Preparation of α-aminophosphines on solid support: Model studies and parallel synthesis

Article abstract of DOI:10.1016/S0040-4020(02)00471-4

On-resin assembly of phosphine ligands represents a formidable challenge. Following model solution studies, we developed two synthetic routes for α-aminophosphine synthesis on solid support. The ligands are prepared via the Mannich reaction of the resin-bound aldehyde, amine and diphenylphosphine with very good yield and purity. Alternatively, the amine can serve as the anchoring building block. The ligands were qualitatively and quantitatively analyzed using gel-phase ³¹P and ¹³C NMR techniques. The studies culminated in the parallel synthesis of a 40-member library of borane-protected α-aminophosphines.

Full text of DOI:10.1016/S0040-4020(02)00471-4

TETRAHEDRON
Pergamon
Tetrahedron 58 32002) 5147±5158
Preparation of a-aminophosphines on solid support:
model studies and parallel synthesis
B. Bar-Nir Ben-Aroya and Moshe Portnoyp
School of Chemistry, Raymond and Beverly Sackler Faculty of Exact Sciences Sch. Chem., Tel-Aviv University, Ramat Aviv,
Tel-Aviv 69978, Israel
Received 17 October 2001; revised 27 March 2002; accepted 25 April 2002
AbstractÐOn-resin assembly of phosphine ligands represents a formidable challenge. Following model solution studies, we developed two
synthetic routes for a-aminophosphine synthesis on solid support. The ligands are prepared via the Mannich reaction of the resin-bound
aldehyde, amine and diphenylphosphine with very good yield and purity. Alternatively, the amine can serve as the anchoring building block.
The ligands were qualitatively and quantitatively analyzed using gel-phase 31P and 13C NMR techniques. The studies culminated in the
parallel synthesis of a 40-member library of borane-protected a-aminophosphines. q 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
those ligands with secondary nitrogen moieties. Recently,
we reported an ef®cient route to resin-bound b-amino-
phosphines.8 a-Aminophosphines represent an especially
attractive target since they can be readily assembled by a
multicomponent Mannich condensation 3Eq. 31)). Multi-
component reactions are highly valuable for combinatorial
synthesis since they introduce a number of diversity
elements into the target molecule in one step.9 Ligands
with potential hemilabile chelating ability 3like a-amino-
phosphines) were effective in a number of catalytic
processes.10 Numerous synthetic schemes, based on varia-
tions of the Mannich condensation, were used for the
preparation of a-aminophosphines in solution.11 While
this work was in progress, solution-phase Mannich conden-
sation of secondary phosphines, aldehydes and secondary
amines was exploited for the preparation of a ®rst library
of soluble a-aminophosphines.12 However, an analogous
solid-phase route has never been explored.
The combinatorial approach has been applied to catalysis
research with increasing frequency over the last ®ve years.1
Arrays of ligands and catalysts have been tested using
various high-throughput techniques. A signi®cant number
of these studies were based on parallel solid-phase synthesis
of ligands. Solid-phase organic synthesis 3SPOS), a vital
tool of combinatorial research, experienced a dramatic
revival over the last decade.2 This technique can lead to
the ef®cient and expeditious assembly of ligands on a
solid support in a single or parallel format. Ligand libraries,
prepared in this way, can be screened while still bound to
the support, as has been demonstrated in recent years.3 The
majority of ligands assembled in a library format on solid
supports are peptide or Schiff base ligands.3b,4 No parallel
assembly of phosphine ligands on solid support have ever
been reported, except for the phosphine-containing peptides
of Gilbertson.5 One of the reasons for this fact is the general
underdevelopment of solid-phase synthetic methods,
especially those for the preparation of air-sensitive
compounds. Thus, while the attachment of phosphine
ligands, pre-synthesized in solution, to a reactive polymer
through a remote functionality is well known,6 the multi-
step/multicomponent assembly of such ligands on resin has
hardly been investigated and still represents a formidable
challenge.7
ꢀ1†
Herein we report our studies of a-aminophosphine assembly
on solid support, an examination of anchoring building
blocks and the synthesis and analysis of the ®rst supported
a-amiophosphine libraries. Some of these results have been
published in a preliminary communication.13
We have become interested in the parallel assembly of phos-
phorus±nitrogen ligands on solid support and, particularly,
2. Results and discussion
2.1. Model studies in solution
Keywords: solid-phase synthesis; aminophosphines; Mannich reaction;
combinatorial chemistry.
Corresponding author. Fax: 1972-3-640-9293;
e-mail: portnoy@post.tau.ac.il
p
Our primary goals in this work were the development of
0040±4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020302)00471-4

5148
B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
substituents and was recently published as a communication.16
ꢀ4†
2.2. Method development on solid support
Following the model solution studies, one of the possible
approaches on solid support was implemented. According to
this approach, the aldehyde partner of the Mannich conden-
sation carries out the anchoring function. The model ligand
6, analogous to soluble compound 1, was assembled on
Merri®eld or Wang Bromo support according to Scheme
2. Immobilization of 4-carboxy-benzaldehyde, via standard
nucleophilic substitution,17 was followed by the conversion
of the aldehyde to the imine with aniline using a TMOF-
based procedure.18 The ®nal step formed the a-amino-
phosphine 6 via Ph2PH addition to the imine. Alternatively,
the immobilized aldehyde could be converted into the
product 6 in one step via a 3-component reaction. In the
latter case, double condensation is unlikely, due to the
pseudodilution principle.
Scheme 1.
solid-phase synthetic routes to a-aminophosphine ligands
3PCN) with secondary amino groups and the preparation
of a diverse library of the ligands through parallel synthesis.
Initially, model studies in solution 3ligand 1) demonstrated
that the stepwise approach 3Eq. 32)) must be favored over
the one-pot three-component reaction 3Eq. 31), R1ˆR3ˆ
R4ˆPh, R2ˆH) due to the formation of the double conden-
sation by-product. In solution, traces of the imine and
diphenylphosphine are always observed 3by NMR) together
with product 1. This demonstrates the slight disassembly of
1 to its precursors since the reaction between the imine and
the secondary phosphine is in fact an equilibrium process
3Eq. 32a)).14
In order to analyze the products along the synthetic path-
ways, we mainly used nucleophilic cleavage, followed by
proton NMR of the cleaved mixture. Unfortunately, this
method was not applicable for analyzing the ®nal product
which is unstable under the cleavage conditions. 3Equi-
librium 32a) is notably shifted to the left under basic con-
ditions.14) Therefore, all phosphine-containing products
were analyzed qualitatively and quantitatively using gel-
phase 31P NMR.8
ꢀ2†
The gel-phase 31P NMR spectrum of 6b 3Fig. 13a)) exhibits
a single peak at 6.6 ppm 3for comparison, 1 absorbs at
ꢀ2a†
The synthesis of ligand 1 suffers from its high sensitivity to
oxygen, as was repeatedly observed. Without special
precautions, some a-aminophosphine oxide 2 was always
formed along with 1. Benzylidene aniline reacts very
quickly and irreversibly with diphenylphosphine oxide or
diphenylphosphine under air, forming pure 2 3Eq. 33)).
Thus, it is quite possible that the oxidation of 1 does not
proceed directly but rather through the aforementioned
equilibrium 3Scheme 1).
ꢀ3†
The abovementioned observations led us to use a phos-
phine±borane adduct in the Mannich condensation
3Eq. 34)), forming 3 irreversibly. Compound 3 can readily
be converted to 1 by deprotection with amines.15 The syn-
thesis according to Eq. 34) was examined for a number of
Scheme 2. 3i) For 4a: 4-formylbenzoic acid, Cs2CO3, KI 3cat.), DMF, 808C,
24 h. For 4b: 4-formylbenzoic acid, DIPEA, CsI, DMF, rt, 18 h; 3ii) PhNH2,
AcOH 3cat.), DCM/TMOF, rt, 18 h; 3iii) Ph2PH, CHCl3, rt, 18 h;
3iv) PhNH2, Ph2PH, DCM/TMOF±CHCl3, AcOH 3cat.), rt, 18 h.

