Optimisation of the preparation and isolation of 5-amino-2,4,6-triiodoisophthalic acid dichloride

Article abstract of DOI:10.1021/op980055+

A scaleable process for the chlorination of 5-amino-2,4,6-triiodoisophthalic acid to the corresponding acid chloride, a building block for the synthesis of iodinated X-ray contrast agents, has been developed. Two dimeric byproducts have been isolated and assigned structures consistent with an amide and an anhydride. Hydrolysis of the N-sulfinyl intermediate and simultaneous crystallisation of the phthalic acid chloride have been optimised by applying fractional factorial design.

Full text of DOI:10.1021/op980055+

Organic Process Research & Development 1999, 3, 38−43
Optimisation of the Preparation and Isolation of
5-Amino-2,4,6-triiodoisophthalic Acid Dichloride
Harrie J. M. Gijsen, Herman C. C. K. van Bakel, Wim Zwaan, and Lumbertus A. Hulshof*
DSM-Andeno, P.O. Box 81, 5900 AB Venlo, The Netherlands
Abstract:
A scaleable process for the chlorination of 5-amino-2,4,6-
triiodoisophthalic acid to the corresponding acid chloride, a
building block for the synthesis of iodinated X-ray contrast
agents, has been developed. Two dimeric byproducts have been
isolated and assigned structures consistent with an amide and
an anhydride. Hydrolysis of the N-sulfinyl intermediate and
simultaneous crystallisation of the phthalic acid chloride have
been optimised by applying fractional factorial design.
Introduction
The 5-amino-2,4,6-triiodoisophthalic moiety is the basic
backbone responsible for the contrasting capability of many
water-soluble, nonionic iodinated X-ray contrast agents1 for
imaging soft tissues, such as Iopamidol2 (1), Iohexol3 (2),
and Ioversol4 (3). Consequently, several examples of the
synthesis of the key intermediate building block 5-amino-
2,4,6-triiodoisophthalic acid chloride (5) have been repor-
ted.5-12 Conversion of 5-amino-2,4,6-triiodoisophthalic acid
(4) to the corresponding acid chloride 5 is usually carried
out using thionyl chloride as the chlorinating agent.
Most of the procedures described for the chlorination of
4 either have a low space-time yield or are difficult to scale-
up because of safety issues, generation of unacceptably high
levels of impurities, and problems associated with isolation.
After we had encountered problems during scale-up of
isolated crystals, we embarked on an extensive study to find
our standard reaction conditions (see Experimental Section)
due to poor crystallisation and high moisture content of the
the optimal conditions necessary for a robust process to
produce ton quantities of 5-amino-2,4,6-triiodoisophthalic
acid dichloride.
(1) Radiocontrast Agents, Handbook of Experimental Pharmacology, Vol. 73;
Sovak, M. Ed.; Springer-Verlag: Berlin 1993.
(2) (a) Felder, E. InVest. Radiol. 1984, 19, S164. (b) Felder, E.; Grandi, M.;
Pitre, D.; Vittadini, G. Anal. Profiles Drug Subst. 1988, 17, 115
(3) (a) Haria, M.; Brogden, R. N. CNS Drugs 1997, 7, 229. (b) Haavaldsen, J.;
Nordal, V.; Kelly, M. Acta Pharm. Suec. 1983, 20, 219.
(4) Thesen, R. Pharm. Ztg. 1992, 137, 36.
Optimisation Studies
To achieve acceptable reaction rates and to drive the
reaction to completion, the formation of the acid chloride 5
reported in the literature is generally carried out at elevated
temperatures and in the presence of a catalyst such as DMF.7
The reaction has been performed in an excess of thionyl
chloride, acting both as a reagent and as the solvent,5-7 but
recent patents describe the use of cosolvents to render the
reaction more controllable and to avoid the use of an excess
of thionyl chloride.8-10 In earlier patents, Mallinckrodt
researchers describe the use of ethyl acetate,11 but this gives
low yields and partial breakdown of the solvent. Other esters,
e.g., isopropyl acetate, lead to similar problems. Chlorinated
solvents such as methylene chloride10 are not acceptable for
environmental reasons and, moreover, result into prolonged
reaction times and give the product as a finely divided
(5) Smith, K. R. Ger. Offen. DE 2710730, 1977; Belg. Pat. BE 852418, 1977
(to Mallinckrodt).
