Valeronitrile hydrolysis in supercritical water

Article abstract of DOI:10.1002/cssc.200900154

Nitriles act as important intermediates for the chemical industry and are accessible on a large scale through hydrocyanation or ammonoxidation. Nitrile hydrolysis yields amides and acids used in various applications. The conventional nitrile hydrolysis process relies on stoichiometric amounts of mineral acid or base which inherently deliver great amounts of waste brine. Improving this process towards green chemistry would require reaction conditions which can provide technically significant results without the use of catalysts. Under these conditions, the hydrolysis of valeronitrile in pure supercritical water was investigated. The experiments were performed in a continuous high pressure laboratory-scale apparatus at a temperature between 400 and 5008C, 30 MPa pressure and maximal residence time of 100 s. Nitrile conversion and valeric acid selectivity greater than 90% were achieved.

Full text of DOI:10.1002/cssc.200900154

DOI: 10.1002/cssc.200900154
Valeronitrile Hydrolysis in Supercritical Water
Michael Sarlea, Sabine Kohl, Nina Blickhan, and Herbert Vogel*[a]
Nitriles act as important intermediates for the chemical indus-
try and are accessible on a large scale through hydrocyanation
or ammonoxidation. Nitrile hydrolysis yields amides and acids
used in various applications. The conventional nitrile hydrolysis
process relies on stoichiometric amounts of mineral acid or
base which inherently deliver great amounts of waste brine.
Improving this process towards green chemistry would require
reaction conditions which can provide technically significant
results without the use of catalysts. Under these conditions,
the hydrolysis of valeronitrile in pure supercritical water was in-
vestigated. The experiments were performed in a continuous
high pressure laboratory-scale apparatus at a temperature be-
tween 400 and 5008C, 30 MPa pressure and maximal residence
time of 100 s. Nitrile conversion and valeric acid selectivity
greater than 90% were achieved.
Introduction
Results and Discussion
In the past few years sustainability has gained in importance in
the chemical industry. Supercritical (sc) fluids have awakened
interest as new, environmental friendly solvents and processes
and methods of synthesis have been evaluated.[1–3] Sc fluids
and especially water offer many advantages in comparison to
ordinary organic solvents.[4,5] Supercritical water is cheap, non-
flammable, and environmentally friendly. Furthermore, it poses
no health threat and is easily available worldwide .[6] The out-
standing characteristic of water as a reaction medium is the
possibility of tuning properties, such as polarity or ionic prod-
uct, by changing temperature and pressure.[7–10]
Kinetic tests on valeronitrile, valeramide and valeric acid were
performed at temperatures between 400–5008C, 30 MPa pres-
sure, and maximal residence time of 100 s. The collected data
and findings are presented below.
Usually hydrolysis reactions under hydrothermal conditions
are performed at near-critical temperature to take full account
of the high dissociation rate and thus high catalytic potential
exhibited by water. However, valeronitrile demonstrates a fair
level of temperature stability and therefore higher tempera-
tures had to be applied for the kinetic investigations.
The reaction mechanism is well-known and described in lit-
erature.[20–22] Nitriles react with water yielding amide intermedi-
ates which react in a second step to form the corresponding
carboxylic acid (Scheme 1).
Nitriles are of great interest with regard to sustainability
both because of their appearance in industrial waste-streams
and their familiar usage as starting materials for commodity
chemicals production.[11,12] The default hydrolysis process is
based on the addition of strong acids, bases, or enzymes. A
good example is the biotechnological hydrolysis of acryloni-
trile, which shows high selectivity under mild reaction condi-
tions (20–408C, pH 7, ambient pressure).[13,14] The deficiency of
the acid or base-dependant process is the high amount of
waste brine which results during the extraction of the desired
products.[15,16] This disadvantage can be circumvented by the
use of supercritical water which serves as a catalyst, solvent,
and reactant.[17–19]
Under the aforementioned experimental conditions, valeric
acid can undergo a decarboxylation reaction to butane.[23,24]
The goal of this work was the confirmation of the reaction
network suggested in literature by hydrolyzing both valeroni-
trile and the intermediate valeramide.[20] Furthermore, optimal
reaction conditions for maximizing the valeric acid yield were
determined, which are significant for potential industrial appli-
cation. This has proven difficult due to product consumption
by decarboxylation reactions. Hence the stability of the main
product, valeric acid, was investigated. Kinetic results were
gathered and compared with modeled findings provided by
the program Presto-Kinetics.
