Highly functionalised sulfur-based silica scavengers for the efficient removal of palladium species from active pharmaceutical ingredients

Article abstract of DOI:10.1021/op7000172

The use of multidentate sulfur-based silica scavengers 1,2, and 3 as highly effective adsorbents for the removal of precious metals, specifically palladium residues in this paper, from highly functionalised synthetic intermediates and APIs is described. The synthesis and purification of the polar and electron-rich reaction products, containing multiple functional groups, from palladium-catalysed removals of commonly used protecting groups such as benzyl, benzyloxycarbonyl, and allyloxycarbonyl and Sonogashira, Suzuki, Heck, and Buchwald-Hartwig coupling reactions is reported. The significant levels of residual palladium species, typically associated with these reaction products, are successfully and rapidly removed to below acceptable regulatory levels, of less than 5 ppm, by simple, unoptimised treatment with the designed silica scavengers at room temperature. Performance aspects, including broad solvent compatibility, excellent stability, and high metal affinity, combined with large-scale availability, ease of handling, and minimal loss of API make these silica scavengers particularly useful to process development groups.

Full text of DOI:10.1021/op7000172

Organic Process Research & Development 2007, 11, 406−413
Highly Functionalised Sulfur-Based Silica Scavengers for the Efficient Removal
of Palladium Species from Active Pharmaceutical Ingredients
Nicola Galaffu, Siud Pui Man, Robin D. Wilkes, and John R. H. Wilson*
PhosphonicS Ltd., 114 Milton Park, Abingdon, Oxfordshire OX14 4SA, England
Abstract:
methods include distillation, extraction, adsorption, and
crystallisation. However, the most commonly utilised meth-
ods for metal removal from APIs often suffer from significant
disadvantages; for example the use of either activated carbon
or recrystallisation can result in significant product loss, and
often with these methodologies very low levels of residual
metal content cannot be achieved.
Functionalised materials are an attractive option for
scavenging residual metals in products or for capturing
catalyst residues from waste streams. One of the key
advantages is that the functionality can be designed to have
a very high affinity for the residual metal.
The structures of most pharmaceutical intermediates and
products contain invariably a number of functional groups
in close proximity. These groups or ligands can bind strongly
to the precious metal catalyst and its residues. Thus, a key
feature in the design of an effective metal scavenger is for
the material to possess ligands that have an overall higher
affinity for the metal compared to the pharmaceutical
product. In addition, given the structural diversity of APIs
and related intermediates, a wide range of multifunctional
and complex metal scavengers are required.
The use of multidentate sulfur-based silica scavengers 1, 2, and
3 as highly effective adsorbents for the removal of precious
metals, specifically palladium residues in this paper, from highly
functionalised synthetic intermediates and APIs is described.
The synthesis and purification of the polar and electron-rich
reaction products, containing multiple functional groups, from
palladium-catalysed removals of commonly used protecting
groups such as benzyl, benzyloxycarbonyl, and allyloxycarbonyl
and Sonogashira, Suzuki, Heck, and Buchwald-Hartwig cou-
pling reactions is reported. The significant levels of residual
palladium species, typically associated with these reaction
products, are successfully and rapidly removed to below
acceptable regulatory levels, of less than 5 ppm, by simple,
unoptimised treatment with the designed silica scavengers at
room temperature. Performance aspects, including broad
solvent compatibility, excellent stability, and high metal affinity,
combined with large-scale availability, ease of handling, and
minimal loss of API make these silica scavengers particularly
useful to process development groups.
Introduction
In this contribution we would like to report the use of a
number of multifunctionalised materials for the removal of
palladium residues from the products of a selection of
reactions widely used in medicinal chemistry. These reactions
are commonly used in the synthesis of compounds that
contain typical functional groups found in most APIs. Upon
completion of a metal-catalysed reaction, the metal residue
contained in the reaction product can exist in a variety of
oxidation states, such as Pd (0) and Pd (II) in these cases,
and to be effective, functionalised materials need the ability
and inherent functionality to remove metals in these various
oxidation states from the reaction product. The silica
scavengers described here are able to achieve this for a wide
range of common palladium-catalysed processes.
In this study we have used materials based on a silica
framework, which offers both operational and performance
advantages, such as no swelling requirement prior to use,
broad solvent compatibility, and excellent stability properties
(thermal, mechanical, and chemical). The broad solvent
compatibility means existing process streams can be directly
treated with the supported materials, allowing costly solvent
switches to be avoided by process development groups.
Customisation of the properties of the silica, including
particle size and shape, active site loading, accessible surface
area, and pore size, is readily achieved, allowing the silica
to be specifically modified for the desired purpose.
Precious metal catalysts are extensively used to generate
a wide range of products across a variety of industries.
Typical precious metal-promoted reactions include carbon-
carbon bond formation, carbon-nitrogen bond formation,
deprotection reactions, and hydrogenation. In medicinal
chemistry, palladium is perhaps the most-widely utilised
precious metal, being used for reactions such as Suzuki
couplings and Buchwald-Hartwig aminations, which rep-
resent key transformations towards the synthesis of new
active pharmaceutical ingredients (APIs).1 This is under-
standable, given that these reactions provide an easy access,
and in high yield, to complex molecules that previously could
only be achieved through multistep synthesis. The disad-
vantage of this metal-catalysed chemistry is that expensive
and toxic metal residues are invariably left bound to the
desired product. For pharmaceutical intermediates and
products there have been a number of reviews on methods
for the removal of precious metals from products.2 These
* Author to whom correspondence should be addressed. E-mail : jwilson@
phosphonics.com.
(1) Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T. Org. Biomol. Chem.
2006, 4, 2337.
