Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Contents lists available at ScienceDirect

Chemical Engineering & Processing: Process Intensification

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

Intensification and performance assessment of the formic acid production T

process through a dividing wall reactive distillation column with vapor
recompression
⁎
Swapnil Sharmaa, Dipesh S. Patlea,b, , Akhil Premkumar Gadhamsettia, Sanket Pandita,
Davide Mancab, G S Nirmalaa
a
Department of Chemical Engineering, School of Civil and Chemical Engineering, VIT University, Tamilnadu, 632014, India
b
PSE-Lab, Process Systems Engineering Laboratory, Dipartimento di Chimica, Materiali e Ingegneria Chimica “Giulio Natta”, Politecnico di Milano, Milan, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: Distillation operation in the production of formic acid (FA) involves a large amount of investment and has high-
Formic acid energy requirements. Thus, process intensification is crucial in minimizing the energy consumption and in-
Process intensification vestment. Dividing wall column (DWC) is one such way of achieving it. Aim of this study is to design a dividing
DWC wall reactive distillation column (DWRDC) with vapor recompression in the production of FA instead of normal
Vapor recompression
reactive distillation – distillation sequence and investigate the performance in terms of energy, cost, and carbon
emission. A patented FA production process is chosen as a base case. The intensification of the base case FA
production process is then carried out through the application of (i) DWRDC, (ii) vapor recompression in a
distillation column, and (iii) vapor recompression in DWRDC. The performance of the new intensified processes
is subsequently investigated for multiple performance indexes. It is observed that an impressive saving of
18.33% in utility cost, 8.4% in total annual cost (TAC), 2.55% in cost of manufacturing (COM), and 17.68% in
carbon emission can be secured with the proposed new intensification as compared to its conventional coun-
terpart for a payback period of 4 years.

1. Introduction continuous stirred tank reactor (CSTR) using a heterogeneous catalyst,

Formic acid (FA) is a strong organic acid that is principally used in
various industries, such as chemical, pharmaceuticals, rubber, textile,
agriculture, and leather. FA is readily bio-degradable, hence it is en-
vironmentally sustainable, which makes it a popular raw material in
many industries for the production of various desirable chemicals [1].
Among several commercial processes for the FA production, namely
acidolysis of formate salts, oxidation of hydrocarbons, hydrolysis of
formamide, mineral acid catalysis, and the hydrolysis of lower alkyl
formates, the hydrolysis of methyl formate (MF) is attractive as it
overcomes the limitations of existing counterparts [2–4]. Limitations of
existing processes include slow reaction rate, unwanted side-products,
high cost of investment, and issues related to the environment. Hy-
drolysis of MF involves two steps: (1) carboxylation reaction in a
(2) hydrolysis of the intermediate product to obtain formic acid. This
process solves the limitations of the conventional process viz. undesir-
able side products, slow reaction time, environmental complications,
and high capital investments [5]. Huang and colleagues [6] designed a
process by combining a CSTR and a simple distillation column in the
MF hydrolysis-based process into a reactive distillation (RD) unit. The
combination of the reaction and separation occurring simultaneously in
RD provides many advantages: (1) enhanced yield and selectivity, (2)
reduced energy consumption, and (3) avoidance of hot spots. In addi-
tion, it is observed that there is a significant reduction in heat duty,
which further reduces the operating cost.
Distillation is the most commonly used separation technique in the
chemical process industry, worldwide, accounting for 95% of all se-
parations [7]. Sixty per cent of energy use in chemical industry is from

Abbreviations: C1-VR, vapor recompression in C1; CO, carbon monoxide; COM, cost of manufacturing; CSTR, continuous stirred tank reactor; DWC, dividing wall column; DWRDC,
dividing wall reactive distillation column; DWRDC-VR, vapor recompression in DWRDC; FA, formic acid; FCI, fixed capital investment; MF, methyl formate; RD, reactive distillation; TAC,
total annual cost; TCDC, thermally coupled distillation column; TCDWC, thermally coupled dividing wall columns; VRC, vapor recompression
⁎
Corresponding author at: Department of Chemical Engineering, School of Civil and Chemical Engineering, Vellore Institute of Technology, Tamilnadu, 632014, India. (Currently a
post-doctoral researcher at PSE-Lab, Process Systems Engineering Laboratory, Dipartimento di Chimica, Materiali e Ingegneria Chimica “Giulio Natta”, Politecnico di Milano, Milan,
Italy).
E-mail address: dipesh.patle@vit.ac.in (D.S. Patle).

