N-acylation reactions performed in aqueous reaction medium: Screening and optimising of a synthetic step of a process for iodixanol

Article abstract of DOI:10.1021/op010078f

This contribution presents results from a development and optimisation project concerning a synthetic process for iodixanol. Iodixanol is the "active ingredient" in Visipaque (Nycomed Amersham now Amersham Health) a medical X-ray imaging contrast agent. T

Full text of DOI:10.1021/op010078f

Organic Process Research & Development 2002, 6, 113−119
N-Acylation Reactions Performed in Aqueous Reaction Medium: Screening and
Optimising of a Synthetic Step of a Process for Iodixanol
,
†
‡
Hans-Ren e´ Bjørsvik* and Torgrim Engell
Department of Chemistry, UniVersity of Bergen, All e´ gaten 41, N-5007 Bergen, Norway, and Amersham Health,
P.O.Box 4220 Nydalen, N-0401 Oslo, Norway
Abstract:
This contribution presents results from a development and
optimisation project concerning a synthetic process for iodix-
anol. Iodixanol is the “active ingredient” in Visipaque (Nycomed
Amersham now Amersham Health) a medical X-ray imaging
contrast agent. The process step that was developed and
optimised comprised a reaction between a tetraacid chloride
and 3-aminopropane-1,2-diol for obtaining the corresponding
tetraamide. This reaction was originally performed using DMA
as reaction medium, a reaction that gave a medium-to-high yield
of the tetraamide. Solvent screening was performed followed
by process optimisation using response surface methodology.
Surprisingly, a solvent mixture composed by 1:2 v/v THF:water
was determined to be a superb reaction medium. Tuning the
experimental condition according to the developed multivariate
empirical model gave an excellent yield of the target molecule,
nearly 94% with only small quantities of the main side product,
monoacid triamide, which indicates a rather low hydrolytic
degradation of the substrate. The current synthetic process step
to iodixanol showed also a general validity when a small variety
of amino alcohols and acid chlorides were reacted to successfully
obtain the desired amides in high yields, concomitant with low
outcomes of the most likely side products, that is the hydrolysis
products of the acid chlorides.
also influence the chemotoxicity of the entire compound. The
triiodo aromatic compounds are classified as first- or second-
generation X-ray contrast agents, depending on whether the
compound contains ionic or nonionic groups, respectively.
Third generation contrast agents are nonionic agents com-
posed of two nonionic triiodo aromatics, such as compound
2,3
1
, which are connected by a C
3
or C
4
bridge. Iodixanol
2
(
Scheme 1) is an example of a compound with a C
3
bridge.
Clinical tests of formulations of iodixanol 2 have shown
many improved advantages over earlier generations of X-ray
contrast agents, moreover iodixanol 2 is isotonic and
4
isoosmotic with blood.
Many of the X-ray contrast agents commercially available
1
today have in common that two of the side chains (R and
2
R ) are carboxamides. The carboxamide group is found to
be very suitable in X-ray contrast agents due to in situ
stability as well as low contribution to the final chemotoxicity
of the compound. However, the N-acylation or amidation
reaction step in the synthesis of these compounds involves
reacting water-soluble amines (the side chains) with water-
insoluble substrates (the triiodinated moiety). This situation
makes it difficult to select a suitable solvent for the reaction,
especially if aprotic solvents such as dimethylformamide
Introduction
Generally, X-ray contrast agents for medical imaging
1
diagnostics are compounds that contain at least one triiodo-
benzene moiety 1. This moiety can be considered as the main
(DMF) and dimethylacetamide (DMA) have to be omitted
“
pharmacophore” in such compounds, since it is this part
due to (1) their high boiling points and the consequent
difficulties of removal from the bulk drug substance, (2) their
toxicity and thus the environmental aspect, (3) the costs, and
that gives the desired effect, the X-ray absorption. However,
iodine has lipophilic properties, which imply that it is
necessary to introduce some other moieties that can coun-
(4) the fact that the solvents DMA and DMF are not inert
1
2
3
teract that property. Hence, the R , R , and R groups of the
molecule must then have a high degree of hydrophilic
character. Examples of such groups are carboxyl, carbox-
amide, hydroxyl, alkoxyl, and glycosyl, and so forth.
under the applied reaction conditions, where for example
5-7
DMF may participate in a Vilsmeyer-Haack,
reaction.
