Screening of potential complexing agents in methyl formate hydrolysis

Article abstract of DOI:10.1016/j.molliq.2014.04.007

More than 70 individual compounds, mainly amines, natural products, organophosphates and sulfur-containing organic compounds, deemed as having potentially suitable complexing strength were screened in order to determine their ability to improve the equilibrium conversion of methyl formate hydrolysis. In fact, alkylphosphates, natural products and organosulfur compounds gave lower methyl formate conversions. After careful screening and investigation of the different classes of amines, aromatic heterocyclic and some tertiary amines were chosen for further investigation. However, it was discovered that imidazoles and substituted piperazines came closest to fulfilling most of the selection criteria as a good complexing agent due to their ability to form a reversible ion pair with the acid product, in addition to their high boiling point, low nucleophilic strength, less bulkiness and thermal stability. A general procedure was proposed for the screening of potential complexing agents. This approach was proven as a rapid one, consumed fewer chemicals and could be used to estimate the complexation power of different candidate compounds in the hydrolysis-complexation process when accurate calculation tools are not available.

Full text of DOI:10.1016/j.molliq.2014.04.007

Journal of Molecular Liquids 196 (2014) 334–339
Contents lists available at ScienceDirect
Journal of Molecular Liquids
journal homepage: www.elsevier.com/locate/molliq
Screening of potential complexing agents in methyl formate hydrolysis
a
Olatunde Jogunola a, , Tapio Salmi , Jyri-Pekka Mikkola
⁎
a,b,⁎⁎
a
Laboratory of Industrial Chemistry and Reaction Engineering, Process Chemistry Centre, Åbo Akademi University, FI-20500 Åbo/Turku, Finland
Technical Chemistry, Department of Chemistry, Chemical and Biochemical Centre, Umeå University, SE-90187 Umeå, Sweden
b
a r t i c l e i n f o
a b s t r a c t
Article history:
More than 70 individual compounds, mainly amines, natural products, organophosphates and sulfur-containing
organic compounds, deemed as having potentially suitable complexing strength were screened in order to
determine their ability to improve the equilibrium conversion of methyl formate hydrolysis. In fact,
alkylphosphates, natural products and organosulfur compounds gave lower methyl formate conversions.
After careful screening and investigation of the different classes of amines, aromatic heterocyclic and
some tertiary amines were chosen for further investigation. However, it was discovered that imidazoles
and substituted piperazines came closest to fulfilling most of the selection criteria as a good complexing
agent due to their ability to form a reversible ion pair with the acid product, in addition to their high boiling
point, low nucleophilic strength, less bulkiness and thermal stability. A general procedure was proposed for
the screening of potential complexing agents. This approach was proven as a rapid one, consumed fewer
chemicals and could be used to estimate the complexation power of different candidate compounds in
the hydrolysis–complexation process when accurate calculation tools are not available.
Received 23 December 2013
Received in revised form 3 April 2014
Accepted 9 April 2014
Available online 24 April 2014
Keywords:
Screening
Complexing agents
Complexation power
Decomposable ion pair
Imidazoles
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
The complexing agent, present as an initial charge, starts to capture
the acid as more H+ accumulates in the solution. The complexing
The relatively low equilibrium conversion of industrially applied
reactions such as esterification, etherification and ester hydrolysis
has been a bottleneck for a long time. These processes are often slow,
reversible and thus inevitably limited by chemical equilibrium.
In essence, the equilibrium can be improved by increasing the concen-
tration of one of the reactants, but this approach does come with its
own limitations. An example is the hydrolysis of methyl formate
(MeFo), which accounts for more than 70% of the global formic acid pro-
duction process [1]. In this case, an excess amount of water is added to
drive the equilibrium forward, with the associated problem of finding
an energy-efficient way of removing the water [2]. Alternatively, the
formic acid product can be extracted from the reaction mixture to
another phase with N-formylmorpholine and subsequently distil from
the solvent [3].
Another approach to process intensification is to shift the equilibri-
um towards the products by adding a complexing agent (X), which
interacts with one of the products [4]. The reacting molecules start to
feel the system as an irreversible one and consequently, higher amounts
of products are formed. In effect, as the formic acid is produced from
methyl formate hydrolysis, it tends to dissociate and H+ is produced.
agent, in fact serves a dual role: it catalyzes the reaction slightly and
improves the equilibrium conversion. The reaction is a two-step
process, which involves the slow, endothermic hydrolysis reaction and
the fast, exothermic complexation step.
