Sterically demanding, sulfonated, triarylphosphines: Application to palladium-catalyzed cross-coupling, steric and electronic properties, and coordination chemistry

576

Organometallics 2008, 27, 576–593

Sterically Demanding, Sulfonated, Triarylphosphines: Application to

Palladium-Catalyzed Cross-Coupling, Steric and Electronic

Properties, and Coordination Chemistry

Lucas R. Moore, Elizabeth C. Western, Raluca Craciun, Jason M. Spruell, David A. Dixon,*

Kevin P. O’Halloran, and Kevin H. Shaughnessy*

Department of Chemistry and Center for Green Manufacturing, The UniVersity of Alabama,

Tuscaloosa, Alabama 35487-0336

ReceiVed August 28, 2007

Tri(2,4-dimethyl-5-sulfonatophenyl)phosphine trisodium (TXPTS · Na₃) and tri(4-methoxy-2-methyl-

5-sulfonatophenyl)phosphine trisodium (TMAPTS · Na₃) both provide more active catalysts for Suzuki

and Sonogashira couplings of aryl bromides in aqueous solvents than tri(3-sulfonatophenyl)phosphine

trisodium (TPPTS · Na₃). In the Heck coupling, TXPTS · Na₃provides the most effective catalyst system.

Cone angles determined from DFT-optimized structures show that both TXPTS · Na₃(206°) and

TMAPTS · Na₃(208°) are signiﬁcantly larger than TPPTS · Na₃(165°). The identity of the counterion

had a signiﬁcant effect on the calculated cone angles for these ligands. The electronic properties of these

ligands determined by the CO stretching frequencies of trans-RhL₂(Cl)CO complexes were identical,

although calculated electronic parameters suggest subtle differences between these ligands. Similar to

TPPTS · Na₃, both TXPTS · Na₃and TMAPTS · Na₃react with Pd(OAc)₂in aqueous solvents to give

L_nPd⁰complexes and the corresponding phosphine oxide. The reduction of palladium(II) by TXPTS · Na₃

is signiﬁcantly slower than is seen with TMAPTS · Na₃or TPPTS · Na₃at room temperature. Evidence of

palladacycle complexes derived from TXPTS · Na₃and TMAPTS · Na₃by activation of an ortho-methyl

substituent was also observed in ligand coordination studies and under catalytic reaction conditions.

Introduction

Water is an attractive solvent for organic synthesis due to its

low toxicity, nonﬂammability, low cost, and limited environ-

mental impact.^1,2In addition, the use of an aqueous–organic

biphasic solvent system in conjunction with a water-soluble

catalyst provides a route to recover and recycle the catalyst from

organic product streams. Water-soluble catalyst systems also

provide efﬁcient methods to modify hydrophilic substrates, such

as nucleosides and peptides.^3–8A variety of metal-catalyzed

reactions can be successfully run in aqueous solvent systems,

with palladium-catalyzed cross-couplings being one of the most

widely developed.^9,10The ﬁrst example of aqueous-phase,

palladium-catalyzed cross-coupling was reported by Casal-

nuovo,¹¹who used Pd(TPPMS · Na)₃(TPPMS · Na ) diphen-

yl(3-sulfonatophenyl)phosphine sodium, Figure 1) as the cata-

Figure 1. Sulfonated triarylphosphines.

lyst. This catalyst system gave good yields with aryl iodides

and a few examples of aryl bromides, but required high catalyst

loadings. Since Casalnuovo’s initial report, a wide variety of

water-soluble phosphine ligands have been developed.⁹Like

TPPMS · Na, nearly all of these ligands are triarylphosphines

substituted with a variety of anionic, cationic, or neutral

hydrophilic substituents. Catalysts derived from these ligands

generally give good yields with aryl iodides and are effective

in some cases with aryl bromides at elevated temperatures.

* Corresponding authors. E-mail: dadixon@bama.ua.edu; kshaughn@

bama.ua.edu.

(1) Li, C.-J.; Chen, L. Chem. Soc. ReV. 2006, 35, 68–82.

(2) Li, C.-J. Chem. ReV. 2005, 105, 3095–3165.

(3) Western, E. C.; Daft, J. R.; Johnson, E. M., II; Gannett, P. M.;

Shaughnessy, K. H. J. Org. Chem. 2003, 68, 6767–6774.

Over the past decade, sterically demanding and electron-rich

ligands have been shown to give highly effective catalysts for

cross-coupling reactions under mild conditions. Bulky, electron-

rich ligands, such as tri-tert-butylphosphine,^12–152-biphenyldi-

alkylphosphines,^16–19and N-heterocyclic carbenes,^20–22provide

ˇ

(4) Capek, P.; Pohl, R.; Hocek, M. Org. Biomol. Chem. 2006, 4, 2278–

2284.

(5) Collier, A.; Wagner, G. Org. Biomol. Chem. 2006, 4, 4526–4532.

(6) Dibowski, H.; Schmidtchen, F. P. Angew. Chem., Int. Ed. 1998, 37,

476–478.

(7) Dibowski, H.; Schmidtchen, F. P. Tetrahedron Lett. 1998, 39, 525–

(12) Hartwig, J. F.; Kawatsura, M.; Hauck, S. I.; Shaughnessy, K. H.;

Alcazar-Roman, L. M. J. Org. Chem. 1999, 64, 5575–5580.

(13) Littke, A. F.; Dai, C.; Fu, G. C. J. Am. Chem. Soc. 2000, 122,

4020–4028.

528.

(8) Thoresen, L. H.; Jiao, G.-S.; Haaland, W. C.; Metzker, M. L.;

Burgess, K. Chem.—Eur. J. 2003, 9, 4603–4610.

(9) Shaughnessy, K. H.; DeVasher, R. B. Curr. Org. Chem. 2005, 9,

585–604.

(14) Littke, A. F.; Fu, G. C. J. Am. Chem. Soc. 2001, 123, 6989–7000.

(15) Littke, A. F.; Schwarz, L.; Fu, G. C. J. Am. Chem. Soc. 2003, 124,

6343–6348.

(10) Shaughnessy, K. H. Eur. J. Org. Chem. 2006, 1827–1835.

(11) Casalnuovo, A. L.; Calabrese, J. C. J. Am. Chem. Soc. 1990, 112,

4324–4330.

(16) Burgos, C. H.; Barder, T. E.; Huang, X.; Buchwald, S. L. Angew.

Chem., Int. Ed. 2006, 45, 4321–4326.

10.1021/om7008606 CCC: $40.75

2008 American Chemical Society

Publication on Web 02/02/2008

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 577

active catalysts for coupling of aryl bromides and chlorides at

ambient temperatures in many cases. Steric bulk is necessary

to promote the formation of coordinatively unsaturated pal-

ladium species, which are the presumed catalytically active

species. Electron-rich ligands help to promote oxidative addition,

which is particularly important when using less reactive aryl

bromide or chloride substrates. Ligand size appears to be the

critical factor in determining catalyst activity in cross-coupling

reactions of aryl bromides, while electron-donating ability

becomes more important with the less reactive aryl chlorides.^23,24

Only recently have hydrophilic ligands with large steric

demand and strong electron-donating ability been applied to

aqueous-phase cross-coupling reactions. Hydrophilic trialkyl-

phosphines^25,26and 2-phosphinobiphenyl ligands^27–29have been

shown to give highly effective catalysts for cross-coupling

reactions of aryl bromides and chlorides. Dialkylphospinous

acids have also been applied to aqueous-phase cross-coupling

of aryl bromides and chlorides, although at elevated tempera-

tures.^30–34Tri-(2,4-dimethyl-5-sulfonatophenyl)phosphine tri-

sodium (TXPTS · Na₃), a sterically demanding analogue of

TPPTS · Na₃, was ﬁrst prepared by Bakos and co-workers.³⁵We

have previously communicated the use of TXPTS · Na₃and

TMAPTS · Na₃(tri-(4-methoxy-2-methyl-5-sulfonatophenyl)phos-

phine trisodium) in aqueous Heck and Suzuki couplings.³⁶

Herein we report the utility of TXPTS · Na₃and TMAPTS · Na₃

as ligands in palladium-catalyzed coupling reactions, their steric

and electronic properties, and the coordination chemistry of these

ligands with palladium.

prepared by condensation of the corresponding Grignard

reagents with PCl₃, with fuming sulfuric acid at room temper-

ature. The electron-releasing substituents on TXP and TMAP

make these phosphines more readily sulfonated than tri-

phenylphosphine, which requires long reaction times and/or

harsh conditions to achieve complete sulfonation.³⁷These harsh

reaction conditions result in competitive oxidation of the

phosphorus center during the sulfonation of triphenylphosphine.

In contrast, TXP and TMAP can be readily sulfonated in a few

hours at room temperature to give the desired ligands in good

yield with little oxide impurity.³⁸

Catalytic Applications. TXPTS · Na₃and TMAPTS · Na₃

both gave effective catalysts for the Suzuki coupling of aryl

bromides at 50 °C.³⁶Catalysts derived from these ligands were

moderately more active than the catalyst derived from

TPPTS · Na₃. Using 0.05 mol % of a catalyst derived from the

ligand and Pd(OAc)₂(2.5:1 L:Pd) for the coupling of 4-bro-

motoluene and phenylboronic acid, the catalyst derived from

TXPTS · Na₃and TMAPTS · Na₃gave 78% and 68% yields,

respectively, as determined by GC after 2.5 h. The catalyst

derived from TPPTS · Na₃had given only a 48% yield over the

same time period. Pd(OAc)₂without ligand gave no conversion

under these conditions. The TXPTS · Na₃/Pd(OAc)₂catalyst

system also gave slow conversion for the coupling of 4-bro-

motoluene and phenyl boronic acid at room temperature,

requiring 70 h to give an 87% yield of product using 2.5 mol

% catalyst. Both the TXPTS · Na₃- and TMAPTS · Na₃-derived

catalysts gave excellent yields in the Suzuki coupling of a variety

of aryl bromides (eq 1, Table 1). Electron-deﬁcient and -neutral

aryl halides gave excellent yields at 50 °C (entries 1–5).

4-Methoxy-substituted aryl bromides also gave excellent yields

at 50 °C with TMAPTS · Na₃, but required 80 °C with

TXPTS · Na₃. Couplings to give ortho-substituted biphenyls gave

good yields at 50 °C, but the yields could be improved by

running the reaction at 80 °C (i.e., entries 10, 11).

Results

35

Ligand Synthesis. The syntheses of TXPTS · Na₃and

36

TMAPTS · Na₃have previously been reported. Both ligands

are prepared by sulfonation of the corresponding phosphines

(tri-(2,4-dimethylphenyl)phosphine (TXP) and tri(4-methoxy-

2-methylphenyl)phosphine (TMAP)), which can be readily

(17) Barder, T. E.; Walker, S. D.; Martinelli, J. R.; Buchwald, S. L.

J. Am. Chem. Soc. 2005, 127, 4685–4696.

(18) Walker, S. D.; Barder, T. E.; Martinelli, J. R.; Buchwald, S. L.

Angew. Chem., Int. Ed. 2004, 43, 1871–1876.

(19) Milne, J. E.; Buchwald, S. L. J. Am. Chem. Soc. 2004, 126, 13028–

13032.

(1)

(20) Viciu, M. S.; Navarro, O.; Germaneau, R. F.; Kelly, R. A., III;

Sommer, W.; Marion, N.; Stevens, E. D.; Cavallo, L.; Nolan, S. P.

Organometallics 2004, 23, 1629–1635.

(21) Hadei, N.; Kantchev, E. A. B.; O’Brien, C. J.; Organ, M. G. Org.

Lett. 2005, 7, 3805–3807.

(22) Marion, N.; Navarro, O.; Mei, J.; Stevens, E. D.; Scott, N. M.;

Nolan, S. P. J. Am. Chem. Soc. 2006, 128, 4101–4111.

(23) DeVasher, R. B.; Spruell, J. M.; Dixon, D. A.; Broker, G. A.;

Grifﬁn, S. T.; Rogers, R. D.; Shaughnessy, K. H. Organometallics 2005,

24, 962–971.

Catalysts derived from TXPTS · Na₃and TMAPTS · Na₃also

proved to be superior to the catalyst derived from TPPTS · Na₃

for the Sonogashira coupling of aryl bromides at 50 °C. In the

coupling of 4-bromotoluene and phenylacetylene, the catalysts

derived from TXPTS · Na₃and TMAPTS · Na₃gave 81% and

92% yield, respectively (3 mol % Pd(OAc)₂, 8 mol % ligand,

5 h). The TPPTS · Na₃-derived catalyst gave only a 47% yield

of diphenylacetylene under identical conditions. In the absence

of ligand, no product formation occurred. Signiﬁcantly, the

reaction proceeded in the absence of copper cocatalyst. In fact,

the use of CuI (2 mol %) resulted in lower yields due to

competitive homodimerization of phenylacetylene. Excellent

yields were obtained in the coupling of electron-deﬁcient aryl

halides with phenylacetylene using both TXPTS · Na₃and

(24) Hill, L. L.; Moore, L. R.; Huang, R.; Craciun, R.; Vincent, A. P.;

Dixon, D. A.; Chou, J.; Woltermann, C. J.; Shaughnessy, K. H. J. Org.

Chem. 2006, 71, 5117–5125.

(25) DeVasher, R. B.; Moore, L. R.; Shaughnessy, K. H. J. Org. Chem.

2004, 69, 7919–7927.