(2) Garrett, C. E.; Prasad, K. AdV. Synth. Catal. 2004, 346, 889; Welch, C. J.;
Albaneze-Walker, J.; Leonard, W. R.; Biba, M.; DaSilva, J.; Henderson,
D.; Laing, B.; Mathre, D. J.; Spencer, S.; Bu, X.; Wang, T. Org. Process
Res. DeV. 2005, 9, 198; Bien, J. T.; Lane, G. C.; Oberholzer, M. R. Top.
Organomet. Chem. 2004, 6, 263.
406
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development
10.1021/op7000172 CCC: $37.00 © 2007 American Chemical Society
Published on Web 04/28/2007

Scheme 1. Deprotection of amino acids and scavenging of
residual palladium
synthesis. Among the most commonly used metal catalysts,
those based on palladium are routinely employed for the
removal of protecting groups and for cross-coupling reac-
tions, as well as having application for a number of other
synthetic operations such as carbon monoxide insertions and
Wacker-type oxidations.4
Figure 1. Multidentate sulfur-based silica scavengers.
These functionalised silicas have molar loadings of
typically 1.0 mmol/g, although the multiple functional groups
present within scavengers such as 3 result in higher effective
loadings. Whilst scavengers based on alternative materials
such as polystyrene (typically 2 to 3 mmol/g) and grafted
fibres may have higher nominal molar loadings, the affinity
of the functionality, the spacial arrangement of functional
groups, and access of the API to the scavenging groups all
contribute significantly to the overall effectiveness of the
scavenger in use.
The functionalised silica scavengers are prepared by a
highly scaleable process in which a range of trialkoxysilyl
compounds are employed as key components, allowing
variation of both the spacer to which scavenging functionality
is attached and the range and combination of functional
groups which can be attached. The inorganic support may
also be varied, and alumina and silicones may be employed
as well as silica described herein. The use of these materials
for process-scale applications has been demonstrated to date
by multikilogram slurry batch processing. The development
and introduction of large-scale cartridges for use in process
development is ongoing.
Sulfur-based ligands are widely known to have a high
affinity for precious metals. As part of our investigations
into the multidentate mechanism we made a number of
functionalised materials, as illustrated in Figure 1. The
synthesis of these functionalised materials is reported
elsewhere.3 In this paper we report our results with the
functionalised silica scavengers 1, 2, and 3. Future publica-
tions will report scavenging progress with the additional
materials shown in Figure 1.
The effectiveness of the designed multidentate scavenging
materials is most challenged, and therefore best demonstrated,
by the removal of residual palladium from highly functiona-
lised organic molecules containing a variety of polar
moieties, each capable of strongly binding palladium residues
and making the metal removal to regulatory-acceptable levels
extremely difficult. APIs and their synthetic precursors
normally contain such diverse polar functional groups, often
imparted as an inherent function of commonly occurring
monomers such as amino acids and cyclic secondary amines.
Amino Acid Deprotections. Orthogonally protected
amino acids are vital intermediates for the synthesis of
complex molecules including many natural products. Selec-
tive deprotections of these key amino acid building blocks
allow required functionality to be revealed at the appropriate
point of a chemical synthesis. Amongst the most common
protecting groups for amino and hydroxyl moieties are
allyloxycarbonyl (Alloc), benzyloxycarbonyl (Cbz), and
benzyl (Bn), which in most cases are all efficiently removed
by palladium-catalysed hydrogenation.5
Scheme 1 depicts the removal of protecting groups from
orthogonally protected serine 4 and lysine 6 derivatives using
palladium-catalysed hydrogenations. In both cases the re-
sidual level of palladium associated with the deprotected
product was analysed after standard reaction work-up, prior
to scavenging and after scavenging by passing a methanolic
solution of either 5 or 7 through a plug of 3-mercaptopropyl
(3) Wilson, J. R. H.; Sullivan, A. C.; Man, S. Substituted Organopolysiloxanes
and Use Thereof. WO2006013060, 2006.
(4) Corbet, J.-P.; Mignani, G. Chem. ReV. 2006, 106, 2651; Hassan, J.; Se´vignon,
M.; Gozzi, C.; Schulz, E.; Lemaire, M. Chem. ReV. 2002, 102, 1359; Tsuji,
J. Palladium Reagents and Catalysts: New PerspectiVes for the 21st
Century, 2nd ed.; John Wiley & Sons: New York, 2004.
Results and Discussion
Recent surveys of the pharmaceutical literature indicate
that a very high percentage of small-molecule therapeutics
now feature at least one metal-catalysed reaction during their
(5) Wuts, P. G. M.; Greene, T. W. Greene’s ProtectiVe Groups in Organic
Synthesis, 4th ed.; Wiley: New York, 2006.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
407

Table 1. Palladium content of products 5, 7, and 13 before
and after scavenging
sation to be just 7 ppm (Table 1). Residual tin levels in 13
after the deprotection reaction were found to be extremely
high, as tributyltin hydride was used in large excess. In
addition to removal of residual palladium to 7 ppm, scaveng-
ing with 2 resulted in removal of >99% of the tin residues
present.
The residual palladium content of deprotection reactions
of commonly used synthetic intermediates such as amino
acids and aminopiperidines, which produce polar, functiona-
lised low-molecular weight molecules, was found to be high
(>200 ppm). By use of a simple scavenging protocol
employing multidentate scavenger 3-mercaptopropyl ethyl
sulfide silica 2 in a cartridge format, levels of palladium were
dramatically reduced to <1 ppm for both amino acid
derivatives, and to 7 ppm for the aminopiperidine. Additional
scavenging efficiency can be obtained by the use of increased
contact time, temperature, concentration, and/or equivalents
of scavenger, although these factors were not investigated
for these examples.
palladium content/ppm
reaction product before scavenging after scavenging with 2
5
7
13
300
200
1000
<1a
<1a
7
a Indicates palladium content below the detection limits of the instrument (5
ppb).