B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
5149
were made: 31) a large excess of diphenylphosphine was
used for the preparation of 6 from the resin-bound imine
or aldehyde; 32) minimal washing of the resin was
performed at the end of the synthesis.
Quanti®cation using the aforementioned reference resin
demonstrated 85% yield, while 13C gel-phase NMR
3Fig. 13b)) exhibited only peaks for the polymer and the
ligand 6b, indicating the high purity of the product.19
Especially characteristic is the signal of the bridge methine
carbon at 57.6 ppm.
We have also used the amine building block as an anchoring
functionality in an alternative approach. For this task,
N-protected amino acids were immobilized on support.
While both aliphatic aminoacids 3Ala, Gly) and aromatic
aminoacid 34-aminobenzoic acid) were used 3Schemes 3
and 4, respectively), only in the latter case was a good
yield of a-aminophosphine observed.
For the model aliphatic amine 3Ala, Scheme 3), both
Merri®eld/Boc and Wang/Fmoc chemistries led to resin-
immobilized amine with high yield. This amine can easily
be converted to imine with benzaldehyde 394% overall yield
for the product on Wang resin). However, the conversion of
the imine into a-aminophosphine never exceeded 30%
yield. The aforementioned equilibrium 32a) is responsible
for this effect 3according to the literature, electron donating
substituents destabilize the a-aminophosphines14). For
resin-bound ligand 9b, two diastereomers are formed, as
determined from the gel-phase 31P NMR spectrum 3Fig.
2). The integration ratio of the signals of the diastereomeric
ligands at 3.3 and 1.0 ppm indicates chiral 1,3-induction
during the formation of the aminophosphine with de
<60%. Interestingly, a similar de was observed for 10, the
soluble model of 9. 1,3-asymmetric induction was recently
reported for the formation of a-aminophosphine oxides and
sul®des with de's within the same range.20
Figure 1. Gel-phase 31P NMR 3a) and 13C NMR 3b) of 6b.
6.2 ppm), the signal of the resin-bound reference phosphine
325.6 ppm), which helped determine the yield of 6b 3see
Section 4.1), and traces of Ph2PH at 240 ppm. The equi-
librium 32a) is responsible for the generation of these traces
of Ph2PH and a resin-bound imine 3although only in a minor
amount) from 6 whenever it is incubated in a solvent 3e.g.
benzene-d6 for the gel-phase measurements).
The aromatic amine 11 was initially generated on solid
support through the immobilization of Fmoc-aminobenzoic
acid, followed by deprotection of the amino group 3Scheme
4). A large excess of benzaldehyde and a reduced amount of
TMOF 3as compared to the related procedure) was used for
the ef®cient conversion of the amine into the imine in order
Due to the equilibrium, and in order to maximize the yield
and purity of 6, two adjustments of the synthetic scheme
Scheme 3. XˆCl: 3i)t-Boc-Ala-OH, Cs2CO3, KI 3cat.), DMF, 808C, 24 h; 3ii) TFA±CHCl3 31:1 vol); 3iii) DIPEA, THF; XˆBr: 3iv) Fmoc-Ala-OH, DIPEA,
CsI, DMF, rt, 18 h; 3v) 20% piperidine in DMF; 3vi) PhCHO, TMOF/DCM, AcOH 3cat.), rt 24 h; 3vii) Ph2PH, CHCl3, rt, 18 h.

5150
B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
Scheme 4. 3i) Fmoc-4-benzoic acid, DIPEA, CsI, DMF, rt, 18 h; 3ii) piperidine, DMF; 3iii) PhCHO, TMOF/DCM, AcOH 3cat.), rt 24 h; 3iv) Ph2PH, CHCl3, rt,
18 h.
to reduce the amount of imidate byproduct 14. Reaction of
the imine with diphenylphosphine yields the ligand with
82% overall yield 3as determined by gel-phase 31P NMR).
Since the number of easily accessible Fmoc-aminobenzoic
acid building blocks is limited, an alternative simple path to
immobilized anilines was explored 3Scheme 5). According
to this scheme, nitrobenzoic acid was immobilized on Wang
Bromo resin and then reduced with stannous chloride. Such
an approach is expected to bene®t from the diversity of
easily accessible nitroaromatic carboxylic acids.
Figure 2. Gel-phase 31P NMR of 9b.
Although polymer-supported ligands are less sensitive to
oxidation than their solution analogues, an attempt was
made to perform the reaction of an imine with the diphenyl-
phosphine±borane adduct on solid support 3Scheme 6)Ðan
analogue of Eq. 34). Unfortunately, reduction of the imine,
which is only a minor side reaction in solution, becomes
a major process on the solid support. The excess of the
phosphine±borane reagent, that must be used for solid-
phase reaction, is the probable reason for this reaction
outcome. Although the synthetic approach of Scheme 6
was abandoned, post-synthetic protection of the resin-
Scheme 5. 3i) p-Nitrobenzoic acid, DIPEA, CsI, DMF, rt, 18 h; 3ii) 2 M
SnCl2´H2O, DMF, rt, 6 h; 3iii) PhCHO, TMOF/DCM, AcOH 3cat.), rt 24 h.
bound ligand with BH3 was later applied 3vide infra).
Scheme 6. 3i) 4-Formylbenzoic acid, DIPEA, CsI, DMF, rt, 18 h; 3ii) PhNH2, AcOH 3cat.), DCM/TMOF, rt, 18 h; 3iii) diphenylphosphineborane complex,
CHCl3, rt, 18 h.