The complexing agent should not interact with the reactants but
form a relatively strong reversible liquid adduct (ion pair) with the
carboxylic acid. This, in turn, will enhance the easy recovery of
the acid and facilitate a more straightforward recycling of the organic
base — a key feature of an industrially feasible process. The choice of a
suitable ‘X’ is vital in terms of the whole process since an ‘X’ with a
too strong complexing power results in difficulties upon distillation,
causing an extra load of product to circulate in the process. Likewise, a
too weak X is incapable of shifting the equilibrium. In the absence of a
complexing agent and with a water-to-MeFo initial molar ratio of 1.8,
the equilibrium conversion reachable is about 0.3 mol/mol [5].
Therefore, it is expected that the complexing agent should improve
the equilibrium conversion considerably.
Until today, no detailed, systematic study of the complexing ability
of organic bases with regard to methyl formate hydrolysis was
presented in open literature. However, a lot of research efforts have
been devoted to analogous processes such as the extraction of
carboxylic acids, including that of formic acid, taking advantage of
amines [6–9] or phosphorus-bonded, oxygen-containing extractants
[10]. For instance, Hu and Adeyiga investigated the extraction
of formic acid produced from the reaction of sodium formate
with carbon dioxide [11]. Furthermore, Salman and co-workers
⁎
Corresponding author. Tel.: +358 2 2154431.
⁎⁎ Correspondence to: J.-P. Mikkola, Laboratory of Industrial Chemistry and
Reaction Engineering, Process Chemistry Centre, Åbo Akademi University, FI-20500
Åbo/Turku, Finland. Tel.: +358 22154457.
E-mail addresses: jolatund@abo.fi (O. Jogunola), jpmikkol@abo.fi (J.-P. Mikkola).
http://dx.doi.org/10.1016/j.molliq.2014.04.007
0167-7322/© 2014 Elsevier B.V. All rights reserved.

O. Jogunola et al. / Journal of Molecular Liquids 196 (2014) 334–339
335
[12] did a calorimetric study of the reaction between formic acid and
N,N,N′,N′-tetramethylethylenediamine with regard to its thermal
energy storage. The major differences between the extraction
and hydrolysis–complexation processes (HCP) are the absence of
diluent/solvent and the presence of water as one of the reactants in
HCP. The earliest work that described the hydrolysis of methyl
formate in the presence of n-pentylimidazole was published as a
patent [13]. Recently, our group studied the kinetics of both ethyl
and methyl formate hydrolyses using butyl imidazole as the complexing
agent [4].
the chemical analysis. In order to estimate the complexing power of the
‘X’, the concept of apparent equilibrium constant, K1,app was introduced
to describe the new value obtained for K1 of the HCP when X was added
as the initial charge into the hydrolysis solution. In the HCP, X picked
up C, resulting in an increase in the yield of D as more A and B reacted
in contrast to the hydrolysis process, where the yield of D is less
under the same reaction conditions. The concentration of the other
components was based on the analysis of the component ‘D’. Therefore,
the apparent equilibrium constant when activity coefficients were
assumed to be unity, resulted in an expression:
The aim of this work is to draw attention to this approach of
screening different organic bases and determining their capabilities
to improve the methyl formate conversion, while using a stoichiometric
amount of water. The procedure is rapid and consumes fewer
chemicals. The description, nature and structure of the adduct
formed between formic acid and the complexing agent are beyond
the scope of this work.
CꢀCCꢀD
K1;app
¼
ð4Þ
CꢀACꢀB
where C* represents the new concentration due to the presence of a
complexing agent in the solution. The fraction of the free formic acid
in the solution may be represented as
2. Complexation equilibria
CC
K1
f ¼
¼
CTOT K1;app
In the current work, the reaction system, i.e. the hydrolysis and
complexation steps can be represented by the following steps,
⁎
where (CC = CTOT).
The equilibrium constant of the complexation step can be
expressed as
HCOOCH3ðAÞ þ H2OðBÞ sl⇌ow HCOOHðCÞ þ CH3OHðDÞ
HCOOHðCÞ þ αX ra⇌pid HCOO−XHþðCXÞ
ðIÞ
f
K2
¼
α
CꢀXð1−fÞð1−fqÞ
ðIIÞ
CTOT
CXꢀ
⁎
where α is the stoichiometric coefficient, which in practice might not
approach unity. The complexation step (II) is the first reaction that
will occur in the presence of a base since it involves the transfer of
proton. When α is less than one, it implies that more than one acid
per X is formed by the complex. The equilibria of Reactions (I) and (II)
respectively, if the activity coefficients are approximated to unity, can
be described as
where CX = CCX + CX and q ¼
.
