Effect of solvents on the product distribution and reaction rate of a Buchwald-Hartwig amination reaction

Article abstract of DOI:10.1021/op050226s

The Buchwald-Hartwig amination reaction between p-bromotoluene and piperazine in the presence of the homogeneous catalytic system Pd(dba) ₂/(±)-BINAP and the base NaO-t-Bu was investigated in two different classes of solvents: aprotic, nonpolar and aprotic, polar. The reaction was carried out using microwaves as the heating source, and it was found that the product distribution was strongly dependent on the class of the solvent. Based on the experimental results the selectivity towards the desired monosubstituted aryl piperazine was calculated, and it was found that the most appropriate solvent for the Buchwald-Hartwig amination reaction under the conditions applied was m-xylene.

Full text of DOI:10.1021/op050226s

Organic Process Research & Development 2006, 10, 762−769
Effect of Solvents on the Product Distribution and Reaction Rate of a
Buchwald-Hartwig Amination Reaction
Henrik Christensen, Søren Kiil,* and Kim Dam-Johansen
Department of Chemical Engineering, Technical UniVersity of Denmark, Building 229,
DK-2800 Kgs. Lyngby, Denmark
Ole Nielsen and Michael B. Sommer
H. Lundbeck A/S, Process DeVelopment, OttiliaVej 9, DK-2500 Valby, Denmark
Abstract:
neous catalytic system and a base. Compared to the other
The Buchwald-Hartwig amination reaction between p-bromo-
toluene and piperazine in the presence of the homogeneous
catalytic system Pd(dba) /(()-BINAP and the base NaO-t-Bu
methods the Buchwald-Hartwig amination reaction offers
a lower reaction temperature, typically between 80 and 100
°C, a higher selectivity with respect to the N-arylated amine,
and the reaction does not include highly reactive reactants,8
which may cause safety problems. Since 1995 the Buch-
wald-Hartwig amination reaction has formed the basis for
a research field with a large number of publications. The
articles published have covered many aspects of the reaction
2
was investigated in two different classes of solvents: aprotic,
nonpolar and aprotic, polar. The reaction was carried out using
microwaves as the heating source, and it was found that the
product distribution was strongly dependent on the class of the
solvent. Based on the experimental results the selectivity
towards the desired monosubstituted aryl piperazine was
calculated, and it was found that the most appropriate solvent
for the Buchwald-Hartwig amination reaction under the
conditions applied was m-xylene.
9
,10
6,11
including the effect of the catalytic complex, the base,
and the steric properties of the reactants (i.e., the aryl halide
1
2,13
and the amine).
Furthermore, much work has been
performed to reveal the chemical reaction mechanism and
1
0,14,15
the kinetics for the reaction.
summarized in reviews
Conclusions have been
1
6,17
illustrating some of the chal-
lenges, which may be met in the application of the Buch-
wald-Hartwig amination reaction. The effect of solvent is,
however, a relatively unexplored area. We have been
interested in developing a continuous reaction for the
monoalkylation of a piperazine and for this reason would
like to run the reaction in a solvent that dissolves sodium
tert-butoxide. From the literature not only is it known that
the reaction yields the desired N-arylated amine but also both
1
. Introduction
N-Arylamines are important building blocks in drugs,
which affect the central nerve system. In 2001 more than
5 antidepressants and more than 15 antipsychotics that
2
contained N-arylated amine blocks were commercially avail-
able. N-Arylated amines can also be identified in the dye
manufacturing industry.2 Several methods of preparing
N-arylamines are available in the literature. These methods
involve nitration, Ullmann condensation, or a benzyne
pathway. These methods have a number of drawbacks
related to highly reactive reactants, a large excess of amine,
high temperature, and formation of isomeric products.