Scheme 2. Synthesis of amino-piperidine 13 and scavenging
of residual palladium
There are many variants of palladium-catalysed carbon-
carbon-coupling reactions described in the literature but the
Suzuki, Stille, Sonogashira, and Heck reactions have received
the most significant synthetic attention. One of the major
benefits provided by cross-coupling technologies is the
increased convergence when compared to classical methods,
as two highly functionalised molecules can be coupled
together under relatively mild conditions. As a result,
coupling reactions are now likely to feature at any stage in
the synthesis of an API, even increasingly as the final
synthetic step.
Regardless of the point of use within the synthetic
sequence, residual palladium, which can encompass the
catalyst itself and associated palladium species generated
from the catalyst under the reaction conditions, needs to be
minimised in the API. Adherence to the increasingly stringent
regulatory requirements for metal content within APIs is
highly desirable in order to facilitate progression of the API
down the drug development pathway. However, removal of
residual palladium from carbon-carbon- and carbon-
nitrogen-coupled products can frequently be a complex issue.
Products from a number of the common carbon-carbon
coupling reactions were investigated for palladium content
before and after treatment with the multidentate scavengers
described.
Carbon-Carbon Couplings: Sonogashira Reaction.7
Compound 17 is related to a calcium entry blocker,8 which
is known to possess a strong ability for chelation of
palladium. Formation of the biaryl acetylene by Sonogashira
reaction at a late stage of the synthetic sequence towards
the drug substance caused significant palladium-removal
problems, and the synthetic sequence required reorganisation.
We investigated palladium removal from our model com-
ethyl sulfide silica 2 under gravity (over ∼5 min). In both
cases, the palladium content after scavenging was found to
be <1 ppm (Table 1).
Aminopiperidine Deprotections. Piperidine fragments
are ubiquitously found in drug molecules, particularly those
targeting G-protein-coupled receptors (GPCRs). The syn-
thesis of a protected aminopiperidine was achieved as shown
in Scheme 2. N-Protection of piperidine derivative 8 was
achieved with allyl chloroformate to give 9.6 Deprotection
of the ketal was achieved very cleanly and in quantitative
yield upon treatment of 9 with a slurry suspension of
PhosphonicS’ heterogeneous phosphonic acid 10. Reductive
amination of the resultant keto-piperidine 11 with phenethyl-
amine gave the N1-protected 4-aminopiperidine 12, which
was purified by column chromatography and subsequently
treated with bis(triphenylphosphine)palladium (II) dichloride
in the presence of tributyltin hydride to furnish the N-
deprotected aminopiperidine 13.
Again, the residual level of palladium associated with the
deprotected product was analysed after standard reaction
work-up prior to and after scavenging by passing a metha-
nolic solution of 13 through a plug of 3-mercaptopropyl ethyl
sulfide silica 2 under gravity (over ∼5 min). The palladium
content of 13 after scavenging was found without optimi-
(7) Takahashi, S.; Kuroyama, Y.; Sonogashira, K.; Hagihara, N. Synthesis 1980,
627; Sonogashira, K. In Metal-Catalyzed Cross-Coupling Reactions, 1st
ed.; Diederich, F., Stang, P. J., Eds.; Wiley-VCH: New York, 1998; p 203;
Negishi, E.; Anastasia, L. Chem. ReV. 2003, 103, 1979.
(8) Maryanoff, C. A.; Mills, J. E.; Stanzione, R. C., Jr.; Hortenstine, J. T. In
Catalysis of Organic Reactions, 1st ed.; Rylander, P. N., Greenfield, H.,
Augustine, R. L., Eds.; Marcel Dekker: New York, 1988; Chapter 18, p
368.
(6) Ferna´ndez-Forner, D.; Casals, G.; Navarro, E.; Ryder, H.; Albericio, F.
Tetrahedron Lett. 2001, 42, 4471.
408
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

Scheme 3. Sonogashira coupling of 16 with phenylacetylene
and scavenging of residual palladium from 17
Scheme 4. Synthesis of Valsartan precursor 21 using a
Suzuki coupling
Table 2. Palladium content of coupling products 17, 21, 31,
and 34 before and after scavenging
palladium content/ppm
drugs which contain the privileged biaryl moiety.12
A
type of
before
after
potential route into Valsartan precursors involves a key
Suzuki coupling, as shown in Scheme 4. Valine methyl ester
hydrochloride 18 was reductively alkylated with 4-formylphen-
ylboronic acid. The resultant secondary amine 19 was
acylated with valeroyl chloride to give amido aryl boronic
acid 20, which underwent Suzuki coupling under typical
literature conditions to give the biaryl amide 21, a precursor
of Valsartan. The residual level of palladium associated with
21 was analysed separately with three different multidentate
silica scavengers 1, 2, and 3 in methanol both prior to and
after scavenging. The palladium content of 21 after scaveng-
ing was found to be 1.6 ppm with scavenger 1 and <1 ppm
with either scavenger 2 or 3 (Table 2).
coupling
product scavenger scavenging scavenging
Sonogashira
Suzuki
Suzuki
Suzuki
Heck
Heck
Heck
Heck
17
21
21
21
30
30
30
30
33
33
33
2
1
2
3
1
2
3
700
2100
2100
2100
307
307
307
307
30
5.0
1.6
<1a
<1a
18.1
24.3
62.2
4.8
1 + 2 + 3b
1
2
3
Buchwald
Buchwald
Buchwald
<1a
<1a
<1a
30
30
a Indicates palladium content below the detection limits of the instrument (5
ppb). b Equimolar amounts of 1, 2, and 3 were combined.