(26) Shaughnessy, K. H.; Booth, R. S. Org. Lett. 2001, 3, 2757–2759.

(27) Nishimura, M.; Ueda, M.; Miyaura, N. Tetrahedron 2002, 58, 5779–

5787.

(28) Konovets, A.; Penciu, A.; Framery, E.; Percina, N.; Goux-Henry,

C.; Sinou, D. Tetrahedron Lett. 2005, 46, 3205–3208.

(29) Anderson, K. W.; Buchwald, S. L. Angew. Chem., Int. Ed. 2005,

44, 6175–6177.

(30) Wolf, C.; Lerebours, R. J. Org. Chem. 2003, 68, 7551–7554.

(31) Wolf, C.; Lerebours, R. Org. Lett. 2004, 6, 1147–1150.

(32) Wolf, C.; Lerebours, R. Org. Biomol. Chem. 2004, 2, 2161–2164.

(33) Lerebours, R.; Wolf, C. Synthesis 2005, 2287–2292.

(34) Wolf, C.; Ekoue-Kovi, K. Eur. J. Org. Chem. 2006, 1917–1925.

(35) Gulyás, H.; Szöllösy, Á.; Hanson, B. E.; Bakos, J. Tetrahedron

Lett. 2002, 43, 2543–2546.

(37) Bartik, T.; Bartik, B.; Hanson, B. E.; Glass, T.; Bebout, W. Inorg.

Chem. 1992, 31, 2667–2670.

(38) TPPTS · Na₃purchased from Strem Chemical has 10–15% oxide

impurity, while TXPTS · Na₃is available in 97% purity.

(36) Moore, L. R.; Shaughnessy, K. H. Org. Lett. 2004, 6, 225–228.

578 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

Table 1. Suzuki Coupling of Aryl Bromides Using TXPTS · Na₃and

TMAPTS · Na₃

(2)

2.5 mol % Pd(OAc)₂(2.5:1 L:Pd), the catalyst derived from

TXPTS · Na₃gave a 77% yield of 4-methylstilbene after 2 h.

The TMAPTS · Na₃-derived catalyst gave only a 40% yield,

while the catalyst formed from TPPTS · Na₃gave only a 28%

yield of the stilbene product. In the absence of ligand, no product

formation was observed. Good to excellent yields were obtained

from the coupling of aryl bromides and styrene or sodium

acrylate (eq 3, Table 3). Styrene consistently gave higher yields

of coupling products than did sodium acrylate. The choice of

aryl bromide did not have a signiﬁcant effect on the reaction

yield.

(3)

Mercury Poisoning Test. In an effort to determine whether

the active species was a soluble palladium/phosphine complex

or heterogeneous palladium particles, a mercury test was

performed.³⁹The rate of coupling of 4-bromotoluene and

phenylboronic acid catalyzed by TXPTS/Pd(OAc)₂(1.25 mol%

Pd, 2.5:1 L:Pd) in 1:1 acetonitrile/water at 50 °C in the presence

and absence of mercury were compared. In one sample, the

mercury (0.5 mL) was added to the dry reagents prior to the

addition of solvent and the aryl halide. These conditions test

the ability to form the active species in the presence of Hg⁰. In

a second sample, the reaction was allowed to proceed for 15

min to allow time for the active catalyst to form before the

mercury was added. Under these conditions, the stability of the

preformed active species in the presence of Hg⁰is determined.

Figure 2 shows the reaction proﬁles for the two mercury-

containing reactions and a control with no mercury. The reaction

proﬁles are similar for all three reactions. The mercury-

containing reactions have slightly lower conversions than the

control after 75 min, but the difference is small. Up to this point,

the reaction proﬁles are nearly identical. These results are

consistent with a homogeneous, molecular species being the

true catalyst rather than heterogeneous palladium particles or

colloids.

^aA ) TXPTS · Na₃, B ) TMAPTS · Na₃. ^bAverage isolated yield of

at least two runs. Reactions were run until judged complete by GC

(typically 2–4 h). Reaction times were not optimized.

Table 2. Sonogashira Coupling of Aryl Bromides Using TXPTS · Na₃

and TMAPTS · Na₃

DFT-Calculated Ligand Structures. In order to gain a fuller

understanding of why TXPTS · Na₃and TMAPTS · Na₃tend to

give more effective catalysts than TPPTS · Na₃, the steric and

electronic properties of these ligands were determined by

experimental and computational methods. We have previously

shown that the geometries and frequencies of transition metal

complexes^23,40can be predicted reliably at the local density

functional theory (LDFT) level.⁴¹Geometries of Pd⁰L₂and Pd⁰L

complexes were optimized at the LDFT level with the polarized

^aA ) TXPTS · Na₃, B ) TMAPTS · Na₃. ^bAverage isolated yield of

at least two runs. Reactions times were not optimized.

TMAPTS · Na₃(eq 2, Table 2). Bromotoluene derivatives gave

good yields of coupled product, although lower than those

achieved with activated aryl bromides. 4-Bromophenol, an

electron-rich aryl halide, was unreactive under these reaction

conditions. Attempted couplings with aliphatic alkynes, such

as 1-hexyne, were unsuccessful with this catalyst system in both

the presence and absence of catalytic CuI.

In the Heck coupling of aryl bromides, TXPTS · Na₃gave a

more active catalyst than either TMAPTS · Na₃or TPPTS · Na₃.

In the coupling of 4-bromotoluene and styrene at 80 °C using

(39) Widegren, J. A.; Finke, R. G. J. Mol. Catal. A: Chem. 2003, 198,

317–341.

(40) Sosa, C.; Andzelm, J.; Elkin, B. C.; Wimmer, E.; Dobbs, K. D.;

Dixon, D. A. J. Phys. Chem. 1992, 96, 6630–6636.

(41) Parr, R. G.; Yang, W. Density-Functional Theory of Atoms and

Molecules; Oxford University Press: New York, 1989.

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 579

palladium center. The structures of the H⁺and Na⁺adducts of

the trianionic ligands were calculated to compensate for the large

charge repulsions that would be present in the gas-phase

structures of the polyanions. In addition, structures of palladium

complexes of a number of neutral structures were calculated to

determine the effect of the sulfonate group on the ligand

geometries.

Table 3. Heck Coupling of Aryl Bromides Using TXPTS · Na₃

The structures of the PdL complexes are strongly affected

by the addition of Na⁺or H⁺to the sulfonate ion (Figure 3).

The structure of Pd(TPPTS^3-) has C-P-C and Pd-P-C bond

angles near 109°, showing a nearly perfect tetrahedral geometry

around the phosphorus atom (Figure 4, Table 4). The dihedral

angle between the Pd-P bond and the plane of the aromatic

ring (Pd-P-C-C) is -60°. This arrangement gives an open

structure that maximizes the distance between the negatively

charged sulfonate groups. The Pd-P bond length for the

trianionic complex is 2.149 Å. When the charge on the sulfonate

groups is balanced with a proton or sodium ion, the geometry

of the ligand changes signiﬁcantly. In the most stable sulfonic

acid structure, the sulfonic acid groups are held in close

proximity by intramolecular hydrogen bonding. The aryl rings

are pulled back away from the Pd, resulting in a larger Pd-C-P

angle compared to the trianionic complex. The Pd-P-C-C

torsion angle is decreased to -46°, which brings the aryl rings

closer to parallel with the Pd-P bond. The Pd-P bond of the

TPPTS · H₃is 2.114 Å, which is shorter than the trianion

structure. We attempted to optimize a structure without H-

bonding between the SO₃H groups, but it collapsed to the

H-bonded structure. When sodium counterions are added to each

sulfonate group, the structure again becomes more compact due

to each sodium ion coordinating to two sulfonate groups (Figure

3C). Again the Pd-P-C angle increases and two of the phenyl

rings rotate so that they are more parallel with the Pd-P bond

(Pd-P-C-C ) -14°, -32°, and -62°). The Pd-P bond of

the Pd(TPPTS · Na₃) complex is longer than that of the TPPTS^3-

complex by 0.035 Å.

^aAverage isolated yield of at least two runs. Reactions were run for

2 h and were not optimized. ^bCrude product was acidiﬁed with H₂SO₄

prior to isolation.

Figure 2. Reaction proﬁle in the mercury poisoning experiment in

the Suzuki coupling of 4-bromotoluene and phenylboronic acid

catalyzed by TXPTS/Pd(OAc)₂. Control with no Hg⁰(diamonds),

Hg⁰added initially (squares); Hg⁰added after 15 min (triangles).

The Pd-P bond distances for the TXPTS complexes are

longer than were observed for the TPPTS complexes, likely due

to the steric repulsion between the Pd center and the ortho-

methyl groups. The C-P-C angle of the TXPTS^3-structure

(109°) is similar to that of the TPPTS^3-complex, while the

torsion angle of the TXPTS^3-complex (-52°) is smaller than

that of Pd(TPPTS^3-). The sulfonic acid structure (TXPTS · H₃)

has strong hydrogen bonds between the SO₃H groups, leading

to the smallest C-P-C angle for this ligand and shortest Pd-P

bond distance (Figure 5B). Consequently the smallest torsion

angle for the TXPTS structures is found for the H-bonded

sulfonic acid structure. The sulfonic acid structure with no

hydrogen bonding between the sulfonic acid groups (Figure 5C)

is 7.8 kcal/mol higher in energy than the structure with hydrogen

bonding. In the free ligand, the H-bonded structure is 18.0 kcal/

mol more stable than the non-H-bonded structure. The Pd-P

bond length is longest for the non-hydrogen-bonded TXPTS · H₃

structure and, in fact, is the longest P-C bond predicted for

any of the structures. The structure with three Na⁺counterions

(Figure 5D) has a smaller CPC angle than the trianion, but there

is only a small decrease in the torsion angle compared to that

seen for TPPTS^3-. Again, the Pd-P bond becomes longer when

sodium ions are added to the sulfonate groups, although the

increase is less than was seen for the TPPTS system.

double-ꢀ DZVP2 basis set on all atoms except Pd.⁴²We used

the Stuttgart relativistic pseudopotential and the associated basis

set for Pd⁴³with 28 electrons in the core of the ECP. The basis

set for Pd was contracted to [6s5p3d]. The LDFT calculations

were done with the potential ﬁt of Vosko, Wilk, and Nusair for

the correlation functional⁴⁴and the exchange functional of

Slater.⁴⁵The calculations were done with the program Gauss-

ian03⁴⁶on Cray XD-1 and Silicon Graphics Altix computers at

the Alabama Supercomputing Center.

Four optimized geometries were calculated for each sul-

fonated ligand and their corresponding Pd complexes: the

trianionic ligand with no counterion (L^3-), the neutral trisulfonic

acid (L · H₃) with and without intramolecular H-bonding of the

SO₃H groups, and trisodium salt of the ligand (L · Na₃). Each

of the PdL complex calculations started with a zerovalent

(42) Godbout, N.; Salahub, D. R.; Andzelm, J.; Wimmer, E. Can.

J. Chem. 1992, 70, 560–571.

(43) Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H. J. Chem. Phys. 1994,

100, 7535–7542.

The Pd(TMAPTS^3-) structure is nearly identical to that of

Pd(TXPTS^3-), except that the Pd-P bond is shorter.⁴⁷Addition

(44) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200–

1211.

(45) Slater, J. C. Phys. ReV. 1951, 81, 385–390.

(46) Frisch, M. J.; et al. Gaussian 03, ReVision C.02; Gaussian Inc.,

2004.

(47) Models of the trianion, sulfonic acid, and sodium salts of the PdL

complexes are included in the Supporting Information.

580 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

Figure 3. Cylindrical bond models of calculated Pd(TPPTS^3-) (A), Pd(TPPTS · H₃) (B), and Pd(TPPTS · Na₃) (C) structures. Carbon, gray;

hydrogen, white; phosphorus, orange; sulfur, yellow; oxygen, red; palladium, blue; sodium, purple.

found previously, the bond energies for the compounds deﬁned

as BDE ) E_Pd+ E_L- E_PdLcover a small range between 32

and 38 kcal/mol except for the P(2,4-t-Bu₂C₆H₃)₃ligand (Table

4). The latter compound has a very large steric hindrance due

to the tert-butyl groups in the ortho-position that largely covers

the lone pair on the P, making it quite inaccessible to the Pd.

The low BDE is consistent with the large cone angle and the

long Pd-P bond. The TPPTS, TXPTS, and TMAPTS ligands

show some variations in the BDE due to the presence of

counterions. The largest effect is found for TXPTS · H₃, where

the difference in energies between the H-bonded and non-H-

bonded structures is about 10 kcal/mol. The Pd-P bond energy

is largest for PPh₃followed by the three anionic ligands. Alkyl

substituents on PPh₃lower the BDE by about 5 kcal/mol for

TMAP, TXP, and P(o-tol)₃with a lowering of about 11 kcal/

mol for P(Mes)₃. As noted above, the tert-butyl substituents in

P(2,4-t-Bu₂C₆H₃) lower the BDE by almost 30 kcal/mol.