Therefore it was of great interest when the Buchwald-
Hartwig amination reaction was presented in 1995. This
reaction provides the formation of a covalent bond between
a carbon atom in an aryl halogen and a nitrogen atom in a
primary or secondary amine in the presence of a homoge-
1
9
18
reduction and homo-coupling of the aryl halogen are
observed. Furthermore, if the reaction is carried out on
amines, which contain more than one nitrogen atom, mul-
3
4
5
3
4
5
(
(
8) Hepperle, M.; Eckert, J.; Gala, D.; Shen, L.; Evans, A.; Goodman, A.
Tetrahedron Lett. 2002, 43, 3359-3363.
6
,7
9) Nishiyama, M.; Yamamoto, T.; Koie, Y. Tetrahedron Lett. 1998, 39, 617-
620.
(
(
(
10) Strieter, E.; Blackmond, D.; Buchwald, S. Journal of American Chemical
Society 2003, 125, 13978-13980.
11) Meyers, C.; Maes, B. U.; Loones, K. T.; Bal, G.; Lemiere, G. L.; Dommisse,
R. A. The Journal of Organic Chemistry 2004, 69, 6010-6017.
12) Zhao, S.; Miller, A.; Berger, J.; Flippin, L. Tetrahedron Lett. 1996, 37,
4463-4466.
*
To whom correspondence should be addressed. Telephone: (+45)45252827.
Fax: (+45)45882258. E-mail: sk@kt.dtu.dk.
(13) Wolfe, J. P.; Buchwald, S. L. The Journal of Organic Chemistry 2000, 65,
1144-1157.
(14) Singh, U.; Strieter, E.; Blackmond, D.; Buchwald, S. Journal of American
Chemical Society 2002, 124, 14104-14114.
(15) Guari, Y.; van Strijdonck, G.; Boele, M.; Reek, J.; Kamer, P.; van Leeuwen,
W. Chemistrysa European Journal 1998, 7, 475-482.
(16) Hartwig, J. Angewandte Chemie, International Edition 1998, 37, 2046-
2067.
(
(
(
1) Prins, L. D. Psychotropics 2000/2001; Herman & Fischer A/S: Denmark,
2
002.
2) Bohnet, M., Ed.; Ullmann’s Encyclopedia of Industrial Chemistry; VCH
Verlagsgesellschaft: Germany, 2005.
3) Smith, M. B.; March, J. March’s AdVanced Organic Chemistry; John Wiley
&
Sons Inc: USA, 2001.
(
(
(
(
4) Lindley, J. Tetrahedron 1984, 40, 1433-1456.
5) Heaney, H. Chemical ReViews 1962, 62, 81-97.
6) Louie, J.; Hartwig, J. F. Tetrahedron Lett. 1995, 36, 3609-3612.
7) Guram, A. S.; Rennels, R. A.; Buchwald, S. L. Angewandte Chemie,
International Edition in English 1995, 34, 1348-1350.
(17) Schlummer, B.; Scholz, U. AdVanced Synthesis & Catalysis 2004, 346,
1599-1626.
(18) Qiang, L.; Juan, N.; Yang, F.; Zheng, R.; Gang, Z.; Tang, J. Chinese Journal
of Chemistry 2004, 22, 419-421.
7
62
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development
10.1021/op050226s CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/09/2006

Scheme 1. Model reaction for formation of substituted
piperazines
tiarylated products are produced. The intention of this work
is to highlight the effect of the solvent on the chemical
reaction rate and the product distribution.
Figure 1. Reaction mechanism for the formation of N-aryl
amide and the reduction of the aryl halide.
2
. Results and Discussion
2.1. The Reaction. The study is based on the reaction
between p-bromotoluene (1) and piperazine (2). The reaction
takes place in the presence of a homogeneous catalytic
system, which consists of bis(dibenzylideneacetone)palla-
2
dium(0) (Pd(dba) ) and 2,2′-bis(diphenylphosphino)-1,1′-
binaphthalene (BINAP). Sodium tert-butoxide (NaO-t-Bu)
was employed as the base. The overall reaction with the
observed products, 1-(4-methylphenyl)piperazine (3), 1,4-
bis(4-methylphenyl)piperazine (4), toluene (5), and 1,1′-
dimethyl-4,4′-biphenyl (6), is shown in Scheme 1.
