanger. The following

assumptions were made: stationary state; countercurrent configuration; only H2 is

transported through the membrane, due to its high concentration in the feed and high

selectivity in relation to the other hydrocarbons present; the mass transfer coefficient

(permeance) is considered constant throughout the separation; the pressure drop along the

equipment is negligible; the equipment is considered thermally insulated and operates at

a constant temperature equal to that of the feed; the driving force for separation is the

difference in partial pressures of hydrogen between the retentate and the permeate.

To validate the model, three separations were simulated under different conditions of

pressure, molar flow and temperature, with typical gases from refineries A, B and C,

which had respectively 90, 85 and 80% molar hydrogen and the remaining percentage of

C1-C4 hydrocarbons. The results were compared with the values obtained in simulations

carried out by Kaldis et al.,6 who used a model whose parameters were adjusted and

validated from experimental data, using a gas stream resulting from a diesel

desulfurization process with composition of 67.5% H2, 16.7% CH4, 4.3% C2H6 and 11.5%

CO2. The value used for the permeation rate of pure hydrogen through the membrane was

the same used by the authors, of 0.00971 mol/m².s.bar, measured experimentally at 40 °C

and 10 bar for a commercial hollow fiber membrane module of polyimide supplied by

UBE Industries Ltd., located in Japan. The gas streams were chosen due to the greater

similarity with the stream that is to be separated in this work. The results of hydrogen

S16
recovery and H2 fraction in the retentate obtained by varying the stage cut were plotted

and graphically represented in Figures S2 and S3, respectively.

Fig. S2. Hydrogen recovery in the membrane separation module vs stage cut

Fig. S3. Hydrogen fraction in the retentate of the membrane separation module vs

stage cut

S17
 It is worth mentioning that, although the model represents the simulated separations

well, its use is limited. This is mainly due to the hypothesis that only considers the passage

of H2 through the membrane, because for it to be true, the flux of other gases that permeate

the membrane must be negligible in relation to the flux of hydrogen. Therefore, it is not

enough that only the selectivity of H2 in relation to the other gases is high, but it is also

necessary that the decrease in the partial pressure of hydrogen in the retentate throughout

the membrane does not significantly reduce the flux of H2 in relation to the flux of the

other gases in a way to invalidate the assumed hypothesis.

In the case of this simulation, this implies that for high stage cut values and low

fractions of hydrogen in the retentate, the flux of hydrocarbons through the membrane

can become significant, in a way that compromises the purity of the permeated hydrogen

and generates fractions of hydrogen greater than expected in the retentate.

Calculation routine

Based on the assumptions and simplifications presented, the mathematical

modeling of the membrane separator was carried out from the mass balance equations

and the hydrogen flux equation that permeates the membrane, whose driving force is

represented by the average logarithmic difference of the partial pressures of hydrogen

between the permeate and the retentate.

% Resolution of a Selective Permeability Gas Separation Membranes Module

% Parameters definition:

A = [200:200:2000]; % equipment area (m²)

J = 0.00971; % H2 permeability (mol/m².s.bar)

% Variables:

% P -> pressure (bar) N -> flow rate (mol/s) x -> mole fractions of permeate
S18
% y -> mole fractions of retentate z -> mole fractions of feed

% Subscripts:

% ve-> feed vs -> retentate

% ls -> permeate 1..n -> components (1:hydrogen and 2:hydrocarbons)

% Specifications:

Pve = 50; % bar

Pvs = 50; % bar

Pls = 1; % bar

z1 = 0.760253; % mole fraction of H2 in ve

z2 = 1-z1; % mole fraction of HC in ve

Nve = 697.3052778; % feed molar flow rate mol/s

x1 = 1; % mole fraction of H2 in ls

% Indicators to be monitored:

n = 1; % counter

sc = [1]; % stage cut (Permeate Flow/ Feed Flow)

yh2 = [1]; % yh2 = y1

rc = [1]; % percentage of hydrogen recovered

yhc = [1]; % yhc = y2

% System resolution

for a = A

i = [1,3,3.5,0.5]; % initial estimate of unknown variables

p = [a,J]; % vector with the parameters

v = [Pve,Pvs,Pls,z1,z2,Nve,x1]; % vector with known variables

s0 = [i,p,v]; % initial estimate

% Solver

S19
 [X, Fval] = fsolve(@(i)system(i,p,v), s0);

