A para-hydrogen investigation of palladium-catalyzed alkyne hydrogenation

Published on Web 05/01/2007

A para-Hydrogen Investigation of Palladium-Catalyzed Alkyne

Hydrogenation

Joaqu´ın Lo´pez-Serrano,^†Simon B. Duckett,*^,†Stuart Aiken,^†

Karina Q. Almeida Len˜ero,^‡Eite Drent,^‡John P. Dunne,^†Denis Konya,^†and

Adrian C. Whitwood^†

Contribution from the Department of Chemistry, UniVersity of York, Heslington, York, U.K.

YO10 5DD, and Shell Research & Technology Centre Amsterdam, Badhuisweg 3,

NL-1031 CM. Amsterdam, Netherlands

Received January 16, 2007; E-mail: sbd3@york.ac.uk

Abstract: The complexes [Pd(bcope)(OTf)₂] (1a), where bcope is (C₈H₁₄)PCH₂-CH₂P(C₈H₁₄), and

t

[Pd(^tbucope)(OTf)₂] (1b), where bucope is (C₈H₁₄)PC₆H₄CH₂P(^tBu)₂, catalyze the conversion of diphenyl-

acetylene to cis- and trans-stilbene and 1,2-diphenylethane. When this reaction was studied with para-

hydrogen, the characterization of [Pd(bcope)(CHPhCH₂Ph)](OTf) (2a) and [Pd(^tbucope)(CHPhCH₂Ph)](OTf)

(2b) was achieved. Magnetization transfer from the R-H of the CHPhCH₂Ph ligands in these species pro-

ceeds into trans-stilbene. This process has a rate constant of 0.53 s^-1at 300 K in methanol-d₄for 2a,

where ∆H^q) 42 ( 9 kJ mol^-1and ∆S^q) -107 ( 31 J mol^-1K^-1, but in CD₂Cl₂the corresponding rate

constant is 0.18 s^-1, with ∆H^q) 79 ( 7 kJ mol^-1and ∆S^q) 5 ( 24 J mol^-1K^-1. The analogous process

for 2b was too fast to monitor in methanol, but in CD₂Cl₂the rate constant for trans-stilbene formation is

1.04 s^-1at 300 K, with ∆H^q) 94 ( 6 kJ mol^-1and ∆S^q) 69 ( 22 J mol^-1K^-1. Magnetization transfer from

one of the two inequivalent â-H sites of the CHPhCH₂Ph moiety proceeds into trans-stilbene, while the

other site shows transfer into H₂or, to a lesser extent, cis-stilbene in CD₂Cl₂, but in methanol it proceeds

into the vinyl cations [Pd(bcope)(CPhdCHPh)(MeOD)](OTf) (3a) and [Pd(^tbucope)(CPhdCHPh)(MeOD)]-

(OTf) (3b). When the same magnetization transfer processes are monitored for 1a in methanol-d₄containing

5 µL of pyridine, transfer into trans-stilbene is observed for two sites of the alkyl, but the third proton now

becomes a hydride ligand in [Pd(bcope)(H)(pyridine)](OTf) (5a) or a vinyl proton in [Pd(bcope)(CPhdCHPh)-

(pyridine)](OTf) (4a). For 1b, under the same conditions, two isomers of [Pd(^tbucope)(H)(pyridine)](OTf)

(5b and 5b′) and the neutral dihydride [Pd(^tbucope)(H)₂] (7) are detected. The single vinylic CH proton in

3 and the hydride ligands in 4 and 5 appear as strong emission signals in the corresponding ¹H NMR

spectra.

catalyst.^3,4More recently, a comprehensive study on the selective

Introduction

semihydrogenation of alkynes to (Z)-alkenes by homogeneous

The metal-catalyzed reduction of unsaturated hydrocarbons

has received much attention in the past because of the vast

number of opportunities that exist to prepare high-value

products. Such reactions therefore feature in many multistep

syntheses. For instance, the hydroformylation of an alkene is

an industrially important reaction of great significance that

produces detergents and plasticizer alcohols.¹The closely related

methoxycarbonylation of ethene to form methyl propanoate, an

intermediate in the manufacture of methyl methacrylate, is

another well-studied example where the reduction of a double

bond by a metal catalyst (e.g., Pd) is required.²Less attention

has been focused, however, on the hydrogenation of triple bonds.

