438 82 (1963) RECUEIL

541.461.6.01:541.593.21 1 :541.594.3
PREPARATION OF ADIPIC ACID BY OXIDATION OF CYCLO-
HEXANOL AND CYCLOHEXANONE WITH NITRIC ACID
Part 111. Reaction kinetics of the oxidation
BY
W. J. van ASSELT and D. W. van KREVELEN
Central Laboratory Staatsmijnen, Geleen and
Laboratory for Chemical Technology, Technological University, Delft

The reaction velocity of the oxidation of cyclohexanol and cyclohexanone

with nitric acid (without a catalyst) can be calculated from the chromato-
graphically measured change in the concentration of some of the reaction
products: nitrolic acid and adipic acid.
The oxidation reaction appeared to be first order in cyclohexanone, which
rapidly forms in the oxidation of cyclohexanol. The oxidation rate can be
calculated by means of a simple empiric formula.
The catalysts NH4V03 and C u ( N 0 3 ) ~d o not affect the oxidation rate proper.
The rate at which Dione decomposes into adipic acid (kv) in the presence of
vanadate can be calculated from the increase of the adipic acid concentration.
If vanadate is absent, an impression of the rate a t which Dione is further
oxidized into by-products (kH) can be obtained from the increase of the glutaric
acid and succinic acid concentrations.

A. Introduction
As described in a previous publication 1, the oxidation of cyclohexanol
can be represented as follows:

cyclohexanol -5 cyclohexanone adipic acid

rapidly
Dione

ghtaric acid
succinic acid

As shown above, there is a principal reaction (double arrow) accompanied

by some side reactions.The determination of k N A in the consecutive principal
reaction has already been described in Part 11. The present part deals mainly

1 W. J. van Asselr and D . W. van Krevelen, Part I, Rec. trav. chim. 82, 51 (1963).
 Table I
Some results of oxidation experiments
~~ -
Products (2)
Adipic
ment acid (1)
Reaction conditions Time (1) Adipic Glutaric Succinic from N
Nitrolic acid acid acid nitrolic
acid acid
(min) (mmoles) (mmoles) (mmoles) (mmoles) (mmoles) (m

51.0% H N 0 3 1.5 1.89 0.56 0.00

Volume = 50ml 4 3.35 0.75 0.02
Temp. = 15.0" 8 4.75 1.20 0.06
Cyclohexanone = 10.48 mmol 15 5.80 2.84 0.14
NHdV03 = lmmol 25 6.17 4.11 0.27
(HN0z)o = 9mmol 40 6.08 4.65 0.47

40.4% H N 0 3 2.5 3.35 0.22 0.0 0.0

Volume = 100ml 6 6.76 0.63 0.16 0.08
Temp. = 25.2" 12 8.40 1.16 0.33 0.41
Cyclohexanol = 20.0 mmoI 25 8.60 1.35 0.42 0.63
Cu(N03)~ = 2 mmol 60 7.60 2.40 0.64 I .55
300 3.53 6.51 1.03 5.87
-
 Tab I e 1 cori/iirired
Some results of oxidation experiments

(roducts (2)

me,
Reaction conditions
I Ij
Time (1) Adipic Glutaric Succinic
Adipic
acid
from
Nitrolic acid acid acid nitrolic
i acid acid
(min) (mmoles) (mmoles) (mmoles) (mmoles) I (mmoles)

1
I

40.4% HNO 3 2.5 0.93 0.37 0.04

Volume = 50ml 7 , 1.55 0.71 0.07
Temp. = 15.0" 15 1.67 1.21 0.08

1
Cyclohexanol = 9.81 mmol 30 2.02 3.36 0.07
NH4V03 = 1 mmol 2.02 7.25 0.16
WN0z)o = 20mmol 90 2.00 7.40 0.30
(HNOz)tinal = 18 mmol 60
I
I
30.4%HNOs 2 I .64 0.15 0.06 0.07 ~ 0.01
Volume = 50ml 7 3.37 0.57 0.28 0.12 I 0.06
Temp. = 30.0" 15 3.90 0.91 0.57 0.15 0.18 4
Cyclohexanone = 10.61 mmol 35 3.42 1.46 1.10
(HN0z)o = ~IXUIIOI 60 3.07 1.77 1.41
Preparation of adipic acid by oxidation of cyclohexanol, efc. 82 (1963) RECUEIL 441
- __
with kox. For the calculation we depend on the chromatographic deter-
mination of nitrolic acid and adipic acid, because Dione and cyclohexanone
cannot be quickly determined in a simple way.
By the side of nitrolic acid, Dione is formed as a by-product, the ratio
in which Dione and nitrolic acid are produced (Dione/NA ratio) changing
with the reaction conditions 1.
If vanadate is used as catalyst, the Dione is converted to adipic acid.
Under the reaction conditions chosen this reaction appears to proceed at
a moderate rate, and in some cases both kox and kv can be calculated.

