Spectroscopy of hydrothermal reactions. 12. Acid-assisted hydrothermolysis of aliphatic nitriles at 150-260 °C and 275 bar

Article abstract of DOI:10.1021/jp984364d

The conversion of four alkyl nitrile compounds, RCN, to the corresponding carboxylic acids and ammonia in fluid H₂O at 150-260 °C under 275 bar is described. The reaction rate was increased by the addition of 0.3-1.0 N HCl. The kinetics and pathways of the reactions were determined in real time by IR spectroscopy with a sapphire-titanium flow cell. Ex situ IR and Raman spectral data following batch reaction aided in establishing the reaction pathways. When R = CH₃-, CH₃CH₂-, and (CH₃)₂CH-, the relatively simple conversion to RCO₂H and NHs occurred and was modeled by two protonation equilibria and two forward reactions. When R=HOC(O)CH₂-, a more complex pH-dependent reaction took place because of the presence of two functional groups. The carboxylate group reacted at high pH, whereas the nitrile group reacted at low pH. The kinetics of these reactions were determined in real time, and the Arrhenius parameters were obtained. The activation energy for the hydrothermolysis of the nitriles qualitatively correlates with v(CN), indicating that the electron-donating power of R is a factor in the rate of reaction. Consistent with this observation is the fact that a Taft plot is linear when the electronic effect of R is weighed more heavily than the steric contribution. This was not the case when R=HOC(O)CH₂-, where a different transition state is proposed.

Full text of DOI:10.1021/jp984364d

3006
J. Phys. Chem. A 1999, 103, 3006-3012
Spectroscopy of Hydrothermal Reactions. 12. Acid-Assisted Hydrothermolysis of Aliphatic
Nitriles at 150-260 °C and 275 bar
A. J. Belsky and T. B. Brill*
Department of Chemistry and Biochemistry, UniVersity of Delaware, Newark, Delaware 19716
ReceiVed: NoVember 9, 1998; In Final Form: February 12, 1999
The conversion of four alkyl nitrile compounds, RCN, to the corresponding carboxylic acids and ammonia
in fluid H
.3-1.0 N HCl. The kinetics and pathways of the reactions were determined in real time by IR spectroscopy
with a sapphire-titanium flow cell. Ex situ IR and Raman spectral data following batch reaction aided in
CH
2
O at 150-260 °C under 275 bar is described. The reaction rate was increased by the addition of
0
establishing the reaction pathways. When R ) CH
3
-, CH
3
2
3 2
-, and (CH ) CH-, the relatively simple
conversion to RCO
2
H and NH
3
occurred and was modeled by two protonation equilibria and two forward
2
reactions. When RdHOC(O)CH -, a more complex pH-dependent reaction took place because of the presence
of two functional groups. The carboxylate group reacted at high pH, whereas the nitrile group reacted at low
pH. The kinetics of these reactions were determined in real time, and the Arrhenius parameters were obtained.
The activation energy for the hydrothermolysis of the nitriles qualitatively correlates with ν(CN), indicating
that the electron-donating power of R is a factor in the rate of reaction. Consistent with this observation is the
fact that a Taft plot is linear when the electronic effect of R is weighed more heavily than the steric contribution.
2
This was not the case when RdHOC(O)CH -, where a different transition state is proposed.
Introduction
On the other hand, a high concentration of acid (e.g., >2 N
HCl) can be used to accelerate this conversion. The consecutive
The reactions of the nitrile functional group in aqueous
solution have potentially important ramifications in nature. One
popular hypothesis is that Strecker synthesis, or a closely related
reaction, provided easy access to R-amino- and R-hydroxycar-
boxylic acids and were essential to the early life-forming
15
pseudo-second-order reactions 1 and 2 occur in acid solution
+
in which the protonated nitrile RCdNH and protonated amide
are believed to be the reactive species.
k1
1-4
process. Aldehydes and/or ketones react with NH3 and HCN
in this reaction to form R-amino- and R-hydroxynitriles, which,
in turn, react with H2O to liberate the corresponding substituted
carboxylic acids. In this manner the hydrolysis of aliphatic nitrile
groups potentially has a role in the origins of life, although
questions have been raised about the role of Strecker synthesis
in a major source of these essential carboxylic acids, which is
RCN + H O H+8 RC(O)NH2
(1)
(2)
2
k2
RC(O)NH + H O
2
8 RCO H + NH3
2
2
+
H
The value of k2 is about 15 times larger than k1 in 4 N HCl.9,17
Protonation of the nitrile apparently enhances the susceptibility
of the electrophilic carbon atom to nucleophilic attack by H2O
and thereby shortens the reaction time at 50-100 °C to a few
hours.
5
in the submarine hydrothermal vent. Typical intrinsic conditions
in these vents are T ) 200-300 °C and P ) 200-300 bar.