B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
5151
Scheme 7.
Following the preparation of model ligands on support, the
method using the aldehyde as the anchoring building block
was preferred for parallel synthesis. A number of aldehydo-
carboxylic acids were examined for this function. While
some proved unsuitable for the chosen loading scheme
3for instance, 5-formyl salicylic acid only exhibited loading
ef®ciency of 10%), a reasonably high ef®ciency of immo-
bilization was observed for others. Thus, in addition to
4-formylbenzoic acid demonstrating 98%-ef®cient loading,
3-formylbenzoic acid and 2-formylphenoxy acetic acid
demonstrated 99- and 78%-yield loading respectively. The
loading yield was determined via the nucleophilic cleavage
method.
imine 5c formed 1H-3-methoxy-2-phenylisoindol-1-one
318) in 84% yield. Formation of 18 is likely to occur by
the mechanism demonstrated in Scheme 8 and reported
for a similar imino-ester in solution.21 This result, as well
as 13C NMR measurements 3showing aldehydic carbons at
188.5 and 191.2 ppm for 2-formylphenoxyacetic and
2-formylbenzoic acids respectively) indicate that both
acids are bound to the resin as aldehydoesters. It is clear
that the actual ef®ciency of loading for the 2-formylbenzoic
acid is substantially higher than the obtained yield of 16.
An additional aldehydic acid that was examined as a pos-
sible building block is 1,8-naphthalaldehydic acid. Similar
to 2-formylbenzoic acid, an acetal ester, 3-methoxy-1H,3H-
naphtho[1,8-cd]pyran-1-one 317) was obtained as a single
product upon nucleophilic cleavage of the resin-bound acid
3Scheme 7).
Interestingly, while for 3- and 4-formylbenzoic acids, the
cleavage products are simply methyl esters 3Scheme 7), for
2-formylphenoxyacetic acid the methyl ester of 2-benzo-
furan carboxylic acid 315, Scheme 7) is obtained. This is a
product of aldol condensation, a secondary reaction in
cleavage solution. It is likely that, due to the additional
transformation, the measured loading ef®ciency, based on
15, is in fact lower than the actual one. A similar situation
was observed for 2-formylbenzoic acid. Thus, nucleophilic
cleavage of the loaded carboxyaldehyde formed 3-methoxy-
3H-isobenzofuran-1-one 316) with 30% yield 3primary
cleavage product or secondary product of intramolecular
acetalization in solution). However, the cleavage following
an attempted conversion of the resin bound aldehyde to
2.3. Parallel synthesis
Following these studies, an attempt to prepare the ®rst
library of resin-supported a-aminophosphines was accom-
plished. Four carboxy-aldehydes and seven amines were
chosen for its construction 3library I, Scheme 9). Our pre-
liminary attempts to perform automated synthesis of
a-aminophosphines according to Schemes 2 and 3 revealed
substantial oxidation of the target ligands. In order to deter-
mine whether the oxidation occurred before analysis, an
Scheme 8.

5152
B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
Scheme 9.
additional stepÐboronation-protectionÐwas applied to
some of the wells during the synthetic scheme used for
the generation of the ®rst library. While in the wells
where boronation was applied, pure a-aminophosphine
borane was formed, in other wells, a variable amount of
oxidation led to the mixture of a-aminophosphine and the
corresponding a-aminophosphine oxide. The reduction of
the already formed phosphine oxide during the boronation
can be ruled out since it requires the presence of CeCl3.22
The results of this experiment clearly show that the oxida-
tion occurs only after the completion of the automated
synthesis and while the resin is handled outside the synthe-
sizer.
the library members from the phosphorus-containing
by-products, which can interfere with the catalytic screening
of the library in the future. This data was determined for all
library members using 31P NMR and is summarized in Table
1 3in parentheses). The 1,8-naphthalaldehydic acid building
block that had a low loading ef®ciency in the preliminary
results, proved unsuitable for parallel synthesis. The same is
true using the 3-aminopyrazole 3not a true amine, but rather
an amidine) as a building block. Somewhat lower yields
were obtained for 2-aminopyridine 3again amidine) and
benzylamine 3aliphatic substituents destabilize a-amino-
phosphines according to the literature14). The majority of
the members were pure from any phosphorus-containing
byproducts.
The results of the analysis of this library are summarized in
Table 1. Since it was established that the atmosphere in the
synthesizer was kept inert throughout the synthesis and
oxidation only occurred thereafter 3during the NMR sample
preparations), the yield given in the table combines the
yields of both a-aminophosphine and its oxidized counter-
part. The purities of the resin-bound compounds generally
re¯ect the yields, since any functionalized site not occupied
by the product is an impurity per se. However, with regard
to our research, it was important to determine how pure are
An interesting observation was made when the 31P chemical
shifts of the library members were compared to a 6.6 ppm
chemical shift of the model ligand 6 3Table 2). It seems
that the substituent on the bridge strongly in¯uences the
chemical shift of the phosphorus atom. Thus, when
2-formylcarboxylic acid was used as the building block,
the chemical shift of the phosphine moved ca. 4 ppm down-
®eld. When 2-formylphenoxyacetic acid was employed as
an aldehydic/anchoring building block, the chemical shift
Table 1. Yield and purity 3in parenthesis) of library I members
Aldehyde
Amine
20a
20b
20c
20d
20e
20f
20g
19a
19b
19c
19d
nda 3100)
50 3100)
Traces 3±)
6 3±)
97 398)
nd 3100)
80 395)
17 3±)
27 3100)
60 3100)
nd 3100)
Traces 3±)
45 392)
0
20 3100)
56 398)
28 392)
nd 395)
0
Traces 3±)
0
18 396)
nd 395)
66 393)
89 379)
Traces 3±)
0
0
Yields as determined using referenced gel-phase 31P NMR; purity: purity from phosphorus-containing byproducts.
a
ndˆnot determined.

B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
5153
Table 2. 31P Chemical shift of library I members
gen substituents span the range of the 31P chemical shifts of
the library over 11 ppm and this difference in chemical
shifts holds promise for the diverse activity of the libraries'
members in the future.
Library members formed from
d 3ppm)
19b/20a
19c/20a
19a/20b
19c/20b
19a/20c
19b/20c
19a/20d
19c/20d
19a/20e
19b/20e
19c/20e
19c/20f
19b/20g
19c/20g
10.6
2.8
6.1
2.9
5.8
10.1
2.7
0.6
4.4
8.9
3.6
20.8
7.6
2.7
Four carboxaldehydes and 10 amines were chosen for the
construction of the improved, extended library 3library II).
Most building blocks of the previous sets were used
3Scheme 10). Naphthalaldehydic acid, which was unsuitable
for parallel synthesis, was replaced by 3-formylbenzoic
acid. In addition to the ®rst ®ve amino building blocks,
which demonstrated reasonable reactivity in the previous
experiments, ®ve new amino building blocks were chosen.
These included electron-rich, electron-poor and sterically
crowded aromatic amines. The formed a-aminophosphines
were protected with borane 3Scheme 10) in order to prevent
oxidation during post-synthetic manipulations. Initial gel-
phase 31P NMR screening of the whole library demonstrated
that all library members, except one 3derived from 19b and
20d), were formed 3Table 3). The same tests also demon-
strated the excellent purity of the library members from
phosphorus-containing byproducts 3exceeding 90% for
95% of the library members) 3Table 4).
moved ca. 3 ppm in the opposite direction. The in¯uence of
the nitrogen substituent can also be seen from Table 2.
Aliphatic substituents 3Bn), as well as 2-pyridyl, shift the
phosphine peak up®eld by 2±3 ppm 3see also the chemical
shifts of 9). The combined in¯uence of the bridge and nitro-
Scheme 10.
Table 3. 31P Chemical shifts, d 3ppm), of library II members
Aldehyde
Amine
20a
20b
20c
20d
20e
20h
20i
20j
20k
20l
19a
19b
19c
19e
28.8
29.7
28.4
28.5
28.4
29.4
26.9
28.2
28.9
29.9
29.6
28.7
25.8
±
25.0
25.8
30.2
31.4
27.4
30.3
28.8
29.7
27.4
28.6
29.7
30.6
28.1
29.6
29.6
30.3
27.9
29.4
29.4
30.3
27.2
29.1
28.8
29.8
26.9
28.3