Figure 2. Possible products from the â-hydride elimination.
From left to right: 1,2,3,6-tetrahydro-pyrazine, 2,3-dihydro-
pyrazine, 1-p-tolyl-1,2,3,6-tetrahydro-pyrazine.
Furthermore, it is possible that an etherification reaction
between 1 and NaO-t-Bu takes place under the formation of
this mechanism is operating, it is expected that three
additional products are formed during the reaction (Figure
2). However, it was not possible to detect these products by
LC-MS.
19,16
tert-butyl-(4-methylphenyl)-ether.
However, this com-
ponent could not be identified under the conditions applied
in this work. The reaction in Scheme 1 has previously been
studied,1 but none of these studies provide knowledge about
the effects of the solvents.
No probable mechanism has been proposed for the homo-
coupling reaction of p-bromotoluene in the presence of NaO-
t-Bu, but mechanisms where tertiary amines have been used
2,9
2
4
2.2. The Reaction Mechanism. The Buchwald-Hartwig
have been presented.
amination reaction proceeds through a catalytic cycle, and
it is generally believed that the formation of the N-arylated
amine involves six different intermediate reaction steps as
illustrated in Figure 1.1
2.3. Reaction Protocol. The classic solvents employed
in the Buchwald-Hartwig amination reaction are nonpolar,
aprotic solvents such as m-xylene and 1,4-dioxane. However,
due to the low polarity of these solvents, it is not possible
to dissolve the base NaO-t-Bu. Consequently it was decided
to use a more polar solvent which is able to create a
homogeneous system and investigate if there were any effects
on the product distribution during the reaction. Initially, we
chose NMP since this solvent has successfully been applied
in the Heck reaction, which uses the same catalytic system.
In the coming part of the article the results for the
experiments will be presented. A detailed description of the
preparation of the reaction mixtures is given in the Experi-
mental Section. However, it has to be specified that each of
the data points presented in Figure 3 to 8 have been obtained
for an individual reaction mixture, which subsequently is
analyzed by HPLC.
4,20,21
In the present work, we have identified two additional
side products not accounted for by the above mechanism,
namely toluene from the catalyzed reduction of p-bromo-
toluene and 1,1′-dimethyl-4,4′-biphenyl from the homo-
coupling of p-bromotoluene. The literature proposes a
mechanism for the reduction of the aryl halide, which
25
22,23
involves a â-hydride elimination reaction from the amine,
which is shown in Figure 1.
The reduction is initiated by a â-hydride elimination from
the amine forming a double bond. Subsequently the imine
and the reduced aryl halide are released from the catalyst. If
(
(
(
(
(
19) Watanabe, M.; Nishiyama, M.; Koie, Y. Tetrahedron Lett. 1999, 40, 8837-
8
840.
20) Negishi, E. Handbook of Organopalladium Chemistry for Organic Synthesis;
John Wiley & Sons Inc: 2002.
21) Urgaonkar, S.; Xu, J.-H.; Verkade, J. G. The Journal of Organic Chemistry
2.4. Aprotic Nonpolar Solvents. To obtain basic knowl-
edge of how the reaction behaves in the classic solvents for
2
003, 68, 8416-8423.
22) Wagaw, S.; Rennels, R.; Buchwald, S. Journal of American Chemical Society
997, 119, 8451-8458.
(24) Alonso, D. A.; N a´ jera, C.; Pacheco, M. C. The Journal of Organic Chemistry
2002, 67, 5588-5594.
(25) Cornils, B.; Herrmann, W. Applied Homogeneous Catalysis with Organo-
metallic Compounds; VCH Verlagsgesellschaft: Germany, 1996.
1
23) Wolfe, J. P.; Wagaw, S.; Marcoux, J.-F.; Buchwald, S. L. Accounts of
Chemical Research 1998, 31, 805-818.