https://doi.org/10.1016/j.cep.2017.11.016
Received 15 July 2017; Received in revised form 14 November 2017; Accepted 21 November 2017
Available online 22 November 2017
0255-2701/ © 2017 Elsevier B.V. All rights reserved.
S. Sharma et al.
the distillation process [8]. A large amount of energy required in FA
production is in the distillation process. Process intensification is cur-
rently considered as one of the major approaches to reduce the number
of pieces of equipment in the plant and even improve process perfor-
mance. For a distillation column system, thermal coupling provides
such an approach to retrofit the existing simple column configurations
through the elimination of either the condenser or the reboiler [9] after
integrating two columns into one. Thermally coupled distillation
column (TCDC) and divided wall column (DWC) in their fully in-
tegrated form have attracted strong attention due to the greater need to
reduce energy requirement and capital costs in distillation operations
[10]. DWC integrates the two columns into a single shell providing an
excellent path for process intensification. There are considerable energy
savings obtained by this process intensification with a reduction of
space for installation by 35–45% due to the replacement of two col-
umns by a single column [11–13]. The benefits of thermally coupled
dividing wall columns (TCDWC) are: (1) integrated design with reduced
capital investment and footprint, (2) high purity products, (3) increased
thermodynamic efficiency due to reduction in remixing effects, (4)
lower energy requirements in comparison to conventional separation
sequences, and (5) low maintenance costs. Studies suggest that DWC
systems can reduce energy requirement by up to 30% over conventional
direct and indirect distillation sequences [14]. However, practical im-
plementation and operation of this intensified structure is difficult due
to the highly integrated structure. In addition, such highly integrated
configurations pose a tough challenge pertaining their optimal design
and control.
Energy consumption can further be reduced by a suitable heat in-
tegration technology such as vapor recompression (VRC) scheme,
where a fraction of the overhead vapor is compressed and used as a
heating medium in reboiler. The conventional VRC technique is more
energy efficient with small temperature span across the different stages
and large heat duty [15]. The VRC scheme, with side heat exchanger
configurations has been applied in both conventional distillation and
reactive distillations for separating wide boiling mixtures [16–18]. In
the intensified configuration, the conventional VRC is added between
the overhead and the bottom of DWC to recover the heat of the vapors
as much as possible. Luo et al. [19] added a VRC heat pump to further
increase the energy savings up to 40% in extractive distillation for the
bioethanol purification process. Johri et al. [20] proposed a variable
speed vapor recompression in batch reactive distillation, in which a
65.85% energy saving was obtained with a payback period of 4 years
[20]. Luyben and coauthors [21] reported that VRC in azeotropic DWC
secures a lower total annual cost (TAC) with a payback period of 5 years
despite the higher capital costs.
Goal of this study is to suitably reduce energy consumption in the
FA production process by performing process intensification through
the application of a dividing wall reactive distillation column (DWRDC)
with vapor recompression respect to the patented base case process for
FA production. At first, the base case process is simulated using Aspen
Plus v8.8. Subsequently, the process is studied for the application of
DWRDC, vapor recompression in distillation column, and vapor re-
compression in DWRDC. The merits of the new configurations through
the application of above heat integrations are discussed in terms of
crucial criteria, namely energy savings, TAC, cost of manufacturing
(COM), and carbon emission. To our knowledge, this is the first study in
the literature that analyses the application of vapor recompression in
DWRDC for the FA synthesis.
2. Process description
For the present study, we chose the patented configuration from
Huang and coauthors [6] for FA production. This process is chosen as it
combines a hydrolysis unit (i.e. a reactor) and a separation unit (i.e. a
distillation column) into a reactive distillation unit to produce high
purity FA, thereby reducing the processing steps and the footprint.
205
Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213
Fig. 1 shows the overall process flow sheet for the base case process.
The FA manufacturing system consists of a reactor (where the car-
bonyl reaction occurs) and three separation columns, which include one
reactive distillation (RD) column for Hydrolysis of Methyl Formate as
well as for separation of products. A high pressure adiabatic CSTR is
used for the reaction between methanol and carbon-monoxide (CO)
(Carbonyl reaction) with sodium methoxide (CH3ONa) as a hetero-
geneous catalyst (Eq. (1)).
CH3OH + CO → HCOOCH3 (1)
Fresh carbon monoxide is fed to the CSTR with a flow rate of
70.3 kmol/h. Make up of methanol is fed to the CSTR to maintain the
excess of methanol. The reactor is operated at 75 °C and 40.53 bar. The
CSTR product is flashed into the phase separator and is then fed to a
distillation column to separate MF from the top of the methanol re-
covery unit (C1). Finally, it is fed to the RD unit. The remaining little
amount of CO is purged using a partial-vapor-liquid condenser in the
separation column. The total CO emission from all purge streams in this
process adds up to only about 90 kg CO/h in base as well as in modified
cases. The unconverted methanol obtained from the bottom of the se-
paration column (C1) is recycled back to the CSTR. The production of
FA and methanol by the hydrolysis of MF takes place in the RD column
along with its separation. Required amount of fresh water is added to
RD on 2nd tray (trays are numbered from top to bottom) and MF from
C1 enters RD on 33rd tray.
HCOOCH3 + H2O → HCOOH + CH3OH (2)
The separation between reactants (i.e. MF and Water) and products
(i.e. FA and Methanol) takes place simultaneously in the RD unit, which
consists of 35 stages. This configuration helps improving the reaction
efficiency as well as the separation efficiency. RD consists of some re-
active trays (1st-19th from top) and a stripping zone below the reactive
zone. The liquid mixture obtained from RD is fed into the separating
column (C2). FA and water are separated from the C2 column as a
bottom product, whereas methanol and MF are recycled back from the
top. Finally, we obtain about 85% FA as a final product with a flow rate
of 92 kmol/h. The distillate from the C2, mainly containing methanol,
is recycled to the CSTR to maximize the yield. Important operating
data, such as flow rates, temperature, heat duty, and pressure are given
in Fig. 1.
3. Formic acid production process modelling and simulation
The UNIQUAC-HOC thermodynamic model of Aspen Plus is selected
to deal with solvation of polar compounds and dimerization of the
vapor phase [22]. This model is suitable as it accounts for strong as-
sociation and solvation effects. The UNIQUAC Functional-group Ac-
tivity Coefficient (UNIFAC) method is applied to estimate the missing
parameters. The rigorous RADFRAC model is adopted to simulate a
distillation unit explicitly by considering the three phase balances.
Fig. 2 depicts T-xy diagram for FA-water at 1.03 and 4 bar. This figure
shows a large deviation from ideality and dependence on the pressure.
At low pressure, i.e. at 1.03 bar, a negative azeotrope (i.e. maximum
boiling) is observed while that azeotrope disappears at higher pressure
i.e. at 4 bar. An equilibrium stage model (RADFRAC) is used for the
simulation of all distillation and RD columns. The catalyst hold up for
the kinetic rate controlled reaction is 0.25 m3 of catalyst on each stage
of RD.
The kinetics of the carbonyl reaction occurring in the CSTR is taken
from [23]. Kinetics is valid for a temperature of 60–110 °C and pres-
sures of about 2–4 MPa. The CSTR volume is 4084 L.
S. Sharma et al. Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Fig. 1. Formic acid production process based on patented process configuration [6] (Base Case).