Earlier investigations of alkyne hydrogenation led to the

development of heterogeneous catalysts such as the Lindlar

R-diiminepalladium(0) catalysts has been carried out by Elsevier

et al.^5-7A palladium hydride, or more generally a metal hydride,

is often proposed to have a role in such reactions,^1,8although

the direct detection of such species is relatively rare.⁹Recently,

[(1,2-(CH₂P-^tBu₂)₂C₆H₄)Pd(H)(MeOH)], a key intermediate in

the methoxycarbonylation of alkenes, was detected by NMR

spectroscopy,¹⁰and further investigations afforded the identi-

fication of the related alkyl species [(1,2-(CH₂P-^tBu₂)₂C₆H₄)-

Pd(CH₂CH₃)]⁺, which features a â-H agostic interaction. This

(3) Lindlar, H. HelV. Chim. Acta 1952, 35, 446.

(4) Lindlar, H.; Dubuis, R. Org. Synth. 1966, 46, 89.

(5) Dedieu, A.; Humbel, S.; Elsevier, C.; Grauffel, C. Theor. Chem. Acc. 2004,

112, 305-312.

(6) Kluwer, A. M.; Koblenz, T. S.; Jonischkeit, T.; Woelk, K.; Elsevier, C. J.

J. Am. Chem. Soc. 2005, 127, 15470-15480.

(7) van Laren, M. W.; Duin, M. A.; Klerk, C.; Naglia, M.; Rogolino, D.;

Pelagatti, P.; Bacchi, A.; Pelizzi, C.; Elsevier, C. J. Organometallics 2002,

21, 1546-1553.

^†University of York.

^‡Shell Research & Technology Centre Amsterdam.

(1) Konya, D.; Almeida Len˜ero, K. Q.; Drent, E. Organometallics 2006, 25,

3166-3174.

(8) del R´ıo, I.; Claver, C.; van Leeuwen, P. W. N. M. Eur. J. Inorg. Chem.

2001, 2719-2738.

(2) Clegg, W.; Eastham, G. R.; Elsegood, M. R. J.; Heaton, B. T.; Iggo, J. A.;

Tooze, R. P.; Whyman, R.; Zacchini, S. Organometallics 2002, 21, 1832-

1840.

(9) Grushin, V. V. Chem. ReV. 1996, 96, 2011-2033.

(10) Clegg, W.; Eastham, G. R.; Elsegood, M. R. J.; Heaton, B. T.; Iggo, J. A.;

Tooze, R. P.; Whyman, R.; Zacchini, S. Dalton Trans. 2002, 3300-3308.

9

J. AM. CHEM. SOC. 2007, 129, 6513-6527

6513

A R T I C L E S

Lo´pez-Serrano et al.

species is proposed to be in equilibrium with the alkene hydride

[(1,2-(CH₂P-^tBu₂)₂C₆H₄)Pd(H)(η²-CH₂dCH₂)]⁺.²However, the

detection of such species alone does not necessarily imply that

they take an active part in catalysis. Truly catalytic species

should have a dynamic existence under the reaction conditions

and must react further with other reagents during catalysis.

Consequently, their concentration at equilibrium is normally very

low.