6
,7
Thermodynamic considerations suggest that the aldehyde and
ketone reactants may have insufficient concentrations in this
hostile environment for amino and hydroxy acids to form
abundantly by this pathway. Carboxylic acids containing the
nitrile group are also known to form in natural waters after
chlorination of humic acid.8
Nitrile-containing compounds are extensively used in the
manufacture of industrial and commercial products. As such,
the hydrothermolysis of organonitrile compounds is important
if waste streams containing these compounds are to be reme-
diated by hydrothermal methods, including wet-air oxidation
and supercritical water oxidation.
A flourish of studies appeared in the mid-century on the acid-
accelerated kinetics and mechanism of hydrolysis of organo-
nitriles, RCN.
hydrolyze to form amides in the neutral solution. For example,
the half-life of acetonitrile in H2O at 65 °C is about 300 days.
Interest in nitrile hydrolysis was rekindled recently by the
potential use of hydrothermal methods to remediate waste
1
8-20
21
streams
and to convert biomass material containing
organonitriles. These studies have been conducted by heating
of the nitrile in water in a batch-mode tube reactor, cooling
after a prescribed reaction time, and determining the products
off-line by various analytical techniques, such as GC and GC/
MS. An attempt has been made in the work described herein to
advance the understanding of the kinetics and mechanisms of
acid-assisted hydrothermolysis of aliphatic nitrile compounds
by using real-time IR spectroscopy in a flow cell at reaction
conditions. Temperatures of 150-260 °C at a pressure of 275
bar were used. The results for acid-assisted hydrothermolysis
of several simple aliphatic nitriles RCN (RdCH -, CH CH -,
9-16
These studies showed that nitriles only slowly
3
3
2
and (CH ) CH-) were used to validate this method for use in
3
2
7
the study of the more complex hydrothermolysis scheme of
cyanoacetic acid, NCCH2CO2H. In contrast to previous flow
22-31
*
Author for correspondence. E-mail: brill@udel.edu.
reactor spectroscopic studies,
this is the first study in which
1
0.1021/jp984364d CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/31/1999

Hydrothermolysis of Nitriles
J. Phys. Chem. A, Vol. 103, No. 16, 1999 3007
an acid catalyst has been deliberately added to the reaction at
hydrothermal conditions.
TABLE 1: Taft Electronic and Steric Parameters for the
Nitrile Substituents
b
R
σ* a
E
s
Experimental Section
CH
CH
3
-
0
0
Cyanoacetic acid (HOC(O)CH2CN), acetonitrile (CH3CN),
propionitrile (CH3CH2CN), isobutyronitrile ((CH3)2CHCN),
isobutyramide, isobutyric acid, acetamide, acetic acid, propyl-
amide, and propionic acid were all obtained from Aldrich
Chemical Co. Each had +98% purity and was used as received,
except for acetonitrile, which was dried over molecular sieves.
Malonamic acid (H2NC(O)CH2CO2H) was prepared by the
method of Galat.32 Solutions were then made from Milli-Q H2O
that had been sparged with Ar to remove atmospheric gases.
Concentrations were 1.0 m except for the less-soluble isobuty-
ronitrile, which was studied at 0.5 m. Added HCl was 0.30 N
as determined by multiple titrations with a KHP-standardized
NaOH solution.
3
CH
CH-
HOC(O)CH
2
-
-0.1
-0.19
1.08
-0.07
-0.47
-0.75
3 2
(CH )
c
2
-
a
Reference 39. b Reference 40. c Reference 31.
3
having an internal volume of 12 cm . These tube reactors were
then heated in a fluidized sand bath for selected time periods
37
of 5 min or more. IR and/or Raman spectroscopy of the cooled
solution was used to determine the products. These IR spectra
were collected on a Nicolet 560ESP spectrometer equipped with
a liquid-N2-cooled MCT-A detector. Raman spectra were
collected using a Kaiser Optical Systems dispersive spectrometer
with a thermoelectrically cooled CCD detector. A 50 mW,
diode-pumped, frequency-doubled, Nd:YAG laser was used for
excitation at 532 nm. Five scans (10 s each) were recorded and
averaged to give each spectrum.
The laboratory-scale flow reactor spectroscopy cell22,23 was
constructed of grade 2 titanium into which two 0° sapphire
windows were seated. Gold-foil washers were compressed at
each interface to form the seals. The solution flowed in a flat
duct that was created between the two sapphire windows by a
slot cut into the gold washer that separated the windows. The
path length of the cell (the thickness of the flat duct) was (30-
3
8
The Taft plots required the use of Sigma Plot (Jandel
Scientific) to solve for the steric (δ) and electronic (F) weighting
parameters. The steric (Es) and electronic (σ*) constants for the
2
8,39,40
R groups of the nitriles are listed in Table 1.
3
5) ( 1 µm as determined by the absorbance of a known
3
3
concentration of aqueous CO2 flowing through the cell. The
temperature was controlled ((1 °C) by using PID controllers,
the pressure was controlled ((1 bar) by a pneumatic bleed-
and-lock system, and the constant flow rate maintained with
an Isco syringe pump.23 A single liquid phase was maintained
at all times.