(6) For example, see: Felder, E.; Pitre, D. Ger. Offen. DE 2547789, 1976 (to
Savac). Felder, E.; Vitale, R. S.; Pitre, D. E. U.S. Patent 4001323, 1977 (to
Savac). Felder, E.; Pitre, D. Eur. Pat. Appl. EP 23992, 1981 (to Bracco).
(7) Dimo, I.; Bonnemain, B.; Hardouin, M. J. C.; Lautrou, J. (S. A., Fr.). Eur.
Pat. Appl. EP 118347, 1984 (to Guerbet).
(8) Mauro, M.; Viscardi, C. F.; Gagna, M. PCT Int. Appl. WO 96/37459, 1996
(to Fructamine).
(9) Villa, M.; Pozzoli, C.; Russo, L.; Costaldi, G. PCT Int. Appl. WO 96/36590,
1996 (to Zambon).
(10) Cannata, V.; Velgi, C.; Barreca, G. PCT Int. Appl. WO 96/16927, 1996 (to
Zambon).
(11) For example, see: Kneller, M. T.; Lin, Y.; Wheatley, J. PCT Int. Appl.
WO 94/05337, 1994 (to Mallinkrodt). Lin, Y. Eur. Pat. Appl. EP 83964,
1983 (to Mallinckrodt).
(12) Pillai, K. M. R.; Diamantidis, G.; Duncan, L.; Ranganathan, R. S. J. Org.
Chem. 1994, 59, 1344.
38
•
Vol. 3, No. 1, 1999 / Organic Process Research & Development
10.1021/op980055+ CCC: $18.00 © 1999 American Chemical Society and Royal Society of Chemistry
Published on Web 12/11/1998

Scheme 1
a hydrolysis step during workup of the reaction mixture. This
can be achieved by washing a solution of the crude N-sulfinyl
compound 6 in a suitable organic solvent with an aqueous
phase for an extended period of time.5,6 The resulting
hydrolysed product 5 also has to dissolve reasonably well
in the organic solvent used, making subsequent crystallisation
in the same solvent difficult to perform without any losses.
A solvent switch has been applied in these cases by others,5,6
making the process more elaborate. An attractive alternative
has been worked out by us to perform the hydrolysis and
precipitation/crystallisation in one step.
precipitate, which is difficult to isolate. A patent by Fruct-
amine SPA describes the use of a range of aliphatic and
aromatic hydrocarbons.8 In our hands, however, the use of
toluene still requires an excess of thionyl chloride, which is
difficult to recycle as a mixture with toluene.
In most solvents, the N-sulfinyl compound 6 is more
soluble than the desired product 5, and addition of water to
a solution of the N-sulfinyl compound 6 in a solvent in which
the product is not very soluble should give precipitation of
5. Both product 5 and N-sulfinyl compound 6 are very poorly
soluble in water, and (partial) precipitation of 5 during
hydrolysis of 6 will inevitably lead to inclusion of 6 in the
precipitate of 5 and hence to incomplete hydrolysis of the
N-sulfinyl group. As a consequence, some impurity levels
of 6 have always been observed in the final product. Only
use of large volumes of organic cosolvent, which keeps the
product 5 in solution until hydrolysis of 6 is complete, can
avoid inclusion of impurity 6 in the product. This, however,
is not attractive from an economic point of view. The
presence of 6 as a major impurity in 5 is often not mentioned
in the literature14 and has probably not been noticed in
previous work, since 6 is hydrolysed to 5 during reversed-
phase HPLC analysis of the product. Therefore, in addition
to this standard HPLC method, a water-free, normal-phase
method was employed to determine the level of N-sulfinyl
compound. In this way, evidence has been found that the
hydrolysis of 6 with concomitant crystallisation/precipitation
of 5 with just water leads to unacceptable levels (above 1%)
of 6 in the final product. To solve this, a search for a suitable
cosolvent with an optimal balance between solubility of 6
and insolubility of 5 was initiated. From previous large-scale
runs, we have learned the vital importance of a good
crystallisation procedure leading to easy centrifugation of
the crystals. This has proven to be relevant not only for a
reduction in run cycle time but also for a low moisture
content in the product. Although the phthalic acid chloride
5 is not very susceptible to hydrolysis, high levels of moisture
(above 10%) in the initial batches led to partial hydrolysis
of the acid chloride groups upon drying. This results in high
levels of 4 in the final product. The use of a two-phase