% Results printing

disp("The retentate has " + X(4) + "% of H2")

disp("The permeate flow rate is " + X(1) + "mol/s")

disp("The hydrogen recovery is " + X(1)/(Nve*z1) + "%")

disp("The stage cut is " + X(1)/Nve)

disp("The H2 purity of the permeate is " + x1*100 + "%")

% Calculation of the indicators to be monitored

sc(n) = X(1)/Nve;

yh2(n) = X(4);

rc(n) = X(1)/(Nve*z1);

yhc(n) = (Nve*z2)/X(3);

n = n+1

end

% Results plotting

subplot(2,2,1)

plot(A,sc)

title("Stage cut increase with the area ")

grid on

subplot(2,2,2)

plot(sc,rc)

title("Stage cut effect on H2 recovery")

xlabel("Stage cut")

ylabel("H2 recovery")

grid on

S20
subplot(2,2,3)

plot(sc,yh2)

title("Fraction of H2 in retentate vs stage cut ")

xlabel("Stage cut")

ylabel("Mole fraction of H2")

grid on

subplot(2,2,4)

plot(sc,yhc)

title("Mole fraction of HC in retentate vs stage cut")

grid on

% System of equations

function Fun = system (i,p,v)

% Parameters definition:

A = p(1);

J = p(2);

% Variables definition:

Pve = v(1);

Pvs = v(2);

Pls = v(3);

z1 = v(4);

z2 = v(5);

Nve = v(6);

x1 = v(7);

% Unknown variables definition:

S21
 Nls = i(1);

Nvs1 = i(2);

Nvs = i(3);

y1 = i(4);

% Auxiliary variables calculation:

Nve1 = Nve*z1;

DP1 = Pve*z1 - Pls*x1;

DP2 = Pvs*y1;

% System of nonlinear equations to be solved

Fun(1) = (J*A*(DP1 - DP2)/log(DP1/DP2)) - Nls;

Fun(2) = Nve1 - Nls - Nvs1;

Fun(3) = Nve*z2 + Nvs1 - Nvs;

Fun(4)= (Nvs1/Nvs) - y1;

end

It is observed that the larger the area of the equipment, the greater is the stage cut

and, consequently, the recovery of hydrogen is improved. However, as the stage cut

increases, the partial pressure of hydrogen in relation to other hydrocarbons in the

retentate is lower. Thus, this increase may invalidate the hypothesis that only hydrogen

permeates the membrane.

Therefore, a maximum stage cut of 0.6 was defined for this hypothesis to be

maintained. This threshold was chosen based on the model validation, as the threshold

stage cut value for the model results to start showing significant deviations from the

reference.6 Thus, the equipment area parameter was varied until a stage cut of 0.6 was

S22
found, and the area and recovery results obtained in this condition were used in the

subsequent analyses.

EQUIPMENT SIZING

This section presents the sizing methods used for reactors, adsorbers, VPSA and

membrane separator. The rest of the equipment was sized according to the conventional

process engineering literature.7,8

Reactors

To size the glycerol hydrogenolysis fixed bed reactor, it was used the mass space

velocity parameter of 5.4 h⁻¹ WHSV, provided by Mota et al.,9 which is equal to the feed

flow rate divided by the mass of catalyst in the reactor bed. With the catalyst mass and

porosity, it was possible to obtain the packing volume and determine the total volume of

the reactor. A 3:1 length-to-diameter ratio was considered for reactor cost estimation. The

properties of a Fe/Mo hydrogenolysis catalyst supported on activated carbon were

obtained from the literature10: porosity of 25.76%, density of 1.33 kg/L and surface area

of 0.28 km²/kg.

For the tubular loop reactors, the cost was considered equivalent to that of 4

vertical vessels, according to the reactor geometry. The horizontal parts were disregarded

because, according to Regitano et al.,11 the dimension of the horizontal region of the

reactor must be minimized to avoid the accumulation of polymer on the walls and,

therefore, it will contribute in a negligible way to the cost. For the first reactor (“baby

loop”), only two vertical vessels were considered, as their smaller size allows a simpler

design. The length-to-diameter ratio in the loop reactors used by Kim et al.2 was assumed

S23
for cost calculation: this ratio is equal to 268:1 for the first reactor and 317:1 for the

following reactors.