⁎
⁎
When q → 1, CX = CCX ≈ CC , therefore the value of K2 can be
estimated from the experimental results:
f
K2
¼
:
αþ1
CꢀCð1−fÞ
CCCD
CACB
3. Experimental section
K1
K2
¼
¼
ð1Þ
ð2Þ
3.1. Pre-screening stage
CCX
CCCαX
:
The experimental work consisted of three parts namely the
pre-screening, screening and the stability tests. The purpose of
the pre-screening was to set selection criteria for the actual screening
stage. The pre-screening was performed in about 6.3 ml scale, air-
tight isobaric mini-reactors (autoclave-type ‘bomb’ reactors). These
reactors (Fig. 1) are made of stainless steel and can withstand high
pressures and temperatures. In the preliminary study, more than 70
organic bases were tested and the pre-screening involved measuring
In the absence of X, the equilibrium conversion of the hydrolysis
step, K1 = 0.2 at H2O/MeFo molar ratio of 1.8 [5]. If the total concentra-
tion of acid in the system, CTOT = CCX + CC and f = mole of acid bound
per mole of total acid (i.e. CCX/CTOT), then
CCX
K2
K2
¼
¼
; which is simplified to
ðCTOT−CCXÞCαX
f
for which fb1:
ð3Þ
CαX ð1−fÞ
The equilibrium constant of the complex formation is a function of f,
which depends on the nature of the complexing agent and the acid–X
interaction, which in turn depends on the initial charge of the
complexing agent. It can be deduced from Eq. (3) that K2 approaches
its maximum value when there is no free acid (CC → 0) and K2 is
small when the proportion of the formic acid that formed the complex
is small. When there is no additive in the system, CC approaches its
maximum value, determined by Equilibrium (I).
It was difficult to get a reliable chemical analysis method for C, X, and
CX in the hydrolysis solution and A was volatile, thus, D (methanol) was
selected as the key component since it was the most reliable one during
Fig. 1. Reactors used for hydrolysis–complexation process (pre-screening stage).

336
O. Jogunola et al. / Journal of Molecular Liquids 196 (2014) 334–339
the exact amount of methyl formate, water and the compound ‘X’ into
these reactors. The 6–8 reactors running simultaneously with a liquid
volume of about 3.5 ml were closed, placed in a heating unit and heated
up to 100 °C. A molar ratio of ester-to-water-to-X of 1:1.8:0.3 was
retained throughout the experimental matrix. It is noteworthy to
mention here that a reference compound of known complexing ability
was always tested alongside other compounds in each experiment.
The reactors were shaken intermittently to improve mixing. The re-
action was allowed to proceed for 5 h and quenched by putting the
reactors in a refrigerator and the sample was analyzed with the aid
of gas chromatography [4]. The results obtained were compared to
the reference compound. In some cases, gas liberation was observed.
The gas-phase composition could not be analyzed in the pre-screening
stage; it was done during the screening experiments.
possible carcinogenic effects related to the substance. Sulfur-containing
organic compounds (e.g. Taurine and its derivatives) formed slurry with
foul smell and were thus discarded. Therefore, nitrogen heteroatom
compounds were considered to be potentially suitable for our purposes.
About 10 primary amines were investigated and most of them gave
good methyl formate conversion, leading to the formation of formam-
ides (side products), which resulted from the formylation of the amines
by methyl formate. The amides formed are considerably stable, thus it
was difficult to recover the complexing agents and this fact rendered
them inappropriate in terms of process intensification ideal. Even,
butylamine and triethylenetetramine started reacting with the hydroly-
sis mixture at room temperature due to faster collision of the reacting
species. Urea yielded a reasonable conversion but was discarded
because it degraded in the hydrolysis solution, releasing ammonia gas.
Most of the primary amines gave a yellowish solution due to oxidation
in air. Another disadvantage of primary amines is their solubility in
water, which might result to loss of the complexing agent and will
require more energy during separation. Henceforth, none of the primary
amines was found to be ideal for the formation of a suitable ion pair.
Therefore, they were discarded at the pre-screening stage.
About a dozen of secondary amines were screened and only three
of them did not result in a reasonable methyl formate conversion.
These were formylpiperidine, butylformamide and 1,3-diethylurea.
This observation was, indeed, expected due to the poor availability of
the nitrogen atoms, since the carbonyl carbon was already carrying a
partial positive charge [14]. The increase in conversion of other second-
ary amines could be attributed to the fact that they were considered to
be more basic than primary amines because the inductive effect of the
alkyl group is more pronounced and they were all well solvated [15].