5154
B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
Table 4. Yield and purity 3in parenthesis) of library II members
Aldehyde
Amine
20a
20b
20c
20d
20e
20h
20i
20j
20k
20l
19a
19b
19c
19e
85 3100)
67 3100)
nd 3100)
nd 3100)
97 3100)
49 397)
70 3100)
nd 3100)
nda 3100)
60 390)
67 3100)
88 3100)
nd 3100)
Traces 3±)
37 3100)
19 395)
nd 399)
18 387)
19 398)
35 391)
nd 3100)
nd 384)
nd 3100)
96 393)
100 399)
100 3100)
nd 3100)
91 3100)
99 3100)
nd 3100)
nd 3100)
91 3100)
80 3100)
66 3100)
nd 3100)
nd 3100)
nd 3100)
nd 390)
nd 3100)
nd 3100)
Yields as determined using referenced gel-phase 31P NMR; purity: purity from phosphorus-containing byproducts.
a
ndˆnot determined.
The yield was determined using the 31P NMR method for 21
library members. 71% of the members were formed with
yields exceeding 60%. As before, 2-amino-pyridine- and
benzylamine-incorporating products were produced with
lower ef®ciency 3vide supra). Generally, lower yields were
observed for products formed from ortho-substituted alde-
hydes 319b and 19c). Surprisingly, sterically hindered amines
320i±k) form the aminophosphines with excellent yields.
benzylic alcohols, which were derived from unreacted
aldehydes upon boronation, occupy the remaining sites of
the resin.23 For another four resins, additional peaks at
43±47 ppm reveal occupation of the remaining `impurity'
sites by aryl-benzyl secondary amines, formed by the
reduction of unreacted imines during the boronation-
protection step.24
The analysis performed on the library indicated that it can
be used for complexation and catalysis studies. Even the
members formed with low yield are present on the resin in
amounts that allow meaningful evaluation 3as demonstrated
by the 31P and 13C NMR). Moreover, even for members
containing resin-bound byproducts, the catalytic sites,
formed upon complexation, will be isolated from the
byproducts due to the pseudodilution principle.
The strikingly different in¯uence of steric hindrance when
associated with the aldehydic and amine partners, can be
attributed to the fact that the steric interactions between the
phosphine moiety and bridge substituent of the a-amino-
phosphine must destabilize the molecule much more than
the steric interactions between the phosphine moiety and the
nitrogen substituent. Consequently, the equilibrium 32a) for
bridge-hindered a-aminophosphines is less favorable.
As aforementioned, for the library of resin-bound
compounds, the general purity re¯ects the yield. Additional
analysis was applied in order to test the purity of the
products. Twenty one library members were examined
using gel-phase 13C NMR. The spectra generally con®rm
our conclusions from the 31P measurements. While all resins
exhibited the peaks assigned to the products 3Table 5), some
of those derived from electron-poor or aliphatic amines, or
hindered aldehydes, exhibited additional signals. For four
members, an additional peak at 64±65 ppm reveals that
3. Conclusion
In conclusion, we developed synthetic routes for solid-phase
a-aminophosphine synthesis. The methods are compatible
with parallel synthesis, as demonstrated by the construction
of the ®rst libraries of a-aminophosphines. Dif®culties in
qualitative and quantitative analysis of the products were
overcome by gel-phase NMR techniques. This research
opens the way to parallel complexation and catalysis studies
with a-aminoposphines.
Table 5. Partial gel-phase 13C NMR data for library II members
Members formed from
d 3ppm)
19a,20b
19a,20c
19a,20d
19a,20h
19a,20k
19a,20l
19b,20a
19b,20c
19b,20j
19b,20k
19c,20b
19c,20h
19c,20i
19c,20j
19c,20k
19c,20l
19e,20a
19e,20b
19e,20h
19e,20i
19e,20j
19e,20k
139.2, 132.8, 132.2, 130.9, 114.4, 69.3, 65.9, 54.5
132.9, 131.1, 114.3, 73.5, 69.4, 63.5, 59.6, 13.7
145.1, 132.5, 114.3, 69.1, 70.0, 58.1, 50.8
145.9, 144.1, 132.8, 131.9, 114.4, 113.4, 69.2, 65.9, 54.2
145.3, 140.5, 132.9, 132.1, 130.8, 118.1, 114.2, 111.8, 109.5, 69.3, 65.9, 63.8, 45.7, 20.54
132.0, 114.3, 69.1, 65.9, 53.5
166.2, 145.6, 138.7, 132.6, 128.3, 117.9, 113.3, 69.2, 65.8, 47.4
165.4, 138.1, 132.6, 114.2, 69.3, 66.1, 59.3, 47.2, 13.7
154.1, 141.6, 132.7, 130.7, 122.4, 121.1, 114.3, 110.7, 69.3, 65.9, 46.5, 19.8, 16.8
149.5, 144.3, 142.0, 139.2, 132.8, 130.0, 117.5, 114.3, 69.3, 68.1, 62.8, 54.8, 47.5, 20.8
152.4, 132.7, 129.9, 114.4, 69.1, 63.8, 54.4, 46.9
154.0, 144.1, 132.6, 122.5, 121.9, 114.4, 110.7, 109.6, 69.3, 66.1, 63.9, 56.7
132.6, 114.1, 69.2, 63.8, 46.1, 16.8
167.2, 154.1, 141.5, 132.7, 130.7, 128.3, 122.5, 121.1, 114.3, 110.6, 109.5, 102.4, 69.3, 65.9, 63.9, 16.5, 19.8, 16.8
154.2, 145.1, 132.7, 130.1, 121.1, 117.4, 114.3, 111.5, 109.4, 104.1, 69.2, 65.9, 63.9, 56.0, 45.5, 20.7
154.3, 143.0, 136.6, 132.7, 122.4, 121.3, 114.4, 110.6, 69.1, 66.2, 64.5, 46.4
165.4, 145.1, 139.7, 135.5, 132.8, 132.0, 130.9, 118.4, 113.6, 69.2, 65.8, 53.7
165.3, 139.1, 135.9, 132.8, 138.8, 129.5, 114.4, 69.2, 67.0, 54.4
143.8, 132.8, 132.0, 131.1, 123.0, 114.6, 69.2, 65.9, 53.9
143.3, 135.5, 132.8, 131.9, 130.9, 122.5, 118.1, 114.2, 110.9, 106.1, 69.0, 65.7, 54.0, 16.6
165.9, 140.9, 135.9, 132.8, 131.9, 130.8, 129.4, 122.5, 114.3, 111.1, 108.3, 105.6, 102.4, 69.2, 65.7, 54.2, 19.8, 16.6
145.2, 135.1, 132.1, 130.7, 129.4, 118.0, 114.3, 111.9, 109.5, 69.2, 65.6, 54.7, 53.6, 20.5