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
763

Figure 3. Concentration versus time for the reaction carried
out in m-xylene. The lower figure is a magnification of the upper
figure.
Figure 4. Concentration versus time for the reaction carried
out in 1,4-dioxane. The lower figure is a magnification of the
upper figure.
the Buchwald-Hartwig amination reaction, the reaction is
first carried out in m-xylene and 1,4-dioxane. These solvents
are able to dissolve all the reactants with the exception of
NaO-t-Bu, resulting in a heterogeneous system. The course
of reaction where m-xylene was used is shown in Figure 3.
It appears that the reaction proceeds with a high degree
of selectivity with respect to the desired product 3. Small
amounts of 4, 5, and 6 are produced, but no significant
change in selectivity is observed during the reaction.
Subsequently the reaction is performed in 1,4-dioxane, and
the concentration profiles are shown in Figure 4.
formation of the desired product, 3, was increased from 56%
to 78% based on the initial amount of 1. Thus aprotic
nonpolar solvents provide high selectivity toward the desired
N-arylated amine. The yield may be further increased by
increasing the amount of 2.
2
.5. Aprotic Polar Solvents. After having obtained
knowledge about the course of reaction for the Buchwald-
Hartwig amination carried out in classic aprotic nonpolar
solvents, it was desired to discover what happens if the
solvent is changed to an aprotic polar solvent. As previously
pointed out these solvents are able to dissolve NaO-t-Bu to
a higher extent compared to the nonpolar solvents. The
concentration profiles for the reaction performed in NMP
are shown in Figure 6.
Compared to the results in Figure 3 the same tendencies
for the reaction profiles are observed (note the difference in
the time scale, which provides higher resolution in the
experiment with 1,4-dioxane). However, the production of
Compared to the reactions described in the previous
section it is observed that the reaction profiles have changed.
The formation of side product 5 is significant, and it is of
the same order as the desired product 3. In the initial phase
of the reaction the formation of 3 is faster than the formation
of 5. However, as the reaction progresses the formation of 5
becomes dominant, and the production of 3 terminates. The
termination may be a consequence of the dominant reduction
of 1 under the formation of 5. On the contrary the change
4
, 5, and 6 at endpoint is higher in 1,4-dioxane.
To reveal how the formation of the products depends on
the ratio between the reactants an experiment with 1 equiv
of 1 and 2.2 equiv of 2 was carried out. The results are
displayed in Figure 5.
In Figure 5 it is seen that the conversion of 1 is complete.
The production of 5 and 6 was reduced, and the total
consumption of 1 for these two products was decreased from
2
1.4% to 6.9% of the initaial concentration of 1. The
764
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development

Figure 6. Concentration versus time for Buchwald-Hartwig
amination carried out in NMP. The lower figure is a magnifica-
tion of the upper figure.
Figure 5. Concentration versus time for the reaction carried
out in 1,4-dioxane with 1.0 equiv of p-bromotoluene and 2.2
equiv of piperazine. The lower figure is a magnification of the
upper figure.
of the effects of the solvents have been summarized. In this
of solvent does not affect the formation of the side products
table the reaction time for the optimal fractional yield with
4
and 6.
30
respect to 3, φ
3
, has been identified from eq 1. The
To clarify this tendency for increased formation of 5,
corresponding conversion of 1, X
1 3
, and the yield of 3, X ,
another aprotic, polar solvent, N,N-dimethylacetamide
DMAC), was used. The results of this reaction are presented
have been calculated according to eqs 2 and 3.
(
in Figure 7. For reasons unknown the results for 4 and 6 are
fluctuating.