It has been shown that para-hydrogen-induced polarization

(PHIP) can be used to enable the direct NMR detection of many

metal hydride signals for complexes in solution that would

otherwise be invisible. This approach therefore overcomes the

inherent low sensitivity of NMR spectroscopy and makes it

applicable to the study of catalytic reactions.^11-15Insensitive

heteronuclei like ¹³C and ³¹P, which are scalar-coupled to the

enhanced hydride signals, can also be detected indirectly via

appropriate one-dimensional^16,17and two-dimensional^18-22cross

polarization experiments. PHIP NMR spectroscopy has also

been shown to permit the direct identification of metal com-

plexes without the need for enhancement of metal hydride

signals, an example being the characterization of intermediates

in cobalt-catalyzed hydroformylation through proton resonances

of ancillary acyl ligands.²³In a preliminary study, we reported

the detection of an alkylpalladium complex that was shown to

be a key intermediate in the palladium-catalyzed alkyne

hydrogenation of diphenylacetylene.²⁴The three alkyl protons

and the two carbons of the CHCH₂backbone of the alkyl ligand

were unambiguously assigned by PHIP NMR spectroscopy. The

minimal structure of this species was therefore confirmed to be

Pd(bcope)(PhCHCH₂). Subsequent studies using Pd(PEt₃)₂-

(OTf)₂as the metal precursor extended this study further and

enabled the detection of related vinyl hydride species.²⁵

Two possible mechanisms were considered to account for

these observations. One, a cationic mechanism, starts with a

palladium monohydride cation. Subsequent coordination and

migratory insertion of the alkyne leads to a vinyl complex, from

which the alkene is released after dihydrogen addition (Figure

1). The second proposed mechanism (Figure 2), a neutral cycle,

requires controlled reduction of a Pd(II) triflate to a Pd(0)

bisphosphine complex. Reduction of Pd(II) precursors to Pd(0)

species is accepted as the starting point of many catalytic

Figure 1. Schematic cationic mechanism for palladium bisphosphine

[M]-catalyzed alkyne hydrogenation.

Figure 2. Schematic neutral mechanism for palladium bisphosphine

[M]-catalyzed alkyne hydrogenation.

reactions,^26,27and a mechanism for this reaction based on Pd-

(0) species has been recently proposed.^5,6Our initial mechanistic

proposal for the bcope hydrogenation system followed the

neutral cycle and was supported by a preliminary theoretical

study (DFT). The proposed reaction cycle also met the key

constraints imposed by PHIP NMR spectroscopy.^20,21

Herein we describe new experimental evidence for the

generation of monohydride precursors which, in conjunction

with an extensive PHIP NMR study, strongly suggests that a

cationic mechanism is preferred for the hydrogenation of alkynes

catalyzed by palladium(II) bisphosphine triflates.

Experimental Section

General Conditions. All manipulations were carried out under inert

atmosphere conditions, using standard Schlenk techniques (with vacuum

of up to 10^-2mbar, with N₂or Ar as an inert atmosphere) or high-

vacuum techniques (10^-4mbar). Dry N₂and Ar were purchased from

BOC Gases. Storage and manipulation of samples was carried out using

standard glovebox techniques, under an atmosphere of N₂, using an

Alvic Scientific Gas Shield glovebox equipped with a freezer

(-32 °C), vacuum pump, and N₂purge facilities. Solvents were

obtained as analytical grade from Fisher. Diethyl ether was dried by

refluxing under nitrogen over sodium wire, while dichloromethane and

methanol were dried over calcium hydride. The deuterated solvents

methanol-d₄and CD₂Cl₂were purchased from Aldrich and stored under

N₂but used without further purification (when dried solvents were

employed, no difference in the speciation was observed in the reactions

described in this paper). Pd(PhCN)₂Cl₂was prepared as described in

the literature (PdCl₂was purchased from Johnson Matthey and PhCN

from Lancaster). AgOTf was obtained from Aldrich and used as

received. NMR spectra were obtained on a Bruker DRX 400, DSX

(11) Bowers, C. R.; Weitekamp, D. P. Phys. ReV. Lett. 1986, 57, 2645-2648.

(12) Natterer, J.; Bargon, J. Nucl. Magn. Reson. Spectrosc. 1997, 31, 293-

315.

(13) Duckett, S. B.; Sleigh, C. J. Prog. Nucl. Magn. Reson. Spectrosc. 1999,

34, 71-93.

(14) Duckett, S. B.; Blazina, D. Eur. J. Inorg. Chem. 2003, 16, 2901-2912.

(15) Blazina, D.; Duckett, S. B.; Dunne, J. P.; Godard, C. Dalton Trans. 2004,

17, 2601-2609.

(16) Duckett, S. B.; Newell, C. L.; Eisemberg, R. J. Am. Chem. Soc. 1993,

115, 1156-1157.