Results and Discussion
Choice of the Catalyst. As noted in the Introduction, the
addition of a strong acid has long been known to accelerate the
+
decomposition of organonitrile compounds. Since H is con-
sumed by the eventual products (e.g., NH3), it is potentially
confusing to refer to the overall reaction as acid-catalyzed, even
To investigate the kinetics of reactions, flow rates were used
2
2,34
+
such that plug-flow conditions could be assumed to exist.
though H is conserved in the initial step. With this in mind,
The flow rates used resulted in residence times of 3-60 s and
degrees of conversion of about 50% or less. The true residence
the convention of referring to the reaction as “acid-catalyzed”
will be retained here. The counterion and concentration of the
acid catalyst are important factors in nitrile hydrolysis on the
basis of past studies of the acid (H2SO4, HNO3, HBr, and HCl)
time was obtained by dividing the internal volume of the cell
3
(
entrance tube and flat duct), which was 0.0819 cm , by the
9
-11,13
expression [volume flow rate × F25°C/FT]. The density correction
and concentration (0.3-22.2 N).
The identity of the
is needed because the density of H2O, FT, depends on the
counterion affects the overall reaction pathway, and large
changes in the acid concentration affect the ratio of the
consecutive rates k1 and k2 (eqs 1 and 2). For example, k2 . k1
at less than 4 N HCl, whereas the reverse is true at much larger
3
3
temperature (e.g., F150°C ) 0.93 g/cm , F260°C ) 0.81 g/cm at
a constant pressure of 275 bar).
A Nicolet 60SX FTIR spectrometer equipped with a liquid-
N2-cooled MCT-A detector was used for real-time IR spectral
measurements in the flow reactor. Thirty-two spectra were
summed at 4 cm-1 resolution (∼10 s collection time) at each
flow rate and temperature. The IR spectra were normalized with
background spectra of pure H2O recorded at the same conditions.
The carbonyl stretching modes of the carboxylic acid and amide
were partially visible but could not be quantified reliably because
of the interference from the O-H bending mode of H2O and
10,11
acid concentrations.
Thus, exploratory work was conducted
to choose the appropriate acid and concentration for this work.
In turn, this choice dictated the cell type.
Of the four commonly used acid catalysts listed above, HNO3
is known to oxidize and hydrolyze aliphatic nitriles to CO2 and
4
1
other products. H2SO4 has been shown to be a less effective
9
nitrile hydrolysis catalyst than HCl and also to act as a
4
2
nucleophile when its concentration is high. Therefore, HCl
was chosen as the catalyst. HCl corrodes stainless steel,
however, necessitating the use of spectroscopy cells constructed
of titanium and gold, which are relatively unreactive toward
-
1
the sapphire cutoff in the vicinity of 1700 cm . The IR-active
-
1
CN stretch of the nitriles (∼2250 cm ) and the CO2 product
-
1 23
(
2343 cm ) in the case of NCCH2CO2H were resolved at
4
3
each reaction condition using Peakfit software (Jandel Scientific)
and a Matlab program. Four-parameter Voigt functions were
employed to fit these bands. Three or more sets of data were
recorded for each nitrile to establish the average concentration
and the standard deviation. Kinetic models requiring a running
the chloride ion. A relatively low HCl concentration (0.3 N)
provides nearly equal rates throughout the nitrile and amide
hydrolysis reactions, but the carboxylic acid and ammonia
products buffer the system and NH3 consumes the catalyst in
the later stages.
+
tally of [H ] during the reaction were solved in small time
Spectral Determination of the Reaction Pathways. Raman
spectroscopy helps identify the products from batch reaction
of the alkyl nitriles after 20 min in 0.3 N HCl (Figures 1-3).
The formation of the corresponding carboxylic acid is revealed
along with the presence of unreacted parent nitrile. The absence
of Raman modes for the amide supports the earlier finding1
that hydrolysis of the amide is faster than the nitrile (i.e., k2 .
increments with Euler’s method in Lotus 1-2-3. A weighted
least-squares regression was used to determine the value and
error of each rate constant and the Arrhenius parameters, with
3
5,36
care taken when converting the error into log space.
0,11
Several batch reaction mode experiments were conducted with
the reactant (1.0 or 0.5 m) and 0.3 N HCl in titanium tubes

3008 J. Phys. Chem. A, Vol. 103, No. 16, 1999
Belsky and Brill
Figure 3. Raman spectra of (a) (CH
b) 0.5 m (CH
30 °C for 20 min; (c) isobutyric acid in 0.3 N HCl. Partial conversion
3
)
2
CHCN in 0.3 N HCl at 25 °C;
Figure 1. Raman spectra of (a) CH
3
CN in 0.3 N HCl at 25 °C; (b)
(
2
3 2
) CHCN in 0.3 N HCl after heating in a tube reactor at
1
.0 m CH CN in 0.3 N HCl after heating in a tube reactor at 230 °C
3
for 20 min; (c) acetic acid in 0.3 N HCl. The heated solution shows
the presence of unreacted nitrile and acetic acid (reaction 3). The peak
of reaction 3 is indicated in (b).
-
1
3
at 2295 cm is an overtone/combination band of CH CN.