155 x at 1.03 bar 3.2. DWRDC

y at 1.03 bar
x at 4 bar
The dividing wall column is a good example of intensification that
y at 4 bar
combines two conventional columns and can vastly reduce both oper-
ating and capital costs of distillation together with the space for in-
stallation. Energy efficiency is improved by modifying both RD and C2
columns to the DWRDC configuration.
Fig. 3 shows a dividing wall column configuration where a dividing
wall is present in the top section of the column. This arrangement helps
minimizing the heat duty in the RD unit. The operating temperature in
0.0 0.2 0.4 0.6 0.8 1.0 RD and C2 allows integrating them into a single DWRDC. A top dividing
wall is used as it is required to re-circulate unconverted MF in the left
Fig. 2. T-xy diagram for FA-water at 1.03 and 4 bar. section of DWRDC by having a total reflux. To implement the DWRDC
configuration in Aspen Plus, two RADFRAC columns are used to re-
present ‘RD’ (i.e. the left section of DWRDC) and ‘C2’ (representing the
−70748
r= (1.414 × 109)exp ⎛ ⎞ [cat]L [MeOH]L [CO]L right section of DWRDC). A single reboiler is used for both sections
⎝ 8.31451T ⎠
while two condensers are used because of the top dividing mechanism.
−92059
− (2.507 × 1012)exp ⎛ ⎞ [cat]L [MF]L Vapor from the reboiler through 33rd tray is split and sent to both
⎝ 8.31451T ⎠ (3)
sections as shown in Fig. 3. Optimal flow rates and tray location are
selected by performing a sensitivity analysis targeting to reduce the
where r is the reaction rate (mol/L min), R is the universal gas constant
reboiler duty. This configuration is different from the one of [24],
(J/mol/K)), L denotes liquid phase concentration,
where left section has 35 trays. The rest of the process remains as it is
[CO]L = α[CO]G = 0.154Pco /RT , and [] denotes the component con-
shown in Fig. 1. The merits of this configuration are discussed in detail
centration (mol/L) [24].
in Section 5.2.
The rate equation of the hydrolysis of MF occurring in the RD is
taken from [25];
3.3. Vapor recompression in C1 (C1-VR)
63100 ⎛ [MEOH][FA] ⎞
r= k oexp ⎛ ⎞ ⎜ [H2 O][MF] − ⎟
Fig. 4 shows the flow sheet of divided wall distillation process with
⎝ 8.3145T ⎠ ⎝ 0.4492exp(− 251/T) ⎠ (4)
vapor recompression in C1 (C1-VR). At the top of the column, the
6.530 * 106 pressure is 4.053 bar and the temperature is 86.5 °C. Overhead vapors
where r is the reaction rate (mol/L min), k o = 1 + 0.0869[H O]2 and [] de-
2 (277 kmol/h) from the top of the column are split in two streams: a
notes the concentration of component (mol/L). The following sections fraction of the vapors (80%) is sent to the condenser and subsequently
present and discuss the base case and its modified cases viz. DWRDC, to the reflux drum; the remaining fraction of vapors (20%) is sent to the
vapor recompression in C1, and vapor recompression in DWRDC. compressor through a heat exchanger, HE-2, where it exchanges the
energy from the stream coming from HE-1.
The split fraction is optimally assigned based on the sensitivity
3.1. Base case analysis to obtain the least reboiler load. The vapor temperature in-
creases in compressor, Comp-2, due to the compression. This vapor
Fig. 1 refers to the base case, which is described in detail in Section stream is then sent to the heat exchanger ‘HE-1’, where the liquid
2. Process intensification is carried out respect to this base case process stream coming from the bottom of the column is heated. In essence, the
through the introduction of (i) DWRDC, (ii) vapor recompression in heat is provided to the reboiler through a hot vapor stream coming from
distillation column, and (iii) vapor recompression in DWRDC. These Comp-2 (51 kmol/h at 166.5 °C and 12.16 bar with a compression ratio
new configurations are explained below. (CR) of 3) with a compression work of 66.2 kW. The vapor stream then

206
S. Sharma et al. Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Fig. 3. Formic acid production process with DWRDC.

leaves HE-1 at 115 °C. The heat transfer areas of HE-1 and HE-2 are
25.42 and 1.18 m2, respectively. Overall heat transfer co-efficient of
0.85 kW/m2/K is assumed and the log mean temperature difference is
found to be 19.87 °C. This provides a certain amount of energy, repla-
cing a part of heating utility that would be required otherwise.
Remaining requirement of hot utility is fulfilled with medium pressure
steam. Stream coming out from HE-1 is then condensed and sent to the
reflux drum, from which a flow rate of 109.4 kmol/h is refluxed back to
the column. The RD feed with a flowrate of 167 kmol/h is finally sent to
the DWRDC for further processing.
3.4. Vapor recompression in DWRDC (DWRDC-VR)
The FA production process with DWRDC including vapor re-
compression is extremely complex to simulate as the flow rates of
several streams must be optimally determined. Special attention is re-
quired to achieve convergence of the solution. Note that commercial
software does not provide any in‐built models for such simulation.
Hence, vapor split, position of dividing wall, compression ratio, and
feed tray need to be optimally found in order to reliably converge the
simulation. These parameters were determined by performing a sensi-
tivity analysis. Some material recycle streams are torn in order to
converge the simulation; the values of downstream stream have to be

Fig. 4. Formic acid production process with DWRDC and vapor recompression in C1.