(17) Haake, M.; Natterer, J.; Bargon, J. J. Am. Chem. Soc. 1996, 118, 8688-

8691.

(18) Duckett, S. B.; Barlow, G. K.; Partridge, M. G.; Messerle, B. A. Dalton

Trans. 1995, 3427-3430.

(19) Duckett, S. B.; Mawby, R. J.; Partridge, M. G. Chem. Commun. 1996,

383-384.

(20) Jang, M.; Duckett, S. B.; Eisemberg, R. Organometallics 1996, 15, 2863-

2865.

(21) Millar, S. P.; Jang, M.; Lachicotte, R. J.; Eisemberg, R. Inorg. Chim. Acta

1998, 270, 363-375.

(22) Messerle, B. A.; Sleigh, C. J.; Partridge, M. G.; Duckett, S. B. Dalton

Trans. 1999, 1429-1435.

(23) Godard, C.; Duckett, S. B.; Polas, S.; Tooze, R. P.; Whitwood, A. C. J.

Am. Chem. Soc. 2005, 127, 4994-4995.

(24) Dunne, J. P.; Aiken, S.; Duckett, S. B.; Konya, D.; Lenero, K. Q. A.; Drent,

E. J. Am. Chem. Soc. 2004, 126, 16708-16709.

(25) Lo´pez-Serrano, J.; Duckett, S. B.; Lledo´s, A. J. Am. Chem. Soc. 2006,

128, 9596-9597.

(26) Streiter, E. R.; Blackmond, D. G.; Buchwald, S. L. J. Am. Chem. Soc.

2003, 125, 13978-13980.

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

9

6514 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

600, or DSX 700 spectrometer. For the hydrogenation studies, solutions

of the palladium catalyst (ca. 5 mM; ca. 2.5 mg of catalyst in 500 µL

of deuterated solvent) in Young’s tap NMR tubes were added to a ca.

40-fold excess of diphenylacetylene-d₁₀(ca. 20 mg; ca. 0.22 M) and

then reacted with p-H₂(3-3.5 atm). GC/MS data were collected on a

Saturn 2000 gas chromatograph-mass spectrometer combination. A

Factor-Form VF-6mg capillary column (30 m × 0.25 mm i.d. and 0.25

µm film thickness) was used for the GC separation. The initial oven

temperature was 100 °C, and two temperature ramps (100-145 °C,

2.5 °C/min; 145-250 °C, 30 °C/min) were employed, with the final

temperature being retained for 50 min. The helium carrier gas flow

rate was 1.0 mL/min. Mass spectra were recorded in the electron

ionization (EI) mode (70 eV, scanning the 30-650 m/z range). For the

GC/MS measurements, dichloromethane or methanol solutions (1 mL)

of the palladium complex (ca. 2 mg) and diphenylacetylene (ca. 50-

fold excess, 22 mg) were degassed and placed under a H₂atmosphere

(4 atm). The mixtures were heated to 37 °C for 1 h, and the sample

was shaken every 15 min. In order to obtain the product proportions,

a series of relative response factors were produced for the different

species via a control sample containing diphenylacetylene (9.3 mg, 98%

purity), cis-stilbene (10 µL, 10.14 mg, 97% purity), trans-stilbene (11.3

mg, 96% purity), and 1,2-diphenylethane (8.9 mg) in CH₂Cl₂(5 mL).

The bcope [(C₈H₁₄)PCH₂-CH₂P(C₈H₁₄)] and ^tbucope [(C₈H₁₄)PC₆H₄-

CH₂P(^tBu)₂] ligands used in this work were prepared at Shell and the

University of Bristol.^28,29

at 130 °C under 540 mmHg of ¹³CO for 5 h. The resulting mixture

was cooled, diluted with HCl (1 M, 200 mL), and extracted with diethyl

ether (3 × 50 mL). The combined ethereal extracts were washed with

water (100 mL) and dried (MgSO₄), and the solvent was removed under

reduced pressure. The residue was chromatographed on silica using

ethyl acetate (10% in hexanes), and the second band was collected to

give N,N-dibutylbenzamide-1-¹³C-d₅(3.16 g, 53%) as a pale orange

1

oil. H NMR (300 MHz, CDCl₃): δ 3.29 (bs, 2 H) 2.99 (bs, 2 H),

1.6-0.5 (m, 14 H). ¹³C NMR (75.45 MHz, CDCl₃): δ 172.160 (¹³-

CO), 137.43 (d, 65 Hz), 129.45, 128.22, 126.47, 49.19, 44.91, 31.21,

30.06, 20.17, 19.56, 14.38, 14.06. MS (EI): m/z 239 (M⁺, 12%) and

111 (M⁺- Bu₂N, 100%).