Carbon-Carbon Couplings: Heck Reaction. Amongst
potential routes to the oncology candidate CP-724,714, a
selective ErbB2 angiogenesis inhibitor, routes involving
pound 17, which employed the key Sonogashira reaction as
the final step of the synthesis, Scheme 3.
Compound 16 was readily accessible from commercially
available building blocks; however, a highlight of the
synthesis was the facile purification of amide 16, by passing
the crude reaction mixture through a cartridge of Phospho-
nicS’ supported phosphonic acid 10. Amide 16 was suc-
cessfully coupled with phenylacetylene under Sonogashira
conditions to provide the aryl-aryl acetylene 17 in good
yield. After standard reaction work-up, the residual level of
palladium associated with 17 was analysed both prior to and
after scavenging with 3-mercaptopropyl ethyl sulfide silica
2 in methanol. The palladium content of 17 after scavenging
was found to be just 5 ppm (Table 2).
Carbon-Carbon Couplings: Suzuki Reaction. The
Suzuki reaction9 is perhaps the most widely utilised of all
palladium-catalysed cross-coupling reactions by process
development chemistry groups, being routinely performed
on multikilogram scale.10 Valsartan,11 a potent, orally active
angiotensin II antagonist is among a number of marketed
(9) Suzuki, A. J. Organomet. Chem. 1999, 576, 147; Miyaura, N.; Suzuki, A.
J. Chem. Soc., Chem. Commun. 1979, 866; Suzuki, A. Pure Appl. Chem.
1991, 63, 419; Suzuki, A. In Metal-Catalyzed Cross-Coupling Reactions,
1st ed.; Diederich, F., Stang, P. J., Eds.; Wiley-VCH: New York, 1998;
Chapter 3; Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457.
(10) Examples include: Ennis, D.; McManus, J.; Wood-Kaczmar, W.; Richard-
son, J.; Smith, G.; Carstairs, A. Org. Process Res. DeV. 1999, 3, 248; Jacks,
T. E.; Belmont, D. T.; Briggs, C. A.; Horne, N. M.; Kanter, G. D.; Karrick,
G. L.; Krikke, J. J.; McCabe, R. J.; Mustakis, J. G.; Nanninga, T. N.;
Risedorph, G. S.; Seamans, R. E.; Skeean, R.; Winkle, D. D.; Zennie, T.
M. Org. Process Res. DeV. 2004, 8, 201; Conlon, D. A.; Drahus-Paone,
A.; Ho, G.-J.; Pipik, B.; Helmy, R.; McNamara, J. M.; Shi, Y.-J.; Williams,
J. M.; Macdonald, D.; Deschenes, D.; Gallant, M.; Mastracchio, A.; Roy,
B.; Scheigetz, J. Org. Process Res. DeV. 2006, 10, 36; Kerdesky, F. A. J.;
Leanna, M. R.; Zhang, J.; Li, W.; Lallaman, J. E.; Ji, J.; Morton, H. E.
Org. Process Res. DeV. 2006, 10, 512.
(11) De Gasparo, M.; Webb, R. L. WO0002543, 2000; Bu¨hlmayer, P.; Furet,
P.; Criscione, L.; de Gasparo, M.; Whitebread, S.; Schmidlin, T.; Lattmann,
R.; Wood, J. Bioorg. Med. Chem. Lett. 1994, 4, 29.
(12) Evans, B. E.; Rittle, K. E.; Bock, M. G.; Dipardo, R. M.; Freidinger, R.
M.; Whiter, W. L.; Lundell, G. F.; Veber, D. F.; Anderson, P. S.; Chang,
R. S. L.; Lotti, V. J.; Cerino, D. J.; Chen, T. B.; Kling, P. J.; Kunkel, K.
A.; Springer, J. P.; Hirshfield, J. J. Med. Chem. 1988, 31, 2235; Patchett,
A.; Nargund, R. P. Annu. Rep. Med. Chem. 2000, 35, 289; Horton, D. A.;
Bourne, G. T.; Smythe, M. L. Chem. ReV. 2003, 103, 893.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
409

Scheme 5. Heck coupling of iodoquinazoline 28 with N-allyl-2-methoxyacetamide 29 and scavenging of residual palladium
from 30
Sonogashira, Suzuki, and Heck couplings were all assessed.13
A convergent synthesis of CP-724,714 featuring a key Heck14
coupling as the final step was selected as the process route,
as shown in Scheme 5.
treatment of the Heck reaction product under slurry condi-
tions with a 1:1:1 mixture of the three scavengers 1, 2, and
3, provided product 30 in which the palladium content was
reduced to below API acceptable palladium levels.
Even after a number of additional synthetic steps from
the point of use, a significant level of tin (1717 ppm) was
found to be present in quinazoline 30. In this case, scaveng-
ing of residual tin was again relatively effective, given the
scavengers were not optimised; for example, scavenger 3
removed 90% of the residual tin content.
Carbon-Heteroatom Couplings: Buchwald-Hartwig
Reaction. Carbon-nitrogen bond formation under pal-
ladium-catalysed conditions15 is one of the most rapidly
advancing areas of organic synthesis, and recently these
Buchwald-Hartwig reactions have been scaled-up (multi-
kilogram; 250 L).16 Compound 33 is a precursor of the
antioxidant component 4-aminodiphenylamine,17 which can
be prepared by amination reaction between aniline 32 and
the aryl chloride 31, as shown in Scheme 6. After work-up
the residual level of palladium associated with 33 was
analysed prior to and after scavenging, separately with three
different multidentate silica scavengers 1, 2, and 3 in
methanol. The palladium content of 33 after scavenging was
found to be <1 ppm, using either scavenger 1, 2, or 3 (Table
2).