[3]t
φ
)
(1)
(2)
(3)
3,t
[
3] + 2[4] + [5] + 2[6]
t t t t
From Figure 7 it is observed that the formation of 5 is
slower in DMAC compared to NMP. The reaction was also
done with 1.0 equiv of 1 and 2.2 equiv of 2, which is similar
to the reaction carried out with 1,4-dioxane. The results are
shown in Figure 8. Again it is observed that the formation
of the desired product 3 becomes more significant when the
ratio of 2 to 1 is increased. However, taking the results
obtained for the aprotic polar solvents into consideration,
the production of 5 is still significant compared to reactions
carried out in m-xylene and 1,4-dioxane. For the aprotic polar
solvents it is observed that two overall reactions are
dominant: formation of the N-arylated amine and reduction
of the aryl halide.
[
1]t
X1,t ) 1 -
[1]t)0
[3]t
X3,t
)
[
1]t)0
Based on Table 1 it is seen that the solvent that provides
the worse fractional yield with respect to 3 is NMP. The
(
26) Beletskaya, I. P.; Bessmertnykh, A. G.; Guilard, R. Tetrahedron Lett. 1999,
40, 6393-6397.
(
(
27) Bandgar, B.; Kasture, S. Journal of Chemical Research 2000, 5, 252-253.
28) Crossland, H. HVordan man bruger statistik p a˚ kemiske bestemmelser; AiO
Tryk as: Odense (Denmark), 1989.
29) Montgomery, D. C. Design and Analysis of Experiments; John Wiley &
Sons: USA, 1997.
(
2.6. Discussion of Key Numbers for the Reaction. A
significant difference in the product distribution has been
observed dependent on the solvent applied. In Table 1 some
(
30) Levenspiel, O. Chemical Reaction Engineering; John Wiley & Sons Inc:
USA, 1999.
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
765

Figure 7. Concentration versus time for Buchwald-Hartwig
amination carried out in DMAC. The lower figure is a
magnification of the upper figure.
Figure 8. Concentration versus time for Buchwald-Hartwig
amination carried out in DMAC with 1.0 equiv of p-bromo-
toluene and 2.2 equiv of piperazine. The lower figure is a
magnification of the upper figure.
fastest conversion of 1 is in DMAC; after 20 min 79% of
the starting material has been converted. Furthermore, the
table reveals that the highest fractional yield of 3 is obtained
by changing the ratio between the two reactants. The change
in the ratio for the reaction performed in 1,4-dioxane does
not affect the fractional selectivity significantly, but the yield
of 3 is increased from 34% to 69%, and 92% of 1 is
converted during the optimal reaction time instead of 42%.
For the reaction performed in DMAC with 1.0 equiv of 1
and 2.2 equivalent of 2 the formation of 5 is lowered
compared to the results obatined with standard conditions,
but no other significant effects have been observed.
Table 1. Performance for the solvents applied; the number
of equivalents refers to piperazine
time of
reaction
[min]
φ
eq 1
3
X
1
X
3
eq 3
solvent
eq 2
NMP (1.1 equiv)
DMAC (1.1 equiv)
m-xylene (1.1 equiv)
30
20
60
20
0.54
0.80
0.75
0.74
0.68
0.79
0.75
0.42
0.37
0.58
0.55
0.34
1,4-dioxane (1.1 equiv)
DMAC (2.2 equiv)
10
50
0.72
0.76
0.74
0.92
0.54
0.69
1,4-dioxane (2.2 equiv)
2.7. Coupling of the Results to the Reaction Network.
Combining the reaction mechanism presented in this article
with a chemical reaction network found in the chemical
reaction engineering literature,30 it has been found that the
reaction network studied in this article can be divided into
two: First, the production of 3 and 4 can be described as
two-step irreversible series-parallel reactions according to 1
increase in the concentration of 3, whereas the concentrations
of the starting materials are decreasing. In most of the
experiment 1.0 equiv of 1 and 1.1 equiv of 2 are employed.
Therefore as the reaction progresses it is expected that
compounds 2 and 3, which both contain unreacted amine
sites, will be competing in the reaction with 1. However,
these characteristics are not observed in the experimental
results. From these results it appears that there is a tendency
for 2 to be more reactive than 3, since almost no 4 is formed.