Scheme 1. Hydrolysis of valeronitrile to valeric acid via the intermediate va-
leramide.
[a] M. Sarlea, S. Kohl, N. Blickhan, Prof. H. Vogel
Ernst-Berl-Institut fꢀr Technische und Makromolekulare Chemie
Technische Universitꢁt Darmstadt
Petersenstrasse 20, 64287 Darmstadt (Germany)
Fax: (+49)6151-163465
E-mail: vogel@ct.chemie.tu-darmstadt.de
ChemSusChem 2010, 3, 85 – 90
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
85

H. Vogel et al.
This consecutive reaction lowers the valeric acid yield
(Scheme 2).
An aged reactor was used for the experiments in order to
avoid potential catalytic wall effects. Particularly decarboxyla-
Scheme 2. Decarboxylation of valeric acid to butane and CO2.
tion reactions can be enhanced by the high nickel content in
the alloys used for high pressure equipment.[25] During a po-
tential scale-up, the ratio of reactor volume to reactor wall sur-
face area will decrease substantially. Assuming catalytic wall ef-
fects, scale-up would provide even higher valeric acid yields
due to slower decarboxylation reactions.
Figure 2. Valeramide/valeric acid selectivity versus residence time (T=4008C
(&), 4258C (*), 4508C (~), 4758C (!), 5008C (^), p=30 MPa,
c=0.07 molLꢀ1 valeronitrile in water).
Valeronitrile hydrolysis
Yields of valeric acid higher than 80% resulted at 4508C and
one minute residence time (Figure 3). For reaction pathway
analysis, the first-rank Delplot method (selectivity versus con-
version plots) was applied, which allows separation of primary
from nonprimary products.[26]
Valeronitrile conversion was investigated at temperatures
higher than 4008C. The reaction rate increased as expected
with higher temperatures. Above 4758C, the rate of increase
became smaller, which is consistent with the course of water
dissociation (at 30 MPa, the pKw measures 14.8 at 4008C, goes
up to 19.7 at 4508C, and proceeds to 20.9 at 5008C). Despite
the considerably greater pKw difference at lower temperatures,
no impact on reaction rate increase was observed. Thus, the
effect of temperature seems to be predominant.
The conversion reached 90% after 53 s residence time and
temperatures higher than 4758C (Figure 1).
Figure 3. Valeric acid selectivity (!) and yield (&) versus valeronitrile con-
version (modeled first-rank Delplot, T=4508C, p=30 MPa, c=0.07 molLꢀ1
valeronitrile in water).
The selectivity versus conversion diagrams (Figure 4) clearly
identify valeramide as the intermediate and valeric acid as the
desired end product of the reaction. Valeramide selectivity
values drop at higher nitrile conversion whereas valeric acid se-
lectivities rise until decarboxylation occurs at higher tempera-
tures.
Figure 1. Conversion of valeronitrile versus residence time (T=4008C (&),
4258C (*), 4508C (~), 4758C (!), 5008C (^), p=30 MPa, c=0.07 molLꢀ1
valeronitrile in water).
The selectivity of the intermediate valeramide dropped at
higher temperatures quickly from approximately 50% at 4008C
as a result of consecutive acid formation. Accordingly, the se-
lectivity towards valeric acid increased with temperature to a
maximum of 90% at 4508C. At higher temperatures, selectivity
dropped due to valeric acid decarboxylation (Figure 2).
Valeramide hydrolysis
The hydrolysis of valeramide was separately investigated under
the same experimental conditions since valeramide is the re-
86
www.chemsuschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 85 – 90

Valeronitrile Hydrolysis in Supercritical Water
90% valeric acid was achieved at 4508C and 53 s residence
time. Delplot analysis illustrates valeric acid as product of the
valeramide conversion, which is consistent with the assumed
reaction network (Figure 6).