Next acetophenone-1-¹³C-d₅was synthesized. For this, methyllithium

(1.6 M, 9.6 mL, 15 mm in Et₂O) was added to a solution of N,N-

dibutylbenzamide-1-¹³C-d₅(3 g, 12.8 mm) in THF (30 mL) at room

temperature under N₂. The solution was stirred at room temperature

for 2 h, and then HCl (50 mL, 2 M) was added. The solution was

extracted with dichloromethane (3 × 50 mL), the organic extracts were

washed with water (50 mL) and dried (MgSO₄), and the solvent was

removed under reduced pressure. The residue was chromatographed

on silica using ethyl acetate (10% in hexanes) to give acetophenone-

1-¹³C-d₅(1.33 g, 86%) as a pale yellow oil. ¹H NMR (300 MHz,

CDCl₃): δ 2.32 (d, 3 H, J ) 8 Hz). ¹³C NMR (75.45 MHz, CDCl₃):

δ 195.94 (¹³CO), 135.72, 131.12, 128-126 (m), 24.98.

The acetophenone-1-¹³C-d₅was then converted to 2-phenylacetophe-

none-1-¹³C-d₁₀. Palladium acetate (22 mg, 1 mol %), Pd₂(dba)₃(48 mg,

1 mol %), and P^tBu₃(64 µL, 2 mol %) were dissolved in THF (20

mL) under N₂and stirred for 10 min. Sodium tert-butoxide (2.09 g, 22

mm) was added, followed by a solution of acetophenone-1-¹³C-d₅(1.33

g, 10.5 mm) and bromobenzene-d₅(1.60 g, 9.9 mm) in THF (20 mL).

The resulting mixture was heated at 60 °C for 4 h, poured into HCl

(50 mL, 1 M), and extracted with dichloromethane (3 × 30 mL). The

combined organic extracts were washed with water (50 mL) and dried

(MgSO₄), and the solvent was removed under reduced pressure. The

residue was used without further purification in the next step, the

preparation of diphenylacetylene-1-¹³C-d₁₀.

A solution of PCl₅(8.2 g, 39 mm) and 2-phenylacetophenone-1-

¹³C-d₁₀in benzene (30 mL) was heated at reflux for 3 h, cooled, poured

into water (50 mL), and extracted with dichloromethane (3 × 50 mL).

The combined organic extracts were washed with water (50 mL) and

dried (MgSO₄), and the solvent was removed under reduced pressure.

The residue was dissolved in ethanolic NaOEt, heated at reflux for 4

h, cooled, poured into water (100 mL), and extracted with dichlo-

romethane (3 × 50 mL). The combined organic extracts were washed

with water (50 mL) and dried (MgSO₄), and the solvent was removed

under reduced pressure. The residue was chromatographed on silica

using hexane as eluent. The solvent was removed under reduced

pressure, and the residue was crystallized from hexane to give

diphenylacetylene-1-¹³C-d₁₀(0.62 g, 33% from 2-phenylacetophenone-

1-¹³C-d₁₀) as colorless needles. ¹³C NMR (75.45 MHz, CDCl₃): δ 89.73

(¹³C-acetylene). MS (EI): m/z 189 (M⁺, 100%).