Adsorbers

The volume of the adsorbers depends on the amount of packing needed to remove

the desired amount of adsorbate during the entire operating time. It was considered an

operating time of 6 months for packing substitution and two vessels. The height-to-

diameter ratio assumed for cost calculation was 3:1.

Knowing the adsorption capacity of the activated alumina used as packing

material, its porosity and density, it is possible to calculate the required adsorbent mass

and the vessel volume to support this amount. These properties were obtained from the

literature12: adsorption capacity of 0.30 kg of packing / kg of adsorbate, porosity of 57%,

density of 0.85 kg/L and surface area of 0.32 km²/kg.

VPSA

The VPSA used had a seven-step cycle, like that suggested by Kuah et al.13 These

steps are pressurization from the feed, continuous adsorption under high pressure,

depressurization in co-current by equalization with another tank, the passage of propylene

product in co-current for washing, the evacuation of the vessel in countercurrent with the

use of vacuum, pressurization by the outlet stream from the equalization of another tank

in co-current, and pressurization by reentry of the product used for washing.

Using the cyclization strategy proposed by the authors, at least two VPSA

batteries in parallel containing 3 vessels each would be necessary to generate a continuous

outlet flow of product. The number of batteries used impacts the outlet flow, so that the

increase in this number leads to smaller flow rate variations and consequently the cost of

S24
heat exchangers and compressors inherent to the VPSA system is reduced, since they

require less capacity. In the work of Kuah et al.,13 an optimization was performed to

minimize the cost of the VPSA system using the number of batteries as one of the

parameters. However, in this work it was chosen to use only two batteries, since the feed

flow rate is considerably lower than that used by Kuah et al. and, therefore, it is estimated

that the absolute flow rate variation would not cause an excessive increase in the cost of

peripheral equipment related to the VPSA system.

It is worth mentioning that the flow rate variation within the VPSA system does

not affect the subsequent equipment in the plant due to the use of a buffer tank that keeps

the product flow rate constant, although this tank is not represented in the simulation.

The sizing of a VPSA system vessel is performed similarly to that for an adsorber,

but the operating time used is one cycle time. A cycle time of 15 minutes was considered,

the used by Kuah et al.13 The properties of the zeolite 4A used as adsorbent were obtained

from the same reference: adsorption capacity of 2.00 kg of packing / kg of adsorbate,

porosity of 43% and density of 1.21 kg/L.

In addition to the cost the packed vessels, the cost of associated equipment - such

as heat exchangers, compressors and vacuum systems - significantly impact the total cost

of the VPSA system. Kuah et al.13 obtained as a result of their optimization a cost

associated with this peripheral equipment of 50% of the total cost. The same percentage

was assumed in this work to determine the cost of this equipment in relation to the VPSA

system.

Membrane separator

The cost of the membrane separator is determined by the required total surface

area. A cost of 35 US$/ft² was considered for the membrane module, as suggested by

S25
Seider et al.8 The membrane area obtained as a result of the simulation to achieve the

desired separation was 1600 m².

Filter

For the filtration of polypropylene, a filtration rate of 6000 lb/ft²day and a vacuum

pressure of 6 inHg were assumed, as suggested by Seider et al.8 The area of the resultant

rotary-drum filter can be calculated as the ratio between the filter feed flow rate and the

filtration rate.

ECONOMIC RESULTS

Table S7. Prices of raw materials, products and supplies

Unit price (US$/kg) Reference

Raw materials
14
Glycerol 0.435
15,16
Hydrogen 2.756*
Adsorbents
17
Activated Alumina 0.690
18
Zeolite 4A 0.350
Catalysts
19
TiCl4 1.100
20
TEAl 1.170
21
Fe/Mo 1.050
Utilities and other resources
22
Fuel Oil B1 0.420
23
Nitrogen Gas 0.007
Low Pressure Steam 0.004165 Aspen Plus® v8.8
Water 0.000004 Aspen Plus® v8.8
24
Electricity 0.1373**
25
Effluent Disposal 0.036
Products
26–28
Propylene 2.480*
29
Polypropylene 1.215
* Price is an average from more than one reference.

