Optimization of oxidative method for synthesis of 1,3-diformyl-2,4,6- triethylbenzene

Article abstract of DOI:10.1023/B:RJAC.0000008300.47315.75

Oxidation of 1,3-diformyl-2,4,6-triethylbenzene with chromium anhydride in diluted H₂SO₄ to form 1,3-di(hydroxymethyl)-2,4,6- triethylbenzene was studied.

Full text of DOI:10.1023/B:RJAC.0000008300.47315.75

Russian Journal of Applied Chemistry, Vol. 76, No. 8, 2003, pp. 1274 1279. Translated from Zhurnal Prikladnoi Khimii, Vol. 76, No. 8, 2003,
pp. 1312 1316.
Original Russian Text Copyright
2003 by Apel’chugin, Krauklish, Punin, Tsypin.
ORGANIC SYNTHESIS
AND INDUSTRIAL ORGANIC CHEMISTRY
Optimization of Oxidative Method for Synthesis
of 1,3-Diformyl-2,4,6-triethylbenzene
S. B. Apel’chugin, I. B. Krauklish, A. E. Punin, and G. I. Tsypin
Tekhnolog Federal State Unitary Enterprise, Special Bureau of Design and Technology,
St. Petersburg State Technological Institute, St. Petersburg, Russia
Received February 13, 2003
Abstract Oxidation of 1,3-diformyl-2,4,6-triethylbenzene with chromium anhydride in diluted H2SO4 to
form 1,3-di(hydroxymethyl)-2,4,6-triethylbenzene was studied.
Synthesis of nitrile oxides [1], which are used as
efficient cross-linking agents, has become industrially
important for production of polymeric formulations
containing porophore-like compounds. 1,3-Di(nitrile
oxide)-2,4,6-triethylbenzene, whose stability is ensured
by three alkyl substituents screening the nitrile oxide
groups, is among such compounds.
panied by further oxidation of aldehyde groups to
carboxy groups [3]. Dimerization and polymerization,
involving the initial and final products of this stage
[4], and also oxidation of alkyl groups to aldehyde
and carboxy groups, may proceed simultaneously [5].
The rates of the above reactions are comparable, and
their extent depends on the reaction conditions, i.e.,
component ratio, temperature, and time.
The synthesis of the above compound includes
the stage of preparation of 1,3-diformyl-2,4,6-triethyl-
benzene (I) as an intermediate [1]. This stage may
proceed with the formation of the largest amount of
impurities and is, in the end, responsible for the qual-
ity and yield of the target product. Industrially,
compound I is commonly obtained by oxidation of
1,3-di(hydroxymethyl)-2,4,6-triethylbenzene (II) with
Cr(VI) in an aqueous solution of sulfuric acid, with
the yield of the reaction product not exceeding 50
55% [2].
Therefore, when formylation is performed under
nonoptimal conditions, the reaction product may
contain a wide variety of impurities, from the starting
diformyl compound to dicarboxylic acid and triethyl-
benzene. In addition, Cr(VI) is apparently transformed
not only into chromium(II) sulfate removable by wash-
ing with water, but also into water-insoluble chromi-
um(III) oxides. The presence of these latter in the re-
action mass hinders isolation of the target product and
results in its contamination.
In this study, the optimal conditions of formylation
of compound I were determined to obtain the target
product in maximum yield with high content of
the main substance.
CH2OH
Et
Et
+ 4CrO3 + 6H2SO4
CH2OH
The complex composition and instability of the fi-
nal reaction product make a kinetic study of the pro-
cess difficult. For this reason, we solved this problem
using the mathematical experiment design [7 9],
which allows construction of a mathematical model of
the process and its optimization with minimum ex-
perimental effort [10 12].
Et
II
CHO
Et
Et
3
+ Cr2(SO)4 + 12H2O.
CHO
Et
I
EXPERIMENTAL
As is known [3], transformation of a hydroxymeth-
yl compound into a formyl compound can be accom-
The whole set of tests was performed under identical
conditions on an installation consisting of a 500-ml
1070-4272/03/7608-1274 $25.00 2003 MAIK Nauka/Interperiodica