The present contribution will describe a process for
introduction of the side chains of the X-ray contrast agent
iodixanol 2. This synthetic process comprises concomitant
formation of four carboxamide side chains giving the
1
2
3
Moreover, the R , R , and R groups of the substances of
type 1 are also designed to “regulate” the pharmacokinetic
profile of the entire compound.
(2) Hansen, P.-E.; Holtermann, H.; Wille, K. (Nyegaard & Co AS, Oslo,
Norway). EP 0 108 638, 1984.
Two desirable features are good water solubility and low
viscosity. However, the composition of the side chains will
(3) Priebe, H.; Dugstad, H.; Gacek, M.; Hagen, E.; Homestad, O. M.; Larsen,
A° .; Sjøgren, C. E.; Thomassen, T. Acta Radiol. 1995, 36, 21-31.
*
To whom the correspondence should be addressed. E-mail: Hans.Bjorsvik@
(4) Karlsson, J. O. G.; Gregersen, M; Refsum, H. Acta Radiol. 1994, 35, 56.
(5) Vilsmeier, A.; Haack, A. Berichte 1937, 60, 119.
(6) Olah, G. A.; Kuhn, S. J. Olah’s Friedel-Crafts and Releated Reactions;
Interscience Publishers: New York, 1964; Vol 3, Part 2.
kj.uib.no.
†
University of Bergen.
‡
Amersham Health.
1) Sovak, M., Ed. Radiocontrast Agents; Springer-Verlag: Berlin, 1984; p 1.
(
(7) Burn, D.; Kirk, D.; Petrow, V. Tetrahedron 1965, 21, 1619.
1
0.1021/op010078f CCC: $22.00 © 2002 American Chemical Society
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
113
Published on Web 02/01/2002

Scheme 1
cost, the boiling point of the solvent(s), the environmental
effects, and so forth. On the basis of these findings we
decided to carry through a process optimisation of the
N-acylation reaction using the solvent/cosolvent mixture of
water and tetrahydrofuran. A thorough examination of the
procedure showed that eight experimental/process variables
were likely to influence the reaction. Figure 1 shows the
variables represented in an Ishikawa cause-effect diagram.
Since the outcome of the introductory synthetic procedure
gave quite a high yield (∼75%) when measured using HPLC,
8
it was decided to use a response surface design for the
further development of the synthetic procedure. Four of the
eight experimental variables (see Figure 1) were expected
to have the major influence on the synthetic reaction, and
were thus further explored. The experimental variables the
amount of water (x
the amount of the reagent 3-aminopropane-1,2-diol 4 (x
and the reaction temperature (x
1
), the amount of tetrahydrofuran (x
2
),
),
3
4
) were used in a circum-
scribed central composite design (CCC) with seven replicates
of the centre point experiment, giving for k ) 4 variables a
k
design composed of 2 + 2k + 7 ) 31 experiments. The
other experimental variables, such as stirring rate, reaction
time, and quantity of HCl scavenger may well all be
influencing variables. By using sufficient long reaction time,
sufficient stirring rate, and a quantity of HCl scavenger that
was at least as much as the HCl that was developed during
the reaction, these variables did not perturb the reaction in
a negatively manner.
intermediate 7. A competing process is the hydrolysis of the
acid chloride under formation of the side product 8, Scheme
1.
Methods and Results
The synthetic transformation 3 + 4 f 7, Scheme 1, is
carried out by reacting racemic 3-aminopropane-1,2-diol 4
with the tetraacid chloride 3 using triethylamine as HCl
scavenger. The target molecule of this reaction 3,3′,5,5′-
tetrakis(2,3-dihydroxypropylcarbamoyl)-2,2′,4,4′,6,6′-hexaiodo-
N,N′-(2-hydroxy-O-acetyl propane-1,3-diyl)diacetanilide 7
was of interest as a final intermediate in a process leading
to 3,3′,5,5′-tetrakis (2,3 - dihydroxypropylcarbamoyl)-
The selected experimental levels for the variables (x
) are given in Table 1. The CCC design allowed estimation
of the regression coefficients for the variables (x , ..., x ),
, ..., x
× x
1
, ...,
x
4
1
4
for the two-factor interactions (x
as well as for the quadratic terms (x , ..., x ).