207
S. Sharma et al. Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Fig. 5. Formic acid production process with DWRDC including vapor recompression mechanism.

Table 1 reflux drum. Reflux flow rate of 1050 kmol/h is sent to the RD from the
Cost of raw materials. reflux drum with a molar composition of 91.35% MF, 6.9% methanol,
and trace amounts of water. A total reflux is provided to convert
Raw material cost
maximum amount of unconverted MF.
Methanol 430 $/t [31]
CO 600 $/t [32]
Formic acid 700 $/t [33] 4. Performance indicators

Process performance is assessed in terms of both environmental and

updated manually as it was done by Luyben and coauthors [21].
economic criteria. Key environmental criteria are energy consumption
Fig. 5 shows the process flow sheet of FA production with DWRDC
and carbon emission; both are indicative of sustainability. Economic
incorporating vapor recompression. The bottom stream from the
criteria include total annual cost (TAC) and cost of manufacturing
stripping section comes out with 92 kmol/h flow rate having little more
(COM). Both of these include the utilities cost.
than 85% of FA. VRC mechanism is similar to that in C1-VR as ex-
plained above. Hence, for the sake of brevity, some details are not
provided in this section. The heat is supplied to the reboiler through a
4.1. Energy analysis
hot vapor stream coming out from Comp-3 compressor (213.4 kmol/h
at 153.2 °C and 8.1 bar) in Fig. 5.
The first performance index for evaluating the process intensifica-
The vapors leaving the top of rectifying section are split and fed to
tion over the conventional process is energy saving. We use the fol-
the condenser and compressor. The vapor is compressed to 8.1 bar (CR
lowing expression to evaluate the percent reduction in energy con-
of 2) requiring 197 kW of compressor work with a discharge tempera-
sumption as:
ture of about 153 °C. This stream condenses in HE-3 and is cooled in
HE-4 and condensed using cooling water to 70 °C before entering the

(a) 120
Fig. 6. RD column in base case. (a) Composition profile and (b)
(b) Temperature profile.

0.75 100
Temperature (oC)

CO
MF
MEOH 80
0.50 H2O
CH2O2
60
0.25

40
0.00 0 5 10 15 20 25 30 35
0 5 15 20 25 30 35

208
S. Sharma et al. Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Fig. 7. C2 column in base case (a) Composition profile and (b)

(a) (b) Temperature profile.
0.8
140

Temperature ( C)
CO

o
MF
MEOH
0.4 H2O
CH2O2 120

0.0 100

0 2 4 6 8 10 12
0 2 4 6 8 10 12

(a) Fig. 8. RD column in DWRDC (a) Composition and (b)

Temperature profile.
160 (b)
0.75 140
CO
Temperature ( c)

MF 120
o

0.50 MEOH
H2O 100
CH2O2 80
0.25
60
0.00 40
0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35

Table 2 second performance index for economic comparison is TAC, as reported

Effect of different compression ratios (values in brackets show the% change against base in Eq. (6). Total installed cost, considered as capital investment or fixed
case).
capital investment (FCI), is calculated by adding the installed cost of
Parameters Compression ratio (CR)
each piece of equipment such as distillation column, heat-exchanger,
heater, cooler, compressor, and flash. Installed cost, which is based on a
2 2.5 3 bare module cost, is obtained from Aspen ICARUS Process Evaluator.
However, there is no single established method in the literature to
Cumulative condenser −13210 −12978 −12839
duty (kW) (−1.97%) (−3.70%) (−4.73%)
determine the cost of a DWC. Many authors have assumed the diameter
Cumulative reboiler 12817 12568 12421 of the DWC as that of the existing distillation column for its cost as-
duty (kW) (−10.2%) (−11.97%) (−13.01%) sessment with a penalty of 20–50% due to the additional cost of the
Utility cost (million 3.96 (−8.76%) 3.90 (−10.20%) 3.86 (−1.06%) involved complexities [26,27]. In this work, the installed cost of
$/yr)
DWRDC is found by subtracting the cost of C2 installation from the total
TAC (million $/yr) 5.01 (−4.83%) 4.94 (−6.02%) 4.91 (−6.73%)
COM (million $/yr) 22.50 22.42 (−1.77%) 22.37 (−1.97%) installed cost of RD and C2, which is found to have a penalty of about
(−1.43%) 30%. This is consistent with the approach presented in the literature
[26,27]. The operational cost in TAC calculation includes the cost of
utilities (i.e. steam, cooling water, and electricity) as is reported in [28].
% Energy saving
Utility in base case process − Utility in modified process Capital investment
= x100 TAC = Operating cost +
Utility in base case process (5) Pay back period (6)

COM, which includes direct manufacturing cost, fixed manu-

facturing cost, and general expenses, is calculated with Eq. (7). This
4.2. Economic analysis includes a 10% depreciation of FCI. The components of COM are: FCI,
labor cost, raw material cost, and utilities cost. Labor cost is determined
To assess the relative economics of different arrangements, it is based on Eqs. (8)–(10), as reported in Turton et al. [29].
pertinent to consider both capital and operational costs. Therefore, the

Fig. 9. (a) TAC vs CR and (b) COM vs CR, for C1-VR.

