Self-assembly of bidentate ligands for combinatorial homogeneous catalysis based on an A-T base-pair model

Article abstract of DOI:10.1002/anie.200462499

The odd couple: Inspired by the principle of DNA base pairing a conceptually new approach for the generation of a heterobidentate-ligand library based on self-assembly through hydrogen-bonding is realized. From a 4 x 4 library a catalyst that shows outstanding activity and excellent regioselectivity could be identified (see scheme; ^FGR = functional group, Do = donor group). (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200462499

Communications
Homogeneous Catalysis
bidentate ligand, as exemplified in a highly regioselective
hydroformylation of n-alkenes into linear aldehydes.[2]
A potential advantage of this self-assembly approach is
the possibility to form libraries of heterodimeric bidentate
ligands (Scheme 2). Combining m different ligands equipped
with donor functions Dox with n different ligands equipped
with donor functions Doy should result in the formation of a
set of m ꢀ n different bidentate ligands without the need for
additional synthetic steps.[3]
Self-Assembly of Bidentate Ligands for
Combinatorial Homogeneous Catalysis Based on
an A–T Base-Pair Model**
Bernhard Breit* and Wolfgang Seiche
Selectivity control in homogeneous metal complex
catalysis relies in many cases on tailor-made biden-
tate ligands. The quest for the ultimate ligand which
gives a catalyst with optimal activity and selectivity
is difficult. Since rational design still does not allow
the ligand of choice for a given reaction and
substrate to be predicted, the combinatorial syn-
thesis of ligand libraries and their subsequent use
has become an additional strategy.[1] However, the
rate-determining step in catalyst development is in
most cases the time-consuming ligand synthesis
required to generate the library.
Herein we report an alternative approach to
generate bidentate ligand libraries that relies on a
self-assembly of bidentate ligands from monoden-
tate ligands and is based on hydrogen-bonding.
From a 4 ꢀ 4 library of self-assembling ligands a
Scheme 2. An AT base-pair model (highlighted in red) as a platform for the self-
assembly of monomeric to mixed bidentate ligands.
catalyst that shows optimal activity and regioselectivity for
the hydroformylation of terminal alkenes could be identified.
We recently showed that 6-DPPon (1) self-assembles
through hydrogen bonding of the hydroxypyridine 1Awith its
pyridone tautomer 1B in the coordination sphere of a late-
transition-metal center, such as platinum(ii) or rhodium(i)
(Scheme 1). Compound 1 displays the typical behavior of a
Unfortunately, this goal cannot be reached relying on the
tautomeric pyridone/hydroxypyridine platform, since mixing
of two pyridones with different donor functions results in a
statistical mixture of the heterodimeric and the two homodi-
meric ligands.[4] Our goal was to ensure the formation of
single defined catalysts based on heterodimeric bidentate
ligands, since this is a prerequisite for the delineation of
structure–activity and structure–selectivity relations. Hence, a
new template that assembles in a well-defined complemen-
tary mode giving exclusively heterodimers was required.
In fact, self-assembly of two complementary species
through hydrogen bonding is possible with high precision as
exemplified by DNA base pairing, for example, between
adenine and thymine (Scheme 2). The physical basis for this
specific complementary heterodimerization process relies on
the inherent “fixation” of the adenine base as the lactim
tautomeric form, and the thymine base as the lactam
tautomeric form.[5] As a model system emulating these
properties of the AT base pair the aminopyridine 3/isoquino-
lone 4 system was selected. Thus 2-aminopyridines exist in the
lactim form[6] whereas the isoquinolone system strongly
prefers the lactam tautomeric form.[7] As a consequence,
homodimer formation similar to the 6-DPPon system should
be suppressed leaving heterodimer formation (!5) as the
exclusive alternative.
Scheme 1. Self-assembly through hydrogen bonding of the 2-pyridone/
2-hydroxypyridine system 1 to generate bidentate ligand metal com-
plexes 2 for homogeneous catalysis.