1
× x
2
, x
1
× x
3
3
4
)
2
1
2
4
The results obtained in the laboratory experiments, HPLC
measurements y (Table 1), were used as the response variable
and were thus regressed onto the experimental variables
2,2′,4,4′,6,6′-hexaiodo- N,N′-(2-hydroxypropane-1,3-diyl)-
including their interaction- and quadratic terms (constitutes
diacetanilide 2 (iodixanol) that is used as a racemic mixture
in Visipaque (Nycomed Amersham).
9
the model matrix), by using the PLSR method, y ) f (x
1
,
2
2
.
.., x
4
, x , ..., x , x
1
× x
2
, x
1
× x
3
, ..., x
3
× x
4
) The estimated
1
4
One of the major challenges of this synthetic transforma-
tion was to find a solvent system suitable for industrial-scale
production, as well as to determine the conditions under
which the reaction should be performed to obtain a selective
high-yielding procedure. During the screening phase of the
organic process development, it was observed that the
substrate 3 was soluble in tetrahydrofuran, but not in water,
while reagent 4 and the reaction intermediates were soluble
in water. Even though water is not considered as a particu-
larly suitable reaction medium for acid chlorides, some very
introductory experiments were performed using water as
solvent/cosolvent for the “amidation” reaction, step 1 of
Scheme 1. Surprisingly, high yields of the target molecule
were achieved with only a small quantity of the tetraacid
chloride 3 being hydrolyzed to the corresponding mono, di-,
tri-, or tetracarboxylic acid, under the aqueous reaction
conditions. The major by-product was the monocarboxylic
acid 8, even if only in small quantities. Besides these
promising results, the solvent mixture of water and tetrahy-
drofuran fulfilled also the requirements with respect to the
regression coefficients (the â’s in Table 2) were plotted in a
CND (cumulative normal distribution) plot, Figure 2, cor-
responding statistical analysis and estimated values for the
regression parameters are listed in Table 2. The CND plot
in Figure 2 and data in Table 2, both lead to the conclusion
that the regression coefficients â
statistically significant and thus contribute to explanation of
the variation in the response y. The estimates â and â22 are,
3
, â
4
, â33, and â44 were
1
however, “borderline cases” and were included in the model.
Moreover, â23 and â24 are not significant on the basis of
strictly statistical assessments, but are included when using
the more rough method of variable selection using CND plot
(Figure 2). The model that was estimated for the optimisation
of the reaction, 3 + 4 f 7, Scheme 1, is given in eq 1. The
model was estimated by using a ) 5 PLS components that
explaining 99.84% of the variation in y when using cross-
(8) Box, G. E. P.; Draper, N. R. Empirical Model-Building and Response
Surfaces; Wiley: New York, 1987; pp 1-669.
(
9) Martens, H.; Næs, T. MultiVariate Calibration; Wiley: New York, 1991;
p 1.
114
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

Figure 1. Ishikawa cause-effect diagram.
Table 1. Central composite design with response y, the yield
of product 6 in the N-acylation reaction
Table 2. Numerical estimates for the model for the yield in
the N-acylation reaction
experimental
experimental
standard
error
confidence
interval
a
responseb
a
responseb
variables
variables
yield
coefficient
P
exp.
no.
exp.
no.