system for the hydrolysis/crystallisation, e.g. methylene
chloride/water11 or toluene/water, gives rise to a finely
divided precipitate, which is difficult to centrifuge, and to a
very high moisture content. From these observations, it has
been concluded that the use of a water-miscible cosolvent
as the organic component in the crystallisation system will
be more advantageous. Fructamine SA claims to have
obtained good results with diglyme.8 Since this solvent has
several disadvantages, e.g., it is highly toxic, relatively
expensive, and difficult to remove due to its high boiling
Although the patents cited above8-10 mention the danger
of using neat thionyl chloride, in our experience it is by far
the best option. Addition of thionyl chloride to the dicar-
boxylic acid 4, even when not completely dry, and subse-
quent heating do not lead to unacceptable exothermic or run-
away behaviour under any conditions evaluated. The reaction
is mildly endothermic, and in the presence of a catalyst such
as DMF, heating above 50 °C is required in order to achieve
an acceptable reaction rate. After the reaction has gone to
completion, excess thionyl chloride can be distilled off under
reduced pressure and reused in subsequent batches. The
remaining traces of thionyl chloride in the crude product can
be removed efficiently by dissolving the residue in toluene
and subsequently distilling the toluene/thionyl chloride
mixture under reduced pressure.
In the literature, several catalysts are described for the
chlorination of 4, including DMF,7 N-methylmorpholine,8
tertiary amines,8 and quaternary ammonium salts.9 Although
DMF normally is a common catalyst for this type of reaction,
we have encountered the formation of a precipitate and a
decrease in reaction rate near the end of the reaction. DMF-
catalysed reactions can be driven to completion only by
increasing the reaction temperature to 80-90 °C for several
hours, leading to unacceptably high levels of byproducts.
Apparently, the catalyst decomposes, probably either by
forming insoluble, unidentified complexes with the iodinated
intermediates in the reaction or by other decomposition
routes, as have been suggested recently.13 Moreover, the use
of DMF as catalyst with thionyl chloride has recently been
questioned on safety grounds owing to the formation of
highly toxic byproducts.13 Therefore, several catalysts, not
covered by patents, have been tested, among which were
N-methylpyrrolidone and tetramethylurea. Both give catalytic
activities similar to that of DMF, but no precipitation or
deactivation during the reaction has been observed.
During chlorination of the carboxylic acid groups, the
amino functionality in 4, which is extremely difficult to
protonate, also reacts quickly with thionyl chloride, giving
the N-sulfinyl compound 6 as the principal product at the
end of the reaction (Scheme 1).8 This intermediate 6 is
conveniently isolated by concentration of the reaction mixture
or direct crystallisation in an aprotic organic solvent, as
described by Mallinckrodt.11 Therefore, isolation of 5 requires
(14) The N-sulfinyl intermediate has been described in a patent by Fructamine:
ref 8.
(13) Levin, D. Org. Proc. Res. DeV. 1997, 1,182.
Vol. 3, No. 1, 1999 / Organic Process Research & Development
•
39

Table 1. Variables and maximum and minimum levels used
point, we have resorted to the use of acetone as a more
attractive alternative. The water, necessary for hydrolysis of
6, is added as a 10% solution in acetone to a solution of 6
in acetone. This avoids immediate precipitation of product
5 and intermediate 6, due to a local high concentration of
water at the point of addition.
in the factorial designa
Xi
variable
(-)
0
(+)
X1
X2
X3
X4
X5
X6
X7
toluene (%)
acetone (g)
temperature (°C)
dosing time (h)
final H2O concn (%)
agitation (rpm)
stirring time (h)b
2
43.5
25
1
30
280
5
8
62.5
45
3
35
400
24
53
35
2
32.5
340
12.5
Apart from the N-sulfinyl intermediate 6 and the starting
material 4, two other impurities have been observed, which
sometimes exceed the 1% content level in the final product.