OPTIMIZATION OF METHOD FOR SYNTHESIS OF 1,3-DIFORMYL-2,4,6-TRIETHYLBENZENE 1275
jacketed glass reactor with bottom drain (see figure).
The apparatus was equipped with a four-blade high-
speed agitator, thermometer, and a thermocouple con-
nected to a KSP-4 recorder. The jacket of the reactor
was connected to a U-2 thermostat and a cold-water
line by two heat-transfer-agent circuits, which ensured
control of the reactor temperature in the range from
10 to 90 C. The volume of the reaction mass was
constant in all the tests (400 ml). Calculated amounts
of compound II and diluted H2SO4 were charged into
the reactor. The mixture was agitated for 30 min to
create a uniform suspension and warmed to 25 C
(fixed temperature of the reaction onset), and then
a 50% aqueous solution of CrO3 was added. After
Schematic of the formylation set-up: (1) reactor; (2) agi-
that, the reaction mass was heated to the required
tator; (3) thermometer; (4) thermocouple; (5) KSP-4 re-
temperature for 5 10 min and kept at 53 55 C for
60 min.
corder; (6) U-2 thermostat; (7) valve; (8) T-joint;, h.w and
c.w stand for hot and cold water, respectively.
In the first stage of the study, the reaction mass
was allowed to stay for 1 h. In the second stage,
the reaction mass was heated, on adding a CrO3 so-
lution, to a temperature determined from the disign
matrix of a full factorial experiment (FFE) and kept
at this temperature for a time also determined by
the FFE matrix. Then, cold water was fed into the re-
actor jacket to cool the reaction mass to 15 20 c, and
the target product I was extracted with carbon tetra-
chloride. The extraction was done with three portions
of the solvent: the volume of the first portion was
numerically equal to the mass of compound II taken
for the reaction. The volumes of the second and third
portions were half the above mass. The extraction time
was 10 15 min, and the separation time, 20 30 min.
The layers were separated in a separating funnel. Since
the target product I is unstable, the joint extract was
immediately used for preparing 1,3-diformyl-2,4,6-tri-
ethylbenzene dioxime (III) (reaction proceeds in
quantitative yield [5, 6]) by the known procedure [3].
The oximes were formed in a 1000-ml three-necked
flask equipped with a high-speed agitator rotating at
fixed rate and a thermometer.
(organic) layer was separated, whereas the aqueous
solution of disodium salt of oxime III was placed in
the reactor again and its pH was adjusted to 6.5 7.0.
For a sediment of product III to be formed, the solu-
tion was kept at room temperature for 1 h. The sed-
iment was filtered off, washed with water, and dried at
50 55 C in a thermostat to constant mass. The mass
of the substance, its boiling point, and the content of
the main substance were determined by liquid chro-
matography (LC).
A quantitative analysis was done on a Milikhrom
liquid chromatograph with a 64 2-mm KAKh-2 col-
umn packed with Separon-C18 (grain size 5 m). As
eluent served a mixture of dioxane, acetonitrile, and
water (volume ratio 1 : 4.5 : 4.5).
An analysis of an a priory information about for-
mylation made it possible to reveal a number of in-
fluencing factors (independent variables), examine
them, and arrange them in order of increasing priority
for the problem to be solved. To these factors we as-
signed, in the first place, the relative amounts of the
starting components and the process temperature and
time.
Water was poured into the reactor, and NaOH (sol)
was added under agitation. After NaOH was dis-
solved, the extract containing aldehyde I and then
(NH2OH2) H2SO4 were added. The pH of the reac-
tion mass was measured every 10 15 min and adjust-
ed to 6.5 7.0 with NaOH(sol) or (NH2OH2) H2SO4.
The reaction mass was heated to 45 50 C and kept at
this temperature for 1 h. After that, NaOH(sol) was
added in an amount necessary for oxime III to be
transformed into the corresponding disodium salt.
The reaction mass was kept at 45 50 C for another
1 h, cooled to 20 25 C, poured into a separating
funnel, and allowed to settle for 0.5 h. The lower
As optimization criterion (response function) served
the yield of the target product, with account taken
of the content of main substance in the target product,
i.e., its quality.
In the first stage of the study, a FFE was carried
out (Table 1, run nos. 1 8). To eliminate, to the max-
imum possible extent, the effect of interaction, the fol-
lowing three factors, determining the relative amounts
of the starting components, were used originally: x1,
the excess of CrO3 with respect to its amount re-
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 76 No. 8 2003