The reason for this may be that 3 is more bulky compared
to 2, which means that 2 more easily coordinates to the
+
6
2 f 3 and 1 + 3 f 4. Second, the production of 5 and
can be described by parallel reactions. Taking the two-
step irreversible series-parallel reactions into consideration,
the theoretical course of reaction will be the following: In
the initial phase of the reaction it is expected to observe an
766
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development

palladium complex and by that enters the catalytic cycle.12
.8. Catalytic Deactivation Process. In the study of the
2
reaction it has been observed that the reaction from time to
time terminates even though unreacted starting material still
is present in the reaction mixture (Figures 3 and 4). This
may imply that the catalyst sometimes is deactivated during
the course of reaction. A search in the literature for
deactivation processes for homogeneous palladium com-
plexes does not reveal any results related to the Buchwald-
Hartwig amination reaction. However, deactivation processes
for the Heck reaction has been found.3
1,32
The catalytic
complexes employed in the Heck reaction can also be used
in the Buchwald-Hartwig reaction. A number of pathways
for the deactivation of the catalyst are presented: Decom-
position of reactive metal alkyl with water or oxygen,
reactions of the carbon-to-metal bond, and formation of
inactive dimers of the catalyst.31 It will be too speculative
to predict which of the mechanisms leads to the termination
in our work since a great number of variables affect the
reaction. In the experiments where the deactivation is
observed the reactions are done in aprotic nonpolar solvents.
This suggests that another possibility for the termination of
Figure 9. Possible reaction mechanism for the reduction of
the aryl halide.
3
. Conclusion
the reaction could be due to the limited solubility of base in
the nonpolar solvents.11
This study illustrated that the choice of solvent affects
the product distribution of the Buchwald-Hartwig amination
reaction. The solvent that provided the best selectivity toward
the desired product 3 and reduces the production of the side
products the most was m-xylene. On the contrary it was
discovered that both NMP and DMAC initiated the unwanted
debromination reaction of p-bromotoluene. In connection to
the debromination reaction it was also revealed that the
formation of the unwanted side products could be suppressed
by increasing the ratio between 2 and 1. Another unexpected
observation was that the second amination reaction of 3 was
very slow and, thus, only traces of 4 were produced. In
connection to the reaction rate it was observed that the fastest
conversion of 1 was in DMAC. In relation to the transforma-
tion from batch to continuous mode a number of observations
lead to the conclusion that this may not be the ideal solution.
According to the literature the most efficient way of operating
a two-step irreversible series-parallel reaction is in either a
2
.9. Mechanistic Reflections. Taking the reaction mech-
anisms provided in section 2.2 and the experimental results
into consideration some mechanistic reflections may be
provided. The reactions performed in NMP and DMAC
reveal that in the initial phase of the reaction (approximately
the first 20 min) the formation of the desired product 3 is
dominant. However, as the reaction progresses the formation
of 5 becomes significant. The question is how NMP and
DMAC initiates this difference in the product distribution
compared to the ones obtained in m-xylene and 1,4-dioxane.
The obvious difference in the chemical properties between
the polar and the nonpolar solvents is the ability for
stabilizing the charges on NaO-t-Bu. Furthermore, the double
bonds or the lone pairs on the amide part in NMP and DMAC
may coordinate to the palladium complex. This has the
consequence that NaO-t-Bu and the palladium complex are
more hindered in polar solvents than in apolar solvents. This
may have the consequence that the palladium complex,
containing the protonated piperazine, in Figure 1, may be in
a resting state in the polar solvents, and this may give the
palladium complex time to perform the rearrangement that
enables the â-hydride elimination. This explanation may be
3
0
batch or a plug flow reactor. Keeping in mind that the
formation of 4 (the bisubstituted piperazine) is very slow,
meaning that the reaction for practical purposes almost can
be neglected and that the reaction in some cases terminates,
it may seem irrational to establish a continuous facility.