To identify and quantify their appearance, they have been
isolated by silica gel column chromatography. In this way,
small amounts of about 90% pure samples of both byproducts
have been obtained. Due to the impurity of the material, the
small quantity, the presence of only quaternary C atoms, and
the presence of atropoisomers,15 NMR analysis gives no
conclusive evidence on their structure. With the additional
use of infrared and mass spectral data, the dimeric amide 7
and dimeric anhydride 8 have been assigned to these
impurities. The formation of both impurities is dependent
on the reaction conditions, with higher levels formed at
elevated temperatures. By performing the bulk of the reaction
at 50-60 °C and increasing the temperature to 80 °C for a
limited interval at the end of the reaction to ensure comple-
tion, the amounts of 7 and 8 have been kept under control.
Depending on the crystallisation procedure, a further reduc-
tion of all impurities to levels below 1% each has been
achieved.
1
a All experiments were performed with the same batch of crude N-sulfinyl
compound 6 (each run with 64.2 g, ∼0.1 mol of 6). b Stirring time between
addition of water and isolation of the product.
Table 2. Factorial design: experimental matrix and results
responses
variables
yield assay NSO
d
exp. X1 X2 X3 X4 X5 X6 X7 (%) (%) (%) (g/mL)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
-
+
-
+
-
+
-
+
+
-
+
-
+
-
+
-
0
-
-
+
+
-
-
+
+
+
+
-
-
+
+
-
-
0
-
-
-
-
+
+
+
+
+
+
+
+
-
-
-
-
0
+
-
-
+
+
-
-
+
-
+
+
-
-
+
+
-
0
+
-
+
-
-
+
-
+
-
+
-
+
+
-
+
-
0
+
+
-
-
-
-
+
+
-
-
+
+
+
+
-
-
0
-
+
+
-
+
-
-
+
+
-
-
+
-
+
+
-
0
87.2 97.9 3.6
84.0 98.2 2.4
88.9 97.9 1.7
83.0 98.3 1.8
88.0 97.8 2.4
86.6 98.1 1.7
78.9 98.0 0.7
85.2 97.5 1.0
85.8 97.7 0.3
84.9 97.9 0.7
85.7 98.1 1.0
86.3 97.9 1.7
81.9 97.6 0.3
84.9 97.2 1.1
85.3 98.3 1.6
85.4 97.8 2.5
83.0 98.6 1.3
85.8 98.4 1.5
1.4
1.4
1.7
1.7
1.2
0.9
2.0
1.9
1.6
1.7
1.3
1.9
1.4
1.3
1.3
1.2
1.4
1.8
0
0
0
0
0
0
0
list of seven expected critical parameters in this process (X1-
X7) has been set up, as shown in Table 1. Their number and
likely interactions between them has made us decide to apply
a fractional factorial design.17
The selection of the levels for the different factors has
been carried out considering previous experiments, large-
scale possibilities and limitations, and economic motives. For
example, the amount of residual toluene in the final crys-
tallisation mixture will be dependent on the efficiency of
the solvent evaporation in the N-sulfinyl intermediate 6 step.
A final concentration of toluene of less than 2% is difficult
to reach; more than 8% can lead to heterogeneous solvent
mixtures during water addition. The selected maximum and
minimum levels of each factor in the factorial design are
shown in Table 1.
With these variables, a scheme of 18 experiments has been
setup with 16 different combinations of maximum and
minimum levels of each variable, and two centre point
experiments where all levels are chosen at an intermediate
value (0 in Table 1). Subsequently, the experiments have
Factorial Experimental Design
As mentioned before, the crystallisation process appears
to be crucial for good-quality end-product, and we have
carefully elaborated the optimal conditions for the hydrolysis/
crystallisation in going from N-sulfinyl compound 6 to the
end-product 5. Good-quality 5 has to meet the following
criteria: level of N-sulfinyl intermediate 6 below 1%, total
purity of 5 above 97%,16 easy to centrifuge, and low moisture
content. With the knowledge of the previous experiments, a
(15) (a) Henrichs, P. M.; Estep, K.; Musza, L. L.; Rodger, C. A. J. Am. Chem.
Soc. 1995, 117, 2058. (b) Anelli, P. L.; Bracchetta, M.; Calabi, L. Secchi,
C.; Uggeri, F.; Verona, S. Tetrahedron 1997, 53, 11919.
(16) The purity of 5 was determined by reversed-phase HPLC and is given in
area %.
(17) Box, G. E. P.; Draper, N. R. Empirical model-building and response
surfaces; John Wiley & Sons: New York, 1987.