22.54
COM (million $)

4.98
22.47

4.92
22.40

4.86 22.33
2.0 2.5 3.0

209
S. Sharma et al. Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Table 3 respectively [34,35].

Effect of different compression ratios (values in brackets show the% change against the
base case).

Parameters Compression ratio (CR) 5. Results and discussion

2 2.5 3 Firstly, the base case results are provided to validate present si-
mulation. Secondly, results corresponding to DWRDC, vapor re-
Cumulative −12300 −12290 (−8.80%) −12268
condenser duty (−8.73%) (−8.96%) compression in C1, vapor recompression in DWRDC, and vapor re-
(kW) compression in both C1 and DWRDC are presented and analyzed for
Cumulative reboiler 12411 12341(−13.57%) −12267 their merits.
duty (kW) (−13.09%) (−14.09%)
Utility cost (million 3.95 (−8.92%) 3.98 (−8.40%) 3.99 (−8.06%)
$/yr)
TAC (million $/yr) 5.05 (−4.01%) 5.08 (−3.59%) 5.09 (−3.31%) 5.1. Base case results
COM (million $/yr) 22.54 22.57 (−1.10%) 22.59
(−1.22%) (−1.025%) Figs. 6 and 7 show the base case results for liquid composition and
temperature profiles of both RD and C2 columns, respectively. As it can
be noticed from the temperature diagram, the stripping section has a
COM=0.28(FCI) + 2.73(operating labor cost) temperature below 90 °C (that spreads widely to the 30th stage) in the
+ 1.23(cost of utilities + cost of raw material) (7) RD column, whereas the top temperature of the C2 column is about
100 °C. Hence, there is a good potential for heat integration.
NOL = (6.29 + 31.7P2 + 0.23Nnp)0.5 (8) It is clear from Fig. 6(a) that the FA composition increases in the
stripping section in RD. FA along with methanol and water coming from
where, Nnp = ∑ all pieces of equipment
RD needs to be separated in another distillation column to reach the
Operating labor = 4.5x NOL (9) 85% FA purity specification. Fig. 7(a) shows top of C2 is rich in me-
thanol, which is then separated and sent back to CSTR for reuse. These
Labor cost = Operating labor × cost for2000 hour year (10)
composition and temperature profiles are consistent with those ob-
where NOL is the number of operators per shift, P is the number of tained by Novita et al. [24]. Cumulative condenser and reboiler duties
processing steps involving particulate solids (which is 0 in current (i.e. for C1, RD, and C2) are −13477 and 14280 kW, respectively. In-
case), Nnp is the number of non-particulate processing steps. stalled cost, utility cost, TAC, and COM are $/yr 3.7, 4.34, 5.27, and
The gross manpower cost for 2000 h/y in 2001 was $52,900. By 22.82 million, respectively. Carbon emissions are 5608 kg of CO2/h.
considering a CEPCI of 623 in 2017 and 394 in 2001, this cost becomes These values are taken as benchmark in Table 4 for further analysis of
$83,647. Cost of raw materials and utilities for 2017 is calculated based performance investigation of process modifications.
on the unit costs given in Table 1. Utility costs are taken from [30].

4.3. CO2 emission analysis 5.2. DWRDC

Another important criterion for the environmental analysis is the
carbon emissions associated with the energy inputs. Carbon emissions
have been a focus of attention due to increasing environmental de-
gradation. Carbon footprint, expressed as kg of CO2 emissions, is an
environmental indicator that expresses quantitatively the effects on the
climate by the greenhouse gases. Low carbon emissions mean that the
involved product gives a low contribution to climate change. The
carbon footprint is a vital environmental indicator that is emerging as a
tool for sustainability, and is increasingly becoming a marketing tool
used by manufacturers to express their commitment to reduce climate
change. Carbon emissions are calculated in this study and we show that
a significant reduction of emitted CO2 is possible with the proposed
process intensifications, as described in Section 5.5. As CO2 is amongst
the major contributors for greenhouse gases, we focus on the CO2
emissions in this study. The carbon dioxide emission indexes of steam
and electrical energy are assumed as 224 kg CO2/t and 51.1 kg CO2/GJ
Fig. 8 shows the liquid composition and temperature profiles of the
RD section (i.e. left section of DWRDC). We can observe that FA is
formed in trays from 2 to 19 (i.e. the reactive trays) at the expense of
MF. The product is stripped in the last bottom stages (roughly in trays
30–35) while lighter component i.e. MF approaches zero towards the
bottom of the column. Negligible amount of CO is present in the
column. FA with mass purity of 85% is produced with a lower heat duty
than in the base case due to the integration of the two columns.
The energy consumption in reboiler for the DWRDC configuration is
11102 kW, which is equivalent to 6.96% energy savings compared to
that in the base case. As presented in Table 4, cumulative condenser
duty and reboiler duty (i.e. for C1, RD, and C2) are −13184 and
13942 kW, respectively representing 2.17% and 2.36% savings. In-
stalled cost, utility cost, TAC, and COM are $/yr 3.5, 4.27, 5.14, and
22.67 million respectively, amounting to 5.41%, 1.65%, 2.35%, and
0.64% savings. Carbon emissions show savings of 2.22%.