2 2
Figure 4. Time series of IR spectra for 1.0 m NCCH CO H in 0.3 N
HCl in the flow cell at 210 °C under 275 bar. The spectra show an
-1
2
increase in CO (2343 cm ) with a decrease in the nitrile stretch (2265
-
1
Figure 2. Raman spectra of (a) CH
b) 1.0 m CH CH CN in 0.3 N HCl after heating in a tube reactor at
30 °C for 20 min; (c) propionic acid in 0.3 N HCl. Partial conversion
3
CH
2
CN in 0.3 N HCl at 25 °C;
cm ); (a) in the insert shows the position of the nitrile in CH CN if
3
step A had occurred. Since the nitrile does not shift, step A cannot be
occurring.
(
2
3
2
by reaction 3 is indicated in (b).
On the other hand, the addition of a strong acid favors the
occurrence of reaction 3 at a lower temperature and has been
used to produce malonic acid from cyanoacetic acid.41 The more
extensive reaction pathway shown in Scheme 1 thereby becomes
possible depending on the reaction conditions.
k1). The global reaction 3, which is the net reaction of reactions
1
and 2, describes the hydrolysis of simple alkyl nitriles (Rd
CH3-, CH3CH2-, and (CH3)2CH-) because the amide formed
has a low steady-state concentration.
The IR spectroscopic data from the flow cell at 150-260 °C
help refine the conditions for Scheme 1. Step A has been proven
by IR spectroscopy to occur in neutral H2O solution.31 On the
other hand, Figure 4 shows that in 0.3 N HCl, the CN stretch
RCN + 2H O f RCO H + NH
3
(3)
2
2
In contrast to the simple alkyl nitriles, the hydrothermal
reaction of cyanoacetic acid is potentially more complicated
because both the cyano and carboxylate functional groups can
react (reactions 3 and 4) depending on the experimental
conditions used. Previous kinetic and mechanistic studies of
-1
-1
of NCCH2CO2H at 2265 cm does not shift to 2256 cm ,
31
which would have indicated the formation of CH3CN, but CO2
forms. This seeming contradiction is explained by the fact that
the CO2 may also be formed by the combination of steps B
and C in which no CH3CN forms. Three results support steps
B and C. First, the intensity of the CN stretch decreased with
time without a shift in the frequency, which indicates that step
B, rather than step A, took place in 0.3 N HCl solution (Figure
31
31
28
cyanoacetic acid, malonamic acid, and malonic acid in
neutral H2O at high temperature and pressure have confirmed
the occurrence of reaction 4.
4
). Second, the rate of formation of CO2 is observed to be slower
RCO H f RH + CO2
(4)
by a factor of 2 after 30 s than is predicted by the kinetics of
2

Hydrothermolysis of Nitriles
J. Phys. Chem. A, Vol. 103, No. 16, 1999 3009
SCHEME 1
SCHEME 2
SCHEME 3
step A.31 Third, the rate of decarboxylation of malonamic acid
strong acid shuts down the decarboxylation process. In summary,
the main point is that step A is stopped by the presence of strong
acid in favor of steps B and C.
(
step C) is faster at the conditions used in this work than the
3
1
alternative hydrolysis pathway to malonic acid (step E).
Discussion of step E is deferred until the following paragraph
in order to complete the discussion of steps A-C here. As noted
above, step A occurs in neutral H2O solution. Indeed, when the
reaction had progressed sufficiently to consume the acid catalyst
and the pH had increased toward neutral, the decarboxylation
of cyanoacetic acid (step A) occurred. This switch from step A
to steps B and C is demonstrated by progressively increasing
The malonamic acid product of step B and has two possible
decomposition pathways, decarboxylation to acetamide (step C)
or hydrolysis to malonic acid (step E). Previous studies at <90
°C showed that malonamic acid only converts to malonic acid
4
1
(step E) in concentrated HCl. At lower HCl concentrations,
however, Hall showed that the rate of decarboxylation of
malonamic acid (step C) was essentially independent of the HCl
4
5
the HCl concentration from 0.01 to 1.0 N as shown in Figure
concentration, a finding that was also confirmed at hydro-
+
31
5
. The explanation at the low [H ] is that the catalyst was
thermal conditions. Therefore, the decarboxylation step C is
quickly consumed by NH3 formed in Scheme 1, causing the
pH to increase, which, in turn, favors step A. At a higher [H ],
ample acid was present at all times so that steps B and C
exclusively occur. This later observation is consistent with the
results of Darensbourg et al.,44 who used metal catalysts to
decarboxylate NCCH2CO2H and noted that the presence of a
the only pathway that occurs at the temperatures and HCl
concentrations used in this work. Since the acetamide product
of step C has only one pathway to hydrolyze to acetic acid (step
D), the acid-catalyzed decomposition pathway of NCCH2CO2H
is complete. Scheme 2 is the simpified version of Scheme 1,
which applies in 0.3 N HCl at hydrothermal conditions. Acetic
+
46,47
acid is refractory at these conditions.
The other product, NH3,
is also refractory at these conditions except that it consumes
+
the H catalyst, which causes step A of Scheme 1 to become
increasingly important as the pH approaches neutral.