Figure 4. Valeramide/valeric acid selectivity versus valeronitrile conversion
(T=4008C (&), 4258C (*), 4508C (~), 4758C (!), 5008C (^), p=30 MPa,
c=0.07 molLꢀ1 valeronitrile in water).
Figure 6. Valeric acid selectivity versus valeramide conversion (first-rank Del-
plot, T=4008C (&), 4258C (*), 4508C (~), 4758C (!), p=30 MPa,
c=0.03 molLꢀ1 valeramide in water).
sulting product of the valeronitrile hydration. Amide hydrolysis
proceeds at a higher reaction rate than nitrile hydration. Total
amide conversion occurs at 4508C and 53 s residence time
(Figure 5).
Valeric acid yield reaches a definite level at higher tempera-
tures (Figure 7) as a result of the consecutive decarboxylation
reactions, which is similar to the outcome of the experiments
displayed above.
Figure 5. Valeramide conversion versus residence time (T=4008C (&),
4258C (*), 4508C (~), 4758C (!), p=30 MPa, c=0.03 molLꢀ1 valeramide
in water).
Figure 7. Valeric acid yield versus residence time (T=4008C (&), 4258C (*),
4508C (~), 4758C (!), p=30 MPa, c=0.03 molLꢀ1 valeramide in water).
At first glance, activation energies of nitrile and amide con-
version do not seem to mirror the experimental results, since
the latter is about 25 kJmolꢀ1 higher. However, the amide con-
version Arrhenius prefactor is two magnitudes greater than
the nitrile prefactor (Table 4); consequently, the reaction rates
enhance. Hence, the activation energy for amide hydrolysis is
higher and temperature has a bigger impact on the reaction
rate. This can be observed when comparing the valeronitrile
conversion and the increments displayed below.
Valeric acid stability
Experiments regarding valeric acid stability were carried out at
a temperature range of 460–5008C and 30 MPa pressure. Vale-
ric acid shows high temperature stability in pure supercritical
water (Figure 8). At 4608C, only slight conversion occurred and
at 5008C, conversion reached 22%. As a result, high yields of
valeric acid can be achieved at 4608C or lower temperatures,
which is consistent with the prior results of the experiments.
The experiments exhibit high selectivity and yield in consid-
eration of the reaction temperatures. The yield maximum of
ChemSusChem 2010, 3, 85 – 90
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
87

H. Vogel et al.
From Figure 9 an activation energy of EA =98 kJmolꢀ1 results
for the valeronitrile hydration. The experiments regarding va-
leramide (Table 2) and valeric acid (Table 3) were evaluated in
the same manner. An activation energy of EA =114 kJ molꢀ1
was calculated for the valeramide hydrolysis. The valeric acid
Figure 8. Valeric acid conversion versus residence time (T=4608C (&),
4758C (*), 4908C (~), 5008C (!), p=30 MPa, c=0.075 molLꢀ1 valeric acid
in water).
Kinetic analysis and reaction modeling
The total recovered carbon (TRC) of the reactions investigated
shows values higher than 90%, especially at temperatures be-
tween 400 and 4508C. At higher temperatures, the TRC values
decreased due to the decarboxylation of valeric acid (the deg-
radation products of valeric acid were labeled as DP in the re-
action network shown in Scheme 3). Hence, high amide and
Figure 9. Arrhenius plot resulting from the valeronitrile hydrolysis data.
Table 2. Reaction rate constants k2(T) deduced from valeramide hydroly-
sis at temperatures between 400–4758C and 30 MPa pressure in pure su-
percritical water.
T
[8C]
Reaction rate constant k1(T)
Regression
factor R2
[10ꢀ2 sꢀ1
]
400
425
450
475
1.58
3.00
8.11
0.9738
0.9991
0.9978
0.9512
Scheme 3. Reaction network used for the modeling of the experimental
data.