SNAr reaction between pyridinol 22 and 2-chloro-5-
nitrotoluene generated biaryl ether 23, the nitro group of
which was reduced using tin in hydrochloric acid. Concur-
rently, quinazolinol 26 was prepared from iodo-anthranilic
acid 25 and formamidine acetate. Chlorination of the resultant
quinazolinol was achieved using phosphoroyl chloride. SN-
Ar reaction of chloroquinazoline 27 with aniline 24 pro-
ceeded well in isopropanol at reflux to provide anilino-
quinazoline biaryl ether 28, which was coupled with N-allyl-
2-methoxyacetamide 29 (prepared from allylamine and
methoxyacetyl chloride in triethylamine) under typical Heck
conditions to give the quinazoline final product CP-724,714,
30. After work-up, the residual level of palladium associated
with 30 was analysed prior to, and after, scavenging.
Scavenging of this Heck-derived product with each of the
functionalised scavengers separately, using the method of
gravity elution through a cartridge, gave a product 30 with
significantly reduced levels of residual palladium (Table 2)
but which still exceeded the maximum regulatory-allowed
level for palladium content of 5 ppm. Gratifyingly, longer
(13) Brown Ripin, D. H.; Bourassa, D. E.; Brandt, T.; Castaldi, M. J.; Frost, H.
N.; Hawkins, J.; Johnson, P. J.; Massett, S. S.; Neumann, K.; Phillips, J.;
Raggon, J. W.; Rose, P. R.; Rutherford, J. L.; Sitter, B.; Stewart, A. M.,
III; Vetelino, M. G.; Wei, L. Org. Process Res. DeV. 2005, 9, 440.
(14) Heck, R. F. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming,
I., Eds.; Pergamon: New York, 1991; Vol. 4, Chapter 4.3; Bra¨se, S.; de
Meijere, A. In Metal-Catalyzed Cross-Coupling Reactions, 2nd ed.; de
Meijere, A., Diederich, F., Eds.; Wiley-VCH: New York, 2004; Chapter
5, page 217; Heck, R. Org. React. 1982, 27, 345; Chung, J. Y. L.; Cvetovich,
R. J.; McLaughlin, M.; Amato, J.; Tsay, F.-R.; Jensen, M.; Weissman, S.;
Zewge, D. J. Org. Chem. 2006, 71, 8602.
(15) Jiang, L.; Buchwald, S. L. In Metal-Catalyzed Cross-Coupling Reactions,
2nd ed.; de Meijere, A., Diederich, F., Eds.; Wiley-VCH: New York, 2004;
Chapter 13, page 699; Schlummer, B.; Scholz, U. AdV. Synth. Catal. 2004,
346, 1599.
(16) Mauger, C. C.; Mignani, G. A. Org. Process Res. DeV. 2004, 8, 1065;
Mauger, C. C.; Mignani, G. A. Aldrichimica Acta 2006, 39, 17; Buchwald,
S. L.; Mauger, C.; Mignani, G.; Scholz, U. AdV. Synth. Catal. 2006, 348,
23.
(17) Giera, H.; Lange, W.; Pohl, T.; Sicheneder, A.; Schild, C. Patent
DE19942394, 2001.
410
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

Scheme 6. Synthesis of antioxidant precursor 33 by
Buchwald-Hartwig coupling and scavenging of residual
palladium
range of solvents (avoiding the need for an often time-
consuming and costly change of process stream), and display
exceptionally fast kinetics even at ambient temperatures. The
materials are very easy to use in either cartridge or slurry
format. Recovery of the precious metal used, often an
economic necessity on process chemistry scales, is also
possible by use of a simple scavenger wash protocol, which
will be described in due course, or via an incineration
process.
The high performance of designed multidentate scaven-
gers and their capability to remove residual palladium to
below regulatory-acceptable levels has been demonstrated
by the effective removal of residual palladium from druglike
products of deprotection, Sonogashira, Suzuki, Heck, and
Buchwald-Hartwig reactions.
The residual palladium content of cross-coupling reaction
products was found to be variable but typically high (up to
>2000 ppm). Use of one of the multidentate silica scavengers
described was typically effective, under simple, nonoptimised
conditions, for the reduction of palladium content to well-
below acceptable levels for drug submission purposes.18 The
synthesis of tyrosine kinase inhibitor 30 with a Heck coupling
as the final step yielded a product containing many hydrogen-
bond donors and acceptors. The electron-rich, highly polar
properties of 30 were illustrated by its propensity for
palladium binding. Scavenging to below regulatory levels
proved difficult, even with the efficient multidentate scav-
engers employed, and necessitated the use of a scavenger
mixture and prolonged scavenger-API contact time to
achieve the key target palladium content of <5 ppm.
Experimental Section
Reagents and solvents were purchased from commercial
suppliers and used without further purification. LC-ES/MS
analyses were carried out on a Shimadzu LC/MS 2010EV
system equipped with an Atlantis C18-column 2.1 mm ×
50 mm. NMR spectra were recorded on a Bruker 250 MHz
DPX NMR running XwinNMR version 3.5 with a 5 mm
standard QNP probe spectrometer in the solvents indicated
at 298 K. Chemical shifts are reported on the δ scale in ppm
and referenced to residual solvent resonances. ICP-OES
analyses were conducted using a Spectro Genesis system.
Whenever possible, the identity of the products was estab-
lished by comparison of the spectral data with literature
precedents or by direct comparison with commercial samples.
Multidentate Silica Scavengers 1, 2, and 3. See ref 3
for the preparation of a range of multidentate silica scaven-
gers including 1, 2, and 3. The molar loading of the
scavengers is typically 1.0 mmol/g.
