Strong evidence of solution-phase catalysis associated with palladium leaching from immobilized thiols during Heck and Suzuki coupling of aryl iodides, bromides, and chlorides

Journal of Catalysis 251 (2007) 80–93

www.elsevier.com/locate/jcat

Strong evidence of solution-phase catalysis associated with palladium

leaching from immobilized thiols during Heck and Suzuki coupling

of aryl iodides, bromides, and chlorides

John M. Richardson, Christopher W. Jones ^∗

School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, 311 Ferst Drive NW, Atlanta, GA 30332, USA

Received 22 May 2007; revised 3 July 2007; accepted 4 July 2007

Available online 21 August 2007

Abstract

A 3-mercaptopropyl-functionalized silica, SH-SBA-15, is used both as a support for palladium acetate, Pd-SH-SBA-15, and as a selective poison

of soluble palladium complexes. Pd-SH-SBA-15 is used as a precatalyst for Heck and Suzuki couplings of an aryl iodide, bromide, and chloride

under a range of reaction conditions. In all reactions in which metal-free SH-SBA-15 is added, the catalysis associated with either homogeneous

palladium acetate or from Pd-SH-SBA-15 ceased. This strongly suggests that (i) SH-SBA-15 can be used as an effective scavenger and poison of

soluble palladium and (ii) catalysis with Pd-SH-SBA-15 is solely associated with leached metal. Supporting evidence is given from hot ﬁltration

tests, self-quenching behavior of partially metalated Pd-SH-SBA-15, and poisoning by other solid-phase poisons such as poly(4-vinylpyridine),

Quadrapure TU, and 3-mercaptopropyl-functionalized silica gel. It is postulated that the SH-SBA-15 poisons by over coordination of soluble

monomeric or dimeric palladium, which rules out catalysis by palladium nanoparticle surfaces. SH-SBA-15 is an effective and selective poison of

homogeneous palladium species and can be used as a test for heterogeneity of palladium catalysts developed for Heck and Suzuki reactions.

Keywords: Heck reaction; Suzuki reaction; Mercaptopropyl silica; Palladium nanoparticle; Three-phase test; Palladium; Leaching; Supported catalysts; Selective

catalyst poison

1. Introduction

that are active at ppm metal concentrations and (ii) immobiliza-

tion of palladium so that it can be recovered and reused [4,5].

The quest for a commercially viable, highly active, recov-

erable, and reusable palladium catalyst for Heck and Suzuki

reactions is still ongoing despite the vast number of attempts

and strategies used to date. Despite the many claims of hetero-

geneous catalysis by supported palladium, when more rigorous

testing is undertaken, it is most often found that the true ac-

tive species are from leached metal [5–17]. In fact, in a thor-

ough review of the literature, we have asserted that there is

no direct evidence for a truly heterogeneous palladium Heck

or Suzuki coupling catalyst-based palladium nanoparticle- or

macroparticle-catalyzed turnovers [5]. However, many authors

do not attempt to assess the nature of the truly catalytic species

and instead appear to assume that the form of palladium added

to the reactor is the active catalyst. When authors do use vari-

ous reaction tests to attempt to discern the location of the active

species (i.e., leached palladium vs solid-phase palladium), re-

Palladium-catalyzed carbon–carbon coupling reactions, such

as the Heck [1] and Suzuki [2] reactions, represent an ex-

tremely important class of chemical transformations. A mul-

titude of soluble and immobilized palladium precatalysts have

been developed that are active and selective, but complications

for commercial operations exist due to high metal cost and

stringent requirements for removal of residual metal from exit

streams. For example, palladium-catalyzed reactions are often

used by pharmaceutical companies in the synthesis of medicinal

molecules and must meet government requirements of <5 ppm

residual metal in product streams [3]. These cost and puriﬁca-

tion pressures have spurred signiﬁcant research in two distinct

areas: (i) development of highly active homogeneous catalysts

*

Corresponding author.

E-mail address: cjones@chbe.gatech.edu (C.W. Jones).

doi:10.1016/j.jcat.2007.07.005

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

81

sults can be easily misinterpreted. This is because many of the

tests historically used to discern heterogeneous catalysis and

recyclability, such as hot ﬁltration and comparing ﬁnal yield

after recycling, give ambiguous results for Pd(0)-catalyzed cou-

pling reactions [5]. This is due to three facts: (i) Even trace

amounts of soluble palladium metal can be very catalytically

active [11,18–21], meaning that only ppm of palladium need to

leach to effect the catalysis; (ii) multiple deactivation pathways

exist for soluble palladium, including redeposition on supports,

formation of palladium black, and overcoordination by strongly

binding ligands; and (iii) due to leaching and deposition (es-

pecially once all the aryl halide is consumed) the palladium

can be partitioned among different solid phases as well as in

solution [22–24]. The lifetime of active palladium species are

therefore very sensitive to reaction conditions and to manipula-

tions of the reaction solution both during and after catalysis.

Commonly used heterogeneity tests must be carefully inter-

preted and supported by other methods. In previous studies,

we introduced solid poisons to show that Pd(II) immobilized

by pincer ligands [12,25,26] and Pd(II) encapsulated in a poly-

meric matrix [9] are not heterogeneous catalysts, but are simply

reservoirs of leached active palladium. In two recent reviews,

we suggested that immobilized thiols metalated with palladium

warranted further investigation as a class of heterogeneous pal-

ladium precatalysts for C–C coupling reactions [5,27]; some

authors have asserted that they leach active species, whereas

others claim they give truly heterogeneous catalysts (vide in-

fra).

Thiol-modiﬁed surfaces have been previously metalated

with palladium and used for Heck and Suzuki reactions. Li

and Jiang immobilized 3-mercaptopropyltriethoxysilane on a

silica surface and then metalated with either palladium ac-

etate or H₂PdCl₄[28]. These materials where then used in the

Heck coupling of iodobenzene with various oleﬁns. Tests for

heterogeneity included recycling the catalyst three times and

comparing ﬁnal yields. Wang and Liu used polymers containing

bidentate mercapto-hydroxyl ligands that were metalated with

palladium acetate, adsorbed onto silica, and used in Heck cou-

plings [29]. Leaching of palladium was detected and attributed

to reduction of Pd(II) by the oleﬁn, followed by oxidative ad-

dition of the aryl halide. The palladium in the postreaction

solution could be recovered by centrifuging away the precata-

lyst material and then adding nonmetalated support material to

scavenge the leached metal, which was then successfully used

in a subsequent Heck reaction. The authors also pointed out that

materials with S:Pd ratios >6 exhibited no catalytic activity.

Cai, Song, and Huang condensed 3-mercaptopropyltrietho-

xysilane with fumed silica and metalated with Pd(II) chloride,

which was reduced to Pd(0) [30]. This material was used as a

precatalyst for the Heck reaction of aryl iodides with styrene

and acrylic acid. Heterogeneity was implicitly tested by study-

ing the activity on two successive runs, which decreased in ac-

tivity by only 3% each time. Later, these same authors reported

the synthesis of a similar material metalated with Pd(PPh₃)₄,

which was used for a Sonogashira coupling with Cu(I) iodide

as cocatalyst [31]. Choudary et al. [32] synthesized a bifunc-

tional catalyst by tethering 3-mercaptopropyltrimethoxysilane

to a silica gel surface and then partially reacted the thiols with a

cinchona alkaloid. The material was metalated with Pd(II) chlo-

ride, which was assumed to only bind with the thiols, and the

cinchona alkaloid was metalated in situ with OsSO₄. Hetero-

geneity tests of the catalyst were performed by testing activity

of ﬁltrates both from partially converted solutions and after ad-

dition of fresh reagents to completed reactions.

All of the above-mentioned publications suggested that pal-

ladium can be bound to an immobilized thiol surface and then

used as a precatalyst for Heck reactions. However, the two pri-

mary tests used to prove or suggest heterogeneity—hot ﬁltra-

tion and ﬁnal yield after recycling—may be ambiguous (vide

supra) and are not conclusive [5]. Up until the point of these

publications (2002), the question had been unanswered as to

whether catalysis from metalated, immobilized thiols is a re-

sult of leached or bound metal. To further complicate things,

three recent, more detailed reports using similar precatalysts

suggest different types of active species; two claim truly het-

erogeneous catalysis, and one claims that all catalysis is from

leached species.

In 2004, Shimizu et al. [33] reported a careful and detailed

characterization of a mesoporous silica, FSM-16, grafted with

3-mercaptopropylsiloxanes and metalated with palladium ac-

etate to form Pd-SH-FSM. This was the ﬁrst work to conﬁrm

that most of Pd atoms are bound to two sulfur atoms and are

in a Pd(II) oxidation state on these types of mercaptopropyl-

modiﬁed silica supports. Pd-SH-FSM was used as a precatalyst

for both Heck and Suzuki reactions of activated bromides. The

bulk of the reactions studied were Suzuki couplings, whereby

the Pd-SH-FSM was reported as active, recoverable, and recy-

clable. EXAFS data indicated that most of the palladium was

still bound to thiols after the reaction, and no detectable amount

of palladium nanoparticles were observed in TEM. Only 0.05%

of the bound Pd was found in the ﬁltrate after reaction, and truly

heterogeneous catalysis was asserted. However, because palla-

dium can redeposit once aryl halide is consumed [22,34,35],

and only traces of Pd are needed for high activity under some

conditions, this is a lower bound on the amount that could have

been in solution during the reaction. Under Heck conditions,

only 0.01% of Pd was reported to have leached, but in this case

small nanoparticles (2–14 nm) were observed by TEM, the ap-

pearance of which is suggestive of palladium leaching followed

by nanoparticle formation and redeposition onto the support.

EXAFS and XANES data indicated that most of the Pd was

similar to that of the fresh catalyst. Tests for heterogeneity in-

cluded hot ﬁltration, in which the authors used 10 times less

initial palladium (0.1%) than under standard conditions (1.0%),

after which no activity was observed. The Pd-SH-FSM catalyst

was recycled up to ﬁve times with no loss in productivity (based

on ﬁnal yield). However, careful comparison of the kinetics be-

tween the ﬁrst and ﬁfth runs show a dramatic decrease in the

initial rate of reaction after recycling. The combined data from

this report were interpreted as supporting the hypothesis that

the thiol-supported palladium was a heterogeneous catalyst.