x
1
x
2
x
3
x
4
y
x
1
x
2
x
3
x
4
y
2.966 10-
0.099
0.420
19
7
(3.291
(1.777
(1.777
(1.777
(1.777
(1.628
(1.628
(1.628
(1.628
(2.177
(2.177
(2.177
(2.177
(2.177
(2.177
â
â
â
â
â
â
â
â
â
â
â
â
â
â
â
0
80.476
-1.471
-0.694
7.017
-2.714
0.620
-1.217
-1.980
1.848
-0.565
-0.847
-0.870
0.944
1.552
0.838
0.838
0.838
0.838
0.768
0.768
0.768
0.768
1.027
1.027
1.027
1.027
1.027
1.027
1
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
-1 -1 -1 -1
+1 -1 -1 -1
-1 +1 -1 -1
+1 +1 -1 -1
-1 -1 +1 -1
+1 -1 +1 -1
-1 +1 +1 -1
+1 +1 +1 -1
-1 -1 -1 +1
+1 -1 -1 +1
-1 +1 -1 +1
+1 +1 -1 +1
-1 -1 +1 +1
+1 -1 +1 +1
-1 +1 +1 +1
+1 +1 +1 +1
78.9
83.7
71.7
67.4
92.3
90.6
89.2
87.7
67.4
71.2
71.9
64.8
89.6
77.7
82.1
80.8
17 -2
18 +2
0
0
0
0
0
0
0
0
0
0
0
0
88.2
80.0
72.0
81.6
59.8
87.6
91.3
86.7
82.0
80.3
80.0
79.0
80.9
80.9
80.1
2
-
3
3.074 10
19
20
21
22
23
24
25
26
27
28
29
30
31
0 -2
4
0.005
0.432
0.133
0.020
0.029
0.590
0.421
0.409
0.372
0.177
0.853
0 +2
0
0
0
0
0
0
0
0
0
0
0
0 -2
0 +2
11
22
33
44
12
0
0
0
0
0
0
0
0
0
0 -2
0 +2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
2
2
3
4
3
4
1.451
-0.194
34
a
2
2
2
Number of objects ) 31, R
) 0.8713, Q
) 0.4179, R
Adj
) 0.7587, RSD
)
4.0723, confidence level ) 0.95.
a
The experiments were performed using 50 g O-acetyl-iodixanol 3 in each
(definition) [levels: -2, -1, 0, +1, +2];
(amount of water) [12.5, 25.0, 37.5, 50.0, 62.5 mL]; x (amount of
tetrahydrofuran) [0, 50, 100, 150, 200 mL]; x (amount of 3-aminopropane-1,2-
diol) [11.443, 13.931, 16.499, 18.906, 21.394 g]; x (reaction temperature) [10,
experiment. Experimental Variables: x
k
projection of the response surfaces shown in Figure 3 actually
describes variation in one response, the yield of 7, in five
dimensions. The five dimensional response surface is given
x
1
2
3
4
3
1
2
8
0, 30, 40, 50, °C]. Mean value for the center experiment (y ) ∑
i)25
i
y /7) is
0.457. b Yield% measured by HPLC analysis.
by the contour lines and the four experimental variables (x
.., x ). This plot is used in the following way: consider the
outer frame which shows the variation of the two experi-
mental variables x and x at three discrete levels -1, 0, and
1 (x [25.0, 37.5, 50.0 mL] and x [50, 100, 150 mL]). Inside
1
,
.
4
validation10 for PLS via NIPALS with random split and 20
iterations. The product statistics for the model were for this
1
2
2
2
model R ) 0.8497 and Q ) 0.5176, that demonstrates a
+
1
2
fairly good model for prediction purpose.
this frame, the nine subplots show the contour projections
2
2
2
of the response surface when the other two experimental
y ) â + (â x ) + â x + â x + â x + (â x ) + â x +
0
1
1
2
2
3
3
4
4
22
33 3
variables x
here -2 and +2 (x
3
and x
4
are varied continuously between the limits,
[11.4 f 21.4 g] and x [10 f 50 °C]).
2
4
â x + (â x x + â x x ) ) 81.0957 - 1.4833 x -
4
4
23
2
3
24
2
4
1
3
4
2
2
The contour plots represented in Figure 3 reveal that to
obtain a high yield of target molecule 7, it is advantageous
to apply a high quantity of 3-aminopropane-1,2-diol 4
0
.6917x + 7.0250x - 2.7167x - 1.2662x - 2.0412x +
2 3 4 2 3
2
1
.7838x + 0.9375x x + 1.4500x x (1)
4 2 3 2 4
(
3
variable x ). However, by also tuning the other experimental
The model of eq 1 was used for the preparation of the
multidimensional contour plot of Figure 3. The isocontour
variables, the combined goal of high yield at less cost may
be achieved. Thus by using a lowered reaction temperature
(10) Wold, S. Technometrics 1978, 20, 397.