2
6
connected to observations made by Beletskaya et al. They
found that the reduction of aryl bromides was dominant in
reactions where the deprotonation reaction is slow, leaving
the palladium complex, containing the protonated piperazine,
in the same situation as the above explained. Beletskaya et
al. suggested that the slow deprotonation enables the reduc-
tion to proceed through amino coordinated complexes
according to Figure 9, making the mechanism more complex.
However, in this work it has not been possible to reveal the
existence of this reaction path.
4
. Experimental Section
.1. Materials. p-Bromotoluene was purchased from
Merck, and piperazine, from Avocado research chemicals
Ltd. 1,4-Dioxane, m-xylene, DMAC, NaO-t-Bu, Pd(dba)
4
2
,
and (()-BINAP were purchased from Aldrich. All the
chemicals and solvents, with the exception of piperazine,
were used without any prior treatment. Piperazine was
crystallized from toluene and subsequently sublimated. This
treatment increased the melting point for piperazine from
approximately 50 to 108-111 °C. The water content in the
solvents applied has not been investigated. However, in a
recent publication it has been concluded that water may affect
(31) van Leeuwen, P. W. Applied Catalysis A: General 2001, 212, 61-81.
32) Shmidt, A.; Smirnov, V. Kinetics and Catalysis 2005, 46, 54-58.
(
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
767

the reaction rate for the Buchwald-Hartwig reaction. The
Table 2. Relative standard deviation for the detectable
compounds
effect has primarily been related to a better solubility of the
33
base in solvents that contain water, which results in a faster
rate of reaction. Reference compounds for analysis calibration
of 1-(4-methylphenyl)piperazine was from Fluka; toluene and
component
1
3
4
5
6
1
,1′-dimethyl-4,4′-biphenyl were from Aldrich. A reference
relative standard
deviation [%]
3.2
7.2
33.4
13.5
7.9
compound of tert-butyl-(4-methylphenyl)-ether was prepared
according to a literature procedure.27 The same applies for
average concentration 0.209 0.112 0.00367 0.0382 0.00881
[mol/L]
1,4-bis(4-methylphenyl)piperazine, which was also produced
8
according to a literature procedure.
.2. Preparation of the Reaction Mixtures. The reaction
mixtures were prepared as follows: In a flask under nitrogen
atmosphere and agitation at room temperature 1, 2, Pd(dba)
BINAP, and NaO-t-Bu were dissolved/suspended in the
solvent. The solvent was previously purged for 10 min with
nitrogen.
4
by HPLC twice, and the standard deviation for the concen-
tration determination was calculated. The relative standard
deviations for the composite are shown in Table 2.
From Table 2 it is observed that the standard deviation
differs for the components. The relatively high standard
deviation for 4 is due to the low concentration of the
compound, which results in ill-defined baseline separation
on the HPLC spectrum.
2
,
The ratio between the reactants in this flask was as
follows: 1.00 g (1.0 equiv) of 1, 0.555 g (1.1 equiv) and
1
.11 g (2.2 equiv) of 2, 0.169 g (0.05 equiv) of Pd(dba)
2
,
4.4. Quantification of Piperazine. To improve the
chemical understanding of the reaction it is desired to obtain
the concentration profiles for as many of the reactants and
products as possible. The compounds 1, 3, 4, 5, and 6 are
detected by HPLC chromatography; however, piperazine
(reactant 2) does not have any chromophore groups allowing
detection by HPLC. It has been attempted to apply gas
chromatography followed by flame ionization detection for
the concentration determination of piperazine, but it has not
been possible to obtain reliable results. Another solution
strategy, though somewhat uncertain, has been to base the
piperazine determination on the knowledge of the reaction
mechanism for the reaction (i.e., a molar balance calculation).
According to Scheme 1 component 2 is consumed in the
production of 3, 4, and 5. The concentration of 3 and 4 can
be correlated directly to the consumption of 2 (Scheme 1).