However, secondary amines without steric hindrance (availability of
nitrogen atoms) such as dipropylamine, diethanolamine and unprotect-
ed piperazine and piperidines underwent side reactions forming mainly
amides and thus their conversions in some cases reached almost 100%.
Compounds such as diisopropylamine and pyrrolidine gave good
MeFo conversion but were considered unsuitable due to their low
boiling points. Nevertheless, secondary amines in which the alkyl
groups prevent the nitrogen atoms from attack gave good conversions
without the formation of amide. Therefore, five of them were
considered for the final screening stage.
3.2. Screening stage
The compounds selected from the pre-screening stage underwent
a proper screening in this stage. The process was performed in a
well-stirred 500 ml reactor [4] at 130 °C using a molar ratio of
MeFo-to-H2O-to-X of 1:1.8:0.5. The liquid mass was about 100 g.
The reaction was allowed to proceed until equilibrium was attained.
The samples were collected by passing them through an ice bath and
analyzed by means of a gas chromatography [4]. It is pertinent to
mention that the experimental results obtained here are based on
methanol analysis only, due to the volatility of methyl formate and
the difficulty in getting a reliable analytical method for formic acid
under these conditions. The experimental matrix in the absence of
the additive ‘X’ is depicted in Table 1.
3.3. Stability test
At this stage, the compounds selected in the screening stage were
exposed to a higher temperature. The goal was to investigate which
compounds could withstand the higher temperature for a longer time;
i.e. the thermal stability of the complexing agent ‘X’. The process
involved allowing equimolar amounts of the hydrolysis mixture
(including X) to react at 160 °C for 5 h in the reactor used in
Section 3.2. The sample was finally analyzed by means of proton
nuclear magnetic resonance (1H NMR) spectroscopy.
About 30 tertiary amines were studied in the preliminary stage
of this work and more than half of them gave good conversions.
However, compounds such as 1,3-dimethyl-3,4,5,6-tetrahydro-2(1H)-
pyrimidinone, 1,3-dimethyl-2-imidazolidinone, N,N-dimethylformamide
and 1-methyl-2-pyrrolidone were among the groups with the lowest
conversions. Also, substituted ureas such as 1,1,3,3-tetramethylurea,
1,1,3,3-tetraethylurea, and 1,1,3,3-tetrabutylurea yielded lower conver-
sions. Furthermore, they hydrolyzed in the acidic medium, releasing car-
bon dioxide gas. 1,3,5-Trimethylhexahydro-1,3,5-triazine decomposed
during hydrolysis, apart from giving a pungent smell. In general, there
was no trace of amide formation in all the tertiary amines investigated
due to the shielding of the nitrogen atoms by the alkyl groups. However,
the solubility of the complexing agent in the reaction mixture became
an issue due to the presence of more bulky groups. For example,
tributylamine formed two layers, whereas hexamethylenetetraamine
formed crystals upon cooling and it was difficult to analyze the sample.
However, the introduction of a polar group (e.g. \OH), improved the
solubility of the hydrolysis solution but the hydroxyl groups underwent
esterification (e.g. triethanolamine). All in all, seven tertiary amines
were considered worthy for the next stage.
4. Results and discussions
4.1. Pre-screening process
Surprisingly, it was observed that organophosphates such as
trimethyl phosphate and triisopropyl phosphate compounds produced
lower MeFo conversions (0.3 mol/mol). On the other hand, proton
NMR studies of the products of the hydrolysis–complexation process
using triisopropylphosphate as the complexing agent revealed that
trans-esterification of the isopropyl group of the phosphate esters
with the co-product, methanol in the reaction mixture is responsible
for the low conversion. Therefore, these compounds did not pass the
pre-screening stage due to low conversion and safety-related issues.
As an example, trimethyl phosphate, according to European Directive
67/548/EEC is known to be toxic and there is limited evidence of
Table 1
Experimental matrix.
About 15 heterocyclic amines such as imidazoles, pyridines, pyrroles
and pyrazoles were considered for the test. Pyrrole gave the lowest
conversion because it is the least basic. The lone pair of electrons is in
a p orbital contributing to the 6π electrons in the aromatic ring, thus it
is not available for protonation. Furthermore, pyrrole was expected to
be less stable in the hydrolysis mixture. In fact, it decomposed and prob-
ably depolymerized. 1,2,5-Trimethylpyrrole formed ox-blood solution,
Components
C0 (mol/kg)
MeFo (A)
H2O (B)
FA (C)
MeOH (D)
X
9.93
17.89
1.10
1.01
0.00
Temperature = 130 °C.