10.86
acid selectivities occurred without byproduct formation, which
allows a kinetic parameter evaluation via a global approach in
a temperature range between 375 and 4758C. Reaction equili-
bria described in literature were investigated.[20,21] They re-
vealed the insignificant formation of reverse-reaction products.
Since water is available in excess, its concentration change can
be neglected. Hence, a pseudo-first-order reaction kinetics
with reference to valeronitrile was used for approximating the
experimental data.[26] The resulting reaction rate constants k(T)
are displayed in Table 1.
decomposition shows a significantly higher activation energy
(EA =143 kJmolꢀ1) than the previously discussed hydrolysis re-
actions.
Parameter modeling was perfomed by Presto-Kinetics. Both
irreversible and equilibria reactions between nitrile, amide, and
acid were tested, showing only slight differences in fit perfor-
mance. The results with the best fit quality depicted below
were obtained with the reaction pathway described in litera-
ture.[20,21]
The gathered experimental concentration versus time data
were processed applying the above mentioned reaction net-
work and the experimental reaction rate constants as starting
parameters. The procedure was carried out on all previously
mentioned temperatures and reactant data sets. A close match
between least squares fits and measured values for all the data
sets was attained (Figure 10). New reaction rate constants were
extracted from the modeled data and used for Arrhenius acti-
vation energy calculation and prefactor determination of each
reaction step, which were compared with the experimental re-
sults (Table 4).
Table 1. Reaction rate constants k1(T) deduced from valeronitrile hydroly-
sis at temperatures between 375–4758C and 30 MPa pressure in pure su-
percritical water.
T
[8C]
Reaction rate constant k1(T)
Regression
factor R2
[10ꢀ2 sꢀ1
]
375
400
425
450
475
0.53
1.23
2.53
4.26
5.85
0.9742
0.9814
0.9809
0.9718
0.9650
88
www.chemsuschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 85 – 90

Valeronitrile Hydrolysis in Supercritical Water
sion.[27,28] However, reaction mixture preheating and cool-down
should be taken into account and solved by use of special
alloys or lined equipment if necessary.
In conclusion, only minor technical difficulties are expected
at first glance concerning process scale-up. However, higher
feed concentrations should be applied in terms of productivity
and heating energy efficiency in order to take full advantage
of the improved solubility of organic compounds in supercriti-
cal water. In addition, different process applications are con-
ceivable. Critical parameters concerning industrial scale-up,
such as feed and water purity will depend on what the process
is used for (e.g., waste water treatment or for the synthesis of
pharmaceuticals).
Figure 10. Modeled concentrations and experimental data of valeronitrile
(&), valeramide (~) and valeric acid (!) hydrolysis at T=4508C and
p=30 MPa.
Conclusions
As shown in Table 4, the activation energies extracted from ex-
perimental and modeled data exhibit good correlation. Nitrile
and amide hydrolysis have lower activation energies than the
decarboxylation of valeric acid, which mirrors the experimental
results. Literature provides a broad range of activation energies
for aliphatic nitrile hydrolysis depending on pressure and tem-
perature. Belsky et al. determined an activation energy of
90 kJmolꢀ1 for propionitrile hydrolysis.[15] Usually lower values
are expected for amide hydrolysis; however, the majority of
the studies were carried out at near-critical conditions. The in-
vestigation carried out by Lee et al. of acetonitrile hydrolysis at
temperatures as high as 5008C and similar pressure provided
Table 3. Reaction rate constants k3(T) deduced from valeric acid decar-
boxylation at temperatures between 460-5008C and 30 MPa pressure in
pure supercritical water.
T
[8C]
Reaction rate constant k1(T)
Regression
factor R2
[10ꢀ3 sꢀ1
]
460
475
490
500
1.5
1.9
4.2
4.6
0.9624
0.8961
0.9500
0.9894
an activation energy of 51 kJmolꢀ1 [29]
The findings presented
.
in this work tend to match the activation energies of aromatic
amides (111 kJmolꢀ1 for benzamide).[30] The decarboxylation
activation energy is well within the range provided by litera-
ture, however, most studies focus on supercritical water oxida-
tion of organic acids, which involves additional reaction path-
ways.[31]
Table 4. Comparison of the activation energies and Arrhenius prefactors
obtained from the experimental and modeled data.