B. Methods
General procedures
The oxidation experiments were carried out in the reactor described in
the previous publications 1, 2, at temperatures varying between 10 and 40",
and at nitric acid concentrations of 67-20 per cent by weight of N H 0 3 .
1 g of cyclohexanol or cyclohexanone was added to 50 ml of nitric acid and
the temperature was kept constant within f 0.1". The nitrous acid concen-
tration is measured before and after the reaction.
Via a siphon, liquid samples of the reaction medium are collected i n an
amount of ice to which some hydrazine sulphate has been added, the
removal of the NH02 and the cooling effect causing the reaction to stop
immediately. These samples are analyzed chromatographically 1.
To study the side reactions, some oxidation experiments, in which pure
Dione was added to the nitric acid, were carried out in the same way; the
Dione dissolved rather slowly.

C. Results and discussion

I . Determination of the reaction rate of the principal reaction (kox)
Some examples of experiments that are of importance with regard to the
following calculations are shown in table I . For complete experimental
data reference is made to 3.
Tn the oxidations performed without a catalyst, the sum of nitrolic acid
and adipic acid becomes constant after some time, whereas the conversion
is apparently not complete. A characteristic example is given in figure 1.
The lacking portion of the products consists of Dione, and small amounts
of glutaric acid and succinic acid, if any. True, the amounts of nitrolic and
adipic acid are still decreasing and increasing resp. but the changes corre-
spond to the decomposition rate of nitrolic acid 2.

W. J. vat1 Asselt and D. W. van Krevelerr, Part TI, Rec. trav. chim. 82, 429 (1963).
W. J . van Asseli Thesis, Delft (1960).
442 W. J. van Assel! and D. W. van Kreveleti

L I S u m of nitrolic acid ond adipic acid

!I
.- Nitrolic acid 0-

Adipic acid
I
0-

-o--o-o
.
z o
0 4 6 8 10 12 14 16 18 20 22

t Fig. 1.
_c Time (rnin)

Oxidation of 10.63 mmoles cyclohexanol in 50 ml of

nitric acid (51.0%) without a catalyst. Temp. 25.0.

However, in experiments where the nitric acid concentration is below

40 %, the chromatographically determined amount of adipic acid appears
to be much larger than can be expected on the basis of the nitrolic acid
decomposition (Table I, Experiment No. 4). The amount of glutaric acid
is also large then, while, unlike in other analyses, the separation between
adipic acid and glutaric acid on the chromatographic coluinn is very poor
and the end-point is obscured.
The impression is gained that some Dione compound or other interferes
with this determination, and the oxidation rate has therefore been calcu-
lated from the change in the amount of nitrolic acid increased by the
calculated amount of adipic acid meanwhile formed from nitrolic acid.
The oxidation rate can now be calculated in two different ways:
a. If it is assumed that the Dione/NA-ratio during the oxidation is con-
stant, then the amount of nitrolic acid (including the amount of adipic acid
formed from nitrolic acid) will, at any moment, be proportional to the
amount of converted cyclohexanol or cyclohexanone.
The reaction-velocity is then given by:
d(
-
dt
1 -( 1 + Dione/NA ) dt-
d(anonp\

so that the reaction-velocity can be directly found from the change in the
nitrolic acid concentration. To this end we introduce the concept potential
nitrolic acid concentration (NApot.). This is the difference between the
concentration at t = m, and the concentration of nitrolic acid (+ adipic
acid from NA) at time t. The NApot.is consequently proportional to the
Preparation of adipic acid by oxidatioii of cyclokexaiiol, e/c. 82 (1963) RECUEIL 443

concentration of unconverted cyclohexanone. Plotting log. (NAPot.) as a

function of time yields straight lines. Some examples of those lines are given
in Figure 2, where it can be seen that the reaction is first order with respect
to the cyclohexanone concentration.

Unit time ( 4 min

____c 1
Fig. 2. Determination of kos at a supposedly constant Dione/NA ratio,

The reaction-velocity constant (kox) can now be calculated from the

slope of the lines. Some of the values found are listed in Table 2, while in
Figure 3 kox for all experiments is given as a function of temperature and
acid concentration.