O. Jogunola et al. / Journal of Molecular Liquids 196 (2014) 334–339
337
Table 2
while 2,4-dimethylpyrrole formed a resultant viscous solution. This
Complexing power of different amines relative to the hydrolysis process.
might be due to acid-catalyzed polymerization of pyrrole [14]. Pyridine,
more basic than pyrrole gave 0.45 mol/mol conversion. However, a lot
of peaks could be seen in the gas chromatogram (GC), indicating the
presence of side reactions. As a matter of fact, this might be due to the
interaction between pyridine and methanol. The nature of the side
products has not yet been ascertained. The conversion was increased
by more than 20% when the hydrogen on position 2 in the pyridine
ring was substituted by a methylaminoethyl group; however the resul-
tant solution was yellowish in color. 3-Methyl pyrazole, more basic than
pyridine but less basic than imidazole, gave good conversion but formed
a brownish solution with the hydrolysis mixture. The coloration might
be due to light. The imidazoles gave the best conversion in this category
by capturing the proton of the acid into the imidazolium moiety, in ad-
dition to possessing good thermal stability at the reaction conditions.
Thus, five imidazoles were recommended for proper screening.
In addition, natural products such as thymine, uracil, xanthine, low
molecular weight chitosan and DL-proline were tested but they were
not suitable due to their low conversion and low solubilities in the hy-
drolysis mixture. Chelating agents such as ethylenediaminetetraacetic
acid, diethylenetriaminepentaacetic acid and nitrilotriacetic acid
formed high melting solids with relatively low conversions. Nitron
(312.4 g/mol), tri-n-octylamine (353.7 g/mol), tridodecylamine
(522 g/mol), 1,5-dicyclohexylimidazole (232.36 g/mol) and 50 wt.%
pentrol solution (393.6 g/mol) in water were discarded based on their
bulkiness, even though most of them gave good MeFo conversion.
Furthermore, ester formation was visible with pentrol.
⁎
Compounds
pKa
XA
K1,app
bpt (°C)
Secondary amines
1. Dicyclohexylamine, l
2. 2,6-Dimethylpiperidine, l
3. 2,6-Dimethylpiperazine, s
4. 2,2,6,6-Tetramethylpiperidine, l
5. N,N′-Diisopropylethylenediamine, l
10.4016
10.9217
9.8618
11.0716
–
0.77
0.75
0.79
0.68
0.72
3.40
2.81
3.80
1.76
2.32
256
127
162
152
169
Tertiary amines
1,4-Diethylpiperazine, l
1,2,4-Trimethylpiperazine, s
Diazabicyclo[2.2.2]octane, s
N,N,N,N-Tetramethylethylenediamine, l
1,2,2,6,6-Pentamethylpiperidine, l
8.3819
8.3618
8.80
0.65
0.64
0.7
0.7
0.65
1.43
1.38
2.02
1.99
1.48
157
150
174
120
187
8.26
11.2516
Heterocyclic amines
1-Propyl imidazole, 1
1,2-Dimethylimidazole, s
Butylimidazole, l (reference)
2-Ethylimidazole, l
7.2220
8.2120
7.2120
7.2620
7.2120
0.60
0.65
0.64
0.60
0.50
1.05
1.40
1.35
1.03
0.57
221
204
235
268
198
1-Methylimidazole, l
X/MeFo = 0.5, H2O/MeFo = 1.8, XA = conversion at 130 °C, l = liquid, and s = solid.
reactants. Water might play an inhibitory role in this case [23], since
the acid is not present as ‘neat’ formic acid. We observed that it was dif-
ficult to distil 2,6-dimethylpiperidine (XA = 0.75) from the formate–
piperidinium adduct, thus rendering these compounds non-relevant.
The inability to recover the 2,6-dimethylpiperidine from the adduct
might be due to dehydration of the formate–piperidinium adduct to
amide under very harsh condition during distillation. In addition, both
2,6-dimethylpiperidine and 2,6-dimethylpiperazine were discarded
due to their high flammability, thus they were considered as industrial
unsafe substances. In addition, both dicyclohexylamine and N,N′-
diisopropylethylenediamine are sensitive to air and therefore, should
be neglected. When 2,2,6,6-tetramethylpiperidine (TMP) was mixed
with a low concentration of formic acid (10–35 wt.% FA), it formed a
crystalline salt (melting point = 115 °C), liberating much heat in the
process. It is possible that recycling of TMP might be a major challenge.