Experimental and Calculated Steric and Electronic

Ligand Parameters. The steric demand of phosphines can be

expressed by a number of parameters.⁵¹The most commonly

used steric descriptor is the cone angle concept developed by

Tolman.⁵²Tolman utilized plastic CPK models in order to

determine his cone angle values. He assumed a static M-P bond

distance of 2.28 Å and placed the phosphine substituents to give

the minimum cone angle. Despite this somewhat simplistic

approach, the Tolman cone angles have proven to correlate well

with catalyst activity and coordination chemistry in a variety

of systems. Cone angles can also be obtained from solid-state

structures using methods developed by Mingos and co-work-

ers.⁵³Using this approach, Mingos was able to obtain cone

angles for a large number of structures that had been deposited

in crystallographic databases. This data provided a picture of

the range of cone angles for a given ligand for a variety of metal

complex types. Another approach to calculate cone angles of

phosphine ligands is the use of solid angles.^54,55An efﬁcient

Figure 4. Key geometry parameters summarized in Table 4.

of sodium ions results in a similar structural change to that seen

with the TPPTS system. The sodium ions pull the sulfonate

groups closer together as seen from the increase in the Pd-P-C

angle and decrease in the C-Pd-C angle. Two of the three

aryl rings show decreased torsional angles, although this

decrease is less dramatic than in the case of TPPTS. This

arrangement allows the sodium ions to each coordinate to two

sulfonate groups. Protonation of the sulfonate groups gives a

structure that is similar to the Pd(TPPTS · H₃) or Pd(TXPTS · H₃)

structures. The structure is again pulled back by strong

SO₃H-SO₃H hydrogen bonds. The Pd-P bond is shortest for

the hydrogen-bonded protonated ligand and longest for the

trianion as found for the other structures. The hydrogen-bonded

sulfonic acid structure is more stable than the non-H-bonded

structure when bonded to Pd by 5.5 kcal/mol. In the free ligand,

the H-bonded structure is more stable by 3.6 kcal/mol.

The introduction of ortho-methyl substituents in the ligands

results in signiﬁcant steric congestion around the metal. As a

result, TXPTS and TMAPTS have close contacts between

hydrogen atoms on each of the ortho-methyl groups and the

palladium center. The closest contacts ranged from 2.35 to 2.60

Å (Table 4). These contacts are smaller than the sum of the

van der Waals radii for H and Pd (2.89 Å), suggesting that there

are agostic interactions in these complexes. The neutral ana-

logues of these ligands (P(o-tol)₃, TXP, and TMAP) also have

close contacts of 2.35–2.36 Å, and trimesitylphosphine has an

even closer contact of 2.13 Å.

(51) White, D.; Coville, N. J. AdV. Organomet. Chem. 1994, 36, 95–

158.

(52) Tolman, C. A. Chem. ReV. 1977, 77, 313–348.

(53) Müller, T. E.; Mingos, D. M. P. Transition Met. Chem. 1995, 20,

533–539.

(54) Hirota, M.; Sakakibara, K.; Komatsuzaki, T.; Akai, I. Comp. Chem.

1991, 15, 241–248.

The bond dissociation energies were calculated at the LDA

geometries with the B3LYP exchange-correlation functional^48,49

and the polarized Ahlrichs triple-ꢀ basis set⁵⁰on the nonmetal

atoms and the above Stuttgart basis set and ECP on Pd. As

(55) White, D.; Taverner, B. C.; Leach, P. G. L.; Coville, N. J.

J. Comput. Chem. 1993, 14, 1042–1049.

(56) Taverner, B. C. J. Comput. Chem. 1996, 17, 1612–1623.

(57) Taverner, B. C., STERIC, 1.11, http://www.ccl.net/cca/software/

SOURCES/C/steric/index.shtml,1995.

(48) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.

(49) Lee, C.; Yang, W.; Parry, R. G. Phys. ReV. B 1988, 37, 785–789.

(50) Schäfer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571–

2577.

(58) Peng, Q.; Yang, Y.; Wang, C.; Liao, X.; Yuan, Y. Catal. Lett. 2003,

88, 219–225.

(59) Gulyás, H.; Bényei, A. C.; Bakos, J. Inorg. Chim. Acta 2004, 357,

3094–3098.

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 581

Table 4. Selected Geometrical Parameters for Calculated Pd⁰L Structures from DFT Calculations at the SVWN5/DGDZVP Level

a

b

f

complex

R_Pd-P(Å)

R_P-C(Å)

Pd-P-C^c(deg)

C-P-C^d(deg)

P-Pd-C-C^e(deg)

R_Pd-H(Å)

BDE^g(kcal/mol)

Pd(TPPTS^3-

)

2.149

2.114

2.184

2.182

2.151

2.242

2.208

2.165

2.141

2.159

2.209

2.132

2.160

2.163

2.166

2.294

2.268

1.823

1.833

1.834

1.817

1.833

1.857

1.827

1.823

1.832

1.817

1.831

1.824

1.818

1.816

1.820

1.821

1.842

111

119

118

110

118

115

118

110

118

114

118

116

114

108

115

108

98

100

109

100

104

100

109

99

-60

3.36

3.29

3.25

2.40

2.35

2.56

2.51

2.60

2.44

2.36

2.52

3.09

2.36

2.35

2.13

2.14

35.6

36.5

37.7

36.4

31.9

42.1

36.5

35.7

33.0

31.1

35.0

37.3

32.0

32.4

32.6

26.3

8.3

Pd(TPPTS · H₃)

-46

Pd(TPPTS · Na₃)

-62,-32,-14^h

-52

Pd(TXPTS^3-

)

Pd(TXPTS · H₃)

Pd(TXPTS · H₃)ⁱ

Pd(TXPTS · Na₃)

-39

-45

-50

Pd(TMAPTS^3-

)

-53

Pd(TMAPTS · H₃)

Pd(TMAPTS · H₃)ⁱ

Pd(TMAPTS · Na₃)

Pd(PPh₃)

-47

105

-47

102,98,98^h

102

-57,-50,-39^h

-38

Pd(P(o-tol)₃)

104

105

111

-48

Pd(TXP)^j

-48

Pd(TMAP)^k

-47

Pd(P(Mes)₃)^l

-45

Pd(P(2,4-t-Bu₂C₆H₃)₃)

103

-33

^aPd-P bond distance. ^bAverage P-C bond distance. ^cAverage Pd-P-C bond angle. ^dAverage C-P-C bond angle. ^eAverage torsion angle

between the Pd-P bond and the aryl ring plane. ^fClosest nonbonded Pd-H distances. ^gBond dissociation energy (E_Pd+ E_L- E_PdL^hIndividual values

)

for each aryl ring. ⁱNon-hydrogen-bonded structure. ^jTXP ) tri(2,4-dimethylphenyl)phosphine. ^kTMAP ) tri(4-methoxy-2-methylphenylphosphine).

^lMes ) mesityl.

Figure 5. Cylindrical bond models of calculated Pd(TXPTS^3-) (A), Pd(TXPTS · H₃) H-bonded (B), Pd(TXPTS · H₃) non-H-bonded (C),

and Pd(TXPTS · Na₃) (D) structures. Carbon, gray; hydrogen, white; phosphorus, orange; sulfur, yellow; oxygen, red; palladium, blue;

sodium, purple.

algorithm for calculating cone angles from solid angles has been

reported by Tavener.⁵⁶We have previously utilized Tavener’s

STERIC program,⁵⁷which utilizes the solid angle algorithm,

to calculate cone angle values based on DFT-optimized geome-

tries of Pd-phosphine complexes.^23,24In contrast to Tolman’s

method, cone angles determined by this method have optimized

Pd-P bond distances, and the phosphine substituents were

allowed to adopt their most stable conformation, rather than

the conformation leading to the smallest cone angle. The

resulting cone angles should provide a more realistic picture of

the true steric demand of ligands.

atom set at the origin.^56,57Previous estimates of the cone angle

for TPPTS have ranged from 145° to 178°. Tolman’s correlation

between ³¹P NMR chemical shift and cone angle was used to

estimate a value of 145° for TPPTS.³⁷A low-level HF/3-21G*

calculation of the cone angle predicted a value of 160°.⁵⁸Cone

angle values ranging from 152° to 178° have been determined

on the basis of X-ray crystal structures of free TPPTS⁵⁹and

iron and tungsten carbonyl complexes.^60,61Our calculated value

for TPPTS^3-is 155° (Table 5), while neutralization of the

(60) Darensbourg, D. J.; Bischoff, C. J.; Reibenspies, J. H. Inorg. Chem.

1991, 30, 1144–1147.

Cone angles (Table 5) for TPPTS, TXPTS, and TMAPTS

were calculated from the LDA-optimized geometries for the

Pd⁰(phosphine) complexes by using the STERIC program and

the volumetric parameters for the atoms therein with the Pd

(61) Darensbourg, D. J.; Bischoff, C. J. Inorg. Chem. 1993, 32, 47–53.

(62) Pickett, T. E.; Roca, F. X.; Richards, C. J. J. Org. Chem. 2003, 68,

2592–2899.

(63) Vastag, S.; Heil, B.; Markó, L. J. Mol. Catal. 1979, 5, 189–195.

582 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

Table 5. Experimental and Calculated Steric and Electronic Parameters for TPPTS, TXPTS, TMAPTS, and Related Nonsulfonated Ligands

free ligand

PdL complex

b

cone angle^a

(deg)

ν_CO

HOMO^d

(eV)

GAP^e

(eV)

HOMO^d

(eV)

GAP^e

(eV)

ligand

TPPTS^3-

TPPTS · H₃

TPPTS · Na₃

TXPTS^3-

(cm^-1

)

q(P)^c

155

164

165

210

219

209

206

194

214

215

208

172

214

215

230

280

0.18

0.32

0.29

0.12

0.25

0.17

0.24

0.12

0.27

0.16

0.26

0.17

0.16

0.12

0.11

1.18

-7.39

-6.94

1.41

-6.82

-6.62

-6.55

1.26

-6.86

-6.61

-6.33

-5.96

-5.85

-5.63

-5.36

-5.24

-5.60

4.65

4.80

4.96

4.74

4.89

5.45

4.87

4.76

4.78

4.83

4.78

5.11

5.02

4.96

4.85

4.45

4.89

0.03

0.17

0.18

-0.04

0.13

0.07

0.09

-0.05

0.12

0.02

0.13

0.15

0.03

0.04

0.02

0.11

1.81

-6.01

-5.64

1.81

-5.50

-5.49

-5.33

1.74

-5.64

-5.52

-5.21

-4.89

-4.77

-4.62

-4.53

-4.31

-4.01

2.47

3.37

3.48

2.40

3.37

3.31

3.33

2.48

3.41

3.27

3.29

3.38

3.17

3.14

3.09

2.87

3.26

1993

TXPTS · H₃

f

TXPTS · Na₃

TMAPTS^3-

TMAPTS · H₃

TMAPTS · Na₃

PPh₃

1993

f

1993

1979

P(o-tol)₃

1972^g

TXP^h

TMAPⁱ

P(Mes)₃

j

1964^k

P(2,4-t-BuC₆H₃)₃

^aCone angle values determined from LDFT-optimized Pd⁰L structures using the STERIC program.^{56,57 b}Carbonyl stretching frequency measured in

solution (CH₂Cl₂) of in situ prepared trans-Rh(L)₂(CO)Cl complex. Tetrabutylammonium salts of TPPTS, TXPTS, and TMAPTS were used. ^cCharge on

phosphorus calculated at the DFT level with the B3LYP exchange-correlation functional at the LDA-optimized geometry. ^dHOMO energy at the

B3LYP level. ^eGAP ) E(HOMO) - E(LUMO) . ^fNon-hydrogen-bonded structure. ^gLiterature value.^{62 h}TXP ) tri(2,4-dimethylphenyl)phosphine.

52

ⁱTMAP ) tri(4-methoxy-2-methylphenylphosphine). ^jMes ) mesityl. ^kEstimated from the Tolman parameter obtained with Ni(P(Mes)₃)(CO)₃using

the correlation reported by Vastag and co-workers.⁶³

TMAPTS^3-than that of TXPTS^3-is consistent with the much

Scheme 1. Synthesis of Rh Complexes of Sulfonated Phos-

larger agostic Pd-H interaction in the TXPTS^3-complex (2.40

phines after Cation Exchange

Å) as compared to the TMAPTS^3-complex (2.60 Å). The

values within the TMAPTS^3-series are consistent with the

variations in the Pd-P bond lengths and torsional angles.

The TMAPTS · H₃structure has a smaller torsional angle and a

shorter Pd-P bond than the TMAPTS^3-structure, both of which

should make the Pd center more sterically hindered. The

TMAPTS · Na₃complex also has a smaller average torsional

angle than the TMAPTS^3-structure, but a slightly longer bond.

sulfonate charge with a proton or sodium ion gave larger cone

angle values of 164° and 165°, respectively. The larger values

for TPPTS · Na₃and TPPTS · H₃compared to TPPTS^3-result

from the smaller Pd-P-aryl ring torsion angles for the sodium

ion and proton adducts. As the aryl ring rotates toward becoming

parallel with the P-Pd bond, the ortho-proton moves closer to

the Pd center (shorter Pd-H interaction), leading to an increase

in the cone angle (Figure 3). Interestingly, the TPPTS · Na₃and

TPPTS · H₃structures have nearly identical cone angles despite

their Pd-P bonds differing by 0.07 Å. The effect of the shorter

bond for the TPPTS · H₃structure is likely compensated by it

having a larger average torsional angle than the TPPTS · Na₃

structure. The values calculated for TPPTS are in the mid range

of previously reported values.

As expected, addition of a methyl group to the ortho-position

of the aryl ring signiﬁcantly increases the steric demand of the

phosphines. Bakos⁵⁹has previously reported cone angle values

of 196° and 210° for TXPTS based on the X-ray crystal structure

of the triguanidinium salt. The cone angle of TXPTS^3-

calculated using the STERIC program is 210°. Protonation

increased the cone angle to 219° for the more stable H-bonded

structure, whereas the addition of sodium ions decreases the

cone angle to 206°. The increase in cone angle on protonation

is consistent with the smallest CPC angle for this structure

(Figure 5). The small decrease on adding three Na⁺ions is

consistent with the increase in the Pd-P bond length and the

small change in the CPC angle.