[*] Prof. Dr. B. Breit, Dipl.-Chem. W. Seiche
Institut fꢀr Organische Chemie und Biochemie
Albert-Ludwigs-Universitꢁt Freiburg
Synthesis of the phosphine functionalized aminopyridine
derivatives 3a–d is readily accomplished starting from 2,6-
dibromopyridine (Scheme 3). Reaction with aqueous ammo-
nia and protection of the amino function as the pivaloate gave
pyridine 7. The introduction of the pivaloate (Piv) protecting
group proved to be beneficial for clean bromine–lithium
exchange with nBuLi in THF at ꢀ1008C. Diversification was
achieved by trapping the thus generated lithiopyridine with
Albertstrasse 21, 79104 Freiburg (Germany)
Fax: (+49)761-203-8715
E-mail: bernhard.breit@organik.chemie.uni-freiburg.de
[**] This work was supported by the Fonds der Chemischen Industrie,
the Alfried Krupp Award for young university teachers of the Krupp
foundation (to B.B.), and BASF. We thank Dr. M. Keller for the X-ray
crystal structure analysis and N. Stꢂcks and G. Leonhardt-
Lutterbeck for technical assistance.
1640
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200462499
Angew. Chem. Int. Ed. 2005, 44, 1640 –1643

Angewandte
Chemie
Figure 1. PLATON plot of the structure of cis-5aa-PtCl2 in the solid
state (H atoms bound to carbon are omitted for clarity). Selected
interatomic distances [ꢃ] and angles [8]: Pt-P1 2.2486(5), Pt-P2
2.2437(5), NH···N 2.932(2), O···HN 2.977(2); P1-Pt-P2 102.896(18), N-
H···N, 163(2) O···H-N 172(2). Green Pt, yellow Cl, orange P, blue N,
red O.
Scheme 3. Synthesis of monodentate aminopyridinyl phosphines 3 and
isoquinolones 4: 1) NH4OH, 1908C, 5 h; 2) pivaloylchloride, NEt3,
CH2Cl2, 08C!RT; 3) n-BuLi (2 equiv), THF, ꢀ1008C, then ClPR2,
ꢀ1008C!RT; 4) KOtBu (1.1 equiv), toluene, 808C, 1 h; 5) n-BuLi
(1 equiv), THF, ꢀ1008C, then ClPR2, ꢀ1008C!RT, then H2O (1 equiv)
and formic acid excess(50–75%).
different chlorophosphines to furnish the aminopyridine type
ligand set 3a–d.
The complementary isoquinolone system 4 was obtained
starting from 1,3-dibromoisoquinoline (8).[8] Nucleophilic
introduction of a tert-butoxy substituent gave bromide 9
which underwent clean bromine–lithium exchange with
nBuLi. Trapping of the resulting lithioisoquinoline with the
chlorophosphines furnished the isoquinolone ligands 4a–d.
Mixing of one equivalent of 6-diphenylphosphino-N-
pivaloyl-2-aminopyridine (6-DPPAP; 3a) with one equivalent
of 3-diphenylphosphinoisoquinolone (3-DPPICon; 4a) in the
presence of [PtCl2(1,5-cod)] (cod = cyclooctadienyl) gave the
heteroleptic cis-complex 5aa-PtCl2 in quantitative yield.
From the X-ray crystal structure[9] of 5aa-PtCl2 (Figure 1) it
is clear that the two cis-coordinated phosphine ligands form
the expected hydrogen-bonding network reminiscent of the
Watson Crick base pairing of A and T in DNA.
Figure 2. 31P NMR spectrum of 5aa-PtCl2 in CDCl3 solution.
a test reaction (see Table 1). A strong chelate effect on the
regioselectivity of this reaction is well established.[12]
Both monodentate ligands, the aminopyridine 3a as well
as the isoquinolone 4a, furnished active rhodium catalysts for
the hydroformylation of 1-octene (Table 1, entries 1 and 2).