General Scavenging Procedure for Removal of Pal-
ladium Residues from Reaction Products/APIs. The
reaction product (typically 1 g) was dissolved in the
minimum amount of methanol, and any remaining insoluble
particles were removed by filtration. The solution was made
up to 100 mL with methanol in a volumetric flask. This
standard reaction product/API solution (10 mL, typically
containing 100 mg of API) was dropped through a 10 mL
cartridge containing 1 g of the appropriate palladium
scavenger 1, 2, or 3, which had been preconditioned with 5
mL of methanol, maintaining a flow rate of ∼1 mL per
minute. Further portions of methanol (3 × 5 mL) were passed
through the cartridge. The organic fractions were combined
and evaporated under reduced pressure. A wide variety of
other solvents such as ethyl acetate, dichloromethane,
toluene, ethers, and water may be used as the scavenging
solvents, depending on the solubility of the reaction product
to be scavenged.
Conclusions
Palladium catalysts have become an indispensable tool
to the synthetic organic chemist, being necessary for the
performance of a variety of key reactions such as deprotec-
tions and cross-couplings en route to APIs. Whilst the use
of palladium catalysts allows significantly quicker routes to
the desired drug molecule, the downside is the very strong
likelihood of high levels of palladium contamination of the
molecule, due to the ability of the polar groups usually
present within APIs to bind palladium effectively. As with
all metals, stringent regulatory guidelines exist for the amount
of residual palladium that a drug candidate is allowed to
contain.
PhosphonicS has designed a portfolio of functionalised
silica scavengers which, amongst other applications, show
enhanced efficiency for the removal of a variety of transition,
heavy, and precious metals (including palladium) from APIs
which themselves contain multiple, diverse polar groups,
including hydroxyl, amino, carboxylic acids, nitriles, and
amino acids. Scavenger designs incorporate multidentate
functional groups, combinations of which provide different
scavenging mechanisms and allow different palladium spe-
cies, in different oxidation states, to be bound with suf-
ficiently high affinity that any binding to the API is
effectively overcome. The result is that the final level of
palladium associated with the API is removed to below
regulatory limits, with minimal loss of the valuable API itself.
These silica scavengers show excellent stability (thermal,
physical, chemical, and mechanical), are amenable to a broad
The scavenging conditions described in this paper are
unoptimised. Scavenging efficiency is affected by a number
of experimental factors, including contact time, temperature,
and concentration, as well as the nature of the structural
groups present within the functionalised silica scavenger. The
importance of each of these factors is determined by the
(18) See U.S. Food and Drug Administration (www.fda.gov) and European
Agency for the Evaluation of Medicinal Products (www.emea.eu.int). At
the time of this writing, for the palladium group of metals, the regulatory
guidance indicates acceptable residue levels of <5 ppm for oral drug
products and <0.5 ppm for parenteral drug products.
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
411

nature of the API, and optimisation studies are typically
required for each metal removal project.
7.12 (m, 5H), 5.85 (tdd, J ) 17.2, 10.4, 5.9 Hz, 1H), 5.21
(qd, J ) 17.2, 1.5 Hz, 1H), 5.12 (qd, J ) 10.4, 1.4 Hz, 1H),
4.50 (td, J ) 5.4, 1.4, 2H), 2.94-2.67 (m, 8H), 2.63-2.50
(m 1H), 1.83-1.73 (m, 2H), 1.27-1.00 (m, 2H); MS
(ES+): m/z ) 289 [M + H]+.
N-Boc-L-serine (5). N-R-Boc-O-Benzyl-L-serine 4 (1 g,
3.38 mmol) was dissolved in 9:1 ethanol/cyclohexene (10
mL). Palladium hydroxide on carbon (20 wt %) (0.119 g,
0.17 mmol) was added, and the suspension was refluxed for
16 h. The heterogeneous mixture was filtered and the product
solution evaporated under reduced pressure. The residue was
purified using silica scavenger 2. The resultant colourless
solution of 5 was evaporated under reduced pressure
(quantitative yield). MS (ES+): m/z ) 206 [M + H]+.
N-Boc-L-lysine (7). N-R-Boc-N-ꢀ-Cbz-L-lysine 6 (1 g,
2.62 mmol) was dissolved in ethyl acetate (20 mL), and
palladium hydroxide on carbon (20 wt %) (0.092 g, 0.131
mmol) was added. The heterogeneous suspension was stirred
under a balloon of hydrogen at rt for 2 h, after which time
the suspension was filtered and the filtrate evaporated under
reduced pressure. The residue was purified using silica
scavenger 2. The resultant colorless solution of 7 was
evaporated under reduced pressure (quantitative yield). MS
(ES+): m/z ) 247 [M + H]+.
N-Phenethylpiperidin-4-amine (13). Allyl 4-(phenethy-
lamino)piperidine-1-carboxylate 12 (0.23 g, 0.8 mmol) was
dissolved in dichloromethane (5 mL), and PdCl2(PPh3)2 (0.02
g, 0.026 mmol) and acetic acid (0.120 g, 2 mmol) were
added. The solution was stirred at rt under nitrogen for 5
min and tributyltin hydride (2.4 mL, 8.9 mmol) was added
dropwise over 5 min. The reaction mixture was stirred for
an additional 2 h at rt. The solution was washed with 2 M
aqueous potassium carbonate solution, dried over MgSO4
and the solvent was evaporated under reduced pressure. The
residue was purified using silica scavenger 2. The resultant
colourless solution of 13 was evaporated under reduced
1
pressure (0.154 g, 75%). H NMR (250 MHz, CDCl3): δ
7.24-7.11 (m, 5H), 3.11-2.44 (m, 9H), 1.89-1.78 (m, 2H),
1.29-1.13 (m, 2H); MS (ES+): m/z ) 205 [M + H]+.