Therefore, the increase in cone angle is less for the

TMAPTS · Na₃complex than for TMAPTS · H₃. The cone angles

for the TXPTS and TMAPTS qualitatively correlate with the

agostic Pd-H interactions, with the stronger interactions (short

Pd-H) correlating with increased cone angles.

The electron-donating ability of the ligands was determined

experimentally from the CO stretching frequency of trans-

RhL₂(CO)Cl complexes generated from the sulfonated phos-

phine tri(tetrabutylammonium (TBA)) salts and [Rh(CO)₂(µ-

Cl)]₂in the presence of trimethylamine N-oxide (Scheme 1).⁶³

The TBA salts were used to allow the sulfonated ligands to

dissolve in methylene chloride. The TPPTS · (TBA)₃complex

in methylene chloride gave a CO stretching frequency of 1993

cm^-1(Table 5). This value is at higher frequency than

previously reported values of 1973–1982 cm^-1for solid samples

of trans-Rh(TPPTS · Na₃)₂(CO)Cl dispersed in KBr.^64–66Higher

frequency values ranging from 1990 to 1994 cm^-1have been

reported for this complex supported in Zn-Al layered double

hydroxide materials,⁶⁶which shows that the CO stretch is

sensitive to the surrounding environment. The rhodium com-

plexes derived from TXPTS · (TBA)₃and TMAPTS · (TBA)₃

gave identical CO stretching frequencies to the TPPTS · (TBA)₃

complex (1993 cm^-1). Thus on the basis of this measurement,

there is no electronic difference among the sulfonated ligands.

(64) Herrmann, W. A.; Kellner, J.; Riepl, H. J. Organomet. Chem. 1990,

389, 103–128.

The cone angle of TMAPTS^3-was calculated to be 194°,

and the value increases to 208° for TMAPTS · Na₃and 214°

for TMAPTS · H₃. The smaller calculated cone angle for

(65) Serp, P.; Hernandez, M.; Richard, B.; Kalck, P. Eur. J. Inorg. Chem.

2001, 2327–2336.

(66) Zhang, X.; Wei, M.; Pu, M.; Li, X.; Chen, H.; Evans, D. G.; Duan,

X. J. Solid State Chem. 2005, 178, 2701–2708.

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 583

Figure 6. ³¹P NMR spectra: (A) Pd(OAc)₂and TXPTS (2.4 equiv); (B) Na₂PdCl₄and TXPTS (2.5 equiv).

The sulfonated ligands are less electron donating than the

analogous nonionic ligands (i.e., PPh₃, TXP, and TMAP).

Because of the range of CO stretching frequencies that have

been reported for trans-Rh(TPPTS · Na₃)₂(CO)Cl, comparisons

of sulfonated and nonsulfonated ligands must be made with

some caution, however.

eV except for P(Mes)₃with a GAP of 2.87 eV. The GAPs for

the TPPTS, TXPTS, and TMAPTS complexes are in a narrow

range between 3.3 and 3.5 eV.

Coordination Chemistry. Amatore and Jutand⁶⁷have shown

68,69

that TPPTS behaves in a similar manner to PPh₃

when

complexed to Pd^IIsalts in the presence of hard anions (^–OAc,

^–OH, F^–). They observed that treatment of Pd(OAc)₂with an

excess of TPPTS · Na₃in aqueous acetonitrile initially resulted

in the formation of a Pd^IIspecies, presumably Pd(TPPTS ·

Na₃)₂(OAc)₂. This Pd^IIspecies underwent reductive elimination

to give Pd⁰(TPPTS · Na₃)₃, or possibly [Pd⁰(TPPTS · Na₃)_n

(OAc)]^-,⁷⁰and TPPTS-oxide (TPPTSO · Na₃, eq 4). In order

to compare the chemistry of TXPTS and TMAPTS with TPPTS,

a series of studies were carried out to determine what complexes

form from the combination of Pd^IIsalts and TXPTS · Na₃and

TMATPS · Na₃under conditions related to the catalytic reactions

described above.

A variety of electronic properties—the charge on the P, the

HOMO, and the GAP deﬁned as GAP ) E(HOMO) -

E(LUMO) (Table 5)—were calculated for the sulfonated phos-

phines and their neutral analogues as both the free ligands and

Pd-bound complexes. The charges on the P are all reduced on

binding to Pd, with the exception of P(2,4-di-t-BuC₆H₃)₃, which

did not change upon complexation probably due to the weak

interaction of this ligand with Pd. If the counterions are not

present, the charges on the P of the Pd complexes of the

sulfonated ligands were near 0 or slightly negative. When the

counterions are present, the charge on the P becomes substan-

tially more positive. The charge on P for the TPPTS · Na₃and

TPPTS · H₃complexes are the largest values, 0.18 and 0.17 e,

respectively, even larger than the charge of 0.15 e on PPh₃.

The charges on the TMAPTS · Na₃and TMAPTS · H₃complexes

are 0.13 and 0.12 e, respectively, both of which are slightly

less than that on Pd(PPh₃). The charge on P for the TXPTS · Na₃

complex is 0.09 e, while the H-bonded TXPTS · H₃complex

gave a similar value of 0.07 e. The less stable, non-H-bonded

TXPTS · H₃and TMATPS · H₃complexes had higher positive

charges on P than the corresponding H-bonded structures. The

lowest phosphorus charges (0.02–0.04 e) were seen for the

ortho-methylated neutral phosphines (P(o-tol)₃, TXP, TMAP,

P(Mes)₃). The sodium sulfonate and sulfonic acid groups

increase the charge on P in all cases, but the effect is largest

for the ortho-methylated ligands (i.e., TXP vs TXPTS · Na₃).

The HOMO energy for the trianions is positive, as expected

for a molecule with such a large negative charge. Addition of

the counterions makes the HOMO energies negative. The

HOMO energies of the neutral sulfonate complexes with the

counterions present are the most negative. The lowest energy

HOMO for the Na₃complexes is that for TPPTS · Na₃followed

by TXPTS · Na₃and TMAPTS · Na₃. In the case of the sulfonic

acids, Pd(TXPTS · H₃) has the highest energy HOMO followed

by Pd(TMAPTS · H₃), while the TPPTS · H₃complex again had

the lowest HOMO energy. PPh₃has the lowest energy HOMO

for nonsulfonated phosphines. Alkyl substituents tend to raise

the HOMO energy, making the orbital more accessible. The

GAPs for most of the compounds are in the range 3.1 to 3.5

(4)

TXPTS · Na₃and Pd^II. A ³¹P NMR spectrum of a solution

of Pd(OAc)₂(0.1M) and TXPTS · Na₃(2.5 equiv) in 2:1 D₂O/

acetonitrile gave a peak for TXPTS · Na₃-oxide (TXPTSO · Na₃)

at 37.6 ppm; a complex set of peaks between 13 and 28 ppm,

with the major resonance being at 19.5 ppm; and free

TXPTS · Na₃at -30 ppm (Figure 6A). TXPTSO · Na₃, which

is present as an impurity (12 mol %) in the TXPTS · Na₃used

in this work, serves as a convenient chemical shift and

integration reference and is not expected to interact signiﬁcantly

with the palladium. The ratio of TXPTSO · Na₃:13–28 ppm:

TXPTS · Na₃was 7:71:22. The percentage of TXPTSO · Na₃

present in this mixture is approximately the same as is present

in the starting TXPTS · Na₃, so no additional oxide appears to

(67) Amatore, C.; Blart, E.; Genêt, J. P.; Jutand, A.; Lemaire-Audoire,

S.; Savignac, M. J. Org. Chem. 1995, 60, 829–6839.

(68) Amatore, C.; Jutand, A.; M’Barki, M. A. Organometallics 1992,

11, 3009–3013.

(69) Amatore, C.; Carré, E.; Jutand, A.; M’Barki, M. A. Organometallics

1995, 14, 1818–1826.

(70) Amatore, C.; Carré, E.; Jutand, A.; M’Barki, M. A.; Meyer, G.

Organometallics 1995, 14, 5605–5614.

(71) Baber, R. A.; Orpen, A. G.; Pringle, P. G.; Wilkinson, M. J.;

Wingad, R. L. Dalton Trans. 2005, 659–667.

584 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

The remaining peaks observed between 13 and 28 ppm cannot

have acetate or chloride coordinated, as they have been observed

in approximately the same relative intensities when using

Pd(OAc)₂or Na₂PdCl₄as the palladium source (Figure 6).

Therefore, these species must have only TXPTS, water,

hydroxide, and acetonitrile as ligands. In addition, the sulfonate

groups of the ligands could bind in a κO fashion in addition to

the expected κP binding. The two most intense features in this

region are the peak at 19.5 ppm and an AB doublet pair at 24.3

and 14.5 ppm (²J_P,P) 506 Hz). We propose that the 19.5 ppm

resonance and the other minor peaks at 22.2, 18.2, and 17.0

ppm are Pd(TXPTS)_nL_mcomplexes (7, n ) 1, 2; m ) 0–3; L

) water, hydroxide, acetonitrile). Possible structures for complex

7 are listed in Figure 8.

Figure 7. Proposed structures for TXPTS-coordinated palladium

chloride complexes.

The peak splitting in the AB doublet pattern remained

constant when the NMR operating frequency was changed from

145.8 to 202.5 MHz, which supports the conclusion that these

peaks are truly coupled. The large coupling constant is consistent

with a ²J coupling between chemically unique phosphorus nuclei

in a trans-relationship. Since TXPTS is the only phosphine

present initially, some of the TXPTS must be modiﬁed in the

course of the reaction, which would most likely occur by

activation of a C-H bond of one of the o-methyl groups to

form a palladacycle.⁷⁴Palladium complexes of (o-methylphe-

nyl)phosphines (i.e., tri(o-tolyl)phosphine) are known to form

palladacycliccomplexesofthistypeundermildconditions.^72,73,75,76

In addition, our DFT-optimized structures showed close contacts

between hydrogens on the ortho-methyl groups and palladium.

Therefore, the AB pattern has been tentatively assigned to

[Pd(κ²C,P-TXPTS)(κP-TXPTS)L] (8, Figure 9, L ) water,

hydroxide, acetonitrile, or no ligand). In a structurally related

palladacycle derived from trimesitylphosphine (9, X ) Cl, Br,

I), the reported chemical shifts of the metalated phosphine

ranged from 22.1 to 22.5 ppm and the ²J_P,Pvalues ranged from

414 to 419 Hz.⁷⁷

have formed. Treatment of this sample with phenyl iodide

resulted in no change in any of the peak positions or intensities.

Any Pd⁰species formed upon complexation with TXPTS · Na₃

would be expected to oxidatively add phenyl iodide. The lack

of reaction suggests that the cluster of peaks between 13 and

28 ppm represents various Pd^IIcomplexes of TXPTS · Na₃and

–

other ligands (^–OAc, OH, H₂O, CH₃CN, etc.).

The complexation of TXPTS · Na₃with Na₂PdCl₄in 2:1 D₂O/

acetonitrile was next explored. Using a 2.5:1 TXPTS · Na₃/

Na₂PdCl₄ratio, a similar pattern of peaks was observed between

13 and 28 ppm to that seen in the Pd(OAc)₂reaction was

observed (Figure 6B, TXPTSO · Na₃:Pd^II:TXPTS · Na₃

)

14:86:0). All of the peaks from the Pd(OAc)₂reaction were

present in similar relative ratios. In addition, two new resonances

at 24.6 and 20.1 ppm were observed, which were not present

when Pd(OAc)₂was used as the Pd precursor. Presumably these

new peaks represent chloride complexes. The other difference

relative to the Pd(OAc)₂studies was that no free TXPTS · Na₃

was present in the sample. When the amount of TXPTS · Na₃

was increased to 5 equiv relative to Pd, the ³¹P NMR spectrum

showed TPPTSO · Na₃, a cluster of peaks between 13 and 28

ppm, and free TXPTS · Na₃(15:53:34). Neither the 24.6 nor

20.1 ppm resonances were observed, but a new peak was

observed at 18 ppm. Other than the addition of the peak at 18

ppm, the pattern was identical to that seen when Pd(OAc)₂was

used as the Pd source.

TXPTS · Na₃, Pd(OAc)₂, and Na₂CO₃. Since base is present

in all of the coupling reactions, a mixture of Pd(OAc)₂(0.2

M), TXPTS · Na₃(2.7 equiv), and sodium carbonate (6.6 equiv)

in 2:1 D₂O/acetonitrile was analyzed by ³¹P NMR spectroscopy.

Immediately after mixing, only three resonances were seen in

the ³¹P NMR spectrum: TXPTSO · Na₃, a peak at 19.5 ppm (7),

and free TXPTS · Na₃in a 12:38:50 ratio. After standing for

48 h, the peaks present were TXPTSO · Na₃, a new peak at 20.1

ppm, 7, and free TXPTS · Na₃in a 14:7:24:55 ratio.