However, as expected for a monodentate phosphine ligand,
regioselectivity for the linear aldehyde was rather low.[2,13] On
mixing ligands 3a and 4a, hydroformylation occurred with a
significantly increased regioselectivity for the linear aldehyde
(entries 3–6). This result indicates that the heterodimeric 3a/
4a ligand system operates as a bidentate ligand 5aa, which is
held together by hydrogen bonding, in the active rhodium(i)
catalyst.
NMR spectra indicate that a similar structure is found in
aprotic solvents, such as CDCl3. The 31P NMR spectrum
2
shows an AB system with a typical J(P,P) coupling of 13 Hz
which confirms the presence of two non-equivalent phosphine
ligands coordinated to the same platinum center.[10] Further-
more, the size of the 1J(P,Pt) coupling (3658 and 3484 Hz) is in
the order expected for a cis-platinum(ii) diphosphine complex
1
(Figure 2).[11] Additionally, H NOE experiments in CDCl3
show a distinct NOE between the amide NH of the pyridine
ligand and the NH function of the isoquinolone unit. This
result suggests that hydrogen bonding occurs in solution
(CDCl3) as well.
To explore whether the heterodimeric chelate bonding
mode that results through hydrogen-bonding of the 3/4 system
is operative throughout a catalytic reaction, the rhodium-
catalyzed hydroformylation of terminal alkenes was chosen as
Hence, the stage was set for the generation of the first self-
assembled bidentate-ligand library based on hydrogen bond-
ing to identify a catalyst that operates with optimal activity
and regioselectivity.
Angew. Chem. Int. Ed. 2005, 44, 1640 –1643
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1641

Communications
Table 1: Results of rhodium-catalyzed hydroformylation of terminal
alkenes with 3a and 4a.
Table 2: 4ꢄ4 ligand matrix of aminopyridine(3a–d)/isoquinolone(4a–d)
derived self-assembled bidentate ligands in the [Rh]-catalyzed hydro-
formylation of 1-octene.[a]
Entry[a]
Ligand
FGR
l:b[b]
Conversion[b]
L(3)
1
2
3
4
5
6
3a
4a
3a+4a
3a+4a
3a+4a
3a+4a
CH3(CH2)5
CH3(CH2)5
CH3(CH2)5
HO(CH2)9
MeO2C(CH2)8
AcO(CH2)4
72:28
76:24
94:6
95:5
94:6
93:7
quant.
quant.
quant.
quant.
quant.
quant.
L(4)
3a
3b
3c
3d
4a
2425 hꢀ1[b]
94:6[c]
1040 hꢀ1
94:6
2732 hꢀ1
96:4
2559 hꢀ1
95:5
4b
4c
4d
2033 hꢀ1
93:7
1058 hꢀ1
92:8
1281 hꢀ1
96:4
1772 hꢀ1
94:6
3537 hꢀ1
94:6
1842 hꢀ1
93:7
1808 hꢀ1
96:4
2287 hꢀ1
94:6
[a] Conditions: [Rh(CO)2(acac)]:ligand:substrate 1:20:1000, 10 bar CO/
H2 (1:1), toluene (c0(alkene)=0.7m), 708C, 20 h. [b] Linear versus
branched (l:b). Determined by GC analysis and or 1H NMR spectroscopy.
7439 hꢀ1
96:4
2695 hꢀ1
95:5
7465 hꢀ1
94:6
8643hꢀ1
96:4
[a] Reaction conditions: [Rh(CO)2(acac)], [Rh]:L(3):L(4):1-octene
=1:10:10:7500, 10 bar CO/H2(1:1), toluene (c0(1-octene)=2.91m),
5 h. Catalyst preformation: 5 bar CO/H2 1:1 , 30 min, RT!808C.
[b] Turnover frequency (TOF) was calculated as (molaldehydes)ꢄ(mol
catalyst)ꢀ1 ꢄ(thꢀ1 ꢀ1 at 20–30% conversion, determined by GC analysis.