B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
5155
4. Experimental
4.3.1. N-Phenyl, a-phenyl aminomethyldiphenylphos-
phine oxide 12). solution of diphenylphosphine
A
4.1. General
3940 mg, 5 mmol, 1 equiv.) in dry toluene 310 ml), was
added to a dry ¯ask, equipped with CaCl2 tube, which
contained a solution of aniline 31.8 ml, 20 mmol, 4 equiv.)
and benzaldehyde 30.5 ml, 5 mmol, 1 equiv.) in dry toluene
320 ml). After 4 h stirring at room temperature, the precipi-
tate was ®ltered, washed with cold toluene and then dried
under vacuum.
All reactions were conducted in oven-dried glassware. THF
was dried over, and distilled from, sodium metal with
benzophenone as the indicator. DCM and chloroform were
dried over, and distilled from CaH2. Anhydrous DMF was
purchased from Sigma±Aldrich. Diphenylphosphine was
purchased from Strem as a 10% solution in Hexanes. Resins
were purchased from Novabiochem and all other reagents
were purchased from Sigma±Aldrich. 1H, 31P and 13C NMR
were recorded using Brucker Avance-200, Avance-400 or
ARX-500 instruments and referenced to TMS 31H), solvent
313C) or external 85% H3PO4 331P) standard signals.
Chemical shifts are reported in ppm. The gel-phase samples
were prepared in benzene-d6. Since the polymeric matrix
obscures some of the compounds signals, partial 13C data
is reported for most compounds. The library was prepared
on an Advanced ChemTech Multiple Organic Synthesizers
440V. Yields were determined using nucleophilic cleavage
30.4 M NaOMe in MeOH±THF 31:4))25 or 31P NMR of the
products with triphenylphosphine±polystyrene resin 3Fluka,
100±200 mesh, 1.6 mmol/g) as a reference.
Yield 1.92 g 3100%). Mp: 2008C. 1H NMR 3500 MHz,
CDCl3): d 7.78 3dd, Jˆ9.8, 5.8 Hz, 2H, P3O)±Ph), 7.48
3dt, Jˆ7.5, 1.5 Hz, 1H, P3O)±Ph), 7.40 3dt, Jˆ7.5, 3.0 Hz,
2H, P3O)±Ph), 7.31 3m, 3H, P3O)±Ph), 7.18 3dt, Jˆ7.5,
3.0 Hz, 2H, P3O)±Ph), 7.04 3m, 5H, C±Ph), 6.99 3t, Jˆ
8.0 Hz, 2H, N±Ph), 6.58 3t, Jˆ7.3 Hz, 1H, N±Ph), 6.51
3d, Jˆ7.5 Hz, 2H, N±Ph), 5.06±5.12 3m, 2H, CH, NH).
13C NMR 350.4 MHz, CDCl3): d 146.2, 135.3, 132.4,
131.9, 131.8, 131.5, 129.3, 129.1, 128.8, 128.5, 128.3,
128.2, 127.8, 118.6, 115.3, 114.1, 57.7 3d, Jˆ75.6 Hz). 31P
NMR 3202.5 MHz, CDCl3): d 33.6 3s). HRMS 3FAB):
found 3m/z) 384.1528; cald for C25H23NOP 3MH1)
384.1517.
4.3.2. N-Phenyl, a-phenyl aminomethyldiphenylphos-
phine borane 13). In a glove box, diphenylphosphine
borane 3400 mg, 2 mmol, 1 equiv.) and benzylideneaniline
3726 mg, 4 mmol, 2 equiv.) were dissolved in 6 ml of dry
toluene, in a pressure tube. The closed tube was heated to
608C for 72 h. After cooling, the precipitate was ®ltered and
dried under vacuum.
Reduction of nitro group and cleavage of protecting groups
was performed according to literature procedures.26,27
4.2. Solution synthesis
4.2.1. N-Phenyl, a-phenyl aminomethyldiphenylphos-
phine 11).20 A solution of diphenylphosphine 3940 mg,
5 mmol, 1 equiv.) in dry toluene 310 ml), was slowly
added under nitrogen atmosphere, to two-necked ¯ask,
equipped with a condenser, and preheated to 658C, which
contained a solution of aniline 31.8 ml, 20 mmol, 4 equiv.)
and benzaldehyde 30.5 ml, 5 mmol, 1 equiv.) in dry toluene
320 ml). The solution was stirred overnight. After cooling,
the suspension was ®ltered, the toluene was evaporated from
the ®ltrate and the excess of aniline was distilled off under
vacuum, yielding crude product 3colorless solid).
1
Yield 687 mg 390%). H NMR 3500 MHz, CDCl3): d 7.78
3t, Jˆ4.4 Hz, 2H, P±Ph), 7.52 3t, Jˆ6.8 Hz, 1H, P±Ph), 7.45
3dt, Jˆ7.8, 2.3 Hz, 2H, P±Ph), 7.39 3dt, Jˆ7.5, 1.5 Hz, 1H,
P±Ph), 7.30 3t, Jˆ8.0 Hz, 2H, P±Ph), 7.24 3dt, Jˆ7.5,
2.5 Hz, 2H, P±Ph), 7.12±7.06 3bm, 5H, 3H C±Ph, 2H N±
Ph), 7.01 3d, Jˆ6.5 Hz, 2H, C±Ph), 6.68 3t, Jˆ12.5 Hz, 1H,
N±Ph), 6.56 3d, Jˆ4.9 Hz, 2H, N±Ph), 5.38 3dd, Jˆ16.0,
8.5 Hz, 1H, CH), 5.07 3t, Jˆ8.0 Hz, 1H, NH), 0.87±1.27
3bm, 3H, BH3). 13C NMR 350.4 MHz, CDCl3): d 145.9 3d,
Jˆ4.4 Hz), 135.0, 133.4 3d, Jˆ7.8 Hz), 132.6 3d, Jˆ
7.7 Hz), 131.6, 129.3, 129.1, 128.9, 128.8, 128.6, 128.4,
128.3, 127.9, 126.7 3d, Jˆ17.5 Hz), 118.5, 113.9, 54.7 3d,
Jˆ39.9 Hz). 31P NMR 3202.5 MHz, CDCl3): d 27.9 3bs).
Anal. Cald for C25H25BNP: C, 78.76; H, 6.61; N, 3.67.
Found: C, 78.52; H, 6.64; N, 3.67.
1
Yield 880 mg 348%). H NMR 3500 MHz, CDCl3): d 7.44
3m, 1H, P±Ph), 7.37 3m, 2H, P±Ph), 7.32 3t, Jˆ7.5 Hz, 2H,
P±Ph), 7.03±7.30 3m, 12H, 5H P±Ph, 5H C±Ph, 2H N±Ph),
6.64 3t, Jˆ8.0 Hz, 1H, N±Ph), 6.49 3d, Jˆ7.5 Hz, 2H,
N±Ph), 5.12 3d, Jˆ3.4 Hz, 1H, CH), 4.25 3bs, 1H, NH).
13C NMR 350.4 MHz, CDCl3): d 147.2, 140.1 3d, Jˆ
10.9 Hz), 135.4 3d, Jˆ13.6 Hz), 135.2 3d, Jˆ8.5 Hz),
133.7, 129.3, 128.7, 128.6, 128.4, 128.3, 127.7, 126.9,
126.0, 121.0, 117.9, 113.8, 57.5 3d, Jˆ13.6 Hz). 31P NMR
3202.5 MHz, CDCl3): d 6.2 3s).
4.4. Solid-phase synthesis
4.4.1. General procedure for acidic cleavage. Resin
350 mg) was stirred in CDCl3±TFA solution 31:1 vol) for
1 h. The ®ltrate was collected into NMR tube. Yields were
established using benzene as an internal reference 311 mM
solution).
4.3. N-Phenyl, a-phenyl aminomethyldiphenylphosphine
11)Ðstepwise procedure
4.4.2. General procedure for the attachment of acids to
Merri®eld resin. The Merri®eld resin 30.72 mmol/g,
1 equiv.) was swollen in a minimal volume of DMF. The
acid 35 equiv.), Cs2CO3 35 equiv.) and KI 31 equiv.) in DMF
310 ml/1 g resin) were added to the suspension. The mixture
was gently stirred for 24 h at 808C, cooled and ®ltered. The
In a glove box, a solution of diphenylphosphine 3940 mg,
5 mmol, 1 equiv.) in dry chloroform 35 ml) was added to a
20 ml vial, which contained a solution of benzylideneaniline
3905 mg, 5 mmol, 1 equiv.) in dry chloroform 35 ml). After
2 h, the solution was ®ltered and evaporated to give the pure
ligand in 92% yield 31.69 g).