TXPTSO · Na₃comprised 14% of the total area, so little

additional TXPTSO · Na₃had been formed under these condi-

tions. The new peak at 20.1 ppm appears to have been formed

from 7, since the peak at 19.5 ppm had decreased in intensity

from the initial spectrum. Addition of phenyl iodide again had

no effect on this mixture. Therefore, the peak at 20.1 ppm does

not correspond to a Pd⁰complex. Since this species is only seen

when carbonate is present, it has tentatively been assigned as

[Pd(TXPTS)_n(CO₃)]Na_3n(eq 5, 10). The presence of carbonate

decreased the amount of TXPTS · Na₃bound to palladium. In

the absence of sodium carbonate, 76% of the TXPTS · Na₃

(TXPTS · Na₃:Pd(OAc)₂) 2.5) was bound to Pd^II, while 24%

We have tentatively assigned the resonances at 24.6 and 20.1

ppm to anti- and syn-[Pd(TXPTS)(µ-Cl)Cl]₂(4, Figure 7). The

fact that these resonances disappear when additional ligand is

added suggests that they correspond to monophosphine com-

plexes. Orpen and Pringle⁷¹have reported the synthesis of a

structurally similar [PdL(µ-Cl)Cl]₂complex (L ) tri(4-methoxy-

2-methylphenyl)phosphine (TMAP)). The ³¹P NMR chemical

shifts for the syn- and anti-isomers of this complex were found

at 24.1 and 22.4 ppm in CD₂Cl₂. Another possible structure

that could be considered is the palladacycle µ-chloride dimer

(6). This possibility appears less likely, since the related

palladacycles derived from TMAP and P(o-tol)₃have chemical

shifts of 29 and 39 ppm, respectively.^72,73The peak observed

at 18 ppm has been assigned to Pd(TXPTS)₂Cl₂(5), since this

resonance was only observed at high TXPTS/Pd ratios and in

the presence of chloride. In addition, Orpen and Pringle reported

a chemical shift of 18.8 ppm for trans-Pd(TMAP)₂Cl₂.

(74) While TXPTSO could potentially coordinate to Pd, the ³J coupling

would be expected to be much smaller than is observed.

(75) Cheney, A. J.; Shaw, B. L. J. Chem. Soc., Dalton Trans. 1972,

860–865.

(76) Herrmann, W. A.; Brossmer, C.; Öfele, K.; Reisinger, C.-P.;

Priermeier, T.; Beller, M.; Fischer, H. Angew. Chem., Int. Ed. 1995, 34,

1844–1848.

(72) Alyea, E. C.; Malito, J. J. Organomet. Chem. 1988, 340, 119–126.

(73) Herrmann, W. A.; Brossmer, C.; Reisinger, C.-P.; Riermeier, T. H.;

Öfele, K.; Beller, M. Chem.-Eur. J. 1997, 3, 1357–1364.

(77) Alyea, E. C.; Malito, G. Gazz. Chim. Ital. 1993, 123, 709–711.

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 585

Figure 8. Possible structures for complexes observed between 13 and 28 ppm (7).

accelerated by heating the reaction mixture at 50 °C. After

heating the reaction for 2.5 h at 50 °C, 11 was the major species

present in solution along with smaller amounts of TXTPSO · Na₃

and free TXPTS · Na₃(TXPTSO · Na₃:11:TXPTS · Na₃

12:76:12).

)

TXPTS · Na₃, Pd(OAc)₂, PhB(OH)₂, and Na₂CO₃. The other

component present in the Suzuki coupling is phenylboronic acid,

which can reduce Pd^IIin the presence of base or oxyanion

–

ligands (^–OAc, OH) by transferring two phenyl groups to the

Pd^IIcomplex followed by reductive elimination of biphenyl.^23,79

A solution of Pd(OAc)₂, TXPTS · Na₃(2.4 equiv), Na₂CO₃(6.6

equiv), and PhB(OH)₂(16.4 equiv) in 2:1 D₂O/acetonitrile was

analyzed by ³¹P NMR spectroscopy immediately after mixing.

The major species present were TXPTSO · Na₃(14%), 7 (18%),

and free TXPTS · Na₃(68%). No Pd⁰(TXPTS)₂(11) was present

initially. After standing for 24 h at room temperature, complex

7 had been replaced by 11 (TXPTSO · Na₃:11:TXPTS · Na₃)

19:22:59). Complex 7 appears to have been converted to the

Pd⁰complex, while there were no signiﬁcant changes in the

amounts of TXPTSO · Na₃and TXPTS · Na₃. Treatment of this

solution with phenyl iodide resulted in the complete loss of 11,

the formation of 7, and an AB pattern at 21.1 and 15.1 ppm

with a coupling constant of 363 Hz, along with some additional

minor peaks between 14 and 28 ppm (TXPTSO · Na₃;7:AB

patern:TXPTS · Na₃) 18:7:19:41).

Figure 9. Proposed palladacycle structure for the AB quartet feature.

was uncoordinated. In the presence of 6.5 equiv of sodium

carbonate, only 43% of the ligand was coordinated to Pd^II.

The ratio of carbonate to Pd is much higher under catalytic

conditions, so the complexation reaction was repeated in the

presence of 27 equiv of sodium carbonate (Pd/TXPTS · Na₃

2.5:1). Under these conditions, TXPTSO · Na₃, the resonance

at 19.5 ppm (7), and free TXPTS · Na₃remained the major

resonances observed immediately after mixing. The peak at 19.5

ppm was much broader than when a lower amount of carbonate

was used, suggesting a possible equilibrium between 7 and a

carbonate-bound complex (10). In addition. a small peak at -9.6

ppm was observed (TXPTSO · Na₃:7:-9.6 ppm:TXPTS · Na₃)

18:27:7:47). The amount of TXPTSO · Na₃was larger than is

present in the starting TXPTS · Na₃(12%), showing that some

additional TXPTS · Na₃had formed under these conditions. After

standing for a day, the same peaks were present with the

exception of 7, which had been replaced by complex 10 (20.1

ppm). The TXPTSO · Na₃:10:-9.6:TXPTS · Na₃ratio was

29:29:15:27. Thus the amount of TXPTSO · Na₃and the -9.6

ppm peak had increased at the expense of the decrease of the

amount of TXPTS · Na₃. The new peak at 20.1 was nearly

identical in size to that of 7 seen in the initial spectrum.

Addition of phenyl iodide to this mixture resulted in the

immediate loss of the peak at -9.6 ppm, indicating that this

species is a Pd⁰complex of TXPTS · Na₃. The free TXPTS · Na₃

peak increased in intensity to 61% of the total integrated value

after the phenyl iodide was added. In addition to TXPTSO · Na₃

(17%), the only other peak present was a broad peak centered

at 20 ppm (22%). On the basis of the similar chemical shift of

bis(tri(2,4-dimethylphenyl)phosphine)palladium(0) (Pd(TXP)₂,

-8.5 ppm),⁷⁸the -9.6 resonance has been tentatively assigned

to Pd⁰(TXPTS · Na₃)₂(11). The formation of 11 could be

When the amount of base was increased signiﬁcantly

(TXPTS · Na₃(2.8 equiv), Na₂CO₃(103 equiv), PhB(OH)₂(3.9

equiv)), there was only a small amount of 7 present (4% of

total integrated area) immediately after the components were

combined. A signiﬁcant amount of Pd⁰(TXPTS)₂(11) had

already formed (20%) after a few minutes. Free TXPTS · Na₃

(61%) and TXPTSO · Na₃(15%) comprised the remainder of

the material. Since little additional TXPTSO · Na₃had formed,

PhB(OH)₂appears to be the reductant under these conditions.

After standing for 24 h at room temperature, only TXPTSO · Na₃

(18%), 11 (62%), and TXPTS · Na₃(20%) were present.

1

The H NMR spectrum of this mixture showed two broad

resonances in the aromatic region due to TXPTS · Na₃that were

shifted upﬁeld from those of free TXPTS · Na₃. The phenylbo-

ronic acid resonances were also shifted upﬁeld, presumably due

to the formation of PhB(OH)₃^–under the strongly basic reaction

conditions. The alkyl region showed a small amount of

TXPTSO · Na₃and two resonances that presumably correspond

to the Pd⁰complex. A peak was observed in the same position

as the para-methyl of free TXPTS (2.48 ppm).⁸⁰Another broad

resonance was observed at 2.60 ppm, which may correspond

to the ortho-methyl groups of the Pd⁰complex. This resonance

was downﬁeld of where the ortho-methyl of free TXPTS · Na₃

was observed (2.24 ppm).

(79) Adamo, C.; Amatore, C.; Cioﬁni, I.; Jutand, A.; Lakmini, H. J. Am.

Chem. Soc. 2006, 128, 6829–6836.

(78) Paul, F.; Patt, J.; Hartwig, J. F. Organometallics 1995, 14, 3030–

3039.

(80) ¹H and ¹³C resonances were fully assigned based on HSQC/HMBC

spectra. See Supporting Information for a full description.

586 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

Figure 10. ³¹P NMR spectra of Pd/TXPTS · Na₃: (A) Pd(OAc)₂and TXPTS · Na₃(2.4 equiv), arrows indicate the AB quartet feature; and

Pd(OAc)₂, TXPTS · Na₃(2.5 equiv), and PhB(OH)₂(4.0 equiv): (B) after 1 h, arrows indicate the new AB quartet; (C) after 24 h; (D) after

5 days.

Although a signiﬁcant amount of 11 could be generated in

the presence of phenylboronic acid and a large excess of base,

the rate was still slow compared to the catalytic reactions. The

catalytic reactions were all performed at 50 °C, or higher,

however. When the complexation of TXPTS · Na₃(4:1 TXPTS ·

Na₃/Pd) with Pd(OAc)₂in the presence of Na₂CO₃(20 equiv)

and PhB(OH)₂(10 equiv) was performed at 50 °C, the

Pd⁰(TXPTS)₂(11) species was the major species after a period

of 3 h (TXPTSO · Na₃:11:TXPTS · Na₃) 12:76:12).

Figure 11. Proposed structure of the palladacycle formed in the

presence of PhB(OH)₂.

AB pattern were observed. The alkyl region of this mixture

showed TXPTSO · Na₃and a trace of free TXPTS · Na₃. Several

very broad resonances in the baseline were observed that

presumably correspond to the complex responsible for the AB

doublet pair. Thus it appears that the TXPTS ligands are

undergoing a dynamic process on the ¹H NMR time scale,

TXPTS · Na₃, Pd(OAc)₂, and PhB(OH)₂. Reaction of

TXPTS · Na₃(2.5 equiv), Pd(OAc)₂, and PhB(OH)₂(4 equiv)

in 2:1 D₂O/CD₃CN in the absence of sodium carbonate gave a

complex mixture of Pd^IIspecies between 14 and 26 ppm (Figure

10), while no free TXPTS · Na₃was observed. The cluster of

peaks between 14 and 26 ppm contained 7, palladacycle 8, and

the other minor species that were also observed in the absence

of PhB(OH)₂(Figure 6). The major species present in this region

was a new AB pair of doublets with chemical shifts of 21.1

and 15.1 ppm and a coupling constant of 363 Hz. The spectral

data for the AB doublet pair were identical to one of the species

formed when 11 was treated with phenyl iodide. After standing

for 24 h at room temperature, the AB pattern was the major

phosphorus-containing species present in solution other than

TXPTSO · Na₃. In addition, some very broad resonances were

observed between 20 and 30 ppm.

1

resulting in very broad resonances. H NMR spectra obtained

at various operating frequencies (360–600 MHz) and over the

temperature range accessible in the solvent system also gave

broad unresolved peaks that did not allow the structure of the

complex to be conclusively determined.

The complex with the 363 Hz coupling has been observed

upon the reaction of phenyl iodide with 11 and when

TXPTS · Na₃is reacted with Pd(OAc)₂in the presence of

phenylboronic acid. We propose that this complex is an ana-

logue of 8 where L ) phenyl-[Pd(κ²C,P-TXPTS · Na₃)(κP-

TXPTS · Na₃)Ph] (Figure 11, 12). Introduction of the phenyl

substituent is consistent with the conditions under which 12 is

generated, as well as the smaller ²J_P,Pvalue for 12 compared to

8. Replacement of a weak cis-inﬂuence ligand (^–OH, H₂O, or

CH₃CN) in 8 with phenyl would be expected to decrease the

1

At room temperature, the aromatic region of the H NMR

spectrum of the reaction mixture containing the AB pattern and

TXPTSO · Na₃showed only phenylboronic acid and a small

amount of TXPTSO · Na₃. No resonances corresponding to the

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 587

Scheme 2. Reaction of TMAPTS · Na₃with Pd(OAc)₂in the

than in the previous case, because this is a complete catalytic

system in which 14 can be continually regenerated.

Presence of Na₂CO₃

(6)

Calculated ³¹P NMR Chemical Shifts. Since we were

unable to isolate pure palladium complexes of sulfonated

ligands, the ³¹P NMR chemical shifts were calculated for the

compounds with the pure DFT gradient-corrected BLYP

exchange-correlation functional⁸¹and the TZP basis set using

the ADF code.^82,83The NMR calculations were done in the

gauge invariant atomic orbital (GIAO) approach at the DFT

level based on the developments in the Ziegler group.^84–87Scalar

relativistic effects were included at the two-component zero-

order regular approximation (ZORA) level for the NMR

calculations (inclusion of the mass-velocity and Darwin

terms).^87–89The standard for the calculations is H₃PO₄with an

absolute chemical shift of 292.2 ppm. We ﬁrst compare the

calculated and experimental value for PPh₃. The isolated

molecule value is 9.4 ppm, 15 ppm larger than experiment.