Fig. 10. (a) TAC vs CR and (b) COM vs CR, for DWRDC-VR.
(a) 22.60
(b)
COM (million $)

5.08 22.58

22.56

5.06
22.54

2.0 2.5 3.0 2.0 2.5 3.0

210
S. Sharma et al. Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213

Fig. 11. Vapor recompression in both C1 (CR = 3) and DWRDC (CR = 2).

Table 4
Comparison between different cases (values represent% change against base case) [‘−’ indicates decrease and ‘+’ indicates increase in cost].

Parameters Base case [A] DWRDC [B] C1-VR (CR = 3) [C] DWRDC-VR (CR = 2) [D] VRC in both C1 (CR = 3) and DWRDC (CR = 2) [E]

Cumulative condenser duty − −2.17 −4.73 −8.73 −11.29

Cumulative reboiler duty − −2.36 −13.01 −13.09 −24.37
Installed cost − −5.41 +13.5 +19.2 +38.4
Utility cost − −1.65 −11.06 −8.92 −18.33
Carbon Emissions − −2.22 −12.72 −12 −17.68
TAC − −2.35 −6.73 −4.01 −8.4
COM − −0.64 −1.97 −1.23 −2.55

(a) TAC
Fig. 12. (a) Comparative illustration of TAC and

COM (b) COM, and (b) Carbon emissions for cases [A] to [E].
5500
22.5
kg of CO2/h

5000
22.0
6

4500
5

4000
4

5.3. Vapor recompression in C1 (C1-VR)
Liquid composition and temperature profiles of RD (i.e. left section
of DWRDC) with respect to different stages in the proposed C1 vapor
recompression (with DWRDC configuration) are depicted in Fig. S6 of
Supporting material. Significant savings in heat duties of both con-
denser and reboiler are observed with this process intensification.
Table 2 shows a comparison of different CR values. CR is varied
between 2 and 3, a typical operating range for compression considering
the limitation of allowable maximum compressor outlet temperature.
Improved savings are obtained with CR of 3.
Fig. 9(a) and (b) shows that there is a decreasing trend of TAC and
COM for increasing CR values. Increasing CR from 2 to 3 allows de-
creasing the heat duties in both condenser and reboiler with an increase
in the electricity consumption in the compressor. Therefore, it is im-
portant to find an optimal compression condition by considering the
cost for heat energy and electricity. It is observed that the cost of
electricity increases from 18.7 to 20.6 $/h when the CR is changed from