40
•
Vol. 3, No. 1, 1999 / Organic Process Research & Development

Table 3. Factorial design: statistical analysis and influence
of variablesa,b
yield
cake density
N-sulfinyl
variables
interactions 6 variables variables
interactions
b0 ) 85.1
b2b5 ) 0.37
b0 ) 1.54
b0 ) 1.50
b1b3 ) -0.07
b1 ) -0.44 b2b6 ) -0.60 b1 ) -0.27 b1 ) -0.06 b2b3 ) 0.07
b2 ) -0.94 b2b7 ) 1.35
b4 ) 0.40
b5 ) 0.66
b2 ) -0.57 b2 ) 0.16
b3 ) -0.35 b3 ) 0.07
b6 ) 0.08
b2b6 ) -0.09
b3b6 ) 0.13
b6 ) -0.86
b7 ) 0.93
a Number of experiments ) 16. Degrees of freedom ) 15. b Only significant
values are given.
Figure 2. Influence of the temperature (X3) and agitation (X6)
on the cake density of 5. Other variables: toluene 5%; acetone
53 g; dosing time 2 h; final H2O conc. 32.5%; final stirring
time 12.5 h.
Table 4. Calculated and found results for different sets of
variables
responses
variables
X1 X2 X3 X4 X5 X6 X7 (%) (%) (%) (g/mL)
yield assay NSO
d
exp.
1 calcd
1 found
2 calcd
2 found
3 calcd
3 found
+
-
+
+
-
+
0
0
-
+
-
-
+
+
+
-
+
-
+
+
78.7 98.0 0.7
78.9 97.8 0.7
87.2 97.7 2.4
89.9 97.6 1.8
84.8 97.4 0.3
85.5 97.8 0.6
1.6
1.5
1.2
1.1
1.8
1.6
+
parameters are the level of N-sulfinyl byproduct 6 (NSO (%))
and the overall purity of 5 (assay (%)). The density of the
filter cake (d (g/mL)) has been chosen as a final response,
since it shows a good correlation with both the ease of
centrifugation and the moisture content of the product.
Moreover, it can be easily determined and related to results
obtained on large scale. In Table 2 the experimental matrix
of the factorial design is shown, including the results.
The statistical analysis and the significant influences of
this factorial designsusing the Modde program18sare sum-
marised in Table 3. The coefficients bi have been calculated
for each response using a polynomial function of the seven
experimental variables as given in the following equation:
Figure 1. (a) Influence of the amount of acetone (X2) and the
agitation (X6) on the yield of 5. Other variables: toluene, 5%;
temperature, 35 °C; dosing time, 2 h; final H2O concn, 32.5%;
final stirring time, 12.5 h. (b) Influence of the amount of acetone
(X2) and final stirring time (X7) on the yield of 5. Other
variables: toluene, 5%; temperature, 35 °C; dosing time, 2 h;
final H2O concn, 32.5%; agitation, 340 rpm.
response ) b0 + SbiXi + SbijXiXj
(1)
Only significant values are given in Table 3. Since the assay
of all samples is above 97% and no variable has a major
influence on this purity, these data are neglected in Table 3.
Based on the data from Table 3, various responses may
be described, at 95% confidence level, as a polynomial
equation with the main significant variables. This gives
been performed at random, and for each experiment the yield
and three responses to determine various quality aspects have
been measured. As mentioned above, important quality
(18) Modde 4.0, Umetri AB, Umea˚, Sweden.
Vol. 3, No. 1, 1999 / Organic Process Research & Development
•
41

Figure 3. Isolated 5-amino-2,4,6-triidoisophthalic acid dichloride (5) before (a) and after (b) optimisation.
yield:
of acetone. Application of longer stirring times should ensure
a high yield.
Y (%) ) 85.1 - 0.44X1 - 0.94X2 + 0.4X4 + 0.66X5 -
0.86X6 + 0.93X7 + 0.37X2X5 - 0.6X2X6 + 1.35X2X7 (2)
Unfortunately, a high yield of above 89% is accompanied
by an increased level of N-sulfinyl 6. Influential parameters
on the level of 6 are toluene (X1), acetone (X2), and
temperature (X3). Increased values for all three parameters
shift the saturation point of 5 to a higher percentage of water.