Variation of the Pd-P-C-C torsion angle by +10° leads to a

value of 0.6 ppm, differing by 6 ppm from experiment. Placing

PPh₃in a self-consistent reaction ﬁeld⁹⁰with the COSMO

parameters^91–93and with a value for the dielectric of DMF leads

to only a slight decrease in the chemical shift to 8.1 ppm. Thus

the variation in the torsion angle is more important in determin-

ing the chemical shift than the surrounding medium. For

Pd(PPh₃)₂, the agreement between the calculated and experi-

mental values is within ∼2 ppm. Similar good agreement is

found for Pd(P(t-Bu)₃)₂and P(t-Bu)₃. Inclusion of solvent

effects, in this case THF at the COSMO level, leads to slightly

worse agreement with experiment (calculated value ) 67.9

ppm). The calculated value for TXPTS^3-differs by almost 50

ppm from the experimental value. When the ³¹P chemical shifts

for the sulfonic acid or trisodium salt derivatives were calculated,

much better agreement with experiment was found, with the

sodium salt providing the closest agreement. The calculated

value for 11 differs by only 14 ppm from experiment even

though no counterions are present.

²J_P,Pvalue due to phenyl’s larger cis inﬂuence. Unfortunately

1

the ambiguous nature of the H and ¹³C NMR spectra of this

complex and our inability to isolate X-ray quality crystals of it

do not allow us to unambiguously assign the structure of 12.

TMAPTS · Na₃Complexation with Pd(OAc)₂. The com-

plexation of TMAPTS · Na₃with Pd(OAc)₂was also explored.

Analysis of a solution comprised of Pd(OAc)₂, TMAPTS · Na₃

(3.5 equiv), and sodium carbonate (20 equiv) in 2:1 water/

acetonitrile by ³¹P NMR spectroscopy gave resonances at 37.3

(TMAPTS-oxide (TMAPTSO · Na₃), 19.3 (broad), and -11.2

ppm (integrated ratio: 30:23:47) immediately after mixing the

reagents. No free TMAPTS · Na₃was observed under these

conditions. The TMAPTS · Na₃used in these studies had only

a trace of TMAPTSO · Na₃present, so all of the oxide formed

is a result of the reduction of Pd^II. No change was observed in

spectra of the reaction mixture over a period of 2 h, other than

a slight increase in the amount of the TMAPTSO · Na₃peak

relative to the -11.2 ppm peak (36:22:42 after 2 h). 4-Eth-

yliodobenzene was added to this mixture. Immediately after

adding the aryl iodide, the -11.2 ppm peak was nearly

completely consumed. The peak at 19.3 ppm remained and a

new pair of doublets at 19.1 and 13.4 ppm (²J_P,P) 376 Hz)

appeared in the spectrum.

The 19.3 ppm resonance is presumably due to a similar

Pd^II(TMAPTS · Na₃)_nL_m(L ) acetonitrile, hydroxide, or water)

species to 7 (Scheme 2, 13). The -11.2 ppm peak disappears

in the presence of an aryl iodide, so it is likely

Pd⁰(TMAPTS · Na₃)₂(14), in analogy to 11, which was observed

at -9.6 ppm in the TXPTS system. The pair of doublets feature

that formed when the Pd⁰complex reacted with the aryl iodide

has similar chemical shift and coupling constant values to 12.

In analogy to 12, the AB pattern has been assigned to a

palladacycle complex ([Pd(κ²C,P-TMAPTS)(κP-TMAPTS)Ar]

Na₆, Ar ) 4-ethylphenyl, 15a). This complex is formed by a

sequence of events involving oxidative addition of the aryl

iodide and C-H activation of an ortho-methyl group on the

ligand.

A solution of Pd(OAc)₂, TMAPTS · Na₃(3.5 equiv), sodium

carbonate (20 equiv), and phenylboronic acid (10 equiv) was

then analyzed by ³¹P NMR spectroscopy (eq 6). Only a trace

of TMAPTSO · Na₃was observed. A similar set of AB doublets

was present, which presumably belongs to the Ph analogue of

15a (eq 3, 15b), along with Pd⁰(TMAPTS)₂(14) and free

TMAPTS · Na₃(-35.5 ppm, broad) (36:47:17 integrated ratio).

This system remained unchanged over the course of 2 h. Upon

addition of 4-ethyliodobenzene, the amount of 14 decreased,

while resonances for 15 and the TMAPTS · Na₃peak became

more intense (15:14:TMAPTS · Na₃) 42:36:21). Upon standing,

complex 14 continued to slowly decay, while the amount of

free TMAPTS · Na₃slowly increased. After standing for 3 h,

the ratio had changed to 41:30:29. The loss of 14 is much slower

(81) Becke, A. D. Phys. ReV. A 1988, 38, 3098–3100.

(82) See also: ADF 2004.01; ADF Users Guide, http://www.scm.com,

SCM, Theoretical Chemistry, Vrije Universiteit, Amsterdam.

(83) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra,

C.; van Gisbergen, J. A.; Snijders, J. G.; Ziegler, T. J. Comput. Chem. 2001,

22, 931–967.

(84) Schreckenbach, G.; Ziegler, T. J. Phys. Chem. 1995, 99, 606–611.

(85) Schreckenbach, G.; Ziegler, T. Int. J. Quantum Chem. 1997, 61,

899–918.

(86) Wolff, S. K.; Ziegler, T. J. Chem. Phys. 1998, 109, 95–905.

(87) Wolff, S. K.; Ziegler, T.; van Lenthe, E.; Baerends, E. J. J. Chem.

Phys. 1999, 110, 7689–7698.

(88) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1993,

99, 4597–4610.

(89) Autschbach, J.; Ziegler, T. In Calculation of NMR and EPR

Parameters: Theory and Application; Kaupp, M., Buhl, M., Malkin, V. G.,

Eds.; Wiley-VCH & Co.: Weinheim, 2004; pp 249–264.

(90) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. ReV. 2005, 105, 2999–

3093.

(91) Klamt, A. J. Phys. Chem. 1995, 99, 224–2235.

(92) Klamt, A.; Schüürmann, G. J. Chem. Soc., Perkin Trans. 2 1993,

799–805.

(93) Klamt, A.; Jonas, V. J. Chem. Phys. 1996, 105, 9972–9981.

588 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

Table 6. Calculated and Experimental ³¹P NMR Chemical Shifts

The calculated chemical shifts for the Pd(κ²C,P-TXPTS)(κP-

TXPTS) complex (16) without counterions were predicted to

be 40.3 and 47.9 ppm, which are 25 to 30 ppm higher than

either of the proposed palladacycle complexes (8 and 12). The

inclusion of Na⁺counterions did not substantially improve the

agreement. It is possible that a fourth ligand is coordinated to

the Pd center, which would affect the chemical shift, although

steric hindrance may preclude the presence of an additional

ligand. GIAO⁹⁴calculations of the chemical shift with the

B3LYP functional and the Ahlrichs triple-ꢀ basis set⁵⁰gave

predicted chemical shifts of 16.0 and 22.7 ppm, which are in

excellent agreement with the observed values for 8 and 12.

However, the agreement between the B3LYP calculations and

experiment for the ³¹P chemical shifts for the other compounds

in Table 6 was generally not as good as that from the GIAO/

ZORA calculations. Overall, the calculations provide estimates

of the ³¹P chemical shifts for these types of complexes that are

only accurate to within about 15 ppm, which are insufﬁcient to

reach any conclusions regarding the proposed structure

assignments.

Discussion

Ligand Structure, Steric, and Electronic Properties. The

presence or absence of a counterion and the identity of the

ion were found to have a signiﬁcant effect on the calculated

structure and cone angle of the sulfonated ligands. Predicted

structures without counterions for the sulfonate groups

resulted in structures adopting an open conformation typiﬁed

by smaller Pd-P-C angles, larger C-P-C angles, and larger

Pd-P-C-C torsion angles. As a result, the sulfonate ions

were able to achieve maximum separation, thus reducing the

charge repulsion. When sodium ions or protons were added

to the sulfonate groups, the geometry of the ligand changed

to bring the sulfonate groups closer together. In the L · Na₃

structures, each sodium ion was coordinated to two sulfonate

groups, while in the case of the L · H₃structures hydrogen

bonding occurred between the sulfonate groups. In order to

allow the proton or sodium ion to interact with more than

one sulfonate group, the aryl rings were pulled back from

the Pd center (larger Pd-P-C and smaller C-P-C angles)

and the torsional angle was decreased for some of the aryl

rings. The result of these conformation changes was an

increase in cone angle by 10–20° from the L^3-structure to

the L · Na₃and L · H₃structures. In the cases where a

hydrogen-bonded structure could be compared to a non-

hydrogen-bonded structure (TXPTS and TMAPTS), the

former was always more stable than the latter for both the

free ligand and the palladium complex. The results suggest

that solvent and counterions can play a role in determining

the actual value of the cone angles.

^aExperimentally determined chemical shifts in 2:1 D₂O/CH₃CN

unless noted. ^bCalculated with the BLYP exchange-correlation

functional. See text. ^cCalculated with the B3LYP exchange-correlation

functional. See text. ^dDMF solvent.^{68 e}Obtained at -80 °C in

toluene-d₈.^{95 f}THF solvent.^{13 g}Non-hydrogen-bonded structure.

results say about the relative steric demand of TXPTS and

TMAPTS under the catalytic conditions? The strong charge

repulsion experienced by the L^3-structures in the gas phase

would be less signiﬁcant in solution where counterions and

hydrogen bonding with the solvent would partially mask the

negative charges. The L · H₃and L · Na₃were calculated to lessen

the inﬂuence of the charge repulsion, but the presence of sodium

ions and protons also induces structural effects due to intramo-

lecular coordination of the sodium and proton between the

sulfonate groups in the TPPTS and TMAPTS structures. In an

aqueous solution, the sodium or hydrogen ion as well as the

sulfonate would be solvated by water. Therefore, the ligands

would likely not adopt the same conformation seen in the gas

phase. The “true” structure of the ligands is probably intermedi-

ate to the trianion and sodium salt structures. The sulfonic acid

The calculated cone angle values show the importance of

carefully considering the structural model to use when ionic

groups are present on the ligand. For the trianionic structures,

the cone angle trend is TPPTS^3-<< TMAPTS^3-<<

TXPTS^3-; the sodium salts follow the trend TPPTS · Na₃<<

TXPTS · Na₃≈ TMAPTS · Na₃; and for the sulfonic acid

structures the trend is TPPTS · H₃<< TMAPTS · H₃

<

TXPTS · H₃. In all three cases, the ortho-methylated structures

have much larger cone angles than TPPTS, but what do these

(94) Wolinski, K.; Hinton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112,

8251–8260.

(95) Urata, H.; Suzuki, H.; Moro-oka, Y.; Ikawa, T. J. Organomet. Chem.

1989, 364, 235–244.

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 589

structures are less relevant to the catalytic results, which were

carried out under basic conditions. Although much more

challenging computationally, it will be of future interest to

calculate the Pd(L^3-) and Pd(L · Na₃) structures with shells of

water molecules and solvated counterions to more closely

simulate the real environment in which these complexes are

found and used. A conclusion that can be drawn from these

results is that the cone angle, and thus catalytic performance,

could be affected by the choice of counterion and solvent

medium.

the large counterion effects seen in the computation results,

it is possible that ion-pairing effects in the trans-

Rh(L · TBA₃)₂(CO)Cl complexes have a strong affect on the

CO stretching frequency. We have previously seen that the

TBA counterions in trans-Rh(TPPTS · TBA₃)₂(CO)(NO₃)

cluster tightly around the metal center in CH₂Cl₂, preventing

displacement of the nitrate anion.⁹⁶A similar tight clustering

in this CO study may have had a dominating effect on the

CO stretching frequency, making the electronic contribution

of the phosphine inconsequential.

Palladium Coordination Chemistry. The coordination

chemistry of TXPTS · Na₃and TMAPTS · Na₃with palladium

is more complex than that of TPPTS · Na₃. Differences were

also observed in the coordination chemistry of TXPTS · Na₃and

TMAPTS · Na₃, despite their similar structural and electronic

characteristics. Coordination of TXPTS · Na₃with Pd(OAc)₂did

not produce a Pd⁰complex or generate TXPTSO · Na₃even after

standing for 24 h at room temperature, in contrast to Amatore’s

results with TPPTS · Na₃.⁶⁷Amatore⁹⁷has reported that tri(o-

tolyl)phosphine will not reduce Pd(OAc)₂, presumably due to

the large cone angle that inhibits attack by acetate on the

phosphine,⁶⁹so the lack of reduction by TXPTS · Na₃is not

necessarily surprising.