Furthermore, according to the literature, 2,2,6,6-tetramethylpiperidine
is toxic and flammable. It is pertinent to state that the toxicity and flam-
mability of the substances investigated were based on Sigma-Aldrich
Material Safety data. Nevertheless, secondary amines with steric hin-
drances performed excellently in the hydrolysis process (gave the best
conversions), but they were not considered suitable candidates in our
attempt towards reaction intensification due to the formation of very
stable complexes with formic acid.
Henceforth, after some painstaking investigation and observation,
we came up with some selection criteria for a suitable X to narrow
down the search: it should have low molecular weight so as to take
less space in the reactor, thereby increasing the molar capacity of
the X; high boiling point to facilitate easy separation of formic acid
and X; low nucleophilic strength to reduce the likelihood of side re-
actions and stronger HCOOH–X complex and good thermal stability to
guarantee better regeneration during distillation.
4.2. Screening process
About 20 compounds were screened at this stage based on the
selection criteria set in the preliminary study. No primary amines
were chosen; some secondary amines with steric hindrance were con-
sidered, including 7 tertiary amines and 5 imidazoles. The complexing
ability of the chosen amines is likely controlled by the relative availabil-
ity of the unshared electrons on the nitrogen atom to proton donor and
by the stabilization of the positively charged nitrogen atom by solvation,
which in turn depends on the pKa of the conjugate acid. That is, the
complexation equilibrium depends on the amine pKa. We expect that
a good ‘X’ will have a low gas-phase pKa and a high solution phase
pKa, which is high in water but low in formic acid. The pKa should be
in the range of 6–11 in water. The pKa of these compounds are obtained
from references [16–20]. For the purpose of the discussion and compar-
ison, the results of the screening process are categorized into three main
classes such as secondary, tertiary and heterocyclic amines. The equilib-
rium constant K1 (Eq. (1)) for the hydrolysis reaction in the absence of a
complexing agent is around 0.20. Some of the results of our investiga-
tion are summarized in Table 2.
4.2.2. Tertiary amines
Among the tertiary amines screened, 1,2,2,6,6-pentamethylpiperidine
(PMP) gave the lowest conversion but the conversion was improved with
higher water concentration. It was discarded due to its toxicity and its
tendency to form crystalline salt (melting point = 50 °C) when heated
with low concentration of formic acid and allowed to cool to room
temperature. Also, substances such as tripropylamine, and quinuclidine
were discarded due to their toxicity even though they gave better conver-
sion. In addition, N,N,N,N-tetramethylethylenediamine and diazabicyclo
[2.2.2]octane gave good conversions but are considered highly flamma-
ble. Another potential threat to the recovery of tertiary amines is the
Hoffman elimination in multiple recycling. After careful consideration,
1,4-diethylpiperazine and 1,2,4-trimethylpiperazine were considered
for further testing.
4.2.1. Secondary amines
Under the reaction conditions studied, the secondary amines
investigated at this stage gave, indeed, very good MeFo conversions.
This was, in fact, evidently proven true by the high pKa and complexa-
tion power values. The formation of amide was not detected; this
might be partly due to steric factor [21,22]. Therefore, it is likely that a
thermally stable formic acid–amine complex was formed under the
reaction conditions. Another possibility that might prevent or slow
down the formation of amide is the presence of water as one of the
4.2.3. Heterocyclic amines
All the imidazoles investigated formed thermally decomposable
ion-pair complexes with formic acid, leading to significant conversions.
Protonation creates a stable aromatic cation with low pKa and a
loose ion pair. 1,2-Dimethylimidazole gave the best conversion,

338
O. Jogunola et al. / Journal of Molecular Liquids 196 (2014) 334–339
while 1-methylimidazole gave the least conversion. Alkylimidazoles are
sparingly soluble in water and the solubility decreases as more alkyl
groups are added to the pyrrolic nitrogen atom. For example, 2-
ethylimidazole is soluble in water in all proportion, while 1,2-
dimethylimidazole is hygroscopic and has an affinity for water.
Thus, more energy will be required during deprotonation and sepa-
ration to regenerate these compounds. The toxicity of these com-
pounds decreases as the alkyl chain increases. Butylimidazole
(BIM) and propylimidazole (PIM) were considered to be among the
best choices in terms of their sparing solubility in water and higher
boiling point.
and is not too flammable and toxic to humans. All in all, the flamma-
bility and toxicity of the candidate compound should be of utmost
important for good laboratory and industrial practices and these
properties should not be compromised at all.
In the absence of theoretical calculating tools, Table 3 should be
considered as a guide for the selection of a good complexing agent.