1276
APEL’CHUGIN et al.
Table 1. Results of realization of OCCD matrix*
The levels of factors and their variation ranges
were chosen on the basis of an a priory information.
The other formylation factors, i.e., the temperature of
the beginning of introduction of a Cr2O3 solution,
time of introduction, and also exposure temperature
and time, were also fixed on the basis of a priory
information at the levels considered to be optimal.
The systematic experimental error was compensated
for by the randomization method. The reproducibility
of the experiment was estimated (Table 2), the coef-
ficients of the regression equation (1) were calculated,
the significance of the regression coefficients was de-
termined, and the adequacy of the model was verified:
Run
no.
Concentration of main substance (%)
in parallel runs
Yield,
%
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
99
99
98
98
97
81
92
86
98
100
98
100
82
88
99
63
63
66
65
56
48
51
55
57
64
59
61
58
54
62
y = 58.31 + 0.83x1
+ 1.36x1x2
0.94x2 + 5.94x3
1.59x1x2x3.
(1)
The high value (5.94) of the coefficient of x3 shows
that the factor x3 (H2SO4 concentration) strongly af-
fects the optimization parameter, i.e., the yield of
the target product I. The sign + means that, with
the given factor increasing, the optimization param-
eter increases too, in good agreement with the data of
[13, 14], according to which the oxidizing capacity
of Cr(VI) strongly depends on the acidity of the reac-
tion mass and on the ability of Cr(VI) ions to form
complex compounds. Thus, the sulfuric acid concen-
tration in the oxidizing system affects rather strongly
the oxidation direction and the purity of the target
product.
*
Values averaged over two parallel runs.
Table 2. Estimate of the experiment reproducibility*
Value of
Characteristic
criterion
Maximum of the estimated reproduci-
bility variances, Ssp
31.75
15.82
Sum of estimated reproducibility
variances, Si2
Application of the steep-ascent method failed to
find a maximum in the region studied. This may be
a consequence of the possible curvature of the re-
sponse surface, which is indicated, despite the formal
adequacy, by the significance coefficients of the dou-
ble and triple interaction terms. For adequate math-
ematical description of the process, we used the sec-
ond-order regression equation derived using the or-
thogonal central composition design (OCCD). The
OCCD matrix is standard for all the three factors [8]
and is not randomized. The design of the first order,
the FFE matrix, was completed to the OCCD plan
by runs carried out at the center points (Table 1, run
nos. 10 15). The results of its realization are listed in
Table 1. These data were used to determine the coef-
ficients of the second-order regression equation (2)
and to estimate the significance of the equation coef-
ficients and its adequacy.
Cochran criterion, Gcal [8, 9]
Cochran criterion Gtab for P = 0.95
and N = 8
Estimated reproducibility
Number of runs N
0.50
0.68
Gtab > Gcal
8
2
1
Number of parallel runs k
Number of degrees of freedom f for Sj2
( f = k
1)
Estimated reproducibility variance Sy2
3.97
8
Number of degrees of freedom for Sy2
[ f = N(k
1)]
Estimated variance of the mean Sy2 value
1.98
15
Number of degrees of freedom
*
Calculations were based on the data of Table 1.
quired by stoichiometry (%); x2, the H2SO4 amount
1
[mol mol of initial compound II]; x3, the acidity
of the reaction mass in terms of H2SO4 concentra-
tion in the total amount of water used in the pro-
cess (%).
y = 60.43
0.95x2 + 4.69x3 + 1.36x1x2
1.59x1x2x3
2.77x23.
(2)
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 76 No. 8 2003