As can be seen i.a. from Table 2 the oxidation rates of cyclohexanol and
cyclohexanone are equal. The conversion of -an01 into -anone consequently
proceeds much quicker than the further oxidation. It has also appeared
that extra addition of HNOz (which, as a matter of fact, is necessary i n the
case of cyclohexanone) does not affect kox. Neither does copper have any
influence on the reaction in this temperature range, as is to be expected1,
while the use of vanadate increases only the amount of adipic acid, without
affecting the conversion of cyclohexanone to nitrolic acid and Dione.
With vanadate being used as a catalyst, kox can be calculated from the
sum of NA and adipic acid formed from nitrolic acid (calculated with the
aid Of kNA).
b. So far, it has been assumed that the Dione/NA ratio remains constant.
Tn reality, however, the HNOs-concentration will decrease owing to forma-
444 W. J. van Asselt and D. W.van Krevelen
____ -. __ _ _ _ ~ _- -

T a b l e I1
Comparison of the reaction velocities (with and without a catalyst)

Experimer Temperature Catalyst kox . lo3

no. Reactant (mmoles/ I ) ( I /set)
(C)

anol 10.0 2. I
anol 14.1 2.5
anol 20.0 4.3

8 anol 51.0 15.0 2.2

9 anone 51 .O 15.0 2.3
I anone 51 .O 15.0 2.5
10 anol 51.0 25.0 5.8
11 anol 51.0 25.1 5.9

12 anol 40.4 20.7 - 2.5

13 anol 40.4 20.1 NH4V03 20 2.5
14 anol 40.4 25.0 - 3.8
15 anol 40.4 25.2 NH4V03 40 3.7

16 anol 30.4 30.6 3.8

17 anol 30.4 35.7 Cu(N0a)a 30 5.0
{ NHiVO3 30
18 anol 30.4 35.1 5.1
Cu(N03)z 30
Preparation of adipic acid by oxidation of cyclohexanol, elc. 82 (1963) RECUEIL 445
______ -~ -- .- - ..__

tion of nitrolic acid, with the consequence that the Dione/NA ratio will
also decrease during the oxidation. The difference is greater as more nitrolic
acid is produced, in other words as the nitric acid concentration is higher
and/or the temperature is lower. The HNO2-concentration at any moment
can be calculated from the initial concentration minus the amount of
nitrolic acid formed. How much Dione has formed beside the nitrolic acid
can now be calculated stepwise over small intervals at the actual HN02-
concentrations; the cyclohexanone concentration as a function of time can
then be found by summation. An example of this procedure is given in
Table 3.
Even in cases as unfavourable as that of the example, the value of kox
calculated from the change in the -anone concentration proves to be only
slightly higher than that found previously (4.6 as compared with 4.3).
Various calculations have shown that the reaction-velocity constants
obtained by these two methods differ by no more than O-lO%. The values
found by the first method have therefore been maintained for the further
calculations.

2. Influence of temperature and nitric acid concentration

From Figure 3 it can be seen that there is no regular change in the temper-
ature dependence as a function of the nitric acid concentration; for this
reason the lines through the observation points have all been drawn at the
same angle of inclination. The activation energy calculated from the course
of these lines amounts to 15 850 cals/grmol.

Fig. 4. kos as function of the nitric acid concentration at T = 20".

 Table I11
Calculation of kox at inconstant Dione/NA ratio

imental conditions: 56% H N 0 3 , volume = 50 ml

Temperature = 20.0 "C
Cyclohexanol = 9.53 mmoles

e
NA + adipic A NA HNOz Dione/NA A Dione +
N A Dione Unconverted
acid cyclohexanone
) (mmoles) (mmoles/l) ratio * (mmoles) (mmoles)
(mmoles) (mmoles)

1.56 1.56 175 0.43 0.67 2.23 7.54

5 3.39 1.83 140 0.40 0.73 4.19 4.98
5.33 1.94 100 0.28 0.54 1.21 2.50 4
6.74 1.41 70 0.21 0.29 8.97 0.80
7.23 0.49 50 0.16 0.08 9.54 0.23 (
7.44 0.21 35 0.13 0.02 9.17 0.0

fig. 8 .
Preparation of adbic acid by oxidation of cyclohexunol, etc. 82 (1963) RECUEIL 447

To examine the influence of the nitric acid concentration, the initial

concentrations as given in Figure 3 have to be corrected, because during
the oxidation HNO3 is consumed and HzO is formed. The average values
during the oxidation are about 1 per cent by weight below the values
mentioned in the tables and the figures.