5156
B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
resin was washed with DMF, DMF/THF 31:1), THF, THF/
H2O 31:1), H2O, THF and DCM and dried under vacuum.
Fmoc-aminobenzoic acid: yield 87%.
4.4.4. Typical procedure for the preparation of imines
from resin bound aldehydes. Under a nitrogen atmos-
phere, aniline 310 equiv.) and acetic acid 30.1 equiv.) were
dissolved in DCM/trimethylorthoformate 31:1, 5 ml/g resin)
and added to resin-aldehyde which was swollen in dry DCM
35 ml/g resin). The mixture was gently stirred overnight and
then ®ltered. The polymer was washed with DCM and dried
under vacuum.
4.4.3. General procedure for the attachment of acids to
Wang Bromo resin. The Wang Bromo resin 31.19 mmol/g,
1 equiv.) was swollen in a minimal volume of DMF. The
acid 35 equiv.) DIPEA 35 equiv.) and CsI 30.5 equiv.) were
dissolved in DMF 310 ml/g resin) and added to the resin
suspension. The mixture was gently stirred for 18 h and
then ®ltered. The resin was washed with DMF, DMF/THF
31:1), THF, THF/H2O 31:1), H2O, THF and DCM and dried
under vacuum.
For the following imines the yield 3based on initial loading
of the Merri®eld or Wang resin) was determined by nucleo-
philic cleavage
For the following acids the yield was determined by nucleo-
philic cleavage:
5a, b: yield 85%.
4-Formylbenzoic acid 34b): yield 98%. Partial gel phase 13
NMR: d 190.7, 165.4, 145.6, 139.4, 108.7, 70.0, 67.1.
C
5b: Partial gel-phase 13C NMR: d 165.7, 158.8, 152.2,
146.8, 128.0, 124.5, 121.4, 70.0, 66.9.
4-Formylcinnamic acid: yield 34%. Partial gel phase 13C
NMR 3128.5 MHz, CDCl3): d 190.6, 165.8, 145.7, 143.2,
124.6, 121.3, 70.0, 66.4.
5c: yield 84%.
1H-2,3-Dihydro-3-methoxy-2-phenyisoindol-1-one 318):32
3product of the nucleophilic cleavage): 1H NMR 3500
MHz, CDCl3): d 7.93 3d, Jˆ7.5 Hz, 1H, Ar), 7.82 3d, Jˆ
8.0 Hz, 2H, N±Ph), 7.66 3t, Jˆ7.4 Hz, 1H, Ar), 7.61 3t,
Jˆ7.5 Hz, 1H, Ar), 7.58 3d, Jˆ6.3 Hz, 1H, Ar), 7.45 3t,
Jˆ7.8 Hz, 2H, N±Ph), 7.22 3t, Jˆ7.5 Hz, 1H, N±Ph), 6.98
3s, 1H, CH), 2.88 3s, 3H, OCH3). 13C NMR 350.4 MHz,
CDCl3): d 166.9, 140.0, 137.6, 133.1, 132.9, 130.5, 129.3,
129.2, 128.5, 125.7, 125.5, 124.1, 121.9, 87.6, 49.3.
5-Formylsalicylic acid: yield 7%.
3-Formylbenzoic acid: yield 99%.
2-Formylphenoxy acetic acid: yield 71%. Partial gel phase
13C NMR 3128.5 MHz, CDCl3): d 188.5, 160.2, 145.3,
135.2, 126.1, 123.9, 121.7, 112.8, 70.0, 66.7, 65.4.
2-Benzofuran carboxylic acid, methyl ester 315):28 3product
of the nucleophilic cleavage) 1H NMR 3500 MHz, CDCl3):
d 7.68 3d, Jˆ7.9 Hz, 1H, Ar), 7.59 3d, Jˆ8.5 Hz, 1H, Ar),
7.45 3t, Jˆ7.8 Hz, 1H, Ar), 7.31 3t, Jˆ7.5 Hz, 1H, Ar), 6.97
3s, 1H, CHvC), 3.98 3s, 3H, O±CH3).
4.4.5. General procedure for the reaction of resin-bound
imines with diphenylphosphine. In a glove box, a solution
of diphenylphosphine 310 equiv.) in chloroform 35 ml/g
resin) was added to resin-bound imine 31 equiv.) which
was swollen in chloroform 35 ml/g resin). The mixture
was gently stirred overnight and then ®ltered. The resin
was brie¯y washed 3just once) with DCM and dried under
vacuum.
2-Formylbenzoic acid: yield 30%. Partial gel phase 13C
NMR 350.4 MHz, CDCl3): d 191.2, 168.8, 145.7, 137.7,
132.3, 130.7, 126.2, 124.6, 70.0, 67.3.
3-Methoxy-3H-isobenzofuran-1-one 316):29 3product of the
nucleophilic cleavage) 1H NMR 3200 MHz, CDCl3): d 7.90
3d, Jˆ7.4 Hz, 1H, Ar), 7.76 3d, 7.2 Hz, 1H, Ar), 7.70 3t,
Jˆ8.0 Hz, 1H, Ar), 7.61 3t, Jˆ7.6 Hz, 1H), 6.31 3s, 1H,
CH), 3.64 3s, 3H, OMe).
Yield for 6a: 85% 3by gel-phase 31P NMR); gel-phase 31P
NMR: d 6.6 3s).
Yield for 6b: 85% 3by gel-phase 31P NMR); gel-phase 31P
NMR: d 6.6 3s). Partial gel-phase 13C NMR: d 146.0, 134.1,
133.4, 124.6, 118.6, 114.0, 70.0, 66.4, 57.6.
1,8-Naphthaldehydic acid: yield 25%.
Yield for 9b: 30%; gel phase 31P NMR 3202.5 MHz,
benzene-d6): d 3.3 3s), 1.0 3s).
3-Methoxy-1H,3H-naphtho[1,8-cd]pyran-1-one
317):30
3product of the nucleophilic cleavage) 1H NMR 3200
MHz, CDCl3): d 8.47 3d, Jˆ7.8 Hz, 1H, Ar), 8.16 3d, Jˆ
8.0 Hz, 1H, Ar), 7.98 3t, Jˆ7.4 Hz, 1H, Ar), 7.59±7.73 3m,
3H, Ar), 6.48 3s, 1H, CH±OMe), 3.74 3s, 3H, O±CH3).
Yield for 13: 82%; gel phase 31P NMR 3202.5 MHz,
benzene-d6): d 5.1 3s).
4.5. General procedure for the multicomponent
synthesis of 6
t-Boc-l-Ala-OH: yield 92%.
4-Nitrobenzoic acid: yield 94%.
In a glove box, diphenylphosphine 310 equiv.) and aniline
310 equiv.), dissolved in a 2:2:1 mixture of DCM, CHCl3
and trimethylorthoformate 35 ml/g resin), were added to
aldehyde-bound resin which was swollen in the same
solvent mixture 35 ml/g resin). The mixture was gently
For the following acids, yields were established by spectro-
photomeric determination of Fmoc group.31
Fmoc-l-Ala-OH: yield 81%.