Synthesis of the Complexes. The triflate complexes used in this

study were synthesized from their chloride analogues, which were

prepared according to literature procedures.³⁰

Synthesis of [Pd(bcope)(OTf)₂] (1a). Pd(bcope)(Cl)₂(120 mg, 0.25

mmol) was suspended in dry, degassed dichloromethane (30 mL) and

AgOTf (158 mg, 0.65 mmol) added to the suspension. The mixture

was stirred overnight in the absence of light. The new suspension was

filtered off and the yellow solution concentrated to ca. 1 mL before

n-hexane was added to precipitate the product, which was dried under

vacuum. Yield: 121 mg or 63%. ³¹P{¹H} NMR (161 MHz, CD₂Cl₂):

δ 73.11 (s). Anal. Calcd for C₂₀H₃₂Cl₂F₆O₆P₂PdS₂‚2.5H₂O: C, 31.99;

H, 4.83; S, 8.54. Found: C, 31.54; H, 4.94; S, 8.54.

Synthesis of [Pd(^tbucope)(OTf)₂]‚5H₂O (1b). AgOTf (320 mg, 1.25

mmol) was added to a CH₂Cl₂(10 mL) solution of [Pd(^tbucope)(Cl)₂]

(200 mg, 0.31 mmol). The reaction mixture was stirred at room

temperature for 72 h in the absence of light. The resulting suspension

was filtered, and then the filtrate was evaporated to produce a colorless

powder. Yield: 158 mg or 59%. ¹H NMR (400 MHz, CD₂Cl₂): δ 7.86-

7.79 (m, 1 H, CHC₆H₄), 7.68-7.58 (m, 2 H, C₆H₄), 7.54-7.48 (m, 1

H, C₆H₄), 3.49 (m, 1 H, CH₂P), 3.35 (broad m, 1 H, C₈), 3.29 (apparent

dd, J_HH) 15.5 Hz, J_HP) 4.5 Hz, 1 H, CH₂P), 2.97-2.61 (several m,

4 H, C₈), 2.34-1.34 (several m, 9 H, C₈), 1.56 (d, J_HC) 15.5 Hz,

C^tBu), 1.22 (d, J_HC) 15.5 Hz, C^tBu). ¹³C{¹H} NMR (100 MHz, CD₂-

Cl₂): δ 137.75 (c, CCH₂PPh₂), 133.29 (s, CH C₆H₄), 132.96 (s, C₆H₄),

132.65 (s, C₆H₄), 129.39 (s, C₆H₄), 123.50 (c, CP), 41.63 (s, C^tBu),

40.53 (d, J_CP) 16.0 Hz, C^tBu), 29.85 (s, C^tBu), 29.10 (s, C^tBu), 24.52

(s, CH₂P). ³¹P{¹H} NMR (161 MHz, CD₂Cl₂): δ 110.0 (d, J_PP) 7.2

Hz, P^tBu), 31.13 (broad s, PC₈H₁₄). Anal. Calcd for C₂₅H₃₈F₆O₆P₂-

PdS₂‚5H₂O: C, 34.47; H, 5.55; S, 7.36. Found: C, 34.34; H, 5.22; S,

7.65.

Results and Discussion

Preparation and Characterization of [Pd(bcope)(OTf)₂]

(1a) and [Pd(^tbucope)(OTf)₂] (1b). The triflate complexes 1a

and 1b [where bcope is (C₈H₁₄)PCH₂-CH₂P(C₈H₁₄) and

^tbucope is (C₈H₁₄)PC₆H₄CH₂P(^tBu)₂; both have been abbreviated

as P₂where necessary] were synthesized from their chloride

analogues by the addition of AgOTf and isolated in good yield

as pale yellow/white solids. The details of these syntheses and

essential NMR data for these species can be found in the

Experimental Section.

In order to prepare the diphenylacetylene-1-¹³C-d₁₀used in this

study, the following procedures were followed. First, N,N-dibutylben-

zamide-1-¹³C-d₅was prepared. To do this, a mixture of bromobenzene-

d₅(4 g, 24.7 mm), dibutylamine (10 mL), ethyldiisopropylamine (10

mL), triphenylphosphine (520 mg, 8 mol %), bisbenzothiazolecarbene-

palladium diiodide (650 mg, 4 mol %), and tetrabutylammonium

bromide (4.5 g) in dimethylamine (30 mL) was degassed and heated

X-ray Diffraction Studies. The crystal and molecular

structures of compounds [Pd(^tbucope)(Cl)₂] (see Figure 3 and

Supporting Information), [Pd(^tbucope)(OH₂)₂](OTf)₂‚2(H₂O)

(28) Eberhard, M. P. Ph.D. Thesis, University of Bristol, 2000.