This results in a more complete hydrolysis of 6 before
crystallisation, and consequently a decreased level of 6 in
the crystalline mass of 5.
N-sulfinyl intermediate 6:
NSO (%) ) 1.54 - 0.27X1 - 0.57X2 - 0.35X3
cake density:
d (g/mL) ) 1.50 - 0.06X1 + 0.16X2 + 0.07X3 + 0.08X6 -
(3)
0.07X1X3 + 0.07X2X3 - 0.09X2X6 + 0.13X3X6 (4)
As mentioned before, the cake density is related to the
ease of centrifugation and moisture content of the isolated
product. In general, a high cake density corresponds to large
crystals, a product that is easy to centrifuge, and a low
moisture content. A similar picture as for the level of
N-sulfinyl 6 emerges from the data obtained for the cake
density. A high amount of acetone and high temperature
during hydrolysis/crystallisation keeps the product 5 longer
in solution and shift the precipitation of 5 more towards
crystallisation. In this case, however, a high percentage of
toluene gives rise to a decrease in cake density (X1, b1 )
-0.06). Unfortunately, coupling of high cake density with
low moisture content and ease of centrifugation cannot be
made in this respect. More toluene leads to a rounder crystal
form, and although this results in a decreased cake density,
it appears to be favourable for the isolation of the product.
Agitation has a pronounced effect on the cake density.
In Figure 2, the influence of temperature and agitation on
the cake density is depicted. A combination of high tem-
perature and high agitation increases the cake density. A high
agitation is necessary to prevent the heavy crystals from
settling on the bottom of the reactor. This results in growth
of the existing crystals instead of the constant formation of
new crystal nuclei. Although generally higher agitation rates
give rise to increased attrition, the formation of smaller
particles through collision does not play a dominant role.
High temperatures can amplify this effect by constant
redissolving of newly formed particles. Here again, precipita-
tion of the product is shifted towards crystallisation.
Combination and incorporation of all the observations
described above lead to a set of process parameters as shown
in Table 4 for experiment 3. The product obtained by
application of this experimental setup meets all criteria
In general, the effects are rather small, but this is to be
expected when an optimisation is carried out close to the
optimal result. Still, interpolation of the data, calculated with
Modde, gives a variation in the yield of the reaction between
78.7% and 87.2%. Experimental backing of the calculated
model has been obtained by performing the reaction using
the variable settings which predicted the lowest and highest
yields, respectively (Table 4, experiments 1 and 2, respec-
tively.). This results in yields of 78.9% and 89.9%, respec-
tively, in excellent accordance with the theoretical yield.
The observed effects of the separate variables are largely
as expected. Increased amounts of organic solvent, toluene
(X1, b1 ) -0.44) as well as acetone (X2, b2 ) -0.94), lead
to a lower yield. More water (X5, b5 ) 0.66), longer addition
time (X4, b4 ) 0.4), and longer stirring time (X7, b7 ) 0.93)
all promote a better crystallisation, and thus a higher yield.
Additional information is collected from the significant
interactive effects between the variables. In Figure 1a, the
influence of the amount of acetone and the agitation on the
yield are depicted graphically. The negative effect of
increased amounts of acetone gets more pronounced with
increased agitation. This can be explained by an increased
hydrolysis of the acid chloride 5 to the phthalic acid 4.
Indeed, a higher level of 4 is detected in the mother liquor
when more acetone is used in combination with high
agitation.
An even stronger interaction is observed between the
amount of acetone and stirring time (Figure 1b). A clear drop
in the yield is observed when an increased amount of acetone
is combined with a short stirring time after addition of the
total amount of water. Apparently, the rate of crystallisation
is decreased in a solvent mixture with a higher percentage
42
•
Vol. 3, No. 1, 1999 / Organic Process Research & Development

regarding purity, yield, and moisture content and is easy to
centrifuge. A picture of the crystalline product is shown in
Figure 3b. Comparison with a sample prepared before
initiation of the optimisation study (Figure 3a) clearly shows
the effect of shifting the process from precipitation towards
crystallisation.
In conclusion, the careful examination of the reaction
conditions has resulted in a scaleable process with an
extremely high productivity. The application of factorial
design for the workup of the reaction has given us the tools
to set up a procedure where the original precipitation of the
product 5, with all its isolation problems, has been shifted
to crystallisation of 5, leading to excellent quality material
which is easy to centrifuge and has a very low (2-4%)
moisture content.
concentrated under reduced pressure to give the crude
5-sulfinylamino-2,4,6-triiodoisophthalic acid chloride (6).