Reaction Experimental EA Modeled EA
Experimental A Modeled A
[sꢀ1 [sꢀ1
step
[kJmolꢀ1
]
[kJmolꢀ1
]
]
]
VN!VA 98
VA !VS 114
VS!DP 143
89
114
143
4.81ꢁ105
1.08ꢁ107
2.03ꢁ107
8.64ꢁ104
1.03ꢁ107
1.83ꢁ107
Valeronitrile conversion, which yields valeramide and valeric
acid occurs with high selectivity and space-time yield (e.g., at
4508C and 53 s residence time, nitrile conversion reaches 87%
and valeric acid selectivity reaches 90%). The key benefits are
the abandonment of acid or base addition, which are used as
default reaction conditions and thus the avoidance of waste
brine generation. Given that industrial scale feed concentration
and raw material quality will strongly differ from laboratory-
scale conditions, it is rather difficult to make a final statement
about any potential scale-up problems like plugging or corro-
sion.
Scale-up assessment
A major drawback of high pressure equipment, especially on
industrial scale, is the high investment cost. In addition, major
issues such as corrosion problems or reactor plugging are
often encountered in supercritical water applications. None of
the drawbacks mentioned above occurred during the experi-
ments carried out. Reactor plugging is of concern due to the
low solubility of salts under supercritical conditions and to side
reactions with products of high molecular weight, commonly
noticed in supercritical biomass conversion. The investigated
reactions show good selectivities towards low molecular prod-
ucts. The concentration of salts was low due to a salt-free feed
and negligible corrosion products caused by valeric acid. Cor-
rosion caused by water occurs typically at near-critical condi-
tions; the high temperature applied in the experiments proves
beneficial both in terms of reactivity as well as of corro-
Additional experiments with low acid concentrations were
carried out for a better understanding of the catalytic impact.
These results will be presented in another publication.
Experimental Section
Testing plant
The high pressure equipment used for the experiments is briefly il-
lustrated in Figure 11. The feed was pumped via a HPLC-pump
through the preheater into the reactor. This tubular reactor was
ChemSusChem 2010, 3, 85 – 90
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemsuschem.org
89

H. Vogel et al.
Keywords: hydrolysis · kinetics · nitriles · supercritical fluids ·
sustainable chemistry
[1] B. Subramaniam, M. A. McHugh, Ind. Eng. Chem. Process Des. Dev. 1986,
25, 1–12.
[2] M. Perrut, Ind. Eng. Chem. Res. 2000, 39, 4531–4535.
[3] A. R. Katritzky, D. A. Nichols, M. Siskin, R. Murugan, M. Balasubramanian,
Chem. Rev. 2001, 101, 837–892.
[4] P. E. Savage, S. Gopalan, T. I. Mizan, C. J. Martino, E. E. Brock, AIChE J.
1995, 41, 1723–1778.
[5] H. Weingꢂrtner, E. U. Franck, Angew. Chem. 2005, 117, 2730–2752;
Angew. Chem. Int. Ed. 2005, 44, 2672–2692.
[6] P. G. Jessop, W. Leitner, Chemical Synthesis Using Supercritical Fluids,
Wiley-VCH, Weinheim, 1999.
[7] R. van Eldik, C. D. Hubbard, Chemistry under Extreme or Non-Classical
Conditions 1 st ed., Wiley-VCH, Weinheim, 1996.
[8] X.-Y. Zheng, Y. Arai, T. Furuya, Trends Chem. Eng. 1996, 3, 205–217.
[9] N. Akiya, P. E. Savage, Chem. Rev. 2002, 102, 2725–2750.
[10] A. Kruse, E. Dinjus, J. of Supercritical Fluids 2007, 39, 362–380.
[11] B. Izzo, C. L. Harrell, M. T. Klein, AIChE J. 1997, 43, 2048–2058.