Meanwhile, the correlation between the formation of 5 and
the consumption of 2 is more uncertain. According to the
mechanism shown in Figure 1 it can be seen that the
formation of 1 mol of 5 causes 1 mol of amine to perform
a â-hydride elimination reaction. The products from this
reaction have been shown in Figure 2. It has not been
possible to identify and quantify these products. For simplic-
ity it is assumed that only 2 is oxidized during the â-hydride
elimination reaction and that it only performs a single
â-hydride elimination. The correctness of this assumption
decreases as 2 is converted to 3 during the course of reaction,
because 3 may be performing the â-hydride elimination
reaction instead of 2. However, 3 is assumed not to perform
any â-hydride elimination reaction. Based on the assumptions
stated above the concentration of 2 may be estimated from
eq 5.
0.274 g (0.075 equiv) of (()-BINAP, and 0.844 g (1.5 equiv)
of NaO-t-Bu. The reactants were dissolved/suspended in 12
mL of solvent. The reaction mixture was distributed to a
number of process vials. Each of the vials contained
approximately 2 mL of the reaction mixture, was equipped
with a magnetic stirring bar, and was purged with nitrogen
prior to the sealing. Subsequently the reactors were loaded
in an autosampler and reacted at 100 °C from 10 to 300
min. Initial investigation of the reaction mixtures left at room
temperature for 24 h showed no conversion of the reactants.
4.3. Reproducibility of the Experimental Results.
Initially in this study the validity of the method of analysis
and the reaction stability were studied. The concentrations
of the components in the reactions mixture have all been
calculated using Lambert-Beers Law in eq 4.
A
C )
(4)
∈
‚ l
Therefore the uncertainties in the final calculation of the
concentration are introduced in the determination of the
expansion coefficient, ∈ ‚ l, and the absorbance, A, of the
reaction mixture. This means that the standard deviation of
a composite measurement needs to be found in order to make
a conclusion about the reproducibility of the experiments.
The standard deviation of a composite measurement is
calculated based on the standard deviation on the directly
28
measured values for ∈ ‚ l and A. The expansion coefficients
2
were determined based on a linear regression. An R -value
higher than 0.997 was obtained for all the components. Due
to the satisfying correlation it was concluded that uncertain-
ties only were introduced to the system from the HPLC
analysis. The standard deviation for each of the following
components was found using the following method: One
mother solution of reaction mixture was prepared. The
solution was subsequently equally distributed to four process
vials. Each of the vials were reacted for 1 h at 100 °C in
NMP. The reaction mixture of each of the vials was analyzed
2
9
[
2] ) [2] - ([3] + [4] + [5] )
(5)
t
t)0
t
t
t
Equation 5 is used to estimate the concentrations of 2 in
the experiments. Note, that eq 5 is valid only if the
mechanism of Figure 1 is correct and the above assumptions
hold.
4
.5. Effect of Stirring. No studies were conducted with
(
33) Dallas, A. S.; Gothelf, K. V. Journal of Organic Chemistry 2005, 70, 3321-
3
323.
the aim of identifying the effect of the stirring velocity in
768
•
Vol. 10, No. 4, 2006 / Organic Process Research & Development

our work. Mass transfer limitations are only relevant due to
a poor solubility of NaO-t-Bu. All the other components are
completely soluble at the temperature where the reaction is
carried out. According to the literature the physical structure
of the weak base Cs CO can have an impact on the reaction
2 3
rate,11 but stronger and more soluble bases such as sodium-
Acknowledgment
This work was cofounded by the Technical University
of Denmark and the Danish Research Agency. This particular
project is also founded by H. Lundbeck A/S, and all the
experimental work was carried out in their laboratories.
Supporting Information Available
Additional experimental information. This material is
available free of charge via the Internet at http://pubs.acs.org.
tert-amylate have been reported not to influence the reaction
14
rate. However, it is beyond the scope of this work to clari-
fy the effect of stirring in the case where NaO-t-Bu is ap-
plied.
Received for review November 21, 2005.
OP050226S
Vol. 10, No. 4, 2006 / Organic Process Research & Development
•
769