In our investigation, it was observed that a complexing agent with
XA N 0.8 at a MeFo:X:H2O molar ratio of 1:0.5:2 might be difficult
to regenerate from the hydrolysis–complexation solution. Compounds
with XA = 0.4–0.8 should be considered a good candidate. However,
if XA b 0.4, the compound should be discarded and no further test
should be performed on it. After successful screening, an equimolar
amount of the hydrolysis–complexation solution should be subjected
to higher temperature (160–180 °C) to ascertain the thermal stabil-
ity of the base. Finally, the base should be recovered from the
HCOOH–X adduct by distillation. A higher temperature will facilitate
an easy recovery of the complexing agent. However, distillation is an
energy-intensive process and alternative technologies such as mem-
brane separation processes, with associated cost savings, should be
considered.
Alternatively, a simple concept might be considered by mixing the
compound ‘X’ and HCOOH in a desired ratio and placing the samples
in small uncapped vials in a vacuum oven for a certain period. Potential
candidates might be identified by comparing the relative weight losses
of all the samples as a function of time with respect to that of HCOOH–X
mixture. It is not out of place to try out a mixture of bases in order to en-
hance the complexing power/molar mass ratio. However, the endeavor
is to take advantage of natural species as potential candidates to reduce
the risk of toxicity, assumed biodegradability and affordable cost.
4.3. Thermal stability test
In the final test, butylimizadole, 1,4-diethylpiperazine and 1,2,4-
trimethylpiperazine were all subjected to stability tests. The process
involved allowing an equimolar amount of the hydrolysis mixture
(including the component ‘X’) to react at 160 °C for 5 h. This approach
was taken to reveal which chemical structures have the necessary
thermal stability, which is needed when aiming at the regeneration of
compound ‘X’ and recovery of the formic acid. It is important to note
that one test method for thermal stability might not be good enough.
1,4-Diethylpiperazine decomposed and formed by-product, as con-
firmed by 1H NMR. The selective elimination of the ethyl group is of con-
siderable interest. The product was ethylated and as a result, the ethyl
groups were visible in other positions. However, both imidazoles and
1,2,4-trimethylpiperazine did not give rise to any side product. For an
imidazole species, the situation is unique: its structure is stabilized
by resonance and only substituents that enhance the resonance stabili-
zation are present [24]. However, the imidazoles are preferred to 1,2,4-
trimethylpiperazine (boiling point = 150 °C) due to their high boiling
point. They have the potential for extensive intermolecular hydrogen
bonding and a stronger dipolar association emanating from very
marked permanent charge separation [15]. In general, the properties
of a good complexing agent for methyl formate hydrolysis are
summarized in Table 3.
5. Conclusions
Most of the compounds screened for their ability to complex formic
acid were eliminated due to their high toxicity, flammability, bulkiness
and low boiling point. Primary and secondary amines without steric
hindrances formed side products, mostly formamides. However, sec-
ondary amines with steric hindrances gave the best conversions but
formed formic acid–amine adduct, which dehydrated to amide under
the very harsh conditions encountered during distillation. Therefore,
it was difficult to separate the formic acid from the solution and thus
regenerate the complexing agent. Tertiary amines were among the
most flammable and toxic compounds tested even though they have
good complexing capabilities. Nevertheless, 1,2,4-trimethylpiperazine
is a potential candidate. Most compounds tested in the screening
stage improved methyl formate conversion, although both butyl
and propyl imidazoles were considered to be the most suitable
compounds because they fulfilled nearly all the selection criteria
for a good complexing agent.
4.4. General methodology
The whole screening procedures were based on the reaction steps,
toxicity and flammability of the compounds and in some cases the
regeneration of the base. The choice of a good complexing agent in-
volved the interplay of the complexation power, the recyclability and
the health impacts. It is always a challenge to get a compound that
will fulfill all the selection criteria. For example, additives with better
complexing ability were difficult to regenerate, while those with good
complexing power and regenerability were flammable or/and toxic to
human or too bulky. There could be a compromise between complexing
ability and bulkiness if the candidate compound could be regenerated
A general methodology was proposed for the selection of a potential
complexing agent. The methodology is based on the pre-screening of
suitable complexing agents based on their physico-chemical proper-
ties (Table 3), screening of the selected candidates based on their
complexing power, determination of the thermal stability of the
choice complexing agents and investigation of the recovery process
for both the formic acid and the complexing agent from the ionic-liquid
like ion pair.
Table 3
Physico-chemical properties of a good complexing agent.