3-(2-Bromophenyl)propanoyl Chloride (15). 3-(2-Bro-
mophenyl)propanoic acid 14 (2.29 g, 10 mmol) was dis-
solved in dichloromethane (5 mL), and thionyl chloride (10
mL, 137 mmol) was added dropwise at rt over 15 min. The
mixture was refluxed for 2 h, cooled, and stirred at room
temperature for a further 4 h. The excess of thionyl chloride
was removed under reduced pressure to give 15 (2.29 g,
99%). 1H NMR (250 MHz, CDCl3): δ 7.45 (d, J ) 7.8 Hz,
1H), 7.16 (d, J ) 4.3 Hz, 2H), 7.08-6.96 (m, 1H), 3.18-
3.11 (m, 2H), 3.05-2.99 (m, 2H).
N-Alloc-4-piperidone Ethylene Acetal (9). 4-Piperidone
ethylene acetal 8 (0.501 g, 3.5 mmol) and potassium
carbonate (0.725 g, 5.2 mmol) were dissolved in deionized
water (10 mL). Allyl chloroformate (0.632 g, 1.5 mmol) was
added dropwise over 15 min at 0 °C. The reaction mixture
was stirred at 0 °C for 1 h, allowed to warm slowly to rt,
and stirred for a further 2 h. The solution was extracted with
ethyl acetate, washed with 2 M aqueous potassium carbonate
solution and water, and dried over magnesium sulfate. The
solvent was evaporated under reduced pressure to give 9 (1.2
3-(2-Bromophenyl)-N-phenethylpropanamide (16). 3-(2-
Bromophenyl)propanoyl chloride 15 (2.28 g, 9.23 mmol),
was dissolved in dichloromethane (20 mL), cooled to 0 °C
and phenethylamine (5.8 mL, 30 mmol) was added dropwise
over 15 min. The reaction mixture was warmed to rt and
stirred for an additional 4 h. The solution was passed into a
plug of heterogeneous phosphonic acid scavenger 10 (1.5
g) in order to remove the excess amine. The resultant solution
was evaporated under reduced pressure to give 16 (2.8 g,
1
g, 90%). H NMR (250 MHz, CDCl3): δ 5.96 (tdd, J )
17.2, 10.5, 5.5 Hz, 1H), 5.27 (qdd, J ) 18.1, 10.4, 1.5 Hz,
2H), 4.61 (td, J ) 5.5, 1.4 Hz, 2H), 3.99 (s, 4H), 3.63-3.58
(m, 4H), 1.79-1.68 (m, 4H); MS (ES+): m/z ) 228 [M +
H]+.
N-Alloc-4-piperidone (11). N-Alloc-4-piperidone ethyl-
ene acetal 9 (0.7 g, 3.3 mmol) was dissolved in acetone (10
mL) and water (0.5 mL). Heterogeneous phosphonic acid
10 was added, and the slurry was stirred at reflux for 6 h.
The slurry was filtered through a sintered funnel, and the
clear filtrate was evaporated to dryness under reduced
pressure to give 11 (0.6 g, quantitative yield). 1H NMR (250
MHz, CDCl3): δ 5.99 (tdd, J ) 17.2, 10.4, 5.6, Hz, 1H),
5.31 (qdd, J ) 16.2, 10.4, 1.4 Hz, 2H), 4.67 (td, J ) 5.6,
1.4 Hz, 2H), 3.82 (t, J ) 6.3 Hz, 4H), 2.50 (t, J ) 6.3 Hz,
4H); MS (ES+): m/z ) 184 [M + H]+.
1
94%). H NMR (250 MHz, CDCl3): δ 7.56 (dd, J ) 7.92,
1.05 Hz, 1H), 7.32-7.24 (m, 5H), 7.15-7.09 (m, 3H), 3.55-
3.47 (m, 2H), 3.10 (t, J ) 7.5 Hz, 2H), 2.77 (t, J ) 6.90 Hz,
2H), 2.46 (t, J ) 7.5 Hz, 2H); MS (ES+): m/z ) 332/334
[M + H]+.
N-Phenethyl-3-(2-(phenylethynyl)phenyl)propana-
mide (17). 3-(2-Bromophenyl)-N-phenethylpropanamide 16
(0.51 g, 1.53 mmol) was dissolved in triethylamine (15 mL)
and DME (15 mL). PdCl2(PPh3)2 (0.035 g, 0.046 mmol),
copper (I) iodide (0.02 g, 0.1 mmol), and phenylacetylene
(0.17 mL, 1.5 mmol) were added, and the solution was heated
at 120 °C for 16 h with stirring. The reaction mixture was
evaporated under reduced pressure, and the residue was
dissolved in ethyl acetate and filtered. The filtrate was
washed with saturated aqueous brine solution and dried over
MgSO4. The residue was purified using silica scavenger 2,
at which point the palladium content was measured by ICP-
OES. Further purification by column chromatography using
gradient elution from 0.5% to 60% ethyl acetate/hexane gave
Allyl 4-(phenethylamino)piperidine-1-carboxylate (12).
N-Alloc-4-piperidone 11 (0.51 g 2.8 mmol) was dissolved
in methanol (5 mL); phenethylamine (0.4 mL, 3 mmol) and
glacial acetic acid (5 drops) were added. The reaction mixture
was stirred at rt for 1 h. Sodium borohydride (0.113 mg, 3
mmol) dissolved in methanol (5 mL) was added dropwise
over 5 min. The reaction mixture was stirred for an additional
2 h at rt, and the solvent was evaporated under reduced
pressure. The resultant oil was purified by column chroma-
tography using 20% ethyl acetate/heptane as eluant to give
1
12 (0.3 g, 40%). H NMR (250 MHz, CDCl3): δ 7.25-
412
•
Vol. 11, No. 3, 2007 / Organic Process Research & Development

1
17 (0.53 g, 93%). H NMR (250 MHz, CDCl3): δ 7.48-
Methyl 2-(N-((2′-Cyanobiphenyl-4-yl)methyl)pentana-
mido)-3-methylbutanoate (21). 4-((N-(1-Methoxy-3-methyl-
1-oxobutan-2-yl)pentanamido)methyl)phenylboronic acid 20
(0.75 g, 2.2 mmol) and 2-iodobenzonitrile (0.54 g, 2.35
mmol) were dissolved in 7:3:2 DME/ethanol/water. Pd-
(PPh3)4, (0.022 g, 0.02 mmol) and 2 M aqueous sodium
carbonate solution (4.7 mL, 9.4 mmol) were added with
stirring under nitrogen. The solution was refluxed for 12 h,
after which time the solvent was evaporated under reduced
pressure (0.82 g, 91%). The residue was purified using silica
scavenger 1, 2, or 3. MS (ES+): m/z ) 407 [M + H]+.