Kinetic Determinations. Scheme 3 is the proposed acid-
catalyzed hydrothermolysis mechanism of the simple alkyl
nitriles, RCN (RdCH3-, CH3CH2-, and (CH3)2CH-). It is
generally agreed that protonation of the nitrile and amide
facilitates nucleophilic attack by H2O and that this is the function
48
of the acid catalyst. Equilibria K1 and K2 both favor the neutral
(
left-hand) form because the nitrile group is only weakly
4
2,49
basic.
The rate expression for the disappearance of RCN
(
eq 5) is considered to be pseudo-second-order with respect to
+
13,18-20
[H ] and [RCN].
+
+
-
-
d[RCN]/dt ) k K [H O][H ][RCN] ) k ′[H ][RCN] (5)
1 1 2 1
+
d[RC(O)NH ]/dt ) k K [H O][H ][RC(O)NH ] )
2
2
2
2
2
+
k ′[H ][RC(O)NH ] (6)
2
2
Figure 5. Plot of the concentration of the nitrile and CO
2
measured in
Likewise, the disappearance of the amide (eq 6) is pseudo-
second-order, but these expressions disguise the subtle com-
plexities of the reaction.
the IR spectroscopy flow cell of NCCH CO H at 230 °C and 275 bar
2
2
showing that the reaction slows due to the shift from step A to step B
+
as the [H ] concentration is increased.

3010 J. Phys. Chem. A, Vol. 103, No. 16, 1999
Belsky and Brill
TABLE 2: Pseudo-Second-Order Rate Constants for the
Hydrothermolysis of RCN at 275 bar in the IR Flow Reactor
k, kg/(mol s) × 100
R ) CH3- R ) CH3CH2- R ) (CH3)2CH- R ) HOC(O)CH2-
T, °C
(1.0 m)
(1.0 m)
(0.5 m)
(1.0 m)
1
1
1
1
1
2
2
2
2
2
2
2
50
60
70
80
90
0.22 ( 0.06
0.39 ( 0.03
0.72 ( 0.05
1.3 ( 0.1
0.051 ( 0.008
0.12 ( 0.02
0.27 ( 0.03
0.66 ( 0.08
1.5 ( 0.2
1.4 ( 0.3
1.8 ( 0.1
2.5 ( 0.10
3.5 ( 0.5
7.3 ( 4.2
14 ( 3
1.8 ( 0.5
00 0.79 ( 0.19 3.5 ( 0.7
3.4 ( 0.5
10 1.6 ( 0.4
20 3.0 ( 1.0
30 6.5 ( 1.7
40 12 ( 3
50 18 ( 2
60 31 ( 4
4.7 ( 1.4
10 ( 4
7.5 ( 1.4
16 ( 3
14 ( 6
18 ( 4
38 ( 7
20 ( 7
24 ( 5
29 ( 9
31 ( 6
43 ( 11
40 ( 8
TABLE 3: ν(CN) and Arrhenius Parameters for
Decomposition of RCN
Figure 6. Calculated and actual concentrations for 0.5 m isobutyroni-
trile in 0.3 N HCl (eq 5) at 240 °C under 275 bar.
q
∆S
E
a
ln(A, kg/
(mol s)
cal/(Kmol)
(T ) 200 °C)
ν(CN)
1
-
R
kcal/mol
cm
As the reaction progresses, the acid catalyst is gradually
consumed by the association of the H with NH3 (reaction 7).
+
CH -
3
30.4 ( 0.7 27.7 ( 0.7
-6.4
-14
-22
-15
-26
5.4
2254
a
2
6.5
21.5 ( 0.3 19.7 ( 0.4
25.6b
23.2
23.8
Kb
CH CH -
2247
3
2
+
+
NH + H O y
\
z NH4 + H O
(7)
(8)
(9)
3
3
2
(
CH
3
)
2
CH-
19.7 ( 0.9 17.8 ( 1
2241
2265
Ka
HOC(O)CH -
2
34.7 ( 0.5 33.7 ( 0.5
-
+
26.3a
RC(O)OH + H O y
\
z RC(O)O + H O
22.8
-16
2
3
a
Calculated from ref 11. b Calculated from ref 10.
Kc
-
+
+
HCl + H O y
\
z Cl + H O
The time derivative of [H ] (eq 12) based on the charge
balance provides [H ] throughout the reaction. The rate
2
3
+
The acid-base equilibrium involving the carboxylic acid,
constants at each temperature were determined by assuming the
steady-state condition for the amide concentration.
+
ammonia, and [H ] buffers the solution from the initial pH of
about 0.5 to about pH ) 3 in the later stage. For comparison
d[H+]
dt
[H ]
+
d[BT]
Ka
0
.5 m NCCH2CO2H has a pH of 1.5 at 200 °C. As part of
)
-
establishing the running pH value during the reaction, the
equilibrium constants Ka, Kb, and Kc (eqs 7, 8, and 9) at the
reaction temperature were determined from literature values at
{
+
+
}
dt
[
]
K + [H ] K + [H ]
/
a
b
2
+
B
B
[H ]
K
W
T
T
5
0,51
1 +
-
+
+
+ 2
lower temperature
by the iso-Coulombic extrapolation
+
+
2
(
K + [H ] (K + [H ])
[H ]
method.5 This method takes into account the effect of changing
KW of H2O on the ionization constants of the acids and bases.