211
S. Sharma et al.
[A]
5.7 [B]
[C]
5.4 [D]
[E]
5.1
4.8
4.5
3.0 3.5 4.0 4.5 5.0
Fig. 13. TAC vs Payback period (CR = 3 for C1-VR and CR = 2 for DWRDC-VR).
2 to 3, respectively. However, the cost of steam reduces from 400 to 388
$/h, and this produces a drop of both TAC and COM. A payback period
of four years is considered in the TAC assessment. The energy con-
sumption in the condenser and reboiler of C1 due to the vapor re-
compression with a CR of 3 shows a reduction of 4.73 and 13.01%
respectively. These savings result in reduced TAC (by 6.73%) and COM
(by 1.97%). Similarly, reduced carbon emissions (by 12.72%) are ob-
served.
5.4. Vapor recompression in DWRDC (DWRDC-VR)
Fig. S7 in Supporting information presents the liquid composition
and temperature profiles of RD section with respect to different stages
of the proposed DWRDC vapor recompression. The changes in the
composition along the stages are clearly illustrated by these profiles.
This configuration allows reaching an 85% FA purity. We observe here
higher energy savings as compared to those of previous configurations.
Table 3 shows a comparison of different CR values. Contrary to C1-
VR configuration, by increasing the CR value, we observed an increase
in utility cost. This is due to the increased electricity consumption in
compressor ‘Comp-3’ (in Fig. 5) from 40.7 kW to 53.94 kW when CR is
varied from 2 to 3. Even though some savings are obtained in condenser
and reboiler duties for higher CRs, a CR of 2 is beneficial due to the
savings in electricity that is more expensive compared to other utilities.
Therefore, a CR value of 2 is considered for final comparison in Section
5.6.
TAC and COM variations with respect to CR are presented in Fig. 10
(a) and (b) respectively. The total energy consumption for RD vapor
recompression in DWRDC configuration is found to be 12,267 kW,
which is equivalent to about 14% compared to the base case. These
savings result in reduced TAC (by 4.01%) and COM (by 1.23%). Simi-
larly, reduced carbon emissions (by 12%) are observed.
5.5. Vapor recompression in both C1 and DWRDC
Subsequently, a study has been conducted to investigate the effect
of vapor recompression in both C1 and DWRDC (see Fig. 11). Compo-
sition and temperature profiles remain similar to those obtained in
above sections; hence they are not presented here for the sake of
brevity. Table 4 reports cumulative condenser and reboiler duties of
−11955 and 10800 kW with 11.29% and 24.37% savings, respectively.
On the other hand, installed cost is found to increase to $5.12 million
i.e. 38.38%. Despite this, savings in the utility cost, TAC, and COM are
calculated as 18.33% ($/yr 3.54 million), 8.4% ($/yr 4.82 million) and
2.55% ($/yr 22.24 million), respectively. Reduction in carbon emission
is found to be 17.68%, which is impressive as it denotes a significantly
reduced environmental impact.
5.6. Comparative analysis among above cases for their economic and
environmental merits
Table 4 compares the merits of process retrofitting against the base
212
Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213
case by focusing on energy, economic, and environmental criteria.
Overall, the energy integration configurations in distillation processes
provide significant energy savings. However, installed cost is higher for
retrofitted cases, namely C1-VR, DWRDC-VR, and combination of both.
For the sake of clarity, only a comparative analysis is conducted to
assess the best case in this section. Clearly, cases [B] to [E] are bene-
ficial as compared to base case [A]. However, maximum savings in
terms of energy, carbon emission, TAC, and COM are obtained for case
[E], where vapor recompression is applied to both C1 and DWRDC.
On the contrary, this case shows higher installation costs (i.e. FCI).
In essence, the benefits of energy savings outperform the increased
installed cost. Case [E] shows considerable savings of 18.33% in utility
cost, 17.68% in carbon emission, 8.4% in TAC, and 2.55% in COM with
respect to the base case. Fig. 12(a) depicts variations of TAC and COM
for cases [A] to [E] from which it is clear that case [E] has the least TAC
as well as COM. Between cases [C] and [D], case [C] provides relatively
better savings. This points out that VRC mechanism is more effective in
C1 than that in DWRDC due to the higher consumption of electricity in
compressor. Similarly, Fig. 12(b) shows the carbon emissions for all five
cases. Although carbon emissions are considerably reduced in case [B]
to [E] against the base case, case [E] shows the maximum saving, and
hence it can be termed as the more environmentally sustainable among
all cases.
Finally, an analysis has been carried out to determine the effect of
considered payback period on TAC. From Fig. 13, it can be seen that
TAC is smaller for a longer payback period. It is clear that all process
modifications provide better performance in terms of TAC as compared
to the base case for payback periods of 3–5 years. However, case [E] is
superior among all cases as it shows the least TAC. Therefore, we can
conclude that FA production using DWRDC with vapor recompression
in both C1 and DWRDC is more beneficial. It can be noticed from
Fig. 13 that the difference in TAC is greater for increasing payback
periods. This is because increased capital investment, say in case [E],
requires a longer period to recover the investment costs.
In summary, significant energy savings are obtained due to the
process intensifications that make them extremely intricate. However,
this extent of process intensification is likely to introduce considerable
challenges in its operation. Controlling inventories and maintaining
product purity may be tough for considerable disturbances and
throughput changes. Additional controllers or different control struc-
ture may be warranted as the results of added components such as
compressors and higher coupling of some process variables. Control of
RD itself is demanding as temperature profile along the tray has to be
maintained to obtain optimal conversion and separation. The integra-
tion of two columns and incorporation of vapor recompression may
result in higher challenges for process control, which call for further
future investigation.
6. Conclusions
A new dividing wall reactive distillation column (DWRDC) config-
uration with vapor recompression was applied to enhance the energy
efficiency of the FA production process. A comparative analysis of
modified processes and their conventional counterpart (i.e. the base
case) was presented, assessed, and discussed in terms of multiple per-
formance indicators, namely energy consumption, carbon emissions,