The residue was dissolved in acetone (584 g, 738 mL) and
warmed to 45 °C. At this temperature, a 90/10 w/w acetone/
water mixture (449 g, 550 mL) was added dropwise to the
solution over a 1-h period. After the solution was stirred for
an additional hour and cooled to 25 °C, water (487 mL) was
added dropwise over a 2-h period. Stirring was continued
for 24 h, and the crystals were recovered by filtration (small
scale) or centrifugation (large scale). The cake was washed
with 400 mL of acetone/water 50/50 (w/w), followed by 400
mL of water, to give a wet product with a moisture content
of 2-4%. Drying was performed at 40 °C under vacuum to
give 547 g of 5 (85.5%). Analytical and spectroscopic data
were identical to those described in ref 12.
Isolation and Identification of Major Byproducts. The
mother liquor, obtained from the previous step and enriched
in byproducts, was concentrated under reduced pressure, and
a few grams of the residue was dissolved in a minimal
amount of toluene/EtOAc 6/1. Silica gel chromatography of
this solution (eluent hexane/EtOAc, 70/30 to 50/50) provided,
in order of elution, samples of 7 and 8.
Experimental Section
Commercially available solvents and reagents were used
without further purification. 5-Amino-2,4,6-triiodoisophthalic
acid was prepared according to literature procedures.12
Reversed-phase HPLC elutions were performed on a
Lichrospher 100 RP-18 column (5 mm) with H2O/CH3CN
mixtures. For straight-phase HPLC, a Lichrospher 100DIOL
column (5 mm) was used, with eluents hexane/EtOAc/
trifluoroacetic acid 70/30/0.1. 13C NMR spectra were deter-
mined using a Bruker ACF 200 (50.31 MHz) spectrometer.
Infrared spectra were obtained using a Perkin-Elmer Paragon
1000 FT-IR. MS data were obtained via ESI on a Finnigan
Navigator LC-MS.
Representative Procedure for the Synthesis of 5-Amino-
2,4,6-triiodoisophthalic Acid Chloride (5). Thionyl chloride
(1.43 kg, 1.2 mol) was added to 5-amino-2,4,6-triiodoiso-
phthalic acid (600 g, 1,07 mol). After addition of tetra-
methylurea (2.9 g, 25 mmol), the slurry was stirred at 50-
60 °C for 6 h, during which time the solids dissolved. After
the reaction mixture was heated for 30 additional minutes
at 80 °C, the excess of SOCl2 was distilled off under reduced
pressure, thus enabling recovery of 806 g of SOCl2, which
could be used in subsequent batches. To remove the
remaining SOCl2 from the residue, the solid was dissolved
in 500 mL of toluene (at 50-60 °C) and subsequently
N-5′-Amino-2′,4′,6′-triiodo-3′-chlorocarbonylbenzoyl-
2,4,6-triiodo-3,5-di(chlorocarbonyl)aniline Amide (7). 13
C
NMR (acetone-d6) (clearly distinguishable signals): δ 59.23,
60.60, 67.03, 72.90, 76.44, 79.40, 146.15, 149.75, 150.07,
150.44, 161.72, 170.04. IR (cm-1): ν 1761 (br s), 1673 (s),
1596 (s), 1528 (m), 1487 (s), 1393 (m), 1354 (s), 1046 (m),
999 (s), 962 (m). MS (M - H+): m/z 1152, with a Cl isotope
pattern of 3:3:1.
5-Amino-2,4,6-triiodo-3-chlorocarbonylbenzoic Anhy-
dride (8). 13C NMR (DMSO-d6) (clearly distinguishable
signals): δ 61.03, 73.13, 95.89, 142.52, 149.78, 169.60,
170.40. IR (cm-1): ν 1806 (s), 1765 (br s), 1594 (s), 1515
(m), 1398 (s), 1312 (m), 1268 (m), 1074 (s), 1005 (s), 917
(s). MS (M - H+): m/z 1134, with a Cl isotope pattern of
9:6:1.
Received for review June 25, 1998.
OP980055+
Vol. 3, No. 1, 1999 / Organic Process Research & Development
•
43