(29) Marsh, P. Ph.D. Thesis, University of Bristol, 2003.

(30) Stang, P. J.; Cao, D. H.; Saito, S.; Arif, A. M. J. Am. Chem. Soc. 1995,

117, 6273-6283.

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6515

A R T I C L E S

Lo´pez-Serrano et al.

Figure 3. Molecular structure of [Pd(^tbucope)Cl₂]: ORTEP view showing

50% probability ellipsoids, and selected bond lengths (Å) and angles (°).

Hydrogen atoms have been omitted for clarity.

(Figures 4 and 5) and [Pd(bcope)(MeOH)₂](OTf)₂‚2(MeOH)

(Figure 6) have been determined by X-ray diffraction studies.

Selected crystallographic data for these complexes are presented

in Table 1; complete crystallographic data for all three com-

plexes are given in the Supporting Information. For all these

structures, the hydrogens were placed using a riding model,

except for the hydroxyl hydrogens, which were placed by

difference map after all the other atoms had been positioned

and refined.

Single-crystal X-ray-quality crystals of [Pd(^tbucope)(Cl)₂]

were obtained by addition of diethyl ether to dichloromethane

solutions of the solid. The palladium atom in this species is in

the center of a square-planar arrangement of two chlorine and

two cis-phosphorus atoms, and the P-Pd-P (90.497(19)°) angle

is close to the ideal 90°. The Pd-Cl(1) and Pd-Cl(2) bond

lengths are similar and comparable to Pd-Cl distances in other

palladium bisphosphine dichloride complexes. While the posi-

tion of the carbon atoms of one of the tert-butyl groups was

disordered, two positions could be successfully modeled, the

relative occupancy of which was refined to 45 and 55

respectively.

Figure 4. Molecular structure of the cation [Pd(^tbucope)(OH₂)₂]²⁺: ORTEP

view showing 50% probability ellipsoids, and selected bond lengths (Å)

and angles (°). Hydrogen atoms other than those of the H₂O ligands have

been omitted for clarity.

are hydrogen-bonded to the same triflate and to two independent

water molecules. In addition, these waters of crystallization are

each bonded to a bridging triflate anion and to another

independent triflate, as shown in Figure 5. The OsH‚‚‚O

distances are in agreement with those of related complexes,^31-36

and similar hydrogen-bonding patterns have been described for

other palladium aqua complexes with triflate counteran-

ions.^31,37,38The hydrogen-bond networks constitute infinite

chains parallel to the crystallographic a-axis.

When 1a was successfully crystallized from cooled methanol

solutions, the metal center contained two methanol solvates that

were bound directly to palladium, in addition to a further pair

which extended out to the two triflate counterions via appropriate

hydrogen bonds. This species, [Pd(bcope)(MeOH)₂](OTf)₂‚

When 1b was successfully crystallized by addition of diethyl

ether to dichloromethane solutions, the cation actually proved

to be the di-aqua complex, [Pd(^tbucope)(OH₂)₂]²⁺. The pal-

ladium atom proved to be located in a distorted square-planar

environment, with a phosphine bite angle P(1)-Pd(1)-P(2) of

90.049(16)°, which is very similar to that of the chloride

precursor [Pd(^tbucope)(Cl)₂]. The corresponding phosphorus-

palladium distances are slightly shorter than those of the

chloride, at 2.2343(4) and 2.2619(4) Å, rather than 2.2485(5)

and 2.2953(5) Å, respectively, and confirm that the H₂O ligands

have a poorer trans influence than chloride ligands.

(31) Qin, Z.; Jennings, M. C.; Puddephatt, R. J. Inorg. Chem. 2001, 40, 6220-

6228.

(32) Ruiz, J.; Florenciano, F.; Vicente, C.; De Arellano, M. C. R.; Lo´pez, G.

Inorg. Chem. Commun. 2000, 3, 73-75.