TAC and COM. With the intensified process, a significant amount of
savings in energy, TAC, and COM were obtained as compared to the
base case. Largest savings were obtained for case [E] where vapor re-
compression is carried out in both C1 and DWRDC columns with re-
ductions of 18.33% in utility cost, 17.68% in carbon emission, 8.4% in
TAC, and 2.55% in COM. These savings denote economic and en-
vironmental merits of the proposed new process configuration. Other
advantages are compactness and reduced capital investment. However,
these process modifications may pose tough challenges in process
control, which require further assessment and future research activity.
S. Sharma et al.
Indeed, we are currently assessing the plantwide controllability of the
process as a function of the suggested retrofitting.
Acknowledgements
We gratefully acknowledge Erasmus Mundus (Intact) funding. Also,
we are grateful to Professor G.P. Rangaiah (NUS Singapore) and
Professor Wang (Zhejiang University China) for the insightful discus-
sions in due course of this study.
References
[1] Applications of Formic Acid. http://www.intermediates.basf.com/chemicals/
formic-acid/applications. [Accessed December 16].
[2] O. Jogunola, T. Salmi, J. Wärnå, J.-P. Mikkola, E. Tirronen, Kinetics of methyl
formate hydrolysis in the absence and presence of a complexing agent, Ind. Eng.
Chem. Res. 50 (2011) 267–276.
[3] J.F. Eversole, Manufacture of formic acid. Patent- US2160064 A; 1939.
[4] O.A. Homberg, J.B. Lynn, A.H. Singleton, Formic acid synthesis by lower alkyl
formate – hydrolysis. Patent- US3907884 A; 1975.
[5] J.D. Leonard, Preparation of formic acid by hydrolysis of methyl formate. Patent-
US4299981 A; 1981.
[6] H.-P. Huang, M.-J. Lee, H.-Y. Lee, J.-H. Chen, Formic acid manufacturing system
and method of manufacturing formic acid. Patent- US20120123157 A1, 2010.
[7] A.K. Jana, Heat integrated distillation operation, Appl. Energy 87 (2010)
1477–1494.
[8] E. Díez, P. Langston, G. Ovejero, M.D. Romero, Economic feasibility of heat pumps
in distillation to reduce energy use, Appl. Therm. Eng. 29 (2009) 1216–1223.
[9] M. Errico, B.-G. Rong, G. Tola, I. Turunen, Process intensification for the retrofit of a
multicomponent distillation plant- an industrial case study, Ind. Eng. Chem. Res. 47
(2008) 1975–1980.
[10] J. Shin, S. Lee, J. Lee, M. Lee, Manage risks with dividing-wall column installations
A simple auxiliary configuration and an extensive modeling study can mitigate the
implementation risks of DWCs, Hydrocarb. Process. 90 (2011) 59–62.
[11] Ö. Yildirim, A.A. Kiss, E.Y. Kenig, Dividing wall columns in chemical process in-
dustry: a review on current activities, Sep. Purif. Technol. 80 (2011) 403–417.
[12] N. Asprion, G. Kaibel, Dividing wall columns: fundamentals and recent advances,
Chem. Eng. Process. Process Intensif. 49 (2010) 139–146.
[13] F.I. Gómez-Castro, M.A. Rodríguez-Ángeles, J.G. Segovia-Hernández, S. Hernández,
C. Gutiérrez-Antonio, A. Briones-Ramírez, et al., Analysis of dynamic performance
for multiple dividing wall distillation columns, Ind. Eng. Chem. Res. 52 (2013)
9922–9929.
[14] A.A. Kiss, R.M. Ignat, Enhanced methanol recovery and glycerol separation in
biodiesel production – DWC makes it happen, Appl. Energ. 99 (2012) 146–153.
[15] O. Annakou, P. Mizsey, Rigorous investigation of heat pump assisted distillation,
Heat Recovery Syst. CHP 15 (1995) 241–247.
213
Chemical Engineering & Processing: Process Intensification 123 (2018) 204–213
[16] L.R. Lynd, H.E. Grethlein, Distillation with intermediate heat pumps and optimal
sidestream return, AIChE J. 32 (1986) 1347–1359.
[17] A.K. Jana, A. Mane, Heat pump assisted reactive distillation: wide boiling mixture,
AIChE J. 57 (2011) 3233–3237.
[18] V. Kumar, B. Kiran, A.K. Jana, A.N. Samanta, A novel multistage vapor re-
compression reactive distillation system with intermediate reboilers, AIChE J. 59
(2013) 761–771.
[19] H. Luo, C.S. Bildea, A.A. Kiss, Novel heat-pump-assisted extractive distillation for
bioethanol purification, Ind. Eng. Chem. Res. 54 (2015) 2208–2213.
[20] K. Johri, G.U.B. Babu, A.K. Jana, Performance investigation of a variable speed
vapor recompression reactive batch rectifier, AIChE J. 57 (2011) 3238–3242.
[21] W.L. Luyben, Control of an azeotropic DWC with vapor recompression, Chem. Eng.
Process. Process Intensif. 109 (2016) 114–124.
[22] Aspen Physical Property System, Version Number V8.8, Aspen Technology, 2016
July.
[23] L. Bai, Kinetics of the carbonylation of methanol in a mechanically agitated reactor,
Fuel Energy Abstr. 38 (1997) 120.
[24] F.J. Novita, H.-Y. Lee, M. Lee, Self-heat recuperative dividing wall column for en-
hancing the energy efficiency of the reactive distillation process in the formic acid
production process, Chem. Eng. Process. Process Intensif. 97 (2015) 144–152.
[25] C.X. Wang, Study on hydrolysis of methyl formate into formic acid in a catalytic
distillation column, Gao Xiao Hua Xue Gong Cheng Xue Bao/Journal of Chemical
Engineering of Chinese Universities 20 (2006) 898–903.
[26] D. An, W. Cai, M. Xia, X. Zhang, F. Wang, Design and control of reactive dividing-
wall column for the production of methyl acetate, Chem. Eng. Process. Process
Intensif. 92 (2015) 45–60.
[27] L. Xu, M. Li, X. Yin, X. Yuan, New intensified heat integration of vapor re-
compression assisted dividing wall column, Ind. Eng. Chem. Res. 5 (2017)
2188–2196.
[28] G.U.B. Babu, E.K. Pal, A.K. Jana, An adaptive vapor recompression scheme for a
ternary batch distillation with a side withdrawal, Ind. Eng. Chem. Res. 51 (2012)
4990–4997.
[29] R. Turton, R.C. Bailie, W.B. Whiting, J.A. Shaeiwitz, Analysis, Synthesis and Design
of Chemical Processes, Pearson Education, 2008.
[30] D.S. Patle, S. Sharma, Z. Ahmad, G.P. Rangaiah, Multi-objective optimization of two
alkali catalyzed processes for biodiesel from waste cooking oil, Energy Convers.
Manage. 85 (2014) 361–372.
[31] https://www.methanex.com/our-business/pricing. [Accessed January 2017].
[32] https://cleantechnica.com/2012/04/10/zero-carbon-cement-production-with-
solar-thermal/. [Accessed 17 January].
[33] http://www.icis.com/resources/news/2006/07/26/2015258/chemical-profile-
formic-acid/. [Accessed 17 January].
[34] A.O. Oni, D.A. Fadare, M.A. Waheed, A. Adewumi, O.J. Adejobi, M.A. Sulaiman,
Exergetic assessment of a crude oil distillation plant, Proceeding of the Science,
Engineering and Technology Conference, OSUN State University, 2011, pp.
151–158.
[35] M.A. Waheed, A.O. Oni, S.B. Adejuyigbe, B.A. Adewumi, D.A. Fadare, Performance
enhancement of vapor recompression heat pump, Appl. Energ. 114 (2014) 69–79.