Figure 6 shows the apparent rate of disappearance of 0.5 m
2
b
b
5
3
KaBT
K C
c
T
+
(12)
+
2
+
2
)
(
K + [H ])
(K + [H ])
c
a
(CH3)2CHCN at 230 °C in the IR flow reactor. The rate clearly
decreases as the reaction proceeds as a result of the removal of
Figure 6 shows the fitted values for the disappearance of
CH3)2CHCN (k1′) and the calculated values for the appearance
+
[
H ] by reaction 7. This change is the basis for determining
(
+
the apparent nitrile hydrolysis rate, k1′, in eq 5. [H ] is known
of the carboxylic acid and ammonia products at 240 °C. The
model demonstrates that as [H ] decreases, reaction 3 slows
by paramatrizing eq 10 at all times and temperatures used.
+
down and eventually stops. Table 2 gives the rate constants for
each nitrile at the temperatures studied. The Arrhenius param-
eters are listed in Table 3 and will be discussed further below.
The kinetics of acid-catalyzed hydrothermolysis of cyanoace-
tic acid were solved in the same way as for the alkyl nitriles
with a few exceptions. The concentration of acetamide in
Scheme 2 was assumed to be small and in steady state. Because
the rate of decarboxylation of malonamic acid was found to be
the same with and without added HCl, the uncatalyzed rate
+
B [H ]
KW
KaB
K C
c T
+
T
T
[
H ] +
)
+
+
(10)
+
+
+
+
K + [H ] [H ] K + [H ] K + [H ]
b
a
c
Equation 10 is determined from the conservation of charge (eq
1) during the reaction.
1
+
+
-
-
-
[
H ] + [NH ] ) [Cl ] + [RCO ] + [OH ] (11)
4 2
3
1
determined previously was used.
The charge balance for Scheme 2 is given by eq 13.
The total carboxylic acid concentration is AT ()RCO2H +
-
RCO2 ), the total concentration of ammonia is BT()NH3 +
+
-
-
NH4 ), and the total of concentration HCl is CT. Because the
[
Cl ] + [H NC(O)CH CO (MMA)] +
2 2 2
carboxylic acid and ammonia are produced simultaneously in
reaction 2, AT ) BT. Therefore, only BT appears in eq 10.
Equilibration reactions 7-9 occur much faster than the rate-
determining step of Scheme 1, so acid-base equilibrium was
assumed at all times.
-
-
-
[
NCCH CO (CAA)] + [CH CO ] + [OH ] +
2 2 3 2
-
+
+
[HCO ] ) [H ] + [NH ] (13)
3 4
In arriving at eq 13, it was determined that the contribution of

Hydrothermolysis of Nitriles
J. Phys. Chem. A, Vol. 103, No. 16, 1999 3011
2
Figure 7. Calculated and observed concentrations for 1.0 m NCCH -
CO H in 0.3 N HCl at 190 °C under 275 bar.
2
MMA and CAA anions had to be included even though their
concentrations were small. CO2 is produced, which necessitates
Figure 8. Arrhenius plot of the alkyl nitriles and cyanoacetic acid
compared to the literature values for the same compounds.
-
the inclusion of HCO3 in the charge balance, even though the
-
solution is acidic and the HCO3 contribution is very small.
+
When the charge balance equation is redefined in terms of [H ],
+
the time derivative of [H ] is given by eq 14 and is slightly
more complicated than eq 12.
d[H+]
dt
d[CAA] KCAA
d[MMA] KMMA
)
+
+
{
dt
(
+
)
dt
(
+
)
[
H ]
[H ]
K
+
d[CO ] CO2
d[B ]
Ka
[H ]
2
T
+
-
(
+
)
dt
[
+
+
]
}
dt
[
H ]
K + [H ] K + [H ]
/
a
B
2
+
B
B [H ]
KW
KaB
T
T
T
1
+
-
+
+
+
2
(
+
+
2
+ 2
+
K + [H ] (K + [H ])
[H ]
(K + [H ])
b
b
a
[
CAA]KCAA [MMA]KMMA
K C
c
T
+
+
(14)
+
2
+ 2
+
2
)
[
H ]
[H ]
(K + [H ])
c
+
As before BT ) AT. With d[H ]/dt now available, the nitrile
hydrolysis rate constants can be determined at each temperature.
Since the reactant nitrile band and CO2 were both visible in the
Figure 9. Taft plot for the rate of hydrothermolysis of the alkyl nitriles
compared to CH CN at 200 °C in which F ) -7.2 and δ ) 1.4.
3
IR flow cell, each temperature was solved minimizing the error
k0 is the rate constant of the reference compound, which in this
case was chosen to be acetonitrile. The substituent constants
used for the R groups are given in Table 1. δ and F are the
weighting coefficients that were adjusted to optimize the fit
between the substituent constants and the rate constants. Figure
2
(ø ) on both compounds (Figure 7). The focus was on fitting
the concentration data for the nitrile with the most accuracy
because the rate for the nitrile hydrothermolysis, k1′ (step B in
Scheme 2), is the most important unknown for this work.