OPTIMIZATION OF METHOD FOR SYNTHESIS OF 1,3-DIFORMYL-2,4,6-TRIETHYLBENZENE 1277
Table 3. Results of tests of potentially optimal modes
Product
Product yield, %
experimental
Run
no.
x1
x2
x3
content
Box’s
*
Tmp
,
of main
substance, %
recalculated to calculation
100% content
raw
1
2
3
4
5
6
19.7
0.3
0.3
3.5
3.5
3.5
2.0
2.0
3.0
3.0
3.0
3.0
15.0
15.0
15.0
15.0
22.0
26.4
105
<60
74
87
100
81
69
72
110.5
115.5
120
113
109
67
69
69
67
67
50
60
69
54
47
67
66
70
77
79
*
The melting point of recrystallized sample is 121 C.
The equation is adequate to the experimental data
within the range of the influencing parameters, studied
here.
chromium anhydride, which is confirmed by the pres-
ence of unreacted bishydroxymethyl derivative of tri-
ethybenzene in the reaction mass.
The second-order model developed was optimized
using Box’s complex method [12], which made it
possible to calculate the coordinates of the maximum
of the target function.
In the next run (Table 3, run no. 4), chromium an-
hydride was taken in a 3.5% excess with respect to
the stoichiometry at the same amount of sulfuric acid
and its same concentration in the reaction mass. As
a result, a 99.5% pure product was obtained in 69%
yield (recalculated to 100% content of the main sub-
stance). In run nos. 5 and 6 (Table 3), we studied
whether an increase in the x3 factor (H2SO4 concen-
tration) outside the factor space to, respectively, 22
and 26% affects the yield of the target product. The
raw product was obtained in rather high yield, but
with drastically impaired quality. This result can be
accounted for by the fact that, under these conditions,
the rate of side oxidation reactions and the fraction
of their products in the raw product grow, in good
agreement with the data on how the oxidation depth
and the purity of the target product depend on the sul-
furic acid concentration [13, 14].
The data of Table 3 show that the maximal yield is
achieved at the boundary of the factor space. An at-
tempt to perform the experiment at the point obtained
in optimization (Table 3, run no. 1) failed owing to
the impossibility to separate the reaction mass in ex-
traction.
Apparently, at large excess of chromium anhydride
and stoichiometric amount of sulfuric acid (Table 3,
run no. 1), we presumably obtain a considerable
amount of water-insoluble chromium(III) oxide, which
serves as emulsifier, facilitates formation of a stable
emulsion, and impedes layer separation.
As a result, only inconsiderable portion of a poor-
quality final product could be separated from the re-
action mass in this run (Table 3, run no. 1). With
the amount of chromium anhydride diminished to
a nearly stoichiometric value under otherwise invari-
able conditions (Table 3, run no. 2), the yield of
the raw product increased, but its quality remained
poor (content of the main substance 74%).
The first stage of the study resulted in the follow-
ing conclusion: to perform formylation with high
yield and to obtain a high-quality product, it is neces-
sary to use CrO3 in a 3 4% excess with respect to
the stoichiometry, and H2SO4 in an amount of 3 mole
per 1 mole of initial compound II, and to maintain
its concentration in the reaction mass within the range
14 15 wt %.
In the next run (Table 3, run no. 3), at a minimum
excess amount of chromium anhydride and the same
x3 factor (H2SO4 concentration), the amount of sul-
In the second stage, a mathematical model of for-
mylation was constructed, which describes the rela-
tionship between the process temperature and time,
on the one hand, and the optimization parameter (re-
sponse function), i.e., the yield of the target product,
on the other (with account taken of the content of ini-
tial substance in it).
1
furic acid increased [to x2 = 3.0 mol mol of initial
compound II] with respect to stoichiometric amount
(2.0 mole). The comparatively poor quality of the raw
product at its sufficiently high (69%) yield can be
apparently accounted for by the insufficient amount of
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 76 No. 8 2003