** Unit price in US$/kWh.

S26
Table S8. Equipment costs of glycerol hydrogenolysis for propylene production

Name Equipment Type Value (US$)

HT1 Heat Exchanger 155,677.29
HT2 Heat Exchanger 15,421.00
HT3 Heat Exchanger 10,361.00
COOLER1 Heat Exchanger 29,654.00
COOLER 2 Heat Exchanger 25,706.00
COOLER 3 Heat Exchanger 1,101.64
HX Heat Exchanger 649,596.00
COMP1 Compressor 2,382,049.18
COMP2 Compressor 1,283,349.63
COND1 Condenser 62,453.19
COND2 Condenser 17,515.46
ADS Adsorber 31,978.68
REACTOR Reactor 29,599.11
PSA Pressure Swing Adsorption 124,903.86
MEMB SEP Membrane Separator 84,739.09
Total Equipment Cost 4,904,105.12

S27
 Table S9. Production costs in the glycerol hydrogenolysis

Base Scenario Scenario 1 Scenario 2 Scenario 3

Direct Production Costs (US$/year)

Raw Material (CRM)

Glycerol 4,118,576.15 0.00 4,118,576.15 0.00
Hydrogen 1,206,921.76 1,206,921.76 1,206,921.76 1,206,921.76
Operating labor (COL) 242,953.04 242,953.04 242,953.04 242,953.04
Supervisory and clerical labor (CSC = 0.15 COL) 43,731.55 43,731.55 43,731.55 43,731.55
Utilities (CUTIL)
Fuel oil 139,447.95 139,447.95 139,447.95 139,447.95
Steam 39,430.49 39,430.49 39,430.49 39,430.49
Cooling water 45,627.87 45,627.87 45,627.87 45,627.87
Electricity 3,965,642.05 3,965,642.05 3,965,642.05 3,965,642.05
Adsorbents (CADS) 3,881.44 3,881.44 3,881.44 3,881.44
Effluent disposal (CED) 8,363,055.49 8,363,055.49 0.00 0.00
Maintenance and repairs (CMR = 0.06 IFixed) 1,483,001.39 1,483,001.39 1,483,001.39 1,483,001.39
Operating supplies (COS = 0.15 CMR) 222,450.21 222,450.21 222,450.21 222,450.21
Laboratory charges (CLC = 0.15 COL) 36,442.96 36,442.96 36,442.96 36,442.96
Patents and royalties (CPR = 0.03 CTP) 806,588.90 655,909.28 500,623.45 349,943.84
Subtotal 20,717,751.30 16,448,495.50 12,048,730.30 7,779,474.55

Indirect production costs (US$/year)

Overhead, packaging and storage (COPS = 0.60 (COL + CSC + CMR)) 1,061,811.59 1,061,811.59 1,061,811.59 1,061,811.59
Local taxes (LT = 0.015 IFixed) 790,934.07 790,934.07 790,934.07 790,934.07
Insurance (IN = 0.005 IFixed) 123,583.45 123,583.45 123,583.45 123,583.45
Subtotal 1,976,329.11 1,976,329.11 1,976,329.11 1,976,329.11

General expenses (US$/year)

Administrative costs (CAD = 0.25 COPS) 159,271.74 159,271.74 159,271.74 159,271.74

Distribution and sellings costs (CDS = 0.10 CTP) 2,957,492.62 2,405,000.70 1,835,619.33 1,283,127.40
Research and development (CRD = 0.05 CTP) 1,344,314.83 1,093,182.14 834,372.42 583,239.73
Subtotal 4,461,079.19 3,657,454.58 2,829,263.49 2,025,638.87

Total production cost (CTP) (US$/year) 27,155,159.60 22,082,279.19 16,854,322.90 11,781,442.53

S28
Table S10. Equipment costs of propylene polymerization

Name Equipment Type Value (US$)

HT Heat Exchanger 12,164.00
FP Heat Exchanger 17,825.00
COOLER1 Heat Exchanger 10,592.00
COOLER2 Heat Exchanger 11,929.00
PUMP Pump 4,567.49
COMP Compressor 64,468.23
FLASH Flash Vessel 22,466.42
R1 Reactor 13,724.23
R2 Reactor 71,520.03
R3 Reactor 71,520.03
FILTER Filter 117,536.63
DRYER Dryer 238,924.27
STEAMER Flash Vessel 55,523.13
Total Equipment Cost 712,760.46