4
(variable x ) and a concentrated reaction mixture (lowered
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
115

synthetic method was scaled up and fitted to pilot runs. The
3
test run was performed in a 0.5 m stirred tank reactor. The
reactor was charged with 40 L of water, 25.7 kg (254.7 mol)
of triethylamine, 33.7 kg (370.8 mol) of the reagent 4, and
8
0 L of tetrahydrofuran. The reaction mixture was cooled
to a temperature of 0 °C. The tetraacid chloride 3, 79.6 kg
57.94 mol), was added to the solution over a period of 2.5
(
h. The temperature was always maintained below 10 °C
during the addition. The reaction mixture was continuously
stirred for another 15 h at a temperature between 11 and 16
°
C. At completion of the reaction, the liquid phases, the
organic and the water phases, were separated. The water
phase was diluted to a total volume of 400 L and the pH
adjusted to 6.9. Sodium chloride was added to salt out the
organics. The isolated yield was >80% with a strength of
9
9%.
Evaluation of the Generality of the Developed Method.
To evaluate the generality of the developed synthetic method,
a small variety of substrates and reagents were selected.
Schemes 1-3 summarise the experimental results. For each
of the acid chlorides, three different amino alcohols were
tested, namely 3-aminopropane-1,2-diol 4, 2-aminopropane-
1,3-diol (serinol) 5, and 2-amino ethanol 6. The acid chlorides
Figure 2. CND plot.
that were selected were the tetraacid chloride 3, benzoyl
chloride 14, and 3-phenylpropionyl chloride 15. The obtained
results for the experiments that were performed with the
tetraacid chloride 3 and the three amino alcohols (4-6) are
given in Scheme 1. Only in the case where the serinol 5
were used, a somewhat elevated quantity of the correspond-
ing triamide monoacid is observed (9.4%). The results for
the reactions between the amino alcohols (4-6) and benzoyl
chloride 14, and the amino alcohols (4-6) and 3-phenyl-
propionyl chloride 15, are given in the Schemes 2 and 3,
respectively. In all of the experiments for the simple acid
chlorides 14 and 15, only traces of the two corresponding
carboxylic acids, benzoic acid and 3-phenylpropionic acid,
respectively.
volumes of the two solvents water and tetrahydrofuran,
variables x and x ) the quantity of the reagent 4 may be
1 2
lowered to obtain a high yield of target molecule 7. Thus,
the combined goals for a good process are achieved, high
yield, increased reactor volume efficiency, and improved
environmental performance by solvent replacement.
Optimized and Scaled Up Experiments
Optimized Experiments. The multivariable model ob-
tained from the statistically designed experimental plan was
finally used for predicting the optimal conditions that should
preferably give a quantitative yield. The procedure predicted
by the model: the tetraacid chloride 3 (50 g, 0.0364 mol)
was dissolved in a mixture of tetrahydrofuran (50 mL), water
Conclusions
(
25 mL), and triethylamine (16.5 g, 0.163 mol, ∼4.5 equiv)
12
Even though most textbooks of organic chemistry report
that acid chlorides are rather labile in the presence of water,
we have developed a simple selective and high-yielding
synthetic procedure for N-acylation by reacting acid chlorides
with amines using a reaction medium composed of water
and tetrahydrofuran. Due to the high rate of the N-acylation
reaction and the relatively lower rate of the hydrolysis
reaction of the acid chloride in water, only minor traces of
the hydrolysis product are detected.
at a temperature of 10 °C. The reagent 3-aminopropane-1,2-
diol 4 (21.4 g, 0.2349 mol, ∼6.5 equiv) was then added.
The reaction rate was observed to converge over a rather
long period, and was determined to be “optimal” after 24 h.
If the temperature was decreased further below 10 °C, the
reaction was almost blocked-up. The obtained yield of
compound 7 of this experiment was 94%. The remaining
6% was determined to be the main by-product monoacid
triamide 8, but traces of other by-products were also
observed. These results were somewhat surprising, since acid
chlorides in the presence of water are usually expected to
The method was originally developed for synthesis of
3,3′,5,5′-tetrakis(2,3-dihydroxypropyl-carbamoyl)-2,2′,4,4′,6,6′-
hexaiodo-N,N′-(2-hydroxy-O-acetyl propane-1,3-diyl) di-
acetanilide 7, an intermediate for iodixanol, reacting acetic
acid 2-[acetyl[3,5-bis-chlorocarbonyl- 2,4,6 - triiodo -
phenyl)amino]-1-{[acetyl[3,5-bis-chlorocarbonyl-2,4,6-tri-
iodophenyl)methyl]}ethyl ester 3 with 3-aminopropane-1,2-
diol 4. The derived multivariate empirical model, based on
11
undergo a Schotten-Baumann -type reaction, yielding the
corresponding carboxylic acid of the acid chloride.