1278
APEL’CHUGIN et al.
Table 4. Results of realization of FFE design matrix
lated by Box’s method, a 10-min exposure time and
a 69.5 70.0 C temperature, were carried out. As a re-
sult, the product containing 99.5% main substance
was obtained in 70% yield, which is rather close to
Box’s value (71%).
Run
no.
Content of main
substance, %
Yield,
%
1
2
3
4
5
91
90
87
89
94
56
49
56
41
60
The recommended relative amounts of the reagents,
and the process temperature and time worked very
well under pilot plant conditions: oxime III was ob-
tained with high content of the main substance (95
98%) in good yield (66 68%), close to the expected
value. This result shows that the process conditions
chosen ensure predominant oxidation of the initial
compound II to dialdehyde I.
The FFE was also performed. As indicated above,
the yield of the target product was used as response
function, with account taken of the content of main
substance in it. We chose two influencing parameters,
z1 and z2, i.e., the reaction temperature ( C) and re-
action time (min), respectively. The level of factors
and their variation range were chosen on the basis of
an a priory information. The FFE standard matrix
for two factors was randomized.
CONCLUSIONS
(1) Synthesis of 1,3-diformyl-2,4,6-triethylbenzene
from 1,3-di(hydroxymethyl)-2,4,6-triethylbenzene by
oxidizing the latter with CrO3 in diluted H2SO4 was
studied using a mathematical method of experiment
design. The model of the process was constructed and
optimized.
The results of FFE are listed in Table 4. These data
were used to calculate the coefficients of the regres-
sion equation, determine their significance, and verify
the adequacy of the model. A first-degree regression
equation adequately describing the response surface
was obtained:
(2) Under the optimal process conditions [CrO3
taken in 3 4% excess with respect to stoichiometry,
1
and H2SO4 in an amount of 3.0 mol mol of initial
compound II, 14 15 wt % H2SO4 concentration in
the reaction mass, and 10-min keeping of the reac-
tion mass at 69.5 70 C the yield of the final product
is as high as 67 70%.
y = 61.57 + 5.47z1 + 2.06z2
2.25z1z2.
(3)
Analysis of Eq. (3) shows that the coefficient of
the z1 factor exceeds that of z2, which confirms
the positive effect exerted by a rise in temperature.
The significant coefficient of the pair-interaction term,
having minus sign, indicates that for the optimiza-
tion parameter to grow, the z1 factor must increase,
rather than decrease.
REFERENCES
1. Yakubov, A.P., Tsyganov, D.V., Belen’kii, L.I., and
Krayushkin, M.M., Izv. Akad. Nauk SSSR, Ser. Khim.,
1991, no. 5, pp. 1201 1203.
2. Krauklish, I.V., Apel’chugin, S.B., Tsypin, G.I., and
Chernyshova, E.Yu., Abstracts of Papers, Konferen-
tsiya po khimii i khimicheskim produktam (Conf. on
Chemistry and Chemical Products), Moscow: Ross.
Khim-Tekhnol. Univ. im. D.I. Mendeleeva, 2002,
p. 102.
To optimize the process, we used the steep-ascent
method. However, it failed to find a maximum in
the region studied. This may be a consequence of
the curvature of the response surface, as evidenced,
despite the formal adequacy, by the significant coef-
ficients of the pair-interaction term and by the differ-
ence between the free term of the regression equation
and the yield at the center of the design (Table 4, run
no. 5).
3. Weygand, K. and Hilgetag, G., Organischchemische
Experimentierkunst, Leipzig: Barth, 1970.
4. Buehler, C.A. and Pearson, D.E., Survey of Organic
Syntheses, New York: Wiley-Interscience, 1970.
The use of Box’s complex method for optimizing
the resulting first-order equation made it possible to
avoid additional experiments to search for local ex-
trema. This enabled calculation of the coordinates of
the maximum of the target function.
5. Becker, H., Berger, W., and Domschke, G., Organi-
kum. Organisch-chemisches Grundpraktikum, Berlin:
Deutscher Verlag der Wissenschaften, 1967.
6. Houber, J. and Weyl, W.A., Methoden der organi-
schen Chemie, Stuttgart: Tieme, 1953.
The maximum yield is achieved at the boundary of
the factor space. An experiment with factors calcu-
7. Charykov, A.K., Matematicheskaya obrabotka rezul’
tatov khimicheskogo analiza (Mathematical Proces-
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 76 No. 8 2003

OPTIMIZATION OF METHOD FOR SYNTHESIS OF 1,3-DIFORMYL-2,4,6-TRIETHYLBENZENE 1279
sing of Chemical Analysis Data), Leningrad: Khimiya,
1984.
khimicheskaya tekhnologiya (Computer World and
Chemical Engineering), Leningrad: Khimiya, 1991.
11. Markova, E.V. and Rokhvarger, A.E., Matematiches-
koe planirovanie khimicheskogo eksperimenta (Math-
ematical Design of Chemical Experiments), Moscow:
Znanie, 1971.
8. Sautin, S.N., Planirovanie eksperimentov v khimii i
khimicheskoi tekhnologii (Experiment Design in
Chemistry and Chemical Engineering), Leningrad:
Khimiya, 1975.
12. Box, M., Comp. J., 1965, vol. 8, no. 1, pp. 42 52.
13. Dickerson, R.E., Gray, H.B., and Haight, G.P., Chem-
ical Principles, New York: de Gruyter, 1988.
14. Horton, A., Terrll, K., Haffman, J.C., et al., J. Inorg.
Chem., 1997, vol. 36, no. 21, pp. 4875 4882.
9. Ruzinov, L.P., Statisticheskie metody optimizatsii
khimicheskikh protsessov (Statistical Methods for Op-
timization of Chemical Processes), Moscow: Khimiya,
1972.
10. Sautin, S.N. and Punin, A.E., Mir komp’yuterov i
RUSSIAN JOURNAL OF APPLIED CHEMISTRY Vol. 76 No. 8 2003