Also in 2004, Crudden et al. [36] studied mercaptopropyl-

modiﬁed SBA-15, both co-condensed and postgrafted with 3-

mercaptopropyltrimethoxysilane, both of which were metalated

82

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

Table 1

Summary of results obtained using mercaptopropyl-modiﬁed silica materials as supports for Pd(OAc) precatalysts in Heck and Suzuki couplings

2

Author

Solvent

Heck

Halide

Heck

2nd reagent

Heck

Base

Catalysis

Heck

Suzuki

Heck

KOAc

Suzuki

Surface

Shimizu

NMP

(403 K)

DMF (403 K)

4-Bromoaceto 4-Bromoanisole

phenone

Ethyl

acrylate

PhB(OH)

K CO

2

Surface

2

3

Pd-SH-FSM-16

(S:Pd = 2.8)

a

Crudden

DMF

DMF H O

4-Bromoaceto 4-Chlorobenzene

Styrene

PhB(OH)

NaOAc K CO

Surface

2

3

Pd-SH-SBA-15

(S:Pd = 4)

(393 K) DMF/H O

phenone

4-Chloroacetophenone

4-Bromopyridine

4-Bromotoluene

2

(353 K, 363 K,

373 K)

4-Bromoanisole

4-Bromobenzaldehyde

Davis

DMF

N/A

Iodobenzene

N/A

n-Butyl

N/A

NEt

N/A

Leached N/A

3

Pd-SH-SBA-15

(S:Pd = 1.6)

(353 K,

373 K)

acrylate

a

Surface catalysis is implied Heck reactions in [36] and more explicitly stated in [39] where catalysis is claimed as “leach-proof.”

with palladium acetate. Most of the reactions tested involved

Suzuki couplings of either aryl bromides or aryl chlorides, but a

smaller number of Heck couplings also were performed. It was

found that only the co-condensed material exhibited effective

catalytic behavior (S:Pd = 4:1), whereas the grafted material

was inconsistent, with most of the prepared materials showing

no activity. A signiﬁcant achievement was the ability to catalyze

the Suzuki reaction of chloroacetophenone (2 mol% catalyst,

H₂O as solvent) with phenyl boronic acid, with the reaction pro-

ceeding to 93% (yield = 80%) at 80 ^◦C and 99% (yield = 96%)

at 90 ^◦C. To test for heterogeneity for the Suzuki reaction,

the activity of a small amount of soluble palladium acetate

(0.5 ppm) was measured. Only 5% conversion of bromoace-

tophenone with phenylboronic acid was achieved as compared

with 99% with Pd-SH-SBA-15. This was taken as evidence

that the ppm levels of thiol-ligand–free palladium leached from

the Pd-SH-SBA-15 could not have been the source of cataly-

sis. A hot ﬁltration test was also performed, with only a small

amount of conversion observed in the ﬁltrate. Finally, a three-

phase test [37,38] was performed, in which an aryl bromide or

aryl chloride was immobilized. Soluble aryl halide was added,

and conversions of both soluble and supported reagent were

estimated. Most of soluble reagent was consumed, however

only small amounts of surface-bound reagent were converted

in some cases. Thus, in this case, the combined data also sup-

ported the conclusion that the vast majority of catalysis was

from surface-bound palladium. Later reports further strength-

ened the ﬁndings, describing these materials as “leach-proof

catalysts [39].”

most all catalytic activity was ceased. A series of hot ﬁltra-

tion experiments were conducted; in contrast to previous work,

a signiﬁcant amount of activity was observed for ﬁltrates free

of solid precatalysts. In the case of Pd-SH-SBA-15 after 1 h,

the conversion of the ﬁltrate (76%) was nearly the same as

the regular reaction (80%). The authors found that iodoben-

zene was required for leaching of active palladium, which im-

plies that oxidation of immobilized Pd(0) is a cause of leach-

ing. This is consistent with previous observations that aryl io-

dides can pull Pd(0) into solution from Pd(0) nanoparticle sur-

faces [6,11,14,41]. Finally, efﬂuents from a continuous reaction

system were found to be catalytically active, conﬁrming the

presence of leached active metal. The ﬁnal conclusion was that

activity of Pd-SH-SBA-15 for the Heck reaction under these

conditions is solely from leached palladium.

Shimizu, Crudden, and Davis each studied similar types

of palladium-immobilized catalysts using mercaptopropyl-

modiﬁed silicas as supports but under markedly different con-

ditions from each other (Table 1). Conclusions differed as to

whether the catalysis was from supported or leached metal. In

each case, the conclusions drawn by the authors appear rea-

sonable based on the data presented. Unfortunately, a direct

comparison of the results is difﬁcult due to variations in reac-

tion conditions from author to author. Shimizu and Crudden

primarily looked at Suzuki reactions with activated bromides

whereas Davis focused on Heck reactions of iodobenzene. The

more general question as to whether mercaptopropyl-modiﬁed

materials can act as a supports for heterogeneous catalysis for

both Heck and Suzuki reactions was thus still unanswered.

Toward this end, here we report on the use of a mercaptopro-

pyl-modiﬁed silica, SH-SBA-15, as a tethered ligand for palla-

dium and introduce its use in unmetalated form as a selective

poison of soluble, active palladium. Metalation with palladium

acetate yields an immobilized palladium precatalyst, Pd-SH-

SBA-15, which was used for Suzuki and Heck reactions under

the same conditions reported by Shimizu [33], Crudden [36],

and Davis [40] both with and without SH-SBA-15 added as poi-

son. In all cases in which bare SH-SBA-15 was used as poison,

no catalysis was observed. This is quite remarkable as SH-

In contrast, another report suggested exactly the opposite—

that all catalysis associated with thiol-supported palladium

species was associated with leached, soluble species [40]. In

this 2005 contribution by Ji et al., the Heck reaction of iodoben-

zene with n-butyl acrylate in DMF with triethylamine as base

was investigated using both Pd(II) and prereduced Pd(0) forms

of the mercaptopropyl-silica supported precatalyst. These au-

thors performed various tests to determine whether catalysis

was heterogeneous or from leached metal. In the presence of

insoluble poly(4-vinylpyridine) (2% cross-linked) (PVPy), al-

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

83

Scheme 1.

SBA-15 is the exact same material used as the catalyst support

and should not affect the catalysis occurring from truly hetero-

geneous palladium species on separate but identical particles.

Thus, our poisoning results strongly suggest that under all con-

ditions tested here, catalysis is solely associated with leached

palladium species. Poisoning by SH-SBA-15 also implies that

the active soluble catalytic species is not from nanoparticle sur-

faces but rather from molecular or dimeric palladium, which is

consistent with the work of de Vries [20,42,43], Schmidt [24],

Rothenberg [44] and Fiddy [45].

(TMB) as a swelling co-solvent [46]. The as-prepared material

was calcined using the following temperature program: (1) in-

creasing the temperature (1.2 K/min) to 473 K, (2) heating at

473 K for 1 h, (3) increasing the temperature (1.2 K/min) to

823 K, and (4) holding at 823 K for 6 h. Before functionaliza-

tion, the SBA-15 was dried under vacuum (5 mTorr) at 473 K

for 3 h and stored in a dry box (BET surface area = 780 m²/g;

average pore diameter = 110 Å). Complete removal of the or-

ganic surfactant was conﬁrmed by TGA analysis.

Organic modiﬁcation of SBA-15 was achieved by mixing

a toluene (120 mL) suspension of SBA-15 (4.1 g) and 3-mer-

captopropyltrimethoxysilane (10 g) at reﬂux temperature for

48 h under an argon atmosphere. Water (1.8 mL) was then

added to promote the cross-linking and the mixture was heated

at reﬂux for an additional 24 h. The solids were then ﬁltered

and washed with copious amounts of toluene, hexanes, and

methanol to remove unreacted silanes. Recovered solids were

Soxhlet-extracted with dichloromethane at reﬂux temperature

for 3 days. The resulting white solids were collected, kept un-

der vacuum (5 mTorr) at room temperature overnight, dried at

423 K for 3 h under vacuum (5 mTorr), and stored in a nitro-

gen glove box. FT-Raman analysis exhibited a strong peak at

2571 cm⁻¹, which is in the normal range found for S–H stretch-

ing. A loading of 7.5 wt% of sulfur (2.3 mmol S/g solids) was

found by elemental analysis (BET surface area = 320 m²/g;

average pore diameter = 80 Å).

2. Experimental

2.1. General

All organic materials were purchased from commercial

sources. Palladium on carbon (10 wt%), Quadrapure TU (thio-

urea-functionalized polystyrene, 3.3 mmol/g), and 3-mercapto-

propyl-functionalized silica gel (1.2 mmol/g) were purchased

from Sigma Aldrich and used without further treatment. N,N-

dimethylformamide and butyl acrylate were dried with calcium

hydride for 24 h, distilled under an argon atmosphere, and

stored at 278 K until use. Acetone used for metalations was

dried over sieves for 24 h at room temperature, distilled under

argon atmosphere, and stored in a N₂dry box (Plas Labs, Lan-

caster, MI). Nitrogen physisorption was performed using a Mi-

crometrics ASAP 2010 with pore size distributions determined

using the BJH method applied to the adsorption isotherms. Sur-

face areas from N₂physisorption were determined using the

BET method. Hexagonal ordering of the mesopores was veri-

ﬁed X-ray diffraction (PANalytical X’Pert PRO) using CuKα

radiation. Loadings of organically modiﬁed supports were de-

termined from simultaneous thermal analysis (Netzsch STA

409 PG Luxx) and veriﬁed by elemental analysis (Desert Ana-

lytics, Tucson, Arizona). Organically modiﬁed silica materials

were stored in a N₂dry box to prevent moisture adsorption. FT-

Raman analyses were performed on a Bruker FRA-106 with

1028 scans collected for each sample with a 100 kW laser

source. Gas chromatography was performed on a Shimadzu

GC-17A equipped with a ﬂame ionization detector and a HP-5

column. The column treatment program was the same as previ-

ously reported [12].

2.3. Metalation of organically modiﬁed SBA-15

Palladium acetate (450 mg, 2 mmol) was dissolved in 50 mL

of dry acetone. This solution was added to SH-SBA-15 (1.1 g),

resulting in an orange-brown suspension, which was stirred at

room temperature for 24 h (Scheme 1). The solids were ﬁl-

tered using a 150-mL fritted funnel (medium-sized frit) and

washed with 3 × 100 mL each of boiling acetone, toluene,

acetonitrile, and dichloromethane. The recovered solids were

orange-brown in color. The solids were then exposed to high

vacuum (5 mTorr) for 16 h to remove any physisorbed sol-

vent, resulting in orange-brown solid particles. A loading of

11.06 wt% palladium (1.04 mmol Pd/g solids) was determined

by elemental analysis based on the average of two EA tests for

Pd (10.9 and 11.21 wt%), giving a S:Pd ratio of 1.7. This ma-

terial was used for most of the experiments. Two other partially

metalated batches were made by mixing (i) 9 mg palladium ac-

etate, 110 mg SH-SBA-15, and 5 mL acetone and (ii) 45 mg

palladium acetate, 200 mg SH-SBA-15, and 25 mL acetone.