When Pd(OAc)₂was complexed with TXPTS · Na₃, two

major Pd complexes are observed, along with several other

minor components. One was a single resonance at 19.5 ppm

and the second is an AB pattern with a ²J_P,Pcoupling of 506

Hz. On the basis of this very large coupling constant, we

have concluded that the AB pattern is due to a palladacycle

complex (8, Figure 9) formed by C-H activation of one of

the ortho-methyl groups of the ligand. These types of

complexes are well precedented for ortho-methyl-substituted

arylphosphines.^72,73,75,76The 19.5 ppm resonance is likely

On the basis of the CO stretching frequency of the trans-

Rh(L)₂(CO)Cl complexes, no measurable difference in the

electron-donating ability of the sulfonated phosphines could be

determined. The sulfonated phosphines were considerably less

electron-donating than PPh₃or P(o-tol)₃on the basis of these

measurements. Comparison with nonsulfonated ligands is

complicated, however, by the dependence of the CO stretching

frequency on the local environment (solvent, ions, etc.), in

addition to the inherent electronic properties of the ligands. In

contrast, the computational results show that the sulfonated

phosphines do have different electronic properties. For example,

if we compare the charge on phosphorus for either the free or

bound ligands, we see that TPPTS · Na₃has a more positive

charge than does PPh₃, which suggests that PPh₃is a better

electron donor than TPPTS · Na₃. TXPTS · Na₃, which has two

electron-donating methyl groups, has a less positive charge than

TPPTS · Na₃or PPh₃, although it is more positive than the P of

P(o-tol)₃. Interestingly, TMAPTS · Na₃has a more positive P

atom than does TXPTS · Na₃, which suggests that the inductive

withdrawing effect of the methoxy group has a larger effect on

the phosphorus charge than its resonance-donating ability. This

effect was not observed in the nonsulfonated analogues (TXP

and TMAP), however.

to be due to

a complex with the general formula

All three sulfonated ligands have lower (more negative)

HOMO energies than the nonsulfonated ligands for both the

free and Pd-bound ligands. Higher energy HOMOs are more

available to react, thus PdL complexes with higher energy

HOMOs are expected to be more reactive toward oxidative

addition. Of the Pd complexes (PdL) of the sulfonated

ligands, the HOMO energy trend is Pd(TPPTS · Na₃) <

Pd(TXPTS · Na₃) < Pd(TMAPTS · Na₃), all of which have

lower energy HOMOs than Pd(PPh₃). Thus based on the

HOMO energy levels, TMAPTS · Na₃is a better electron

donor than TXPTS · Na₃, while both are signiﬁcantly more

electron donating than TPPTS · Na₃. This trend was mirrored

in the case of the nonsulfonated analogues: PPh₃< TXP <

TMAP. The opposite trend was observed in the case of the

sulfonic acid derivatives, however. The HOMO energy level

of the Pd(TXPTS · H₃) was found to be at higher energy than

that of the Pd(TMAPTS · H₃) complex, while TPPTS · H₃had

the lowest PdL HOMO energy.

–

Pd(TXPTS · Na₃)_nL_m(L ) H₂O, OH, CH₃CN, 7), although

the exact structure of this complex, along with the other minor

species present, could not be determined. This cluster of peaks

remained essentially the same as the TXPTS · Na₃:Pd ratio

was varied. The same species were also observed when

Na₂PdCl₄was used as the precursor, along with new

resonances corresponding to chloride complexes.

TXPTS · Na₃is a much less efﬁcient reductant of Pd(OAc)₂

than is TPPTS · Na₃. While TPPTS · Na₃reduces Pd^IIto Pd⁰

efﬁciently at room temperature without additional reagents,

TXPTS · Na₃does not form Pd⁰complexes in the absence of

other reagents. Traces of Pd⁰(TXPTS · Na₃)₂(11) were formed

in the presence of large excesses of sodium carbonate, but

elevated temperatures were required to form 11 in signiﬁcant

amounts. Use of phenylboronic acid as an external reductant

increased the rate of reduction, but several hours were

required to form 11 as the major species in solution. The

Pd⁰(TXPTS · Na₃)₂complex could be formed as the major

species in solution on a time scale that is comparable to the

catalytic reactions at 50 °C in the presence of sodium

carbonate or phenylboronic acid and sodium carbonate. The

slow formation of the PdL₂complex at room temperature

may explain the low activity of the TXPTS · Na₃catalyst

system below 50 °C.

Overall, the computational results show that TXPTS · Na₃

and TMAPTS · Na₃are better electron donors than

TPPTS · Na₃and are comparable to PPh₃. These results are

in contrast to what is seen experimentally using the CO

stretching frequency of trans-Rh(L)₂(CO)Cl, where all three

sulfonated ligands are much less electron donating than PPh₃.

The fact that all three ligands gave identical CO stretching

frequencies is surprising. Tolman reported a 2.3 cm^-1

difference in the CO stretch of the Ni(CO)₃L complexes of

PPh₃and P(o-tol)₃.⁵²Since the CO stretch of the rhodium

system is more sensitive to ligand electronics, we would

expect to see a difference between the methylated (TXPTS

and TMAPTS) and nonmethylated (TPPTS) ligands. Given

In addition to palladacycle 8, palladacycle 12 could be formed

under two different sets of conditions. Complex 12 could be

(96) Sliger, M. D.; P’Pool, S. J.; Traylor, R. K.; McNeil, J., III; Young,

S. H.; Hoffman, N. W.; Klingshirn, M. A.; Rogers, R. D.; Shaughnessy,

K. H. J. Organomet. Chem. 2005, 690, 3540–3545.

(97) Amatore, C.; Jutand, A. Acc. Chem. Res. 2000, 33, 314–321.

590 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

prepared along with other products when 11 was treated with

phenyl iodide. Alternatively, complexation of TXPTS · Na₃with

Pd(OAc)₂in the presence of excess phenylboronic acid in the

absence of sodium carbonate gave 12 as the main phosphorus-

containing species in solution. Since a phenyl source is required

to form 12, we propose that it is a phenyl-substituted analogue

of 9 ([Pd(κ²-C,P-TXPTS^3-)(κ-P-TXPTS^3-)Ph]Na₆). Forma-

tion of 12 in the presence of PhB(OH)₂would involve

phenylation of 7 by phenylboronic acid and the activation of

an ortho-methyl group. Alternatively, palladacycle 12 could be

prepared by phenyl transfer to 8. Formation of 12 by the reaction

of phenyl iodide with Pd(TXPTS · Na₃)₂(11) presumably

involves oxidative addition, C-H activation, loss of HI, and

coordination of a second equivalent of TXPTS · Na₃, although

the order of these steps has not been determined.

the ¹H and ¹³C NMR are dominated by broad peaks that provide

little structural information. Spectra obtained at higher and lower

temperature show changes in peak shape, but the resonances

remain broad over the temperature range accessible in aqueous

acetonitrile. Calculated ³¹P NMR chemical shifts of the proposed

structures were insufﬁciently precise to provide insight into the

structures observed by NMR. Higher accuracy calculations will

be required in order to conﬁrm or disprove the proposed

structures.

Catalysis. TXPTS · Na₃and TMAPTS · Na₃both gave cata-

lysts that are more active for the Suzuki and Sonogashira

couplings of aryl bromides at 50 °C than the catalyst derived

from TPPTS · Na₃. TXPTS · Na₃also gives effective catalysts

for the Heck coupling of aryl bromides, while the catalyst

derivedfromTMAPTS · Na₃hassimilaractivitytotheTPPTS · Na₃-

derived catalyst. The reason for the lower activity of the

TMAPTS · Na₃-derived catalyst compared to the TXPTS · Na₃/

Pd(OA)₂system is not clear, given they have similar structural

and electronic characteristics.

The increased activity of the TXPTS · Na₃- and TMAPTS · Na₃-

derived catalysts can be attributed to the increased steric demand

and electron-donating ability of these ligands compared to

TPPTS · Na₃. Both ortho-methylated ligands are signiﬁcantly

more sterically demanding than TPPTS · Na₃, which allows them

to more readily form highly active PdL-active species.

Computational results also show that TXPTS · Na₃and

TMAPTS · Na₃are stronger electron donors than TPPTS · Na₃.

For example, the HOMO energy levels for the PdL complexes

of TXPTS · Na₃and TMAPTS · Na₃are lower than that of the

Pd(TPPTS · Na₃) complex. Higher energy HOMO orbitals would

be expected to be more reactive reducing agents in the oxidative

addition step.

In order to rationalize catalyst activity in terms of ligand

properties, it is necessary to establish that the proposed Pd-

ligand species is the true active species. It has been established

for many systems, particularly those carried out at elevated

temperatures (>100 °C), that colloidal palladium clusters are

the true active catalysts.^98–100The nature of the active species

is certainly a relevant question in these systems. In particular,

TXPTS · Na₃, which was the most effective ligand for the

catalytic reactions tried, appeared to bind weakly to Pd and was

slow to promote reduction to Pd⁰at room temperature. Under

catalytically relevant temperatures (50 °C), Pd⁰complexes were

formed readily, however. Therefore, it is possible that 11 and

14 are the catalyst resting states and direct precursors to a Pd⁰L

(L ) TXPTS · Na₃or TMAPTS · Na₃) active species. Alterna-

tively, it could be argued that 11 and 14, as well as the other

species observed by ³¹P NMR spectroscopy, are catalytically

irrelevant.

The coordination chemistry of TMAPTS · Na₃is similar to

that of TXPTS · Na₃, although it binds more strongly to Pd and

is a more effective reductant of Pd^IIthan is TXPTS · Na₃. In

the reaction of TMAPTS · Na₃with Pd(OAc)₂in the presence

of sodium carbonate (20 equiv), Pd⁰(TMAPTS · Na₃)₂(14) was

the major species present along with Pd^II(TMAPTS · Na₃)₂(13).

No free TMAPTS · Na₃was observed in this reaction. Under

these same conditions, TXPTS · Na₃gave an analogous Pd^II

complex (7), but only a trace of a Pd⁰complex (11) after sitting

overnight. Furthermore, free TXPTS · Na₃was always present

in signiﬁcant amounts when large excesses of carbonate were

present. It is interesting that TMAPTS · Na₃coordinates more

readily and is a more effective reductant of Pd(OAc)₂than

TXPTS · Na₃, despite their similar electronic and structural

properties.

Amatore has suggested that P(o-tol)₃does not reduce

Pd(OAc)₂due to steric hindrance of attack by acetate on the

Pd-bound phosphorus.⁶⁸It is possible that a similar scenario

may play a role here, depending on which of our calculated

structures is more representative of the solution-state structure.

The cone angle of Pd(TMAPTS^3-) was 16° smaller than the

corresponding TXPTS^3-complex, while the cone angles for

the trisodium salts were within 2°. The calculated cone angle

of TMAPTS · Na₃may be overly large due to the need to

accommodate coordination of each sodium ion to two sulfonate

groups. In aqueous solution, TMAPTS may adopt a conforma-

tion in which it is less hindered than TXPTS. In this scenario,

Pd-bound TMAPTS would be able to coordinate more readily

and be more susceptible to attack by acetate than TXPTS,

leading to reductive elimination.

When the Pd⁰(TMAPTS · Na₃)_ncomplex was treated with

phenyl iodide, an AB pattern very similar to that seen in the

case of TXPTS · Na₃was observed. This same complex was

formed in the reaction of TMAPTS · Na₃with Pd(OAc)₂in the

presence of sodium carbonate and phenylboronic acid. In

analogy with the TXPTS · Na₃complex (12), we propose that

the AB pattern corresponds to [Pd(κ²-C,P-TMAPTS^3-)(κ-P-

TMAPTS^3-)Ar]Na₆(15, Ar ) Ph, 4-ethylphenyl). The palla-

dacycle complex was observed as a major species under catalytic

conditions, which suggests that it may be a resting state along

the catalytic pathway or a byproduct that siphons off palladium

from the catalytic cycle.

For each of the coupling reactions, we saw no activity in the

absence of ligand. In addition, the ligand identity had a

signiﬁcant effect on catalyst efﬁciency. Therefore, we can

conclude that the ligand plays a role, although these results do

not necessarily require that the active species be a monomeric

PdL complex. The effect of aryl halide substrate structure on

catalyst activity can also provide circumstantial evidence

regarding the identity of the active species. Palladium colloids

are generally highly effective with aryl iodide and activated aryl

bromide substrates, but tend to be much less active with

Attempts to isolate the complexes derived from TXPTS · Na₃

or TMAPTS · Na₃have been unsuccessful, which has limited

1

our ability to fully characterize these structures. H and ¹³C

(98) de Vries, J. G. Dalton Trans. 2006, 421–429.

(99) Phan, N. T. S.; Van Der Sluys, M.; Jones, C. W. AdV. Synth. Catal.

2006, 348, 609–679.

NMR spectroscopy of the species formed in situ have also been

largely uninformative due to the ﬂuxional nature of these

complexes and the fact that a mixture of species is generally

present. With the exception of the free ligands and their oxides,

(100) Köhler, K.; Kleist, W.; Pröckl, S. S. Inorg. Chem. 2007, 46, 1876–

1883.

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 591

Scheme 3. Proposed Catalytic Cycle Based on ³¹P NMR Spectroscopic Observations

deactivated aryl bromides and chlorides.⁹⁸With these less active

substrates, ligand-free catalysts typically require elevated tem-

peratures, although a ligand-free catalyst system has recently

been reported to give high activity for Suzuki couplings of aryl

bromides at room temperature.¹⁰¹Catalysts derived from

TXPTS · Na₃and TMAPTS · Na₃show good activity with

deactivated aryl bromides under relatively mild conditions

(50–80 °C).

monometallic species (PdL) is the true active species rather

than colloidal palladium particles. We cannot conclusively

exclude the possibility that a phosphine-free complex or

colloid, which is unaffected by Hg⁰, is the true active species,

however.