)
[c] Regioselectivity: linear to branched determined by GC analysis. The
best TOF and regioselectivity are highlighted in bold.
3a–d had the greatest impact on catalyst performance.
Second, the increasing donor capability of the phosphine
donors of both 3 and 4 decreased catalyst activity; corre-
spondingly, acceptor substituents at the phosphine donors
increased the catalyst activity. The most active catalyst was
the 3d/4d combination. For this catalyst a turnover frequency
of 8653 hꢀ1 was noted which is exceptional for a bidentate
phosphine/rhodium catalyst.[14] Furthermore, an excellent
regioselectivity of 96:4 in favor of n-nonanal was observed.
In conclusion, the first bidentate-phosphine-ligand library
for homogeneous metal-complex catalysis, based on self-
assembly through hydrogen bonding was realized. The basis
for the success of this concept was the use of an AT base-pair
analogous system—the aminopyridine 3/isoquinolone
4
system. The heterodimeric ligands based on this system
operate as bidentate ligands throughout the hydroformylation
of terminal alkenes. From a 4 ꢀ 4 library generated by self-
assembly a catalyst operating with outstanding activity and
regioselectivity upon hydroformylation of terminal alkenes
was identified. Application of this general principle and
related libraries to asymmetric catalysis is a logical step.
Received: November 2, 2004
Published online: February 9, 2005
Keywords: combinatorial chemistry · homogeneous catalysis ·
hydroformylation · rhodium · self-assembly
From four aminopyridyl phosphines 3a–d and four
phosphinoisoquinolone 4a–d a set of 16 bidentate-ligand
combinations was generated by simple mixing of the compo-
nents with the catalyst precursor [Rh(CO)2(acac)] (acac
= 2,4-pentanedione). The resulting catalysts were explored
with respect to their potential to catalyze the hydroformyla-
tion of 1-octene. Table 2 summarizes the results of these
experiments. Some noteworthy trends are apparent. First,
phosphine-substituent modification of the aminopyridines
.
[1] C. Gennari, U. Piarulli, Chem. Rev. 2003, 103, 3071 – 3100.
[2] B. Breit, W. Seiche, J. Am. Chem. Soc. 2003, 125, 6608 – 6609.
[3] For alternative approaches to ligand/catalyst libraries through
self-assembly see: J. M. Takacs, D. S. Reddy, S. A. Moteki, D.
Wu, H. Palencia, J. Am. Chem. Soc. 2004, 126, 4494 – 4495; V. F.
Slagt, M. Rꢁder, P. C. J. Kamer, P. W. N. M. van Leeuwen,
1642
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 1640 –1643

Angewandte
Chemie
J. N. H. Reek, J. Am. Chem. Soc. 2004, 126, 4056 – 4057; V. F.
Slagt, P. W. N. M. van Leeuwen, J. N. H. Reek, Angew. Chem.
2003, 115, 5777 – 5781; Angew. Chem. Int. Ed. 2003, 42, 5619 –
5623; V. F. Slagt, P. W. N. M. van Leeuwen, J. N. H. Reek, Chem.
Commun. 2003, 2474 – 2475; K. Ding, H. Du, Y. Yuan, J. Long,
Chem. Eur. J. 2004, 10, 2872 – 2884.
[4] M. T. Reetz, T. Sell, A. Meiswinkel, G. Mehler, Angew. Chem.
2003, 115, 814 – 817; Angew. Chem. Int. Ed. 2003, 42, 790 – 793.
[5] J. D. Watson, F. H. C. Crick, Nature 1953, 171, 964 – 967; M. D.
Topal, J. R. Fresco, Nature 1976, 263, 285 – 289.
[6] H. I. Abdulla, M. F. El-Bermani, Spectrochim. Acta Part A 2001,
57, 2659 – 2672.
[7] G. Pfister-Guillouzo, C. Guimon, J. Frank, J. Ellison, A. R.