Materials were washed using a scaled-down ﬁltration procedure

(60 mL fritted funnel and 3 × 50 mL each of solvent). Final Pd

2.2. Synthesis of organically modiﬁed mesoporous silica

Large-pore SBA-15 (110 Å) was synthesized by literature

methods, using the triblock poly(ethylene oxide)–poly(propyl-

ene oxide)–poly(ethylene oxide) (EO–PO–EO) nonionic surfac-

tant as the structure-directing agent and 1,3,5-trimethylbenzene

84

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

loadings and S:Pd ratios were (4.1%, 5.7) and (8.4%, 2.6), re-

spectively.

dissolved in 1.5 mL of acetone, passed through a small plug of

silica gel loaded in a cotton plugged pipet, and analyzed by GC.

The ﬁnal conversion was determined by comparing the ratio of

aryl halide and DGDE to the calibration ﬁle.

2.4. Heck reactions

Various reaction conditions for Heck reactions were stud-

ied. For most of the reactions, the precatalyst Pd-SH-SBA-15

(1.2 mg, 1.25E−03 mmol Pd, 0.05 mol% Pd to aryl halide)

was added to a three-necked ﬂask and connected to a con-

denser linked to a Schlenk line. The system was purged with

argon, and then a solution of N,N-dimethylformamide (5 mL),

aryl halide (2.5 mmol), base (7.5 mmol) and diethylene glycol

dibutyl ether (2.5 mmol, GC standard) was added. The mixture

was magnetically stirred and preheated at reaction temperature

for 25 min, at which time n-butyl acrylate (3.75 mmol) was

added to initiate reaction. Periodically, 0.1-mL samples were

withdrawn, dissolved in 1.5 mL of acetone at room tempera-

ture, ﬁltered through a small plug of silica gel loaded in a cotton

plugged pipet, and analyzed by gas chromatography. In reac-

tions in which homogeneous palladium acetate was used as the

precatalyst, n-butyl acrylate was added before the preheating,

and the reaction was initiated by adding 0.1 mL of a 0.0125 M

solution of palladium acetate in DMF. Exceptions to reaction

conditions described above are noted where necessary.

2.6. Hot ﬁltration

A Heck reaction solution was allowed to reach 20–30% con-

version of iodobenzene, and the solids were ﬁltered off under

static vacuum by means of a swivel frit [47] connected to an

empty reaction ﬂask. The ﬁltrate was kept at reaction tempera-

ture, and conversion was monitored by GC.

2.7. Catalyst poisoning

In most of the reactions, the catalyst poison was added to the

reaction ﬂask before the addition of reaction solution. Poly(4-

vinylpyridine) was used in 350 equivalents of pyridine sites to

total palladium unless noted otherwise. When Quadrapure TU,

SH- SBA-15, or 3-mercaptopropyl-functionalized silica gel was

used as a poison, the molar ratio of poison to palladium was

35 to 1 unless stated differently. In reactions in which the poi-

son was added after the initiation of reaction, a glass plug in

the three-necked ﬂask was removed (brieﬂy exposing the ar-

gon ﬁlled vessel to air for approximately 3–5 s) and the poison

particles were quickly added, after which the glass plug was re-

inserted.

2.5. Suzuki reactions of 4-bromoacetophenone or

4-chloroacetophenone

2.5.1. DMF as solvent

3. Results

Aryl halide (1.0 mmol) and hexamethylbenzene (0.5 mmol,

GC standard) were dissolved in 5 mL of DMF. The solution

was added to a glass reactor with precatalyst (1.0% for bro-

moacetophenone and 2.0% for chloroacetophenone), K₂CO₃

(2.0 mmol), and phenyl boronic acid (1.5 mmol). The suspen-

sion was purged with argon, sealed, and immersed in an oil bath

at 353 K and magnetically stirred for 6 or 8 h. Reactions in

DMF were analyzed by taking a 0.1-mL sample, which was dis-

solved in 1.5 mL acetone at room temperature, ﬁltered through

a small plug of silica gel loaded in a cotton plugged pipet, and

analyzed by gas chromatography.

3.1. Aryl iodide conversions

3.1.1. Comparison of activity between soluble and

immobilized palladium acetate

A typical benchmark reaction for Heck couplings involves

reacting aryl iodides with either styrene or an acrylate. Aryl

iodides are more active than aryl bromides or aryl chlorides

due to their rapid oxidative addition to Pd(0), forming a Pd(II)

complex. Unless stabilizing ligands are present, Pd(0) can ag-

gregate to from palladium black, a typically inactive form of

palladium that commonly precipitates out of solution. In the

presence of excess oxidizing agents such as aryl iodides, deac-

tivation by palladium black formation can be mitigated [14,20,

41,48]. However, as the aryl iodide is consumed in the reaction,

less of it is available to complex to Pd, and the distribution of

Pd can shift to favor Pd(0), leading to formation of palladium

nanoparticles or palladium black. The overall catalytic activity

represents a balance between completing turnovers in the Heck

cycle and deactivation of the palladium associated with palla-

dium black formation or other deactivated forms of palladium,

such as overcoordinated palladium with strongly binding mole-

cules or inactive palladium redeposited on surfaces.

2.5.2. H₂O as solvent

Precatalyst (1.0% for bromoacetophenone and 2.0% for

chloroacetophenone), K₂CO₃(2.0 mmol), aryl halide

(1.0 mmol), phenyl boronic acid (1.5 mmol), hexamethylben-

zene (0.5 mmol, GC standard), and 5 mL of double-distilled

H₂O were added to a glass reactor and magnetically stirred at

353 K for 6 to 24 h. In the original report of these conditions,

hexamethylbenzene was used as an internal standard [36]. In

this work, hexamethylbenzene was again included in the reac-

tion, but because the reaction was in water, we developed an al-

ternate method to assess the conversion of aryl halide based on

a GC calibration ﬁle composed of known concentrations of aryl

halide and 4-acetyl-biphenyl (product) with DGDE as standard.

Reactions were extracted with 3 × 10 mL of dichloromethane.

The DCM was then evaporated, and a known amount of DGDE

in 5 mL of acetone was added. A 0.1-mL sample of solution was

To compare the activity between soluble and immobilized

palladium, Heck coupling reactions of iodobenzene with n-

butyl acrylate were performed using two different palladium

sources: homogeneous palladium acetate and immobilized pal-

ladium acetate on mercaptopropyl-modiﬁed mesoporous silica,

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

85

Fig. 2. Application of hot ﬁltration to the Pd-SH-SBA-15 catalyzed Heck cou-

pling of iodobenzene with butyl acrylate at 363 K in 5 mL DMF with triethy-

lamine as base. Hot ﬁltration was applied at 20 min.

Fig. 1. Activity comparisons of Heck coupling iodobenzene with butyl acry-

late using: (A) 0.05% palladium acetate at 363 K, (B) 0.5% Pd-SH-SBA-15

at 363 K, (C) 0.05% Pd-SH-SBA-15 at 363 K, (D) 0.05% palladium acetate at

343 K, (E) 1.0% palladium acetate at 363 K with start of self deactivation noted,

and (F) 0.05% Pd-SH-SBA-15 at 343 K with ﬁnal conversion at 1000 min.

3.1.2. Hot ﬁltration

A strong test for assessing the presence of soluble active pal-

ladium is the hot ﬁltration test, also known as the split test,

in which solid precatalysts are ﬁltered out of the reaction and

the ﬁltrate is monitored for continued activity [50]. A lack of

ﬁltrate activity is traditionally interpreted as a proof of hetero-

geneous catalysis, but this interpretation cannot be applied to

palladium-catalyzed reactions without support from other tests.

Heck reactions can be catalyzed by extremely small amounts

of palladium (as low as 0.00001%), and the self-deactivation at

high loadings described above suggests that very low loadings

are ideal [5,20]. Disruptions such as hot ﬁltration can deacti-

vate the small amount of active metal, which can lead to the

incorrect conclusion that there are no active soluble catalytic

species before the ﬁltration [5,51]. In the present work, when

Pd-SH-SBA-15 (0.05%, 363 K) was hot ﬁltered, after 20% con-

version, the ﬁltrate continued to react, indicating that some of

the palladium was leaching from the surface and catalyzing

the reaction (Fig. 2). A common method for determining the

amount of solution versus surface catalysis is to compare the

rate of activity after hot ﬁltration to that found under normal re-

action conditions. But because some of the active palladium can

become deactivated during hot ﬁltration, this method of com-

parison provides rough estimates at best. Other tests are needed

to determine whether any heterogeneous catalysis is indeed oc-

curring.

Pd-SH-SBA-15 (Scheme 1). N,N-dimethylformamide (DMF)

was used as the solvent, and triethylamine was used as the base.

Comparisons of activity between soluble and immobilized pal-

ladium acetate are displayed in Fig. 1. Reactions using soluble

palladium acetate rapidly went to completion at 363 and 343 K

(Figs. 1A and 1D). Increasing the amount of homogeneous pal-

ladium acetate is known to result in catalyst deactivation due to

the higher concentrations of Pd(0) that rapidly cluster to form

palladium black [20,49]. This self-quenching was also observed

under the reaction conditions studied in this report (Fig. 1E).

A 1.0% loading of palladium acetate was initially faster than

the 0.05% loading, but after about 5 min, the palladium began

to self-deactivate, and the conversion ultimately reached a lower

value of 69% at 120 min, compared with the 100% conversion

at 40 min reached for the lower catalyst loading.

Reactions with Pd-SH-SBA-15 were also active but dis-

played lower rates than reactions using homogeneous palla-

dium acetate at the same temperature and palladium loading.