Scaling Up for Pilot Plant. On the basis of the successful
result achieved during the laboratory development, the
(11) (a) Schotten, C. Berichte 1884 17, 2544. (b) Baumann, E. Berichte 1886,
1
9, 3218. (c) Smith, M. B.; March, J. March’s AdVanced Organic
Chemistry: Reactions, Mechanisms, and Structure, 5th ed.; Wiley: New
(12) Smith, M. B.; March, J. March’s AdVanced Organic Chemistry: Reactions,
Mechanisms, and Structure, 5th ed.; Wiley: New York, 2001; p 468.
York, 2001; p 482.
116
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

Figure 3. Isocontour projection of the response surfaces.
Scheme 2. Reactions between the amino alcohols (4-6)
and benzoyl chloride 14
Scheme 3. Reactions between the amino alcohols (4-6)
and 3-phenylpropionyl chloride 15
The model shows also that the composition of the reaction
mixture is important, and predicts a good composition to be
v:v ) 1:2 (water:tetrahydrofuran).
A variety of amino alcohols and acid chlorides (see
Schemes 1-3) were reacted under the optimised conditions
determined for the synthetic process to iodixanol 2. These
experiments revealed the method as a general approach for
N-acylation reactions, since very high yields were obtained
the statistically designed experimental plan for optimisation
of the N-acylation reaction, revealed that high reaction
temperature favours the formation of the triamide monoacid.
When a higher quantity of 3-aminopropane-1,2-diol 4 is used,
the fourth step, N-acylation, is accelerated, and the formation
of the triamide monoacid 7 is thus lowered.
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
117

for all of the combinations tried. Moreover, except for the
X-ray-absorbing derivatives, only small quantities of the
hydrolysis product of the acid chlorides were found. In the
experiments for the hexaiodinated substrates, the hydrolysed
acid chlorides were determined to be in the range 2.8-9.4%.
Even though the superb yields were achieved when using
different substrates and reagents of Schemes 1-3, the
reaction conditions may not be the optimum for other
reagents and substrates. Such cases will require further
optimization.
9: C37
H NMR (DMSO-d
O, broad); 3.86-3.57 (CH -OH; CH-O-(CdO)CH ; CH-
NH-CO; broad); 3.13-3.06 (m, CH -N); 1.79 (CH3, OH;
H
46
I
6
N
6
O
16, MW 1592.24.
1
6
) 3.34 (CO-NH, broad); 2.12 (CH-
2
3
2
broad).
1
3
C NMR (DMSO-d ) 170.26; 169.90; 168.99 (CdO);
6
151.03 (Cq, arom-N); 147.15 (Cq, arom-CdO); 100.82; 100.55;
99.92 (C-I); 72.06 (CH -(CdO)-O-C-H); 59.19; 58.60
(CH2-O); 54.45; 52.95 (CH -(CdO)-N-CH); 45.78
3
3
(CH -(CdO)-N-CH); 22.74(CH ); 8.54(CH ).
3
3
3
+
MS-ESP+ m/z: molecule ion + H ) 1593 (85%),
molecule ion + Na ) 1614.9 (100%).
Experimental Section
+
Synthetic Procedure for Designed Experiments. In a
1
1: C33
H
38
6
I N
6
O
12, MW 1472.13.
H NMR (DMSO-d ) 8.66 (N-H, broad); 5.07 (broad,
CH-O-(CdO)-CH ) 3.79-3.58 (broad, CH -O, CH -N-
CdO)-CH ); 3.25 (broad, CH -N); 1.86-1.77 (broad, CH
and OH).