Because of the rather low absorptivity of the -CN stretch
9
shows the plot that results for the simple alkyl nitriles. The
(Figure 4), fitting uncertainties lead to some error in the CO2
values of F ) -7.2 and δ ) 1.4 thus obtained suggest that
increasing the electron-donating character of R causes the rate
to increase to a greater extent than increasing its steric bulk.
This observation is consistent with the qualitative correlation
between Ea and ν(CN) in Table 3, which implies that the
electronic inductive effect of R dominates in determining the
reaction rate. Figure 9 does not include NCCH2CO2H because
its hydrolysis rate is much faster at hydrothermal conditions
than would be predicted from the correlation in Figure 9 for
simple alkyl nitriles. A possible explanation is that the proto-
concentration. Arrhenius parameters are given in Table 3 along
with those calculated from previous rate measurements in the
normal liquid range of H2O.10,11 These parameters differ
significantly from the lower temperature regime, but because
of the kinetic compensation effect between Ea and A, the rates
are comparable (Figure 8).
Structure-Reactivity Relations. A rough correlation exists
in Table 3 between the increasing -CN stretching frequency
and the activation energy as a function of R. This observation
suggests that the Taft eq 15 might provide additional insight
into the role of the steric, Es, and electronic, σ*, properties of
R in the nitrile hydrothermolysis rate.38 In eq 15,
nated nitrile group and H O can form a six-membered ring with
2
the carboxylate group in NCCH CO H, which facilitates nu-
2
2
cleophilic attack by water on the electropositive nitrile carbon
atom. The absence of the carboxylate group in the simple alkyl
log(k/k ) ) δE + Fσ*
(15)
0
s

3012 J. Phys. Chem. A, Vol. 103, No. 16, 1999
Belsky and Brill
nitriles precludes this type of stable transition state, which results
in a slower rate of hydrolysis.
(18) Harrell, C. L.; Klein, M. T.; Adschiri, T. AdV. EnViron. Res. 1998,
1, 373.
(19) Izzo, B.; Harrell, C. L.; Klein, M. T. AIChE J. 1997, 43, 2048.
(
20) Harrell, C. L. Investigation and Modeling of Nitrile Hydrothermal
Conclusions
Chemistry: Physical and Chemical Effects. Ph.D. Dissertation, University
of Delaware, Newark, DE, 1998.
The mechanism of hydrothermolysis of simple alkyl nitriles
in the presence of HCl is easily modeled with two forward
reactions and one equilibrium step. When two reactive functional
groups are present in the molecule, the method of real-time
spectroscopy makes it possible to define and follow competitive,
pH-sensitive reactions.
(21) Katritzky, A. R.; Lapucha, A. R.; Siskin, M. Energy Fuels 1990,
4, 560.
(22) Schoppelrei, J. W.; Kieke, M. L.; Wang, X.; Klein, M. T.; Brill, T.
B. J. Phys. Chem. 1996, 100, 14343.
(
23) Kieke, M. L.; Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. 1996,
1
1
00, 7455.
(24) Schoppelrei, J. W.; Kieke, M. L.; Brill, T. B. J. Phys. Chem. 1996,
00, 7463.
The rates of hydrothermolysis of simple alkyl nitriles in water
at the temperature and pressure conditions of a submarine
hydrothermal vent are probably too slow to be of importance
in the absence of a catalyst. Hydrogen ions accelerate the
reaction. Hence the kinetics, as well as previously discussed
(25) Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. A 1997, 101, 2298.
(26) Schoppelrei, J. W.; Brill, T. B. J. Phys. Chem. A 1997, 101, 8598.
(27) Maiella, P. G.; Brill, T. B. Inorg. Chem. 1998, 37, 454.
(28) Maiella, P. G.; Brill, T. B. J. Phys. Chem. 1996, 100, 14352.
29) Maiella, P. G.; Brill, T. B. J. Phys. Chem. A 1998, 102, 5886.
(30) Belsky, A. J.; Brill, T. B. J. Phys. Chem. A 1998, 102, 4509.
31) Belsky, A. J.; Maiella, P. G.; Brill, T. B. J. Phys. Chem. A,
submitted.
(32) Galat, A. J. Am. Chem. Soc. 1948, 70, 2596.
33) Maiella, P. G.; Schoppelrei, J. W.; Brill, T. B. Appl. Spectrosc., in
press.
(
5
concentration considerations, argue against the Strecker reaction
(
as having a major role in the hydrothermal vent. Likewise, the
destruction of nitriles in waste streams will benefit from the
intentional addition of a strong acid.
(
(
34) Cutler, A. H.; Antal, M. J., Jr.; Jones, M., Jr. Ind. Eng. Chem. Res.
Acknowledgment. We are grateful to the Army Research
Office (Robert W. Shaw, Program Manager) for financial
support of this work on DAAG55-98-1-0253 and to Dr. Bill
Izzo for helpful discussions about nitrile kinetics.