The decrease in temperature from 363 to 343 K signiﬁcantly

slowed the activity of Pd-SH-SBA-15, necessitating a much

longer time to start, progress, and ultimately ﬁnish the reaction

(Fig. 1C and F). This dramatic decrease in activity could be ei-

ther due to effects inhibiting reaction on the surface (if it were to

occur there) or represent the temperature inﬂuence on leaching

of immobilized palladium from supported thiols. Increasing the

catalyst loading of Pd-SH-SBA-15 to 0.5% resulted in a faster

reaction rate (Fig. 1B), which stands in contrast to the observed

behavior after homogeneous palladium acetate loadings were

increased.

3.1.3. Tests of heterogeneous catalytic activity

In our work on assessing homogeneity/heterogeneity of

catalysis associated with supported Pd precatalysts in cou-

pling reactions, we introduced the use of solid, insoluble poi-

sons for extinguishing solution-phase catalysis [5,9,12,25,26].

The insoluble metal scavenger poly(4-vinylpyridine), 2% cross-

linked, was ﬁrst used as a tool to determine whether supported

catalysis was or was not occurring for palladium pincer com-

plexes immobilized on silica or organic polymers [12,25,26],

and later using Pd immobilized on amorphous silica [40] and

palladium acetate encapsulated in a poly(urea) matrix [9]. No

activity was observed when PVPy was used to poison Pd-SH-

SBA-15, typically using 350 equivalents of pyridine sites to to-

tal palladium. The use of PVPy as a solid poison is not without

its drawbacks, however. Due to the equilibrium on-off coor-

dination of the palladium with the pyridine sites [52] and the

nonporous nature of the resin, a large excess of pyridine units

are required to overcoordinate the palladium (Fig. 3). This can

The increased activity by raising the Pd-SH-SBA-15 loading

could be explained by one of two possible scenarios: (i) Pd(0)

formed during the Heck cycle cannot aggregate due to its bind-

ing to an immobilized organic surface, and so increased activity

is simply associated with more metal present, or (ii) the amount

of leached palladium from the surface is low, and so the ac-

tual concentration of active catalytic species in solution is below

levels that promote self-deactivation. To understand where the

catalysis is occurring, tests are needed to conﬁrm the presence

of soluble palladium and determining to what extent, if any,

catalysis is occurring on the surface.

86

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

Fig. 3. Plot of iodobenzene conversion showing the effect of increased equiva-

lencies of PVPy for poisoning of homogeneous palladium acetate catalysis of

the Heck coupling of iodobenzene with butyl acrylate. Poisoning by immobi-

lized pyridine on SBA-15 (pyridine:Pd = 35:1) is also shown for comparison.

Conditions are at 363 K under argon with DMF (5 mL) as the solvent and NEt

as the base.

3

sometimes introduce a large amount of solids to the solution

and may result in some pore blocking of the SBA particles from

agglomerated PVPy nanoparticles or partially solvated poly-

mer. A further disadvantage is that polymer swelling might be

needed for access to many of the pyridine sites; this is largely

inﬂuenced by both choice of solvent and reaction temperature.

For comparison to PVPy, 2-(4-pyridylethyl)triethoxysilane was

grafted onto SBA-15, Pyr-SBA-15 (1.7 mmol pyridine/g solids

and pyridine:Pd = 35:1), and substituted for PVPy as a poison

of palladium acetate (Fig. 3). An induction time of 40 min was

observed, followed by signiﬁcant activity, which is consistent

with an equilibrium on-off mechanism by which soluble palla-

dium is ﬁrst rapidly overcoordinated by the immobilized pyri-

dine sites and then eventually leaches back into solution. Under

the same reaction conditions, PVPy (pyridine:Pd = 35:1) ex-

hibited almost no initial slowing of reaction and reached full

conversion by 40 min. This comparison suggests the impor-

tance of PVPy polymer swelling to make pyridine sites acces-

sible for coordination to homogeneous palladium.

For most immobilized palladium precatalysts, the amount

of leached palladium is unknown; thus, the amount of PVPy

needed for poisoning becomes an optimization problem. As

an example, 350 equivalents of PVPy did not completely stop

catalysis when soluble palladium acetate (0.05%) was used,

but did quench activity from Pd-SH-SBA-15. Comparing these

results provides a qualitative estimate for the amount of pal-

ladium leached from Pd-SH-SBA-15 and indicates that not all

palladium was removed from the surface when the supported

precatalyst was used. More importantly, the quenching of activ-

ity by PVPy is strong evidence for catalysis solely by leached

species, but due to its previously mentioned drawbacks, a more

elegant selective poison is desired to provide supporting evi-

dence.

−1

Fig. 4. FT-Raman spectrum of SH-SBA-15. The peak at 2571 cm is assigned

to an S–H stretch.

Fig. 5. XRD patterns of SBA-15 before and after mercaptopropyl-function-

alization. Retainment of long range ordering is demonstrated by a strong (100)

peak.

was veriﬁed by FT-Raman spectroscopy (Fig. 4). The presence

of long-range order and a hexagonal structure, as expected for

SH-SBA-15, was veriﬁed by XRD (Fig. 5).

Bare, palladium-free SH-SBA-15 was used to poison catal-

ysis of the Heck coupling of iodobenzene with n-butyl acrylate

using either homogeneous palladium acetate or Pd-SH-SBA-15

(Fig. 6). In all cases, the presence of an excess of bare SH-

SBA-15 to palladium of 35–1 resulted in complete cessation of

activity. SH-SBA-15 was a more effective poison than PVPy

and required less material (e.g., 20 mg of SH-SBA-15 vs 50 mg

of PVPy), perhaps due to binding site accessibility. As shown in

Fig. 3, 35 equivalents of PVPy had almost no poisoning effect,

whereas Pyr-SBA-15 stopped activity for 40 min, after which

activity was observed. In contrast, SH-SBA-15 completely shut

down activity even after 16 h, demonstrating its superior perfor-

mance as a poison and also demonstrating the better ability of

immobilized thiols to retain palladium compared with immo-

bilized pyridines. Also, because it is the same material as that

used for the catalyst support, no poison-induced pore blocking

occurs. To probe whether poisoning is from the organic func-

tionalization of SBA-15 and not from anything else inherent to

the SBA-15, a control reaction was performed in which 50 mg

of SBA-15 was substituted for SH-SBA-15 and tested as a poi-

son. No difference in the kinetic proﬁle was observed between

reactions with and without SBA-15, conﬁrming that poisoning

Mercaptopropyl-modiﬁed surfaces, such as SH-SBA-15,

have been used as effective palladium scavengers [36,53–55]

as well as scavengers for other metals, such as silver and mer-

cury [56]. SH-SBA-15 is a rigid mesoporous solid that does not

swell under reaction conditions, a previously mentioned draw-

back of ligands immobilized on polymers such as PVPy. We

thus hypothesized that the same SH-SBA-15 material used as a

support for palladium acetate immobilization could also act as a

selective poison of soluble palladium acetate if used in excess.

Before its use as a selective poison, the presence of S–H groups

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

87

SH-SBA-15 could prevent the aggregation of palladium to form

nanoparticles. In this scenario, a demonstrated lack of nanopar-

ticle surface catalysis would not be substantiated by the use

of insoluble poisons. However, the rapid quenching of cataly-

sis by addition of poison once activity has commenced (Fig. 6,

curve C), presumably when nanoparticle formation could have

already occurred, is suggested as evidence that if nanoparticles

are formed, their surface is inactive, as postulated above. Col-

lectively, this evidence suggests Heck catalysis by molecular

or dimeric palladium as postulated by de Vries [5,20,42,43]

and Schmidt [5,24] and supported by Rothenberg [44] and

Fiddy [45]. Further testing to conclusively demonstrate lack of

activity of nanoparticle surfaces by the use of insoluble poisons

is outside the scope of this paper, but is suggested as a poten-

tially interesting future work.

Fig. 6. Plot of iodobenzene conversion with SH-SBA-15 used as a poison and

either palladium acetate or Pd-SH-SBA-15 as catalyst: (A) control reaction at

363 K using SBA-15 instead of SH-SBA-15 as an ineffective poison of pal-

ladium acetate, (B) Pd-SH-SBA-15 activity without any poison, (C) cessation

of Pd-SH-SBA-15 activity by adding SH-SBA-15 poison after 40 min of reac-

tion, (D) cessation of Pd-SH-SBA-15 activity at 363 K by adding SH-SBA-15

at start of reaction, (E) cessation of Pd-SH-SBA-15 activity at 343 K by adding

SH-SBA-15 at start of reaction, and (F) cessation of palladium acetate activity

by adding SH-SBA-15 at start of reaction.

3.1.4. Probing mercaptopropyl ligand leaching

An alternative mechanism for poisoning of Pd-SH-SBA-15

by SH-SBA-15 is that the mercaptopropyl groups leach and

poison surface-bound palladium. In this case, the poisoning ex-

periments using SH-SBA-15 could not conclusively support a

leaching mechanism of Pd-SH-SBA-15. However, if mercapto-

propyl leaching were occurring for SH-SBA-15, then it would

also be expected to occur for the mercaptopropyl groups used

to immobilize palladium acetate on Pd-SH-SBA-15, unless the

presence of palladium acetate somehow prevented this from

happening. Mercaptopropyl leaching from SH-SBA-15 was ex-

plicitly probed by exposing 50 mg of SH-SBA-15 to reaction

conditions for 1 h and then hot-ﬁltering the solids. The ﬁl-

trate was kept at 363 K, and homogeneous palladium acetate

(0.1 mL of a 0.0125 M solution in DMF) was added. Complete

conversion was reached in <30 min; therefore, it is unlikely

that signiﬁcant leaching of thiols occurred under these condi-

tions. An important assumption for this interpretation is that

any leached thiols can survive the hot-ﬁltration process with-

out redepositing on the surface (recall that leached palladium

often redeposits). Otherwise, the observed activity of the ﬁl-

trate could not conﬁrm the absence of thiol leaching. Assuming

nearly full coverage of the silica surface with silanes, small

amounts of leached silane, if they existed, likely would not

ﬁnd bare silica surfaces on which to redeposit. Thus, we would

expect leached silane redeposition to be unimportant. To inves-

tigate the retention of the mercaptopropyl groups on the silica

surface, the recovered solids from the hot ﬁltration were rinsed

with boiling hot toluene, dichloromethane, and acetone and ex-

posed to hi-vacuum (5–6 mTorr) to remove any physisorbed

organic material. Presence of residual S–H bonds was con-

ﬁrmed by FT-Raman analysis of both the untreated SH-SBA-15

and the SH-SBA-15 recovered from hot ﬁltration. A strong FT-

Raman peak at 2175 cm⁻¹was observed for both materials and

assigned as an S–H stretch. The activity of the ﬁltrate and FT-

Raman analysis suggest that thiol leaching does not occur at a

level that would allow poisoning of Pd-SH-SBA-15 if it were

a heterogeneous catalyst. As noted in Section 3.1.3, a large

amount of leached thiol giving a S:Pd ratio in excess of 4:1

would be required to completely poison hypothesized surface-

mediated catalysis.

was caused by mercaptopropyl modiﬁcation of the surface and

is not inherent to the SBA-15 material. A second control exper-

iment was conducted in which 1-propanethiol was mixed with

Pd-SH-SBA-15 (4:1, propanethiol:Pd) under normal Heck re-

action conditions. Kinetic activity was signiﬁcantly slowed but

did reach a maximum of 40% conversion after 20 h. A com-

plementary reaction using SH-SBA-15 (4:1, SH-SBA-15:Pd)

as the poison resulted in complete cessation of activity up to

33 h. These tests highlight the importance of immobilizing the

thiol groups to increase poisoning capacity. This is most likely

explained by a high local sulfur concentration of the bound

thiols in which captured palladium is rapidly overcoordinated

by neighboring thiols before it can be released back into solu-

tion.