S29
Table S11. Production costs in the propylene polymerization

Base Scenario
Direct Production Costs (US$/year)
Raw Material (CRM) 0.00
Catalyst 2,010.39
Hydrogen 80.88
Clerical labor (COL) 231,472.31
Supervisory (CSC = 0.15 COL) 41,665.02
Utilities (CUTIL)
Steam 2,408.72
Cooling water 605.19
Electricity 30,226.07
Effluent disposal (CED) 165,619.67
Maintenance and repairs (CMR = 0.06 IFixed) 215,538.76
Operating supplies (COS = 0.15 CMR) 32,330.81
Laboratory charges (CLC = 0.15 COL) 34,720.85
Patents and royalties (CPR = 0.03 CTP) 44,882.22
Subtotal 801,560.89
Indirect production costs (US$/year)
Overhead, packaging and storage (COPS = 0.60 (COL + CSC + CMR)) 293,205.65
Local taxes (LT = 0.015 IFixed) 114,954.01
Insurance (IN = 0.005 IFixed) 17,961.56
Subtotal 426,121.22
General expenses (US$/year)
Administrative costs (CAD = 0.25 COPS) 43,980.85
Distribution and sellings costs (CDS = 0.10 CTP) 164,568.15
Research and development (CRD = 0.05 CTP) 74,803.70
Subtotal 283,352.70
Total production cost (CTP) (US$/year) 1,511,034.81

S30
REFERENCES

(1) Zheng, Z. W.; Shi, D. P.; Su, P. L.; Luo, Z. H.; Li, X. J. Steady-State and Dynamic

Modeling of the Basell Multireactor Olefin Polymerization Process. Ind. Eng.

Chem. Res. 2011, 50 (1), 322–331. https://doi.org/10.1021/ie101699b.

(2) Kim, S. H.; Baek, S. W.; Lee, J. C.; Lee, W. J.; Hong, S. U.; Oh, M. Dynamic

Simulation of Liquid Polymerization Reactors in Sheripol Process for

Polypropylene. J. Ind. Eng. Chem. 2016, 33, 298–306.

https://doi.org/10.1016/j.jiec.2015.10.017.

(3) Lee, J. C.; Kofi, O. S.; Kim, S. H.; Hong, S. U.; Oh, M. Polypropylene Production

Simulation with CAPE-OPEN Interfacing of PRO/II and gPROMS. J. Eng. Sci.

Technol. 2015, Special Issue 2, 48–61.

(4) Ballarini, G.; Mileo, A.; Soffritti, L.; Soffritti, E. Steaming of a Polyolefin. United

States Patent US8344099B2, 2013.

(5) Souza, I. R. Development of the Unit Operation of Gas-Liquid Contactor with

Membrane in the HYSYS 8.8 Simulator, Federal University of Rio de Janeiro, Rio

de Janeiro, Brazil. 2018. http://www.h2cin.org.br/download/desenvolvimento-da-

operacao-unitaria-contactora-gas-liquido-com-membrana-no-simulador-hysys-8-

8.pdf (accessed December 27, 2022) (in Portuguese).

(6) Kaldis, S. P.; Kapantaidakis, G. C.; Sakellaropoulos, G. P. Simulation of

Multicomponent Gas Separation in a Hollow Fiber Membrane by Orthogonal

Collocation - Hydrogen Recovery from Refinery Gases. J. Memb. Sci. 2000, 173

(1), 61–71. https://doi.org/10.1016/S0376-7388(00)00353-7.

(7) Towler, G.; Sinnott, R. Chemical Engineering Design - Principles, Practice and

Economics of Plant and Process Design; Elsevier, 2008.

(8) Seider, W. D.; Seader, J. D.; Lewin, D. R. Product and Process Design Principles

S31
 - Synthesis, Analysis, and Evaluation, 2nd ed.; John Wiley and Sons, 2003.

(9) Mota, C. J. A.; Gonçalves, V. L. C.; Mellizo, J. E.; Rocco, A. M.; Fadigas, J. C.;

Gambetta, R. Green Propene through the Selective Hydrogenolysis of Glycerol

over Supported Iron-Molybdenum Catalyst: The Original History. J. Mol. Catal.