13C NMR (DMSO-d
arom-N); 147-147.02 (Cq, arom-(CdO)) 101.086-98.86 (C
I); 71.95 (H-C-O-(CdO)-CH ); 59.14 (CH -O); 41.65
500 mL jacketed reaction vessel connected to an external
1
6
temperature-control unit and equipped with a condenser and
mechanical stirrer, O-acetyl-iodixanol (50 g, 0.036 mol) 3
was added to a mixture of triethylamine (16.2 g, 0.160 mol),
aminopropandiol ([11.44 to 21.39] g, [0.126 to 0.234] mol),
water ([12.5 to 62.5] mL), and tetrahydrofuran ([0 to 200]
ml) at [10 to 50] °C. The reaction mixture was stirred at a
temperature of [10 to 50] °C for a period of 18 h. At the
end of the reaction two separate layers, one organic phase
and one water phase were formed, where the water phase
contained amidated product(s). The two phases were sepa-
rated, and the water phase was diluted to a total volume of
3
2
2
(
3
2
3
6
) 169.85-169.17 (CdO); 151.13 (Cq,
q
-
3
2
(
CH
2
-N); 39.75 (CH
2
-N); 21.69 (CH
3
).
+
MS-ESP+ m/z: molecule ion + H ) 1472.8, molecule
+
ion + Na ) 1494.9
1
6: C10
H
13NO
3
, MW 192.22.
H NMR (DMSO-d ) 166.58 (CdO); 134.49 (Cq, arom );
30.95; 128.11; 127.11 (CH, arom); 70.37 (CH-OH); 67.98
CH -OH); 63.90 (CH -NH).
) 8.34 (1H, NH, broad); 7.87-7.85
2H, m, arom); 7.51-7.42 (3H, m, arom) 3.70-3.64 (1H,
1
6
250 mL with purified water and neutralized to pH 6.5-7.0
1
(
with concentrated HCl. One millilitre of the neutralized water
phase was transferred to a 100 mL measurement flask and
diluted to a total volume of 100 mL with purified water.
2
2
1
3
C NMR (DMSO-d
6
(
This solution was injected onto HPLC for quantification
using the method previously described.3
m, CH-OH), 3.39-3.32 (2H, m, CH
m, CH -NH).
2
-OH) 3.25-3.20 (2H,
2
Control Experiments: Synthetic Procedure to N-(2-
Hydroxyethyl) Benzoylamide 18 (Similar Procedure for
Compounds 16, 17, 19, 20, and 21). To a 100 mL jacketed
reaction vessel connected to an external temperature-control
unit equipped with a condenser and mechanical stirring, water
+
MS-ESP+ m/z: molecule ion + H ) 1472.8, molecule
+
ion + Na ) 1494.9.
1
7: C10
H
13NO
3
, MW 195.22.
H NMR (DMSO-d ) 7.85 (1H, d, N-H); 7.53 (5H, m,
arom); 4.58 (2H, t, OH); 3.99-3.94 (1H. m, H-C-N); 3.53
1
6
(12.5 mL), tetrahydrofuran (25 mL), triethylamine (2.00 g,
(
4H, 2CH
2
, t, CH
C NMR (DMSO-d
30.88; 128.02; 127.19 (CH, arom); 60.38 (CH
2
OH).
0.019 mol), and 2-hydroxyethylamine (1.75 g, 0.027 mol)
1
3
6
) 166.23 (CdO); 134.74 (Cq, arom);
-OH); 53.77
were mixed and cooled to 13 °C. To the cooled mixture
benzoyl chloride (2.43 g, 0.017 mol) was added, keeping
the internal temperature below 20 °C during addition. The
reaction mixture was stirred at 13 °C for 20 h.
HPLC analysis of the reaction mixture after 20 h. showed
a conversion of starting material of 99.2% (area % HPLC at
1
(
2
H-C-N).
+
MS-ESP+ m/z: molecule ion + H ) 196.0 (25%),
+
+
molecule ion + Na ) 218.0 (100%), molecule ion + H
O ) 178 (45%).
8: C , MW 165.19.
H NMR (CDCl ) 8.37 (1h, n-h, broad); 7.88-7.86 (2H,
arom, m); 7.53-7.43 (3H, m, arom); 4.69 (1H, t, O-H);
.57-3.52 (2H, q, CH -OH, J ) 6.07 Hz));3.38-3.30 (2H,
m, CH -NH).
C NMR (CDCl
-
H
2
1
9 2
H11NO
254 nm). The chemical yields according to area % HPLC at
254 nm were 98.7% for N-(2-hydroxyethyl)benzoylamide 18,
0.3% for benzoic acid and 0.15% of other reaction products.