When SH-SBA-15 was mixed with Pd-SH-SBA-15 at either

the beginning of the reaction or after the reaction has initi-

ated, no additional conversion was observed (Fig. 6). This is

strong evidence that all catalysis previously observed in this

work with Pd-SH-SBA-15 was solely from leached palladium,

which agrees with observations made by Davis [40]. To deter-

mine the lower bound of required poison, the equivalency of

SH-SBA-15 to Pd was decreased from 35 to 4, and no reac-

tion of iodobenzene with n-butyl acrylate was observed up to

2000 min (0.05% catalyst, 363 K, DMF, NEt₃). Further decreas-

ing the equivalency to 1 resulted in a slow conversion to 58%

after 700 min, which remained constant up to 1900 min. This

indicates that SH-SBA-15 is an effective poison of leached pal-

ladium from Pd-SH-SBA-15 and can be effective in much lower

amounts than PVPy.

The ability of the SH-SBA-15 to effectively poison these

precatalysts implies that the active catalyst species consists of

only a small number of Pd atoms and not from the surface of

Pd nanoparticles. Due to the rigid, cylindrical structure and

large pore size of SH-SBA-15, it is difﬁcult to envision the

mercaptopropyl groups overcoordinating the entire surface of

a Pd nanoparticle, whereas complexes of one to a few Pd atoms

could be completely coordinated. Alternatively, it is also pos-

sible that the capture of free metal ions from solution by the

88

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

Table 2

a

Results of Heck coupling reactions of bromoacetophenone and various reagents

b

Entry

Alkene

Base

Catalyst (mol%)

Temperature (K)

SH-SBA-15

Time (min)

Conversion (%)

Condition

c

1

Ethyl acrylate

Styrene

KOAc

1.00%

0.50%

0.12%

0.05%

403

393

No

Yes

No

Yes

No

Yes

No

Yes

No

Yes

No

Yes

210

300

900

100

0

75

0

87

5

91

0

93

0

27

0

3-Neck

Sealed

3-Neck

c

2

3

4

5

6

7

8

9

10

11

12

NaOAc

Styrene

900

Butyl acrylate

1350

180

2000

165

1400

NEt

3

a

Reactions conditions used 1.0 mmol aryl halide, 1.5 mmol alkene, 2 mmol base, 1.0 mmol DGDE (internal standard for GC), and 5 mL of DMF under argon

atmosphere unless otherwise noted.

b

Reactions using a 3-neck ﬂask were under an atmosphere of argon using a Schlenk line and a mercury bubbler. Reactions under sealed conditions were inside a

glass tube purged with argon and sealed with a Teﬂon-lined screw cap.

c

Amounts used were 3.0 mmol bromoacetophenone, 4.5 mmol ethyl acrylate, 4.5 mmol KOAc, 1.0 mmol benzonitrile (internal standard for GC), and 5 mL

of N-methyl-2-pyrrolidone, NMP [33].

3.2. Aryl bromide conversions

Table 3

a

Results for Heck couplings of aryl halide with butyl acrylate using various

bases

The previous results are strong evidence that Heck catalysis

of aryl iodides from palladium immobilized on a mercapto-

propyl surfaces is solely from leached palladium. This is con-

sistent with the accumulated literature on Heck coupling [5].

However, when aryl bromides are substituted for aryl iodides,

the effect on surface versus solution catalysis is less clear, es-

pecially when thiol-supported precatalysts are used. Previous

works suggested possible heterogeneous catalysis from pal-

ladium immobilized on mercaptopropyl-modiﬁed surfaces for

couplings between bromoacetophenone and either styrene [36]

or ethyl acrylate [33] using inorganic bases instead of a soluble

organic base. Conditions in which heterogeneous catalysis was

suggested were replicated in this work, with the addition of us-

ing bare SH-SBA-15 as a selective poison (Table 2). Additional

experiments were also conducted using the aryl bromide under

the same conditions reported previously in this study for reac-

tions involving iodobenzene (0.05% catalyst, DMF, NEt₃, butyl

acrylate, 363 K). Without a poison present, Pd-SH-SBA-15 was

an effective catalyst for the reactions listed in Table 2. Interest-

ingly, in all cases in which SH-SBA-15 was added as a selective

poison, little or no product formation was observed; conﬁrming

that activity observed for bromoacetophenone reactions was as-

sociated with leached palladium. Therefore, substituting an aryl

bromide for an aryl iodide or substituting an organic base with

an inorganic base does not appear to promote heterogeneous

catalysis in this case.

Entry Aryl halide

Base

Temper-

ature (K)

SH-SBA-15 Time Conver-

(min) sion (%)

1

2

3

4

5

Iodobenzene NEt

393

No

Yes

No

30 100

300 100

3

Iodobenzene NaOAc 393

Iodobenzene NaOAc 383

Iodobenzene NaOAc 363

1600

0

95

31

48

37

93

97

0

285

270

960

300

1600

30

b

6

Iodobenzene NaOAc 363

No

7

8

Iodobenzene MDA

Bromoaceto- MDA

phenone

393

No

530

9

Bromoaceto- KOAc 393

phenone

No

270

87

a

Reaction conditions are 2.4 mmol aryl halide, 3.6 mmol butyl acrylate,

4.8 mmol base, 2.4 mmol DGDE (GC standard), and 5 mL DMF under argon

atmosphere. Precatalyst is Pd-SH-SBA-15 (0.05%).

b

0.00125 mmol of homogeneous palladium acetate was substituted for Pd-

SH-SBA-15.

either palladium acetate or Pd-SH-SBA-15 as the catalyst (Ta-

ble 3). The higher conversion of bromoacetophenone found

when using sodium acetate instead of organic base was not

observed for iodobenzene; instead, the opposite was true, with

higher conversions achieved when NEt₃rather than NaOAc was

used as the base. Using methyl(dicyclohexyl)amine (MDA),

a base previously described as more active for couplings in-

volving activated aryl bromides [57], produced no activity with

bromoacetophenone but signiﬁcant activity for iodobenzene.

Thus, the exact role of the base on activity is currently not

clear with respect to leaching or stabilization of the leached cat-

alytic species. In general, the use of inorganic bases instead of

organic bases promotes bromoacetophenone activity but low-

ers iodobenzene activity. Changing the base does not promote

heterogeneous catalysis; all SH-SBA-15 poisoning experiments

3.3. Impact of the type of base

Comparing the activity under various conditions demon-

strated that reactions using sodium acetate as the base produced

a higher yield than comparable reactions using triethylamine

(Table 2; entries 7, 9, and 11). To probe the effect of base

on both leaching and activity, various bases were used, with

iodobenzene and bromoacetophenone as the aryl halide and

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

89

Table 4

a

b

Results from Heck reactions of chloroacetophenone and Suzuki reactions of chloroacetophenone and bromoacetophenone [36]

Entry

Aryl halide

Reaction

Catalyst loading

Base

Temperature (K)

SH-SBA-15

Time (h)

Conversion (%)

1

2

3

4

5

6

Chloroacetophenone

Bromoacetophenone

Heck

Suzuki

0.12%

1.00%

2.00%

1.00%

NaOAc

393

353

No

Yes

No

Yes

48

24

8

0

50

0

100

0

K CO

2

3

K CO

2

K CO

2

K CO

8

2

a

Reactions conditions are 1.0 mmol chloroacetophenone, 1.5 mmol n-butyl acrylate, 2.0 mmol base, 1.0 mmol DGDE (internal standard), and 5 mL of DMF

under argon atmosphere in a 3-neck ﬂask.

b

Reaction conditions are 1.0 mmol aryl halide, 1.5 mmol phenyl boronic acid, 2.0 mmol base, 0.5 mmol hexamethylbenzene, and 5 mL H O under argon

2

atmosphere in sealed reaction tube.

resulted in no activity (Table 2; entries 2, 4, 6, 8, 10, and 12 and

Table 3; entry 3).

3.5. Aryl chloride conversions and Suzuki reactions

The ability to use aryl chlorides for Heck, Suzuki, and other

aryl couplings is of strong interest because these molecules

are commonly cheaper than analogous iodides or bromides.

However, the carbon–chlorine bond is stronger and more difﬁ-

cult to activate. Many homogeneous palladium complexes have

been found to catalyze Heck and Suzuki reactions of aryl chlo-

rides, but these usually require speciﬁc combinations of ligands,

base, and solvent [5,58]. Attempts to couple chloroacetophe-

none with n-butyl acrylate in the Heck reaction met with no

success using Pd-SH-SBA-15 at either 0.05% or 1.0% loadings

(Table 4). However, a Suzuki reaction with chloroacetophe-

none and phenyl boronic acid (5 mL water, 2.0 mol% cata-

lyst, K₂CO₃, 353 K) after 24 h resulted in 50% conversion,

and Suzuki couplings of bromoacetophenone (1 mol% cata-

lyst) went to completion in 5 h. Like the work described above

on Heck reactions, no activity was observed with additional

SH-SBA-15 present as a poison in any case. Thus, Pd-SH-

SBA-15 is suggested to act as a source of leached palladium

for Suzuki reactions of aryl chlorides and bromides under these

conditions, just as it did for Heck reactions of aryl iodides and

bromides.