A Chem. 2016, 422, 158–164. https://doi.org/10.1016/j.molcata.2015.11.014.

(10) Cerón, J. E. M. Hydrogenolysis of Glycerin to Propene: Study of the Reaction

Pathway, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil. 2017.

http://epqb.eq.ufrj.br/wp-content/uploads/2021/10/d17_17.pdf (accessed

December 27, 2022) (in Portuguese).

(11) Reginato, A. S.; Zacca, J. J.; Secchi, A. R. Modeling and Simulation of Propylene

Polymerization in Nonideal Loop Reactors. AIChE J. 2003, 49 (10), 2642–2654.

https://doi.org/10.1002/aic.690491017.

(12) Keller, G. E.; Anderson, R. A.; Yon, C. M. Adsorption. In Handbook of separation

process technology; Rousseau, R. W., Ed.; John Wiley & Sons, 1988; Vol. 23, pp

644–691. https://doi.org/10.1016/0255-2701(88)80007-2.

(13) Kuah, W. C.; Effendy, S.; Farooq, S. Industrial Scale Propylene/Propane

Separation Using Pressure Vacuum Swing Adsorption. Ind. Eng. Chem. Res. 2018,

57 (18), 6451–6463. https://doi.org/10.1021/acs.iecr.8b00289.

(14) COMEXSTAT. Foreign Trade Statistics. http://comexstat.mdic.gov.br/pt/home

(accessed October 21, 2021).

(15) DiChristopher, T. Experts explain why green hydrogen costs have fallen and will

keep falling. S&P Global Market Intelligence 2021.

https://www.spglobal.com/marketintelligence/en/news-insights/latest-news-

headlines/experts-explain-why-green-hydrogen-costs-have-fallen-and-will-keep-

falling-63037203 (accessed October 25, 2021).

S32
(16) IEA - International Energy Agency. The Future of Hydrogen. 2019.

https://www.iea.org/reports/the-future-of-hydrogen (accessed October 25, 2021).

(17) Zhengzhou Kelin Water Purification Material Co.

https://www.kelincarbon.com/products (accessed October 21, 2021).

(18) Tianjin Credit International Co.

https://creditchemical.diytrade.com/sdp/2391680/4/pd/19377726-

2824480/4A_zeolite.html (accessed October 21, 2021).

(19) Qingdao Tida International Trade Co. https://www.botianchemical.com/supplier-

466092-titanium-tetrachloride (accessed October 21, 2021).

(20) M&J International Trading Co. https://lnmjgjmyyxgs.en.ecplaza.net/products

(accessed October 21, 2021).

(21) Kouga Corporation. https://www.kougacorp.co.jp/ (accessed October 21, 2021).

(22) ANP - Brazilian National Agency of Petroleum, Natural Gas and Biofuels. Prices

of Producers and Importers of Petroleum Products and Biodiesel.

https://www.gov.br/anp/pt-br/assuntos/precos-e-defesa-da-

concorrencia/precos/precos-de-produtores-e-importadores-de-derivados-de-

petroleo (accessed October 21, 2021).

(23) Ningbo DSW Industry Co. http://en.globalimporter.net/product.asp (accessed

October 21, 2021).

(24) ANEEL - Brazilian National Electric Energy Agency. https://www.aneel.gov.br/

(accessed May 1, 2020).

(25) Turton, R.; Bailie, R. C.; Whiting, W. B.; Shaeiwitz, J. A. Analysis, Synthesis, and

Design of Chemical Processes, 3rd ed.; Prentice Hall: Boston, United States, 2009.

(26) Air Gas Eletronica Materials Enterprise Co.

https://www.agem.com.tw/hydrocarbons.html (accessed October 21, 2021).

S33
(27) Beijing Starget Chemicals Co. http://www.stargetchem.com/ (accessed October

21, 2021).

(28) Chengdu Taiyu Industrial Gases Co. https://www.taiyugas.com/industrial-

gases/propylene-c3h6,-industrial-gas.html (accessed October 21, 2021).

(29) Qingdao Hengke Fine Chemicals Co.

http://www.hengkechem.com/product/p1.html (accessed October 21, 2021).

S34