2-Methoxy-N-(3-{4-[3-methyl-4(6-methyl-pyridin-3-
yloxy)phenylamino]quinazolin-6-yl}-E-allyl) Acetamide
(30). (6-Iodoquinazolin-4-yl)-[3-methyl-4-(6-methyl-pyridine-
3-yloxy)phenyl]amine 28 (0.29 g, 0.6 mmol), N-allyl-2-
methoxyacetamide 29 (0.171 g, 1.3 mmol), Pd2(dba)3 (0.040
g, 0.04 mmol), sodium acetate (0.15 g, 1.82 mmol), and
triethylamine (1.3 mL, 9.3 mmol) were dissolved in ethanol
(20 mL) and refluxed for 6 h. The reaction mixture was
filtered to remove the solid particles and the product solution
evaporated under reduced pressure (0.30 g, quantitative
yield). The residue was purified using a 1:1:1 mixture of
silica scavengers 1, 2, and 3 (0.5 g each). MS (ES+): m/z )
470 [M + H]+.
4-Nitro-N-phenylbenzenamine (33). Aniline 32 (4.0 mL,
42.9 mmol), 4-nitro chlorobenzene (1 mL, 8.22 mmol) 31,
and PdCl2(PPh3)2 (28 mg, 0.04 mmol) were stirred at rt for
10 min. Potassium carbonate (1.7 g, 12 mmol) was added,
and the solution was refluxed for 16 h. The mixture was
cooled to rt, diluted with ethyl acetate (20 mL), washed with
water and saturated aqueous brine solution, and dried over
MgSO4. The solvent was evaporated under reduced pressure
to give 33 (1.9 g, quantitative yield). The residue was purified
using silica scavenger 2. MS (ES-): m/z ) 213 [M - H]-.
7.42 (m, 2H), 7.31-7.26 (m, 2H), 7.21-7.11 (m, 7H), 7.03-
6.95 (m, 3H), 3.43-3.33 (m, 2H), 3.11 (t, J ) 7.11 Hz, 1H),
3.01-2.95 (m, 1H), 2.70-2.58 (m, 2H), 2.49-2.43 (m, 1H),
2.38-2.32 (m, 1H); MS (ES+): m/z ) 354 [M + H]+, m/z
) 376 ) [M + Na]+.
4-((1-Methoxy-3-methyl-1-oxobutan-2-ylamino)meth-
yl)phenylboronic Acid (19). 4-Formylphenylboronic acid
(1.0 g, 6.6 mmol) and valine methyl ester hydrochloride (1.1
g, 7 mmol) were dissolved in methanol (10 mL). Sodium
borohydride (0.567 g, 15 mmol) was dissolved in methanol
(10 mL) and added dropwise to the reaction mixture over 5
min. The resulting suspension was stirred at rt for 4 h. The
solvent was evaporated under reduced pressure, and the
resulting white solid was dissolved in water and extracted
with ethyl acetate. The solvent was evaporated under reduced
pressure, and the residue was purified by column chroma-
tography using 40% ethyl acetate/hexane as eluant to give
1
19 (0.88 g, 54%). H NMR (250 MHz, CDCl3): δ 8.11 (d,
J ) 7.8 Hz, 2H), 7.41 (d, J ) 7.8 Hz, 2H), 3.87 (d, J )
13.6 Hz, 1H), 3.60 (d, J ) 13.6 Hz, 1H), 3.67 (s, 3H), 2.97
(d, J ) 6.1 Hz, 1H), 1.94-1.81 (m, 1H) 0.92-0.87 (m, 6H);
MS (ES+): m/z ) 266 [M + H]+.
4-((N-(1-Methoxy-3-methyl-1-oxobutanyl)pentanami-
do)methyl)phenylboronic Acid (20). 4-((1-Methoxy-3-
methyl-1-oxobutan-2-ylamino)methyl)phenylboronic acid 19
(0.634 g, 2.4 mmol) was dissolved in dichloromethane (20
mL). Valeryl chloride (0.6 mL, 4.4 mmol) and DIPEA (0.9
mL, 4.5 mmol) were added dropwise with stirring over 5
min. The reaction mixture was refluxed for 2 h, after which
time the solvent was evaporated under reduced pressure. The
residue was dissolved in ethyl acetate and washed with 2 M
aqueous sodium carbonate solution and water and was dried
1
over MgSO4 (0.75 g, 89%). H NMR (250 MHz, CDCl3):
δ 8.17 (d, J ) 8.5 Hz, 2H), 7.71 (m, 2H), 5.06-4.94 (m,
1H), 4.71 (s, 1H), 4.64 (s, 1H), 3.47 (s, 3H), 3.36-3.32 (m,
1H), 2.42-2.18 (m, 2H), 1.71-1.57 (m, 2H), 1.45-1.20 (m,
2H), 1.03-0.80 (m, 9H); m/z ) 350 [M + H]+.
Received for review January 26, 2007.
OP7000172
Vol. 11, No. 3, 2007 / Organic Process Research & Development
•
413