1
4
1
988, 27, 691.
(35) Cvetanovic, R. J.; Singleton, D. L. Int. J. Chem. Kinet. 1977, 9,
81.
(
36) Cvetanovic, R. J.; Singleton, D. L. Int. J. Chem. Kinet. 1977, 9,
007.
37) Belsky, A. J.; Li, T.; Brill, T. B. J. Supercrit. Fluids 1997, 10,
201.
(
References and Notes
(38) Taft, R. W., Jr. Steric Effects in Organic Chemistry; Newman, M.
(
1) Miller, S. L. Biochim. Biophys. Acta 1957, 23, 480.
S., Ed.; John Wiley and Sons: New York, 1956.
(2) Miller, S. L.; Van Trump, J. E. In Origins of Life; Wolman, Y.,
(39) Perrin, D. D.; Dempsey, B.; Serjeant, E. P. pKa Prediction for
Ed.; D. Reidel Publishing Co.: Dordrecht, The Netherlands, 1981; p 135.
3) Peltzer, E. T.; Bada, J. L.; Schlesinger, G.; Miller, S. L. AdV. Space.
Res. 1984, 4, 69.
Organic Acids and Bases; Chapman and Hall: London, 1981.
(
(
(
(
40) Unger, S. H.; Hansch, C. Prog. Phys. Org. Chem. 1976, 12, 91.
41) Gal, E. M.; Shulgin, A. T. J. Am. Chem. Soc. 1951, 73, 2938.
42) Liler, M. Reaction Mechanisms in Sulfuric Acid and Other Strong
(4) Cronin, J. R.; Chang, S. In The Chemistry of Life’s Origin;
Greenberg, J. M., Pirronella, V., Eds.; Kluwer Academic: Dordrecht, The
Netherlands, 1993; p 209.
Acid Solutions; Academic Press: London, 1971; Vol. 23.
43) Foy, B. R.; Waldthausen, K.; Sedillo, M. A.; Buelow, S. J. EnViron.
Sci. Technol. 1996, 30, 2790.
44) Darensbourg, D. J.; Holtcamp, M. W.; Longridge, E. M.; Khan-
delwal, B.; Klausmeyer, K. K.; Reibenspies, J. H. J. Am. Chem. Soc. 1995,
(
(5) Schulte, M.; Shock, E. Origins Life EVol. Biosphere 1995, 25, 161.
(6) Haymon, R. M.; Macdonald, G. N. Am. Sci. 1985, 73, 441.
(7) Rona, P. A., Bostrom, K., Laubier, L., Smith, K. L., Eds.
(
Hydrothermal Processes at Sea Floor Spreading Centers; Plenum: New
York, 1984.
1
17, 318.
(
(
108.
(
(
45) Hall, G. A., Jr. J. Am. Chem. Soc. 1950, 72, 4709.
46) Meyer, J. C.; Marrone, P. A.; Tester, J. W. AIChE J. 1995, 41,
(
8) De Leer, E. W. B.; Simninghe Damste, J. S.; Erkelens, C.; De Galan,
L. EnViron. Sci. Technol. 1985, 19, 512.
9) Krieble, V. K.; Noll, C. I. J. Am. Chem. Soc. 1939, 61, 560.
2
(
47) Boock, L. T.; Klein, M. T. Ind. Eng. Chem. Res. 1993, 32, 2464.
48) Schaeffer, F. C. In The Chemistry of the Cyano Group; Rappoport,
(
10) Rabinovitch, B. S.; Winkler, C. A. Can. J. Res. 1942, 20A, 221.
11) Rabinovitch, B. S.; Winkler, C. A.; Stewart, A. R. P. Can. J. Res.
(
Zvi, Ed.; John Wiley and Sons: New York, 1970.
(49) Olah, G. A.; Kiovsky, T. E. J. Am. Chem. Soc. 1968, 90, 4666.
(50) Kortum, G.; Vogel, W.; Andrussow, K. Dissociation Constants of
Organic Acids in Aqueous Solution; Butterworths: London, 1961.
(51) Subcommittee on Ammonia. Ammonia; University Press: Balti-
more, MD, 1979.
1
942, 20B, 121.
(
(
(
(
12) Sullivan, M. J.; Kilpatrick, M. L. J. Am. Chem. Soc. 1945, 67, 1815.
13) Kilpatrick, M. L. J. Am. Chem. Soc. 1947, 69, 40.
14) Liler, M.; Kosanovic, D. J. Chem. Soc. 1958, 1084.
15) Mujumdar, G. G.; Dole, K. K.; Karve, D. D. J. UniV. Bombay,
Sci. 1947, 16A (3), 25.
(
16) Tsai, L.; Miwa, T.; Newman, M. S. J. Am. Chem. Soc. 1957, 79,
530.
17) Taylor, T. W. J. J. Chem. Soc. 1930, 2741.
(52) Lindsay, W. T., Jr. Proc. Int. Conf. Eng. W. Pa. 1980, 41, 284.
(53) Marshall, W. L.; Franck, E. U. J. Phys. Chem. Ref. Data 1981, 10,
295.
2
(