B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
5157
stirred overnight and ®ltered. Brief washing with DCM and
drying under vacuum yielded 6 with ca. 85% yield.
phere. CHCl3 31 ml) and DCM 30.5 ml) were distributed to
each well of the block. The block was shaken at 500 rpm for
20 min. Trimethylorthoformate 31 ml) was added to each
well followed by an amine 310 equiv.) solution in 1 ml
DCM and 0.1 equiv. of AcOH. The block was shaken for
20 min and a Ph2PH 320 equiv.) solution in 0.5 ml chloro-
form was added to each well. The block was shaken at
500 rpm for 18 h. The solution was removed and the resin
in each well was washed with 3 ml of DCM. The resin was
dried under an argon stream for 10 min.
4.5.1. General procedure for the preparation of imines
using amine resins. Under nitrogen atmosphere, benzalde-
hyde 320 equiv.) and AcOH 30.1 equiv.) dissolved in DCM/
TMOF 33:1 vol) were added to a suspension of the swollen
amine resin in the minimum volume of the same solvents.
The suspension was gently stirred at room temperature,
overnight and then washed with DCM and dried under
vacuum.
4.6.3. Borane protection. Under nitrogen atmosphere. THF
32.5 ml) was added to each well, followed by 0.5 ml of a
1 M solution of BH3´THF in THF. The block was shaken at
500 rpm for 6 h. The solvent was removed and the resin in
each well was washed with THF 33£3 ml) and DCM
35£3 ml). The resin was dried under a nitrogen stream for
1 h and, after the contents of each well were transferred to a
separate vial, dried again under vacuum.
For the following resin-bound imines the yield was deter-
mined by nucleophilic cleavage:
N-Benzylidene alanine 37b): yield 94%. 1H NMR 3200
MHz, CDCl3): d 8.51 3s, 1H, CHvN), 8.14 3d, Jˆ8.2 Hz,
2H, C±Ar), 7.90 3d, Jˆ8.2 Hz, 2H, C±Ar), 7.38 3m, 2H,
N±Ph), 7.25 3m, 3H, N±Ph), 3.95 3s, 3H, O±CH3).
N-Benzylidene-4-carboxyaniline 312): yield 98%. Partial
gel phase 13C NMR 3128.5 MHz, CDCl3): d 190.7, 165.4,
145.6, 139.4, 108.7, 70.0, 67.1.
Acknowledgements
This research was supported by the Israel Science
Foundation, founded by the Israel Academy of Sciences
and Humanities.
N-Benzylidene 34-carbomethoxy) aniline 3product of the
nucleophilic cleavage): 1H NMR 3200 MHz, CDCl3): d
8.44 3s, 1H, CHvN), 8.08 3d, Jˆ8.6 Hz, 2H, N±Ar),
7.89±7.84 3m, 2H, C±Ph), 8.47±8.52 3m, 3H, C±Ph), 7.21
3d, Jˆ8.4 Hz, 2H, N±Ar), 3.92 3s, 3H, O±CH3).
References
By-product was obtained: O-methyl, N-34-carbomethoxy-
phenyl)formimidate 3product of the nucleophilic cleavage
of 14): 1H NMR 3200 MHz, CDCl3): d 8.25 3s, 1H, CHvN),
8.01 3d, Jˆ8.5 Hz, 2H, N±Ar), 7.11 3d, Jˆ8.4 Hz, 2H, N±
Ar), 3.91 3s, 3H, O±CH3), 3.86 3s, 3H, O±CH3).
1. 3a) Gennari, C.; Nestler, H. P.; Piarulli, B.; Salom, B. Liebigs
Ann. 1997, 637. 3b) Jandeleit, B.; Schaefer, D. J.; Powers, T. S.;
Turner, H. W.; Weinberg, W. H. Angew. Chem., Int. Ed. Engl.
1999, 38, 2494. 3c) Reetz, M. T. Angew. Chem., Int. Ed. Engl.
2001, 40, 284.
2. 3a) Obrecht, D.; Villalgordo, J. M. Solid-supported Combina-
torial and Parallel Synthesis of Small-molecular-weight
Compound Libraries; Pergamon: Oxford, 1998. 3b) Zaragoza
Dorwald, F. Organic Synthesis on Solid PhaseÐSupports,
Linkers, Reactions; Wiley±VCH: Weinheim, 2000.
4.5.2. Postsynthetic protection of 6. The resin 6 3200 mg,
0.192 mmol, 1 equiv.) was suspended in a 1 M solution of
BH3´THF in THF 36 ml, 40 equiv.) diluted with an
additional 4 ml THF. The suspension was gently stirred
for 2 h and then ®ltered. The resin was washed with THF,
DCM and dried under vacuum.
3. 3a) Liu, G.; Ellman, J. A. J. Org. Chem. 1995, 60, 7712.
3b) Sigman, M. S.; Jacobsen, E. N. J. Am. Chem. Soc. 1998,
120, 4901.
Yield 100% 3by gel-phase 31P NMR); gel-phase 31P NMR: d
28.9 3bs). Partial gel-phase 13C NMR: d 159.8, 145.9, 133.7,
132.7, 119.2, 115.0, 70.0, 66.5, 55.0.
4. 3a) Cole, B. M.; Shimizu, K. D.; Krueger, C. A.; Harrity,
J. P. A.; Snapper, M. L.; Hoveyda, A. H. Angew. Chem., Int.
Ed. Engl. 1996, 35, 1668. 3b) Francis, M. B.; Jacobsen, E. N.
Angew. Chem., Int. Ed. Engl. 1999, 38, 937. 3c) Francis, M. B.;
Finney, N. S.; Jacobsen, E. N. J. Am. Chem. Soc. 1996, 118,
8983. 3d) Burger, M. T.; Still, W. C. J. Org. Chem. 1995, 60,
7382. 3e) Natarajan, A.; Madalengoitia, J. S. Tetrahedron Lett.
2000, 41, 5783. 3f) Brouwer, A. J.; van der Linden, H. J.;
Liskamp, R. M. J. J. Org. Chem. 2000, 65, 1750. 3g) Boussie,
T. R.; Coutard, C.; Turner, H. W.; Murphy, V.; Powers, T. S.
Angew. Chem., Int. Ed. Engl. 1998, 37, 3272.
4.6. Parallel synthesis of the libraries 1libraries I, II)
4.6.1. Aldehyde attachment. Under nitrogen atmosphere.
Wang Bromo resin was distributed to the wells of the reac-
tion block of the synthesizer 3150 mg/well). DMF 31.5 ml)
was added to each well and the block was shaken at 500 rpm
for 20 min. A carboxyaldehyde 310 equiv.) solution in 1 ml
of DMF was added to each well, followed by a DIPEA
310 equiv.) solution in 0.5 ml DMF and a CsI 31 equiv.)
solution in 0.5 ml DMF. The block was shaken at 500 rpm
for 17 h. The solution was removed and the resin in each
well was washed with 3 ml portions of DMF, DMF/THF,
THF, THF/H2O 31:1), H2O, THF and DCM. The resin was
dried under a nitrogen stream for 1 h.
5. 3a) Gilbertson, S. R.; Wang, X. Tetrahedron Lett. 1996, 37,
6475. 3b) Gilbertson, S. R.; Wang, X.; Hoge, G. S.; Klug,
C. A.; Schaefer, J. Organometallics 1996, 15, 4678.
6. See, for example, the following 3a) Bayston, D. J.; Fraser, J. L.;
Ashton, M. R.; Baxter, A. D.; Polywka, M. E. C.; Moses, E.
J. Org. Chem. 1998, 63, 3137. 3b) Uozumi, Y.; Danjo, H.;
Hayashi, T. Tetrahedron Lett. 1998, 39, 8303.
7. For rare examples, see the following 3a) Dumont, W.; Poulin,
J.-C.; Dang, T.-P.; Kagan, H. B. J. Am. Chem. Soc. 1973, 95,
4.6.2. Synthesis of aminophosphine. Under argon atmos-