Table IV
Oxidation of pure Dione with nitiic acid without a catalyst

Adipic Glutaric Succinic Sum

Time kit
acid acid acid
(min) (mmoles) (fisec)
(mmoles) (rnrnoles) (mmoles)

6 0.20 0.16 I .52 1.88

16 0.24 0.61 3.33 4.18 1 4 . 10-3
60 0.22 0.89 4.81 5.92

In Figure 4 the values of kox at 20" read from Fig. 3 are plotted versus the
HN03-concentration. It can be seen here that above 30% HNO3 there
exists a simple empirical relation between koxxand the HNO3-concen-
tration.
The equation for the oxidation velocity can then be written as:
dC
dt
- kos .C = 2.8 . 108 C . ( C ~ h . 0-~ 2.9) . e-15850/11T
where: C = concentration of cyclohexanol, or cyclohexanone;
C H N O=~ concentration of H N 0 3 (gmol/I).
In a following publication * it will be shown that the reaction kinetics,
measured in batch experiments, are completely confirmed by experiments
in a continuous reactor.

3. Determination of the reaction rate of the side-reuctions (kv and kH)

Dione decomposes into adipic acid (under the influence of vanadate)
or into lower oxidation products (without a catalyst). The experiments
using vanadate have shown that in most cases the adipic acid concentration
is still increasing when NAtotaI has already become constant. The conversion
of Dione therefore proceeds more slowly than the oxidation proper, the

* Chem. Eng. Sci. In press.

448 W. J . van Asselt and D. W. van Krevelen
- ~ -

difference increasing as the nitric acid concentration is lower. The moment

when NAtotal becomes constant, the conversion of cyclohexanone is com-
plete, so that the amount of Dione present is known. From the increase of
the amount of adipic acid minus the amount formed from nitrolic acid, the
decomposition rate of Dione can be calculated, on the assumption that the
reaction is first order with respect to Dione.
Tf vanadate is absent, the further oxidation of Dione proceeds very
slowly, as has been demonstrated before 1. The oxidation products consist
of appr. 70% of glutaric acid and succinic acid and appr. 30% of lower
decomposition products. The amount of Dione converted per unit time can
then be found from the increase of the amounts of glutaric acid and succinic
acid, while the Dione concentration can be calculated from the mass
balance.
A numerical example of the latter method is given below.
Conveision of Dione without a catalyst (Experiment 2, Table 1).
After 25 minutes NAtotal becomes constant. Fiom the 25th to the 60th minute the
sum of glutaric acid and succinic acid increases by 1.14 mmol.
The average concentrations in this peiiod are: NAtotal = 10.00 mmol, glutaiic acid i-
succinic acid = 1.60 mmol, showing that about 8 mmol of the hemihydrate are still
present in the solution. The reaction velocity can now be calculated as follows:

Besides the values calculated from the standard experiments after com-
pletion of the oxidation proper, some additional data have been obtained
from the oxidation experiments with pure Dione.

1 kox 40%HN03
2kv ,, ,, 2OvI
3kv ,. ,. 1OV/
4kv 35%.. ,,

T1 1054-

Fig. 5. kv and kH as function of the temperature.

Prepuruiion of udipic acid by oxidurion ofcyclolrexunol, efr. 82 (1963) RECUEIL 449
_. _ _ _ _ _ _ _ ~ _ _ _ ~ ~ _ _ _ _ _ _
If vanadate is present, the reaction velocity in the latter experiments
cannot be determined, because the added Dione dissolves so slowly in the
nitric acid that the dissolved portion is immediately oxidized further.
Without vanadate, the reaction proceeds so slowly that only a small
portion is converted during solution. An example is given in Table 4 (solu-
tion time: 6 minutes). The calculation of k H is similar to that of kox.
In Figure 5 all the calculated values of kIi and kv are given as functions
of the temperature. By way of comparison the figure also gives kox for
experiments with 40% nitric acid.

The number of data is too small and, moreover, not accurate enough to
permit formulae to be drawn up also for kv and k H .
From the position of the lines 2 and 3 in Figure 5 (both running parallel
to line 4) the impression is gained, however, that the conversion of Dione
under the influence of vanadate is proportional to the vanadate concen-
tration.
(Received December 14th. 1962).