[12] H. Rçmpp, Chemie Lexikon 9th ed., Thieme Verlag, Stuttgart, 1995.
[13] M. Ress-Loeschke, B. Hauer, R. Mattes, D. Engels (BASF AG), DE
10010149, 2003.
[14] C. O’Reilly, P. D. Turner, J. Appl. Microbiol. 2003, 95, 1161–1174.
[15] D. E. Mccaslin (Atlantic Richfield Company), US 3876691, 1975.
[16] A. J. Belsky, T. B. Brill, J. Phys. Chem. A 1999, 103, 3006–3012.
[17] S. Mittelstꢂdt, Ph.D. Thesis, Technische Universitꢂt Darmstadt, 1997.
[18] S. D. Iyer, M. T. Klein, J. of Supercritical Fluids 1997, 10, 191–200.
[19] B. Izzo, M. T. Klein, Ind. Eng. Chem. Res. 1999, 38, 1183–1191.
[20] J. Clayden, N. Greeves, S. Warren, P. Wothers, Organic Chemistry, Oxford
University Press, Oxford, 2001.
Figure 11. The high pressure plant used for nitrile hydrolysis under supercrit-
ical conditions.
made of Inconel C276 (2.83 cm3 volume) and was spirally embed-
ded into an aluminum cylinder. The cylinder was enfolded by an
aluminum coat and was heated by a 500 W heating device. After
passing the reactor, the reaction mixture was cooled and depres-
surized. The plant was designed for temperatures as high as 5608C
and pressures of 40 MPa. Residence times could be adjusted be-
tween 8 and 120 s by flow variation.
[21] R. Brꢃckner, in Reaktionsmechanismen 2. Auflage, Spektrum Akademisch-
er Verlag, Heidelberg, 2003.
Analysis
[22] L. A. M. M. Barbosa, R. A. van Santen, J. Mol. Catal. A 2001, 166, 101–
121.
The reaction mixture was analyzed using gas chromatography (SHI-
MADZU GC-17 A, capillary column [CW 20M MN fused silica, length
(L) 19 m, inner diameter (I.D.) 0.32 mm, film thickness 0.5 mm],
equipped with a flame ionization detector and nitrogen as carrier
gas) and high-performance liquid chromatography (KNAUER HPLC
equipped with RI and UV detector (l=210 nm), column (ProntoSIL
C18 ACE-EPS, L=250 mm, I.D.=3 mm, particle size=3.0 mm) and
2.5 mm sulfuric acid in 9:1 water/acetonitrile as an eluent).
Chemicals used: valeronitrile (technical, Lonza), valeramide (97%,
Acros), valeric acid (99%, Acros).
[23] S. Matsubara, Y. Yokota, K. Oshima, Org. Lett. 2004, 6, 2071–2073.
[24] M. Okazaki, T. Funazukuri, J. Mater. Sci. 2006, 41, 1517–1521.
[25] P. G. Maiella, T. B. Brill, J. Phys. Chem. A 1998, 102, 5886–5891.
[26] N. A. Bhore, M. T. Klein, K. B. Bischoff, Ind. Eng. Chem. Res. 1990, 29,
313–316.
[27] P. Kritzer, N. Boukis, E. Dinjus, J. of Supercritical Fluids 1999, 15, 205–
227.
[28] P. Kritzer, J. of Supercritical Fluids 2004, 29, 1–29.
[29] D. S. Lee, E. F. Gloyna, Environ. Sci. Technol. 1992, 26, 1587–1593.
[30] A. Krꢂmer, S. Mittelstꢂdt, H. Vogel, Chem. Eng. Technol. 1999, 22, 494–
500.
[31] J. C. Meyer, P. A. Marrone, J. W. Tester, AIChE J. 1995, 41, 2108–2121.
Acknowledgements
Received: June 30, 2009
Revised: September 10, 2009
Published online on November 18, 2009
The authors thank Lonza AG, Visp for their financial support of
this work.
90
www.chemsuschem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2010, 3, 85 – 90