5158
B. Bar-Nir Ben-Aroya, M. Portnoy / Tetrahedron 58 ,2002) 5147±5158
8295. 3b). Brown, J. M.; Molinari, H. Tetrahedron Lett. 1979,
31, 2933.
3b) NovabiochemÐCatalog and Peptide Synthesis Hand-
book; Novabiochem: Laufel®ngen, 1999; p S42.
18. Gordeev, M. F.; Gordon, E. M.; Patel, D. V. J. Org. Chem.
1997, 62, 8177.
8. Mansour, A.; Portnoy, M. J. Chem. Soc., Perkin Trans. 1 2001,
952.
9. 3a) Armstrong, R. W.; Combs, A. P.; Tempest, P. A.; Brown,
S. D.; Keating, T. A. Acc. Chem. Res. 1996, 29, 123.
3b) Domling, A.; Ugi, I. Angew. Chem., Int. Ed. Engl. 2000,
39, 3168. 3c) Beinayme, H.; Hulme, C.; Oddon, G.; Schmitt, P.
Chem. Eur. J. 2000, 6, 3321.
19. Although this is not a quantitative method, our experience
with 13C NMR gel-phase NMR taught us that the upper limit
for undetected impurities is 15%.
20. Andrieu, J.; Camus, J.-M.; Poli, R.; Richard, P. New J. Chem.
2001, 25, 1015.
10. 3a) Nandi, M.; Jin, J.; Rajan Babu, T. V. J. Am. Chem. Soc.
1999, 121, 9899. 3b) Mecking, S.; Keim, W. Organometallics
1996, 15, 2650. 3c) van den Beuken, E. K.; Smeets, W. J. J.;
Spek, A. L.; Feringa, B. L. Chem. Commun. 1998, 223.
3d) Kamikawa, T.; Hayashi, T. Tetrahedron 1999, 55, 3455.
3e) McCarthy, M.; Guiry, P. J. Tetrahedron 1999, 55, 3061.
3f) Gilbertson, S. R.; Xie, D. Angew. Chem., Int. Ed. Engl.
1999, 38, 2750.
21. Heaney, H.; Shuhaibar, K. F. Synlett 1995, 47.
22. 3a) Imamoto, T.; Kusumoto, T.; Suzuki, N.; Sato, K. J. Am.
Chem. Soc. 1985, 107, 5301. 3b) Imamoto, T.; Oshiki, T.;
Onosawa, T.; Kusumoto, T.; Sato, K. J. Am. Chem. Soc.
1990, 112, 5244.
23. Benzylic alcohol impurities accompany the library members
derived from the following pairs: 19a/20c, 19b/20c, 19b/20j,
19a/20k.
11. 3a) Kellner, K.; Tzschach, A. Z. Chemistry 1984, 24, 365.
3b) Kellner, K.; Seidel, B.; Tzschach, A. J. Organomet.
Chem. 1978, 149, 167. 3c) Kellner, K.; Hanke, W.; Tzscach,
A. Z. Chemistry 1984, 24, 193. 3d) Baruluenga, J.; Campos,
P. J.; Canal, G.; Asensio, G. Synlett 1990, 261. 3e) Pudovik,
A. N.; Romanov, G. V.; Pozhidaev, V. M. J. Gen. Chem.
USSR 1978, 48, 920.
24. Benzyl-aryl amines contaminate the library members derived
from the following pairs: 19a/20h, 19c/20h, 19e/20j, 19c/20k.
25. Frenette, R.; Friesen, R. N. Tetrahedron Lett. 1994, 35, 9177.
26. Kiselyov, A. S.; Armstrong, R. W. Tetrahedron Lett. 1997, 38,
6163.
27. 3a) Kocienski, P. J. Protecting Groups; Thieme: Stuttgart,
1994; Chapter 6, p 192. 3b) Kocienski, P. J. Protecting
Groups; Thieme: Stuttgart, 1994; Chapter 6, p 202.
28. Belleney, J.; Verbel, J.; Platzer, N.; Theobald, F. Spectrochim.
Acta 1983, 39A, 429.
12. LaPointe, A. M. J. Comb. Chem. 1999, 1, 101.
13. The preliminary results of this study were published as
a communication: Bar-Nir Ben-Aroya, B.; Pornoy, M.
J. Comb. Chem. 2001, 3, 524.
29. Shim, S. C.; Lee, D. J.; Jiang, L. H.; Kim, T. J.; Cho, S.-D.
J. Heterocycl. Chem. 1995, 32, 363.
14. Andrieu, J.; Dietz, J.; Poli, R.; Richard, P. New J. Chem. 1999,
23, 581.
30. Bowden, K.; Last, A. M. J. Chem. Soc., Perkin Trans. 2 1973,
1144.
31. See Ref. 173b), p S43.
15. Carboni, B.; Monnier, L. Tetrahedron 1999, 55, 1197.
16. Ben Aroya Bar-Nir, B.; Portnoy, M. Tetrahedron Lett. 2000,
41, 6143.
32. Yamamoto, I.; Yanayi, S.; Mamba, A.; Gotoh, H. J. Org.
Chem. 1974, 39, 3924.
17. 3a) Kondo, Y.; Komine, T.; Fujinami, M.; Uchiyama, M.;
Sakamoto, T. J. Comb. Chem. 1999, 1, 123±126.