Properties
Molar mass
Melting point
Boiling point
50–200 g/mol
b200 °C
200–300 °C
Decomposition temperature
Flash point
Density
Solubility
Inertia (no side reaction)
Toxicity
Complexation power @
X:H2O:MeFo = 0.5:2:1
Diffusibility
200–350 °C
N40 °C
0.7–1.2 g/cm3
Sparing soluble
1° and 2° amines excluded
Non-toxic-irritant (Xi) R37/38, R41
XA = 0.4–0.8,
K1,app = 0.5–5.0
10−4–10−7 cm2/s
5–9
Acknowledgments
This work is a part of the activities at the Åbo Akademi Process
Chemistry Centre (PCC), a center of excellence financed by Åbo
Akademi. Financial support from the Academy of Finland, the Kempe
Foundations, the Knut and Alice Wallenberg Foundation, the Bio4Energy
Programme and the Kemira Oyj is gratefully acknowledged.
pKa range

O. Jogunola et al. / Journal of Molecular Liquids 196 (2014) 334–339
339
[12] M. Salman, Y. Girault, N. Balbi, L. Elegant, J. Therm. Anal. 37 (1997) 2389–2394.
References
[13] H. Hohenschutz, J.E. Schmidt, H. Kiefer, Process for Manufacturing Formic Acid.
European Patent 1432 to BASF, April 1979 (Translated From German to English).
[14] T. Eicher, S. Hauptmann, The Chemistry of Heterocycles: Structure, Reactions,
Syntheses and Applications, Second Ed. Wiley-VCH Verlag GmbH & Co. KGaA,
Weinheim, 2003.
[15] J.A. Joule, K. Mills, Heterocyclic Chemistry, Fifth Ed. Wiley Publisher, U.K., 2010
[16] D.R. Lide, CRC Handbook of Chemistry and Physics, Eighty Ninth Ed. CRC Press Inc.,
Boca Raton, FL, 2008.
[1] S.N. Bizzari, M. Blagoev, Chemical Economics Handbook Marketing Research Report:
Formic Acid, SRI Consulting: USA, June 2010.
[2] Ullmann, Encyclopaedia of Industrial Chemistry, Sixth Ed. Wiley-VCH,
Weinheim, 2003.
[3] H. Buelow, H. Hohenschutz, J. Schmidt, W. Sachsze, Purification of Formic Acid by
Extractive Distillation, United States Patent 4076594 to BASF, Feb. 28, 1978.
[4] O. Jogunola, T. Salmi, J. Wärnå, J.-P. Mikkola, E. Tirronen, Ind. Eng. Chem. Res. 50
(2011) 267–276.
[5] O. Jogunola, T. Salmi, K. Eränen, J. Wärnå, M. Kangas, J.-P. Mikkola, Ind. Eng. Chem.
Res. 49 (2010) 4099–4106.
[6] J.A. Tamada, A.S. Kertes, C.J. King, Ind. Eng. Chem. Res. 29 (1990) 1319–1326.
[7] Y.K. Hong, W.H. Hong, D.H. Han, Biotechnol. Bioprocess. Eng. 6 (2011) 386–394.
[8] S.-T. Yang, S.A. White, S.-T. Hau, Ind. Eng. Chem. Res. 30 (1991) 1335–1342.
[9] R. Canari, A.M. Eyal, Ind. Eng. Chem. Res. 42 (2003) 1285–1292.
[10] S. Sahin, S.S. Bayazit, M. Bilgin, I. Inci, J. Chem. Eng. Data 55 (2010) 1519–1522.
[11] L. Hu, A.A. Adeyiga, Ind. Eng. Chem. Res. 36 (1997) 2375–2379.
[17] H.K. Jr., Hall Jr., J. Am. Chem. Soc. 79 (1957) 5439–5441.
[18] D.A. Keyworth, J. Org. Chem. 24 (1959) 1355–1356.
[19] F. Khalili, A. Henni, A.L. East, J. Chem. Eng. Data 54 (2009) 2914–2917.
[20] B. Lenarcik, P. Ojczenasz, J. Heterocycl. Chem. 39 (2002) 287–290.
[21] P.G. Jessop, Y. Hsiao, T. Ikariya, R. Noyori, J. Am. Chem. Soc. 118 (1996) 344–355.
[22] H. Charville, D.A. Jackson, G. Hodges, A. Whiting, M.R. Wilson, European J. Org.
Chem. (2011) 5981–5990.
[23] K. Arnold, A.S. Batsanov, B. Davies, A. Whiting, Green Chem. 10 (2008) 124–134.
[24] I.B. Johns, E.A. McElhill, J.O. Smith, Ind. Eng. Chem. Prod. Res. Dev. 1 (1962) 2–6.