Control Experiments: Synthetic Procedure to Com-
1
3
3
2
pounds 7, 9, and 11. For the experiments with tetraacid
chloride 3, the procedure was similar to that given above,
except that the quantities of the triethylamine were (4 ×
2
13
3
) 166.36 (dO); 134.59 (Cq, arom); 130.92;
128.10; 127.11 (CH, arom); 59.81 (CH
OH).
MS-ESP+ m/z: molecule ion + H ) 166.1 (100%),
molecule ion + Na ) 188.1 (15%).
2
-OH); 42.19 (CH -
2
0.019 mol ) 0.076 mol), and for the amino alcohols (4-6)
+
the quantities (4 × 0.017 mol ) 0.068 mol) were used.
Spectroscopic Data. The NMR spectra were recorded on
a Bruker 400 MHz NMR Spectrometer. The shift data (δ)
are reported in ppm.
+
19: C12
H
17NO
3
, MW 233.27.
1
H NMR (CDCl
3
) 7.25-7.10 (5H, m, arom); 6.96 (1H, t,
7
: C37
H
46
I
6
N
6
O
16, MW 1592.24. Isolated substance was
N-H, J ) 5.93 Hz); 4.73 (2H, broad, OH); 3.69-3.64 (1H,
m, CH-OH); 3.48-3.37 (2H, m, CH -OH); 3.33-3.18
(2H, m, CH -NH); 2.90-2.85 (2H, t, CH2-CdO, J ) 6.04
compared against reference on HPLC. The reference sub-
stance is fully described in ref 3.
2
2
118
•
Vol. 6, No. 2, 2002 / Organic Process Research & Development

Hz); 2.48-2.44 (2H, t, CH
2
-Ph, J ) 6.04 Hz).
) 174.20 (C , CdO); 140.45 (Cq, arom);
28.46;128.40 128.17; 126.24; 126.14 (CH, arom); 70.83
CH-OH); 63.60; 42.05; 37.86; 31.56 (CH ).
(2H, q, CH
2
-OH); 2.94-2.90 (2H, t, CH
-Ph).
C NMR (CDCl ) 173.42 (C
128.46; 128.21; 126.22 (CH, arom); 61.70; 42.27; 38.17;
31.65 (CH ).
2
-CdO); 2.49-
13
C NMR (CDCl
3
q
2.45 (2H, t, CH
2
1
3
1
(
3
q
, CdO); 140.60 (Cq, arom);
2
+
MS-ESP+ m/z: molecule ion + H ) 224.1 (10%),
2
+
+
molecule ion + Na ) 246.0 (100%).
MS-ESP+ m/z: molecule ion + H ) 194.1 (100%),
molecule ion + Na ) 216.0 (60%).
+
2
0: C12
H
17NO
3
, MW 223.27.
C NMR (CDCl ) 172.98 (C
28.56; 128.48; 126.31 (CH, arom); 63.46; 38.50; 31.87
CH ); 52.66 (CH).
13
3
q
, CdO); 140.86 (Cq, arom);
Acknowledgment
1
(
Nycomed Imaging is acknowledged for permission to
publish the data concerning the iodixanol synthesis. Mr. Atle
Aaberg at Department of Chemistry at University of Bergen
is gratefully acknowledged for helping us with recording the
2
1
H NMR (CDCl
CdO, t); 2.54 (2H, CH
.78-3.64 (4H, 2CH -OH).
MS-ESP+ m/z: molecule ion + H ) 224.1 (100%),
3
) 7.31-7.20 (5H, arom); 2.98 (2H, CH
2
-
2
-Ph, t); 3.93-3.86 (1H, CH, m);
3
2
1
13
+
H and C spectra. We also thank Professor George Francis
+
+
for linguistic assistance and discussions concerning the
manuscripts.
molecule ion + Na ) 246.0 (70%), molecule ion + H -
O ) 206.1 (20%).
, MW 193.25.
H NMR (CDCl ) 7.27-7.15 (5H, m, CH, arom); 6.44
1H, broad, NH); 3.59-3.57 (2H, t, CH -NH); 3.32-3.29
H
2
2
1: C11H15NO
2
1
Received for review September 17, 2001.
OP010078F
3
(
2
Vol. 6, No. 2, 2002 / Organic Process Research & Development
•
119