In cross-coupling reactions employing palladium catalysts

derived from sterically demanding phosphine ligands, the

active species is generally thought to be a monophosphine

palladium species that is in equilibrium with a Pd⁰L₂resting

state (Scheme 3). The mercury experiments are consistent

with soluble Pd-phosphine complexes being the true active

species, rather than colloidal palladium. Therefore, the

phosphine complexes observed by NMR spectroscopy may

be catalytically relevant. We have observed a Pd⁰L₂species

for both the TXPTS · Na₃(11) and TMAPTS · Na₃(14) prior

to the addition of aryl halide. When an aryl iodide is added

to provide a complete catalytic system, complex 14 remains

present in the TMAPTS · Na₃system, while 11 is completely

consumed in the TXPTS · Na₃system. For both ligands, the

Pd(κ²C,P-L)(κP-L)Ph complex (L ) TXPTS · Na₃(12),

TMAPTS · Na₃(15)) was present in signiﬁcant amounts after

the aryl iodide was added.

Although palladacycle complexes are widely used as

precatalysts,¹⁰⁶the majority of evidence in the literature

suggests that they are not catalytically viable.^98,107,108Most

commonly, they are thought to be precursors to palladium

colloids, although they can also act as precursors to ligand-

bound Pd⁰species.¹⁰⁹On the basis of this prior evidence,

we believe that palladacycles 12 and 15 are side products

formed in the catalytic cycle. It is possible that 12 and 15

serve as resting states from which catalytically active species

can be regenerated. Alternatively, they may represent un-

productive side processes that consume the active catalyst.

In an effort to more directly probe the identity of the active

catalyst, we have carried out a mercury(0) poisoning test on

the TXPTS · Na₃/Pd(OAc)₂system in the Suzuki coupling of

4-bromotoluene and phenylboronic acid. Mercury(0) readily

forms amalgams with heterogeneous metal particles, such as

palladium, that are catalytically inactive. Thus, deactivation

of a catalyst system upon addition of Hg⁰provides evidence

for the active species being ligand-free metal particles or

colloids.³⁹The mercury test has been applied to a number

of palladium pincer complexes used as precatalyts in cross-

coupling reactions.^102–105In each case, catalyst activity was

completely or signiﬁcantly inhibited when Hg⁰was added

to the reaction mixture, which was interpreted as evidence

for palladium colloids being the true active species. In our

system, addition of a large excess of Hg⁰had no effect on

the rate of product formation when it was added at the

beginning or in the middle of a catalytic reaction. Addition

of Hg⁰at the beginning of the reaction should have resulted

in no activity if colloidal palladium was the true active

species, since these colloids would have been deactivated

by Hg⁰as they were formed. Addition of mercury after the

reaction had started would have stopped the catalytic reaction

if it was catalyzed by colloidal palladium. The fact that no

inhibition was observed suggests that a phosphine-bound,

(101) Deng, C.-L.; Guo, S.-M.; Xie, Y.-X.; Li, J.-H. Eur. J. Org. Chem.

2007, 1457–1462.

(102) Yu, K.; Sommer, W.; Weck, M.; Jones, C. W. J. Catal. 2004,

226, 101–110.

(106) Dupont, J.; Consorti, C. S.; Spencer, J. Chem. ReV. 2005, 105,

2527–2571.

(103) Bergbreiter, D. E.; Osburn, P. L.; Frels, J. D. AdV. Synth. Catal.

2005, 347, 172–184.

(107) Beletskaya, I. P.; Cheprakov, A. V. J. Organomet. Chem. 2004,

689, 4055–4082.

(104) Eberhard, M. R. Org. Lett. 2004, 6, 2125–2128.

(105) Nilsson, P.; Wendt, O. F. J. Organomet. Chem. 2005, 690, 4197–

4202.

(108) Böhm, V. P. W.; Herrmann, W. A. Chem.-Eur. J. 2001, 7, 4191–

4197.

(109) d’Orlyé, F.; Jutand, A. Tetrahedron 2005, 61, 9670–9678.

592 Organometallics, Vol. 27, No. 4, 2008

Moore et al.

mL), and dried with MgSO₄. The crude products were puriﬁed

by ﬂash chromatography (SiO₂).

We propose that the catalytic cycle proceeds in a similar

fashion to others proposed for sterically demanding ligands

with complexes 11 and 14 acting as direct precursors to the

Pd⁰L active catalyst.

Representative Procedure for the Sonogashira Cross-Cou-

pling Reactions. Pd(OAc)₂(8.8 mg, 0.03 mmol) and ligand (0.08

mmol) were added to a 50 mL round-bottom ﬂask equipped with

a stir bar and rubber septum while in the drybox. Upon removing

the ﬂask from the drybox, the aryl bromide (1.00 mmol),

diisopropylamine (242 mg, 2.40 mmol), degassed 1:1 CH₃CN/

H₂O (10 mL), and alkyne (1.20 mmol) were added via syringe.

For the best conversion the alkyne should be added last. The

reaction was placed in a preheated oil bath and allowed to stir

until determined to be complete by GC. The reaction was poured

into saturated sodium carbonate (50 mL), extracted with ethyl

acetate (3 × 30 mL), and dried with MgSO₄. The crude products

were puriﬁed by ﬂash chromatography (SiO₂).

Conclusion

TXTPS · Na₃and TMAPTS · Na₃are sterically demanding,

air-stable hydrophilic phosphines that provide effective catalysts

for aqueous-phase Suzuki and Sonogashira couplings of aryl

bromides under mild conditions. Moreover, they are more easily

prepared in high purity than TPPTS · Na₃. TXPTS · Na₃also

gives an active catalyst for the aqueous-phase Heck coupling,

while the catalyst derived from TMAPTS · Na₃was much less

effective. These ligands provide signiﬁcantly more active

catalysts than the commonly used TPPTS · Na₃. Computational

studies of these ligands were complicated by their polyanionic

nature. Large structural changes were observed as a function

of the presence and identity of the sulfonate counterion. These

effects suggest that choice of solvent and counterion may play

a role in the steric demand of these ligands. The computational

results show that TXPTS · Na₃and TMAPTS · Na₃are signiﬁ-

cantly more sterically demanding than TPPTS · Na₃. The ortho-

methylated ligands are also predicted to be somewhat more

electron donating than TPPTS · Na₃, a feature that could not be

ascertained from measurement of metal–carbonyl stretching

frequencies. On the basis of the computational studies on the

properties of these ligands, we conclude that the increased

activity observed for catalysts derived from TXPTS · Na₃and

TMAPTS · Na₃compared to TPPTS · Na₃is due primarily to

the signiﬁcantly larger cone angles of TXPTS · Na₃and

TMAPTS · Na₃. The increased electron-donating ability of these

ligands also likely contributes to the improved catalytic activity

seen with these ligands compared to the less electron-donating

TPPTS · Na₃.

Complexation studies showed that Pd(OAc)₂was reduced

very slowly by TXPTS · Na₃at room temperature, even in the

presence of phenylboronic acid. In contrast, TMAPTS · Na₃

reduced Pd(OAc)₂more rapidly. Both ligands gave PdL₂

complexes upon reduction of the palladium salt. Palladacycle

complexes were identiﬁed on the basis of ³¹P NMR spectral

data for both ligands. Both the PdL₂and palladacycle complexes

were observed under catalytically relevant conditions. The PdL₂

species is presumed to be the direct precursor of the active PdL

species, while it is unclear if the palladacycle serves as a resting

state or is an unproductive side reaction that consumes the active

catalyst.

Representative Procedure for the Heck Coupling Reac-

tions with Styrene. Pd(OAc)₂(5.60 mg, 0.025 mmol),

TXPTS · Na₃(43.7 mg, 0.075 mmol), and Na₂CO₃(212 mg, 2.00

mmol) were added to a 50 mL round-bottom ﬂask equipped with

a stir bar and rubber septum while in the drybox. Upon removing

the ﬂask from the drybox, the aryl halide (1.00 mmol), styrene

(156 mg, 1.50 mmol), and degassed 1:1 CH₃CN/H₂O (10 mL)

were added via syringe. The round-bottom ﬂask was placed in

an oil bath at 80 °C and allowed to stir until complete (GC).

The reaction mixture was poured into saturated sodium carbonate

(50 mL), extracted with ethyl acetate (3 × 30 mL), and dried

with MgSO₄. The crude products were puriﬁed by ﬂash chro-

matography (SiO₂).

Representative Procedure for the Heck Coupling Reac-

tions with Sodium Acrylate. The coupling was carried out as

described above for styrene, but using sodium acrylate (141 mg,

1.50 mmol) as the alkene. Upon consumption of the starting

materials, the reaction mixture was added to a saturated solution

of sodium carbonate (50 mL). The residual aryl bromide was

extracted with ethyl acetate (3 × 30 mL). The pH of the aqueous

phase was then brought to ca. 1 using concentrated H₂SO₄. The

cinnamic acid product was extracted with CH₂Cl₂(3 × 30 mL),

and the combined organic extracts were dried over MgSO₄.

Removal of solvent under reduced pressure gave the crude

product, which was recrystallized from H₂O/ethanol.

General Procedure for Synthesis of Ligand TBA Salts. The

sodium salt of the trisulfonated ligand (0.17 mmol) was placed

in a ﬂask with tetrabutylammonium chloride (167 mg, 0.60

mmol). Water (4 mL) was added and the clear solution allowed

to stir for 1 h. The mixture was extracted with methylene

chloride. Removal of solvent gave viscous oils that solidiﬁed

after drying under vacuum (0.1 Torr). The products were used

without further puriﬁcation.

General Procedure for Measurement of Ligand Electronic

Parameters. Solutions of L₂Rh(CO)Cl complexes were synthe-

sized according to the procedure reported by Vastag.⁶³The

phosphine (0.050 mmol), trimethylamine N-oxide (0.020 mmol),

and [Rh(CO)₂µ-Cl]₂(0.010 mmol) were added to a vial equipped

with a stirbar and septa while in the drybox. Upon removing

the vial from the drybox 1 mL of distilled methylene chloride

was added to the reaction vessel via a syringe and the reaction

mixture was allowed to stir for 1 h. At this time a sample of the

reaction solution was transferred to an air-free solution cell (NaCl

windows) via a disposable syringe. Rh(PPh₃)₂(CO)Cl: 1979

cm^-1, Rh(TXPTS · (TBA)₃)₂(CO)Cl: 1993 cm^-1; Rh(TMAPTS ·

(TBA)₃)₂(CO)Cl: 1993 cm^-1; Rh(TPPTS · (TBA)₃)₂(CO)Cl: 1993

Experimental Section

General Comments. TXPTS · Na₃and TMAPTS · Na₃were

prepared as previously described.^35,36All other materials were

obtained from commercial suppliers and used as received. Water

(deionized) and aqueous acetonitrile solutions were deoxygenated

by sparging with nitrogen for at least 15 min before each use.

Representative Procedure for the Suzuki Cross-Coupling

Reactions. Pd(OAc)₂(5.60 mg, 0.025 mmol), ligand (0.063

mmol), arylboronic acid (1.20 mmol), and Na₂CO₃(0.210 g, 2.00

mmol) were added to a 50 mL round-bottom ﬂask equipped with

a stir bar and rubber septum while in the drybox. Upon removing

the ﬂask from the drybox, the aryl bromide (1.00 mmol) and

deoxygenated 1:1 CH₃CN/H₂O (10 mL) were added via syringe.

The reaction was placed in a preheated oil bath at 50 or 80 °C

and allowed to stir until determined to be complete by GC

(typically 2–4 h). The reaction mixture was poured into saturated

sodium carbonate (50 mL), extracted with ethyl acetate (3 × 30

cm^-1

.

General Procedure for Ligand Complexation Studies.

Pd(OAc)₂(2.2 mg, 0.01 mmol), ligand (0.025–0.05 mmol),

Na₂CO₃(0–1 mmol), and PhB(OH)₂(0–0.05 mmol) were

combined in a small, septum-sealed vial in the drybox. The vial

was removed from the box and attached to a nitrogen manifold

Sterically Demanding, Sulfonated, Triarylphosphines

Organometallics, Vol. 27, No. 4, 2008 593

by a needle. Deoxygenated 2:1 D₂O/acetonitrile (1 mL) was

added to the vial via syringe. After mixing to dissolve all of the

solids, a portion of this solution (0.4 to 0.5 mL) was transferred

via syringe to an NMR tube that had been sealed under nitrogen

with a septum. Samples were analyzed by ³¹P{¹H}, ¹H, or

¹³C{¹H} NMR spectroscopy immediately after preparation and

then periodically over a period of several hours.

University of Alabama Graduate School (fellowship for

E.C.W.), and the NSF-funded REU program at The

University of Alabama (summer fellowship for K.P.O.)

is gratefully acknowledged. We thank Dr. Kenneth Bel-

more and Dr. Russell Timkovich for assistance with the

NMR experiments reported in this paper.

Supporting Information Available: Detailed experimental data,

product characterization data, calculated optimized molecular

geometries in Å, as well as total and zero-point energies in au, and

NMR spectra from the complexation studies. This material is

available free of charge at http://acs.pubs.org.

Acknowledgment. Financial support for this work by

the National Science Foundation (CHE-0124255), the

School of Mines and Energy Development at The Uni-

versity of Alabama, Chemical Sciences, Geosciences and

Biosciences Division, Ofﬁce of Basic Energy Sciences,

U.S. Department of Energy (DOE) (catalysis center

program), the Robert Ramsay Chair Foundation, The

OM7008606

Article Doi

Sterically demanding, sulfonated, triarylphosphines: Application to palladium-catalyzed cross-coupling, steric and electronic properties, and coordination chemistry

DOI: 10.1021/om7008606

Source and publish data:

Authors:

Article abstract of DOI:10.1021/om7008606

Full text of DOI:10.1021/om7008606

Products guided by the article

R&D Labs maybe for 13938-94-8

Relevant to this article

Hot Product