A series of selective poisoning experiments were conducted

to both probe the effectiveness of other poisons for the Suzuki

coupling of bromoacetophenone with phenyl boronic acid and

to conﬁrm the poisoning results of SH-SBA-15 (vide supra)

(Table 5). Control reactions of homogeneous palladium acetate

were compared with Pd-SH-SBA-15 and to palladium on car-

bon (Pd/C), which is a precatalyst known to operate by a leach-

ing mechanism. Two solvents (DMF and H₂O) were used, and

Suzuki couplings were carried out under normal reaction condi-

tions. In addition to SH-SBA-15, two other poisons were tested,

Quadrapure TU [9] and 3-mercaptopropyl-functionalized silica

gel (SH-SiO₂). Previously, Quadrapure TU was used as a poi-

son in high-temperature Heck reactions (90 and 110 ^◦C) and

extinguished all activity using carbon-supported precatalysts or

supposedly “leach-proof” Pd-EnCat [9]. Interestingly, for most

of the tests with Quadrapure TU, some activity was observed

for all three precatalysts, although this activity was signiﬁcantly

lower than that without the addition of Quadrapure TU. We at-

tribute this ﬁnding to two primary phenomena. First, a small

amount of activity was observed at room temperature before

submersion into an 80 ^◦C oil bath. Quadrapure TU is an or-

3.4. Reinterpretation of literature regarding activated bromide

reactions

In this work, all of the data strongly suggest that observed

Heck catalysis is associated with leached palladium, which

stands in contrast to interpretations based on previous com-

pilations of data [33,36] (vide supra). A careful examination

of these past works from the standpoint of metal leaching as

the primary mode of catalysis leads to some alternative ex-

planations for the previously observed behavior. For the Heck

reaction of 4-bromoacetophenone with ethyl acrylate, Shimizu

observed 0.01% palladium leaching and the appearance of 2–

14 nm palladium nanoparticles. The formation of nanoparticles

is consistent with palladium leaching from the surface, aggrega-

tion during the reaction on formation of Pd(0), and then redepo-

sition back onto the surface. A hot ﬁltration test was performed

on the Pd-SH-FSM precatalyst, but the conditions were with

10 times less catalyst than normal reaction conditions (0.1%

vs 1.0%). It is possible that a smaller total amount of palladium

was leached, which was deactivated during the hot-ﬁltration test

(vide supra). A lack of observed activity could occur as a result

of deactivation of soluble active palladium species [51]. The

authors report conversions up to 95% within 2 h using a load-

ing 0.0013 mol% of immobilized palladium, and after recycling

for 5 times, the authors again were able to achieve around 95%

conversion, but this time in about 8 h. The lower rate of recy-

cling is consistent with increased metal leaching or deactivation

of the catalyst by leaching and redeposition. Overall, the data

from Shimizu can be reinterpreted from the standpoint of metal

leaching as the primary or only catalytic mode. One major dif-

ference between the current study and the work of Shimizu is

the use of SBA-15 instead of FSM-16 as the silica support, but

given the similar surface characteristics, it is unlikely that this

difference in support material would change whether the cataly-

sis is occurring heterogeneously or homogeneously. The impact

of the slight difference in the S:Pd ratios between the study by

Shimizu (2.8) and the present study (1.7) is discussed in more

detail later. To compare our conditions with Shimizu’s condi-

tions, select tests were also conducted with a material with a

2.6 S:Pd ratio in this work.

90

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

Table 5

species were present, but that these were responsible for only

a

Results from selective poisoning of different palladium precatalysts for Suzuki

reactions of bromoacetophenone with phenylboronic acid

a small amount of the catalysis occurring, with most attribut-

able to heterogeneous catalysis. As noted above, the marginal

increase in conversion after ﬁltration also could be a reﬂection

of the relatively small amount of solubilized palladium that sur-

vived the hot-ﬁltration process. A test using a trace amount of

soluble catalyst (0.5 ppm of palladium acetate; 0.00023 mol%

catalyst) resulted in <5% conversion, leading to the conclusion

that small amounts of solubilized palladium could not account

for the observed activity for Pd-SH-SBA-15 [36]. The strongest

evidence for heterogeneous catalysis came from the use of the

three-phase test in which soluble aryl bromides were reacted to

a much greater extent than immobilized aryl bromides. Com-

bining all these previous data, the conclusion that some hetero-

geneous catalysis was occurring appeared reasonable.

Precatalyst Solvent Time Conversion with poison

(h)

None SH-SBA-15 Quadrapure TU SH–SiO

2

Pd(OAc)

DMF

H O

2

6

100

0

0, 20

20

0

2

Pd-SH-

SBA-15

DMF

6

94

0

6

0

8

6

100

80

0

–

30

–

0

H O

2

Pd-C

a

DMF

H O

2

6

48

76

0

–

0

–

Reaction conditions are 1.0 mmol bromoacetophenone, 1.5 mmol phenyl-

boronic acid, 2.0 mmol K CO , 0.5 mmol internal standard (hexamethylben-

2

3

zene for H O and DGDE for DMF), 1.0% catalyst, and 5 mL solvent under

2

Our observation that addition of bare, unmetalated porous

mercaptopropyl-modiﬁed support completely suppressed catal-

ysis in both Heck and Suzuki reactions brings new data to

be considered. These data can best be interpreted by the hy-

pothesis that leached, soluble palladium plays a key role in

the catalysis. Like all other ligand–metal complexes, there

exists an equilibrium between bound and free metal species

in solution. Based on the results presented here, alkyl–thiol

groups bind Pd more strongly than pyridyl groups, and thus the

mercaptopropyl-modiﬁed support is effective in signiﬁcantly

limiting the amount of soluble species that are in solution. One

hypothesis explaining the combined data is that solely leached

palladium species are active for the observed coupling reac-

tions. This is the most likely explanation for the data in our

assessment, because it is consistent with the literature [5]. An

alternate hypothesis is that only traces of palladium at defect

sites on the surface are active, and these species are more likely

to leach than most of the other sites. Thus, the addition of

solid poison captures these species as they leach into solu-

tion, removing them from the solid precatalyst and rendering

it inactive. In the context of the current catalyst system, one

might envision palladium species bound to a single thiol as

more likely to leach (and perhaps be active on the solid sur-

face), whereas species bound by two or more thiol ligands are

inactive and less likely to leach. This latter hypothesis seems

less likely, because addition of a solid poison at a midpoint of

a reaction should not result in instantaneous deactivation of ac-

tive solid supported species. Instantaneous deactivation is most

consistent with all of the catalysis being associated with reac-

tions promoted by soluble catalytic species (Fig. 6c). The trend

of increased deactivation with an increasing amount of excess

surface thiol on partially metalated surfaces supports the notion

that in the presence of sufﬁcient bare, accessible thiol, all pal-

ladium species can be overcoordinated and thus be unavailable

for leaching and catalysis.

argon atmosphere in sealed reaction tube. The molar ratio of poison to palla-

dium was 35 for all poisons.

ganic polymer resin that requires swelling of the pore structure

to increase access of poisoning sites to solubilized palladium.

Therefore, observed activity could be due to soluble palladium

catalysis before the polymer sufﬁciently swelled sufﬁciently to

allow Pd diffusion into the polymer matrix. Second, when water

is used as the solvent, the bromoacetophenone melts, creating a

separate organic liquid phase that forms a small globule. Most

of the Pd-SH-SBA-15-based particles were observed to prefer-

entially collect inside the bromoacetophenone globule whereas

the Quadrapure TU beads (average particle size ∼500 µm) were

suspended in the water or stuck to the outer surface of the or-

ganic phase. Under these conditions, the diffusion of palladium

from the Pd-SH-SBA to the Quadrapure TU may be further

hampered, thereby allowing some catalysis to occur. Thus, un-

der these conditions, application of Quadrapure TU as a poison

is not ideal. Indeed, choosing the proper solid poison to assess

catalysis requires forethought and sometimes screening, and

ambiguous or marginal information may be obtained in some

cases. For example, when using supported precatalysts that are

soluble under reaction conditions (e.g., polymer-supported pre-

catalysts), solid-phase poisons such as Quadrapure TU may

give ambiguous results [27,59]. However, in this work, no activ-

ity was observed with all tests using SH-SBA-15 or SH-SiO₂,

conﬁrming the selective poisoning ability of tethered thiols and

indicating that the catalysis occurring without the presence of

tethered poisons is associated with leached metal.

3.6. Reinterpretation of literature regarding activated bromide

reactions

Crudden used a number of techniques to assess the hetero-

geneity of the Suzuki reactions catalyzed using mercaptopro-

pyl-functionalized silica-supported palladium precatalysts. One

of these was a hot-ﬁltration test on a Suzuki coupling of

bromoacetophenone with phenylboronic acid. When care was

taken to charge the ﬁltrate with fresh base and phenylboronic

acid, an increase in conversion from 12 to 17% was observed in

the ﬁltrate. This was interpreted as evidence that some leached

Can the previously reported data suggesting heterogeneous

catalysis [36] be rationalized in light of these new results? Both

the ﬁltration test and three-phase test data were used to sug-

gest catalysis by solid-supported species. We suggest that a

positive result for each test—observations of activity in the ﬁl-

trate after ﬁltration of a solid precatalyst and of conversion of

solid-bound reactants in the three-phase test—can serve as de-

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

91

ﬁnitive evidence of some leaching of catalytic species when no

background reaction exists. But in our opinion, the negative of

each test is potentially ambiguous when applied to palladium-

catalyzed coupling reactions. As noted above and in previous

work [5,9], the competing rates of palladium deactivation, re-

deposition, and aggregation all can inﬂuence the clarity of these

tests. For example, both the ﬁltration and three-phase tests gave

results suggesting limited (but not zero) solution-phase activity

in past studies of Suzuki couplings of bromoacetophenone [36].

In the three-phase test using bromoacetophenone, the soluble

bromoacetophenone was at 50% conversion, whereas only 3%

of the supported aryl bromide was converted after 5 h. Increas-

ing the reaction time to 13 h increased the soluble component

yield to 97% and the immobilized yield to 7%. Combined with

the observation that a trace amount of soluble catalyst (0.5 ppm

of palladium acetate; 0.00023 mol% catalyst) resulted in <5%

conversion of bromoacetophenone in a Suzuki coupling, it was

suggested that small amounts of solubilized palladium could

not account for the observed activity for Pd-SH-SBA-15. Fur-

thermore, in a three-phase test using soluble chloroacetophe-

none, after 24 h, the yield of soluble aryl chloride was 80%

and no conversion of the tethered chloride was observed. Thus,

it was suggested that heterogeneous sites were responsible for

the aryl chloride activity. However, this analysis based on pre-

vious data needs to be reexamined in light of our new results.