Katritzky, Liebigs Ann. Chem. 1981, 366 – 375.
[8] A. R. Osborn, K. Schofield, L. N. Short, J. Chem. Soc. 1956,
4191 – 4206.
[9] CCDC-254720 contains the supplementary crystallographic data
for this paper. These data can be obtained free of charge from
the Cambridge Crystallographic Data Centre via www.ccdc.ca-
m.ac.uk/data_request/cif.
Crystal
data
for
5aa-PtCl2
C47H49Cl2N3O3P2Pt; Mr = 1031.82; T= 100(2) K; l = 0.71073 ꢂ;
¯
crystal system: triclinic; space group: P1; a = 11.4445(2), b
= 12.6925(3), c = 16.1777(3) ꢂ, a = 92.8636(12),
= 95.4611(11), g = 109.5054(10)8; V= 2196.66(8) ꢂ3; Z = 2;
1calcd = 1.560 Mgmꢀ3 absorption coefficient: 3.433 mmꢀ1
b
;
;
F(000) = 1036; crystal size: 0.3 ꢀ 0.2 ꢀ 0.1 mm; q range for data
collection: 1.71–27.498; limiting indices: ꢀ14 ꢁ h ꢁ 14, ꢀ16 ꢁ k ꢁ
16, ꢀ20 ꢁ l ꢁ 21; 25457 reflections collected, 10036 reflections
unique (Rint = 0.0312); completeness to q = 25.008:99.9%;
absorption correction: semi-empirical from equivalents; max.
and min. transmission: 0.711 and 0.640; refinement method: full-
matrix least-squares on F2; data/restraints/parameters: 10036/0/
536; GoF on F2: 1.076; final R indices [I > 2s(I)]: R1 = 0.0207,
wR2 = 0.0467; R indices (all data): R1 = 0.0239, wR2 = 0.0477;
largest diff. peak and hole: 0.919 and ꢀ1.430 eꢂꢀ3
.
[10] “31P and 13C NMR of Transition Metal Phosphine Complexes”:
P. S. Pregosin, R. W. Kunz in NMR Basic Principles and Progress,
Vol. 16 (Eds.: P. Diehl, E. Fluck, R. Kosfeld), Springer, Heidel-
berg, 1979, pp. 115 – 122, and references therein; P. S. Pregosin,
S. N. Sze, Helv. Chim. Acta 1978, 61, 1848 – 1855.
[11] see “31P and 13C NMR of Transition Metal Phosphine Com-
plexes”: P. S. Pregosin, R. W. Kunz in NMR Basic Principles and
Progress, Vol. 16 (Eds.: P. Diehl, E. Fluck, R. Kosfeld), Springer,
Heidelberg, 1979, pp. 94 – 95; H. G. Alt, R. Baumgartner, H. A.
Brune, Chem. Ber. 1986, 119, 1694 – 1703.
[12] P. W. N. M. van Leeuwen, C. P. Casey, G. T. Whiteker in Rho-
dium Catalyzed Hydroformylation (Eds.: P. W. N. M. van Leeu-
wen, C. Claver), Kluwer Academic Publishers, Dordrecht, 2000,
chap. 4, pp. 76 – 105.
[13] see P. W. N. M. van Leeuwen, C. P. Casey, G. T. Whiteker in
Rhodium Catalyzed Hydroformylation (Eds.: P. W. N. M. van
Leeuwen, C. Claver), Kluwer Academic Publishers, Dordrecht,
2000, chap. 4, pp. 63 – 75; B. Breit, W. Seiche, Synthesis 2001, 1 –
36.
[14] Under identical conditions Rh/triphenylphosphine (TOF:
1312 hꢀ1, 76:24), Rh/6-DPPon (1; TOF: 3284 hꢀ1, 98:2), for
tBu-Xanthphos see ref. [2]. See also ref. [12].
Angew. Chem. Int. Ed. 2005, 44, 1640 –1643
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1643