The earlier analysis neglects the role that the immobilization

of palladium plays in both protecting palladium from aggre-

gation and controlling how palladium is released into solution

(e.g., at what rate, what temperature), making the trace sol-

uble Pd(OAc)₂test difﬁcult to compare with reactions using

immobilized palladium. If the palladium leached from Pd-SH-

SBA-15 is from oxidative addition of aryl halide, then the initial

homogeneous source of Pd(II) into solution will be different

from homogeneous palladium acetate, which is important with

respect to reduction and deactivation pathways. Also, it is pos-

sible that the trace amount of homogeneous palladium acetate

was deactivated during the heating of the reaction solution be-

fore catalysis could occur, whereas in an immobilized state, the

palladium might remain dormant until the temperature is high

enough to promote both the controlled leaching and subsequent

catalysis. In addition, this analysis does not consider the role

that tethering of a reactant in the three-phase test plays on its

reactivity.

ported aryl halide could slow the reaction of the immobilized

reagent until most of the soluble boronic acid component is con-

sumed by both Suzuki coupling and self-coupling, but even then

the palladium could be inaccessible due to Pd–Pd agglomera-

tion or redeposition. Therefore, whereas the three-phase data

clearly show the presence of leached, active palladium, they

do not conclusively show, in our eyes, that most of catalysis

is occurring on the surface. As demonstrated in this study, the

ability of SH-SBA-15 to poison catalysis under similar condi-

tions provides strong evidence that the catalysis of the soluble

component is more consistently interpreted, in light of the com-

bined data of Shimizu, Crudden, Davis and the present work, as

resulting from leached palladium.

3.7. Effect of S:Pd ratio on activity

One additional variable warranting investigation is the role

of the S:Pd ratio. Previous reports have suggested that this ratio

is important to the observed reactivity when using solid pre-

catalysts. In 1988, Wang and Liu reported their ﬁndings that

macromolecular palladium chelates exhibited no activity in the

Heck coupling of iodobenzene with ethyl acrylate when a S:Pd

ratio of ꢀ6 was used [29]. Reducing the ratio to 3:1 produced a

moderate increase in activity, and further reduction to 2:1 gave

the highest yields. These ﬁndings highlight the importance of

the S:Pd ratio for catalyst activity and demonstrate that overco-

ordination by sulfur can slow or quench catalysis. It is important

to note that these ligands were ﬁrst complexed to palladium be-

fore immobilization and thus do not provide any insight into

how immobilized thiols coordinate to free palladium.

In 2005, Crudden synthesized two different mercaptopropyl-

modiﬁed SBA-15 supports. The ﬁrst of these was made by post-

grafting the mercaptopropyltrialkoxysilane with the preformed

SBA-15. Out of 10 batches, only 3 exhibited any level of cat-

alytic activity. Reasons for the lack of activity were not known

at that time. The S:Pd ratio was not reported, but, based on the

reported thiol loading and assuming that all available palladium

acetate was bound to the surface during metalation, a ratio of

8.8 was calculated. A possible explanation for the lack of ac-

tivity is that the large excess of available sulfur sites could con-

ceivably bind leached palladium (or prevent any leaching) and

self-quenches the catalysis. Indeed, when we reduced our palla-

dium content of Pd-SH-SBA-15 from 11.06 wt% (S:Pd = 1.7)

to 8.45 wt% (S:Pd = 2.6) and 4.6 wt% (S:Pd = 5.3) we ob-

served a decrease in the Heck coupling of bromoacetophenone

with butyl acrylate with each successive increase in the S:Pd ra-

tio (Table 6). For the Heck coupling of iodobenzene, an increase

in S:Pd ratio from 1.8 to 5.3 resulted in a small amount of Heck

coupling of iodobenzene for one run (22% after 1300 min)

and no activity for two replicates (standard conditions) (Ta-

ble 6). This catalyst also was inactive for a Heck coupling of

bromoacetophenone with n-butyl acrylate (standard conditions

with data taken to 24 h). Decreases in activity resulting from

increases in the S:Pd ratio were also observed for Suzuki reac-

tions of bromoacetophenone and phenylboronic acid. To probe

the maximum capacity of SH-SBA-15 to quench catalysis by

homogeneous palladium acetate, a reaction was performed in

The ability of aryl halides to stabilize solubilized palladium

increases according to the ease of oxidative addition (I > Br >

Cl). With regards to the three-phase test, the ease of oxidative

addition and the availability of aryl bromide may play a crucial

role not only in promoting leaching, but also in preventing pal-

ladium deactivation and in inhibiting redeposition. In addition,

the rate of reaction of supported aryl halide may be signiﬁcantly

slower than the homogeneous counterpart for steric reasons. As

noted by Crudden, the aryl bromide conversion data show quite

clearly that there is at least some active leached metal. The rise

in yield of supported aryl halide with time could alternatively

be suggestive of a slow rate of reaction resulting from its im-

mobilization on surface. Unfavorable competition for the small

amount of homogeneous palladium between soluble and sup-

92

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

Table 6

of palladium remained in solution after reaction [60]. These

a

Effect of S:Pd ratio on activity of Pd-SH-SBA-15 precatalysts

low levels of palladium leaching suggest that either the leached

species are very active (i.e., if all conversions are catalyzed by

the amount of palladium in solution after the reaction), or that

a redeposition process is at play, whereby larger amounts of

leached palladium are active during reaction, with some or most

of the palladium redepositing on the SH-SBA-15 surface af-

ter reaction. After recovery, this redeposited palladium may be

available for further reaction, as was similarly found for other

immobilized palladium precatalysts [16,52,61,62]. However, if

the S:Pd ratio is >2:1, then palladium capture and overco-

ordination by excess surface thiols may result in a partial or

complete deactivation of catalysts, as was observed for partially

metalated SH-SBA-15 surfaces (vide supra). Either way, the

amount of palladium in the products after reaction appears to

be sufﬁcient to make the precatalyst useful for producing nearly

palladium-free Heck and Suzuki coupling products.

S:Pd

Halide

Reaction

Solvent

Conver-

sion (%)

Time

(min)

1.8

Iodobenzene

Heck

Suzuki

DMF

100

93

100

80

130

165

480

360

Bromoacetophenone

H O

2

b

2.6

Bromoacetophenone

Heck

Suzuki

DMF

H O

2

14

42

31

1130

360

5.3

a

Iodobenzene

Bromoacetophenone

Heck

DMF

0, 0, 22

0

1300

1440

Experiments conducted under normal Heck and Suzuki reaction conditions.

Addition of 35 equivalents of SH-SBA-15 poison resulted in no activity

b

under similar Heck and Suzuki conditions when using a Pd-SH-SBA-15 pre-

catalyst with a S:Pd ratio of 2.6.

A very recent report¹of thiol-supported palladium precata-

lysts also suggests only traces of palladium are found in solu-

tion after reaction [63]. Although the nature of the true catalytic

species was not addressed in this work, our results presented

here strongly suggest that this system also is simply a precata-

lyst that liberates soluble catalytic species in solution.

4. Conclusion

In summary, a mercaptopropyl-modiﬁed mesoporous silica,

SH-SBA-15, was successfully synthesized and used as a solid

support for immobilization of palladium acetate, Pd-SH-SBA-

15. The metalated material was used as a catalyst for both Heck

and Suzuki reactions under a range of conditions. Metal-free

SH-SBA-15 was shown to be an effective poison of both homo-

geneous palladium acetate and Pd-SH-SBA-15 used as precat-

alysts for Heck and Suzuki reactions. Therefore, we assert Pd-

SH-SBA-15 is simply a reservoir of leached palladium, which

is in agreement with previous ﬁndings by Davis et al. [40] with

regard to Heck reactions using iodobenzene and suggests that

previous analysis of heterogeneous catalysis with these precata-

lysts in Heck and Suzuki couplings may need to be reinterpreted

in light of the new data presented here [33,36].

This work introduces SH-SBA-15 as a new selective poison

for elucidating solution versus surface catalysis by palladium

and demonstrates its use as a more effective and versatile poison

than poly(4-vinylpyridine) and Quadrapure TU. Thus, selective

poisons tethered onto silica substrates, such as SH-SBA-15 and

SH–SiO₂, are a better class of materials for the selective poison-

ing of palladium than are poisons tethered to insoluble organic

◦

Fig. 7. Plot of iodobenzene conversion (DMF, 90 C, NEt ) with SH-SBA-15

3

used as a poison with a 0.1 mL dose of 0.063 mmol/L of Pd(OAc) in DMF

added at time zero (S:Pd = 7.3) followed by additions of 0.02 mL added at

2

90 min (S:Pd = 6.1) and 180 min (S:Pd = 5.2).

which palladium acetate was dosed into a reaction containing

SH-SBA-15 (Fig. 7). Activity was not observed until the S:Pd

ratio was decreased to 5.2, after which rapid conversion fol-

lowed by a quenching of catalysis was observed. This cessation

is likely a reﬂection of the rate at which the partially metalated

SH-SBA-15 is able to bind additional palladium against the rate

of catalysis by the unbound palladium.

3.8. Utility of Pd-SH-SBA-15 as a practical Heck and Suzuki

precatalyst

Although these new data suggesting catalysis by leached

species speak to nature of the true catalytic species, they do not

address the functional utility of these precatalysts. The previ-

ous data on Suzuki couplings promoted by catalysts of this type

strongly suggest that the amount of palladium in solution after

reaction is quite small. Shimizu found that after reaction, the

palladium levels in solution for Pd-SH-FSM were 0.5 ppm for

Suzuki reactions and 0.1 ppm for Heck reactions of activated

bromides [33]. Crudden reported that for Suzuki reactions us-

ing Pd-SH-SBA-15, the palladium leaching levels ranged from

0.003 to 0.75 ppm, depending on reaction conditions, and that

for Heck coupling of bromoacetophenone with styrene, the

palladium in solution was 0.27 ppm [36]. However, the total

amount of palladium leached was generally <0.2 ppm at a S:Pd

ratio of 2:1. Another recent report also suggests that only traces

1

Note that the authors measured palladium solution content after passing

the reaction solution through celite, washing with water, which was discarded,

and drying the organic with MgSO . This procedure was most likely chosen

4

to mimic procedures used in the pharmaceutical industry and thus gives a good

representation of how much metal contamination would be found in the organic

product stream. However, the used workup procedures could have removed a

signiﬁcant amount of soluble metal from the reaction solution through depo-

sition of palladium metal on celite and dissolution in the aqueous phase. Thus

palladium levels of the ﬁnal workup are not reﬂective of palladium levels in the

ﬁnal reaction solution.

J.M. Richardson, C.W. Jones / Journal of Catalysis 251 (2007) 80–93

93

polymers. The ability of insoluble, selective poisons to poison

homogeneous palladium is dependent on the ratio of binding

sites to soluble palladium. Therefore, control reactions in which

known amounts of homogeneous palladium complexes are poi-

soned should be performed ﬁrst to (i) verify poisoning ability

and (ii) determine the proper amount of selective poison re-

quired to quench catalysis if all available immobilized Pd were

to leach into solution. The three-phase test can provide positive

conﬁrmation of catalysis from leached Pd, but lack of activity

may not conﬁrm strictly heterogeneous catalysis. Quenching of

palladium-catalyzed reactions by SH-SBA-15 suggests that the

active catalyst comprises at most a few atoms of palladium and

reactivity is not from catalysis on palladium nanoparticle sur-

faces.

[25] W.J. Sommer, K.Q. Yu, J.S. Sears, Y.Y. Ji, X.L. Zheng, R.J. Davis,

C.D. Sherrill, C.W. Jones, M. Weck, Organometallics 24 (2005) 4351.

[26] K.Q. Yu, W. Sommer, M. Weck, C.W. Jones, J. Catal. 226 (2004) 101.

[27] M. Weck, C.W. Jones, Inorg. Chem. 46 (2007) 1865.

[28] X.A. Li, H.F. Liu, Y.Y. Jiang, J. Mol. Catal. 39 (1987) 55.

[29] Y.A. Wang, H.F. Liu, J. Mol. Catal. 45 (1988) 127.

[30] M.Z. Cai, C.S. Song, X. Huang, Synthesis (1997) 521.

[31] M.Z. Cai, C.S. Song, X. Huang, Synth. Commun. 27 (1997) 1935.

[32] B.M. Choudary, N.S. Chowdari, K. Jyothi, N.S. Kumar, M.L. Kantam,

Chem. Commun. (2002) 586.

[33] K. Shimizu, S. Koizumi, T. Hatamachi, H. Yoshida, S. Komai, T. Kodama,

Y. Kitayama, J. Catal. 228 (2004) 141.

[34] F.Y. Zhao, B.M. Bhanage, M. Shirai, M. Arai, Chem. Eur. J. 6 (2000) 843.

[35] S.S. Prockl, W. Kleist, M.A. Gruber, K. Kohler, Angew. Chem. Int. Ed. 43

(2004) 1881.

[36] C.M. Crudden, M. Sateesh, R. Lewis, J. Am. Chem. Soc. 127 (2005)

10045.

[37] J. Rebek, F. Gavina, J. Am. Chem. Soc. 96 (1974) 7112.

[38] J.P. Collman, K.M. Kosydar, M. Bressan, W. Lamanna, T. Garrett, J. Am.

Chem. Soc. 106 (1984) 2569.

Acknowledgments

[39] C.M. Crudden, K. McEleney, S.L. MacQuarrie, A. Blanc, M. Sateesh,

J.D. Webb, Pure Appl. Chem. 79 (2007) 247.

[40] Y.Y. Ji, S. Jain, R.J. Davis, J. Phys. Chem. B 109 (2005) 17232.

[41] J.G. de Vries, Dalton Trans. (2006) 421.

The US DOE Ofﬁce of Basic Energy Science is acknowl-

edged for ﬁnancial support through Catalysis Contract No. DE-

FG02-03ER15459.

[42] A. Alimardanov, L.S.V. de Vondervoort, A.H.M. de Vries, J.G. de Vries,

Adv. Synth. Catal. 346 (2004) 1812.

[43] G.P.F. van Strijdonck, M.D.K. Boele, P.C.J. Kamer, J.G. de Vries, P. van

Leeuwen, Eur. J. Inorg. Chem. (1999) 1073.

[44] M.B. Thathagar, J.E. ten Elshof, G. Rothenberg, Angew. Chem. Int. Ed. 45

(2006) 2886.

[45] S.G. Fiddy, J. Evans, T. Neisius, M.A. Newton, N. Tsoureas, A.A.D. Tul-

loch, A.A. Danopoulos, Chem. Eur. J. 13 (2007) 3652.

[46] J.V. Nguyen, C.W. Jones, Macromolecules 37 (2004) 1190.

[47] B.J. Burger, J.E. Bercaw, ACS Symp. Ser. 357 (1987) 79.

[48] A.H.M. de Vries, F.J. Parlevliet, L.S. van de Vondervoort, J.H.M. Mom-

mers, H.J.W. Henderickx, M.A.M. Walet, J.G. de Vries, Adv. Synth.

Catal. 344 (2002) 996.

[49] I.P. Beletskaya, A.V. Cheprakov, J. Organomet. Chem. 689 (2004) 4055.

[50] R.A. Sheldon, M. Wallau, I. Arends, U. Schuchardt, Acc. Chem. Res. 31

(1998) 485.

References

[1] R.F. Heck, J.P. Nolley, J. Org. Chem. 37 (1972) 2320.

[2] N. Miyaura, K. Yamada, A. Suzuki, Tetrahedron Lett. (1979) 3437.

[3] C.E. Garrett, K. Prasad, Adv. Synth. Catal. 346 (2004) 889.

[4] V. Farina, Adv. Synth. Catal. 346 (2004) 1553.

[5] N.T.S. Phan, M. Van Der Sluys, C.W. Jones, Adv. Synth. Catal. 348 (2006)

609.

[6] A.F. Shmidt, L.V. Mametova, Kinet. Catal. 37 (1996) 406.

[7] F.Y. Zhao, M. Arai, React. Kinet. Catal. Lett. 81 (2004) 281.

[8] K. Kohler, R.G. Heidenreich, J.G.E. Krauter, M. Pietsch, Chem. Eur. J. 8

(2002) 622.

[9] J.M. Richardson, C.W. Jones, Adv. Synth. Catal. 348 (2006) 1207.

[10] S.J. Broadwater, D.T. McQuade, J. Org. Chem. 71 (2006) 2131.

[11] A. Bifﬁs, M. Zecca, M. Basato, J. Mol. Catal. A Chem. 173 (2001) 249.

[12] K.Q. Yu, W. Sommer, J.M. Richardson, M. Weck, C.W. Jones, Adv. Synth.

Catal. 347 (2005) 161.

[51] B.H. Lipshutz, S. Tasler, W. Chrisman, B. Spliethoff, B. Tesche, J. Org.

Chem. 68 (2002) 1177.

[52] W. Solodenko, K. Mennecke, C. Vogt, S. Gruhl, A. Kirschning, Synthesis

(2006) 1873.

[53] T. Kang, Y. Park, J.C. Park, Y.S. Cho, J. Yi, Stud. Surf. Sci. Catal. 146

(2003) 527.

[13] A. Corma, D. Das, H. Garcia, A. Leyva, J. Catal. 229 (2005) 322.

[14] C.C. Cassol, A.P. Umpierre, G. Machado, S.I. Wolke, J. Dupont, J. Am.

Chem. Soc. 127 (2005) 3298.

[15] M. Nowotny, U. Hanefeld, H. van Koningsveld, T. Maschmeyer, Chem.

Commun. (2000) 1877.

[54] T. Kang, Y. Park, J. Yi, Ind. Eng. Chem. Res. 43 (2004) 1478.

[55] N. Galaffu, S.P. Man, R.D. Wilkes, J.R.H. Wilson, Org. Process Res.

Dev. 11 (2007) 406.

[56] X. Feng, G.E. Fryxell, L.Q. Wang, A.Y. Kim, J. Liu, K.M. Kemner, Sci-

ence 276 (1997) 923.

[57] C. Gurtler, S.L. Buchwald, Chem. Eur. J. 5 (1999) 3107.

[58] A.F. Littke, G.C. Fu, Angew. Chem. Int. Ed. 37 (1998) 3387.

[59] W. Sommer, M. Weck, Adv. Synth. Catal. 348 (2006) 2101.

[60] J.D. Webb, S.L. MacQuarrie, K. McEleney, C.M. Crudden, (2007), per-

sonal communication.

[61] S.S. Prockl, W. Kleist, K. Kohler, Tetrahedron 61 (2005) 9855.

[62] A.F. Shmidt, V.V. Smirnov, O.V. Starikova, A.V. Elaev, Kinet. Catal. 42

(2001) 199.

[16] K. Kohler, W. Kleist, S.S. Prockl, Inorg. Chem. 46 (2007) 1876.

[17] A. Bifﬁs, M. Zecca, M. Basato, Eur. J. Inorg. Chem. (2001) 1131.

[18] I.P. Beletskaya, A.V. Cheprakov, Chem. Rev. 100 (2000) 3009.

[19] A.S. Gruber, D. Pozebon, A.L. Monteiro, J. Dupont, Tetrahedron Lett. 42

(2001) 7345.

[20] A.H.M. de Vries, J. Mulders, J.H.M. Mommers, H.J.W. Henderickx,

J.G. de Vries, Org. Lett. 5 (2003) 3285.

[21] R.K. Arvela, N.E. Leadbeater, M.S. Sangi, V.A. Williams, P. Granados,

R.D. Singer, J. Org. Chem. 70 (2005) 161.

[22] F.Y. Zhao, K. Murakami, M. Shirai, M. Arai, J. Catal. 194 (2000) 479.

[23] F.Y. Zhao, M. Shirai, Y. Ikushima, M. Arai, J. Mol. Catal. A Chem. 180

(2002) 211.

[63] X. Jiang, J. Sclafani, K. Prasad, O. Repic, T.J. Blacklock, Org. Process

Res. Dev. 11 (2007) 769.

[24] A.F. Schmidt, V.V. Smirnov, J. Mol. Catal. A Chem. 203 (2003) 75.

Article Doi

Strong evidence of solution-phase catalysis associated with palladium leaching from immobilized thiols during Heck and Suzuki coupling of aryl iodides, bromides, and chlorides

DOI: 10.1016/j.jcat.2007.07.005

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.jcat.2007.07.005

Full text of DOI:10.1016/j.jcat.2007.07.005

Products guided by the article

R&D Labs maybe for 173464-57-8

Relevant to this article

Hot Product