(33) Ferrer, M.; Mounir, M.; Rossell, O.; Ruiz, E.; Maestro, M. A. Inorg. Chem.

2003, 42, 5890.

(34) Gusev, O. V.; Kalsin, A. M.; Peterleitner, M. G.; Petrovskii, P. V.; Lyssenko,

K. A.; Akhmedov, N. G.; Bianchini, C.; Meli, A.; Oberhauser, W.

Organometallics 2002, 21, 3637-3649.

(35) Vicente, J.; Arcas, A.; Bautista, D.; Jones, P. G. Organometallics 1997,

16, 2127-2138.

(36) Stang, P. J.; Cao, D. H.; Poulter, G. T.; Arif, A. M. Organometallics 1995,

14, 1110-1114.

(37) Stang, P. J.; Cao, D. H.; Saito, S.; Arif, A. M. J. Am. Chem. Soc. 1995,

117, 6273-6283.

Supramolecular Structure of [Pd(^tbucope)(OH₂)₂](OTf)₂‚

2(H₂O). Both water ligands in the cation [Pd(^tbucope)(OH₂)₂]²⁺

(38) Vicente, J.; Arcas, A. Coord. Chem. ReV. 2005, 249, 1135-1154.

9

6516 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

Figure 5. Views perpendicular to the b (left) and c (right) crystallographic axes of the supramolecular structure of [Pd(^tbucope)(OH₂)](OTf)₂‚2(H₂O),

showing the hydrogen-bonding pattern. The cation units are represented by the metal center (big pink sphere), the phosphorus atoms, and the water ligands.

Crystallization water molecules participate in the hydrogen-bonding pattern that build the infinite chains found along the crystallographic a axis.

The formation of palladium aqua complexes from the

corresponding chlorides and silver or thallium salts is not rare,

but there are relatively few reported structures of palladium

triflate complexes, and there is only one reported example of

the crystal structure of a bis-triflate complex.³⁹However, it is

well known that the triflate ligand can be easily replaced by

water molecules present in or added to the solvent.^36,38-40

Moreover, the formation of hydrogen-bonded networks clearly

stabilizes such species with respect to the bis-triflate.^41-43To

the best of our knowledge, only one structure for a palladium-

methanol complex has been described previously.⁴⁴

Hydrogenation of Diphenylacetylene in Methanol: Mecha-

nistic Implications. Typical NMR experiments were completed

in CD₂Cl₂or methanol-d₄containing the complex (1a or 1b)

and a 40-fold excess of diphenylacetylene-d₁₀. Three atmo-

spheres of para-hydrogen was then introduced over the solution

at 295 K, and the ensuing reaction was monitored by multi-

nuclear NMR spectroscopy between 265 and 333 K. During

these studies, the utilization of p-H₂resulted in the observation

of enhanced proton resonances for reaction products containing

two hydrogen atoms that were originally present in a single p-H₂

molecule (Scheme 1). This effect provides information about

the mechanism of reaction and is central to much of the work

now reported.

In summary, the corresponding ¹H NMR spectra show weakly

Figure 6. Molecular structure of [Pd(bcope)(MeOH)₂](OTf)₂‚2MeOH,

enhanced signals at the beginning of the experiments which

correspond to the formation of cis-stilbene at δ 6.66, as well as

unenhanced trans-stilbene at δ 7.18. These signals increase in

intensity, relative to the background, during observation and

hence correspond to reaction products that are essentially stable

on the NMR time scale. Since the two alkene protons of the cis

isomer are chemically and magnetically equivalent, the observa-

tion of a p-H₂-enhanced emission signal at δ 6.66 is consistent

with the involvement of an intermediate where two hydrogen

atoms of the p-H₂molecule become inequivalent.⁴⁵The fact

that the proton signal for trans-stilbene grows in intensity even

including the hydrogen-bonding pattern: ORTEP view showing 50%

probability ellipsoids, and selected bond lengths (Å) and angles (°).

Hydrogen atoms other than those of the MeOH units have been omitted

for clarity.

2(MeOH), therefore possesses a coordination environment of

two cis phosphorus atoms and two oxygen atoms which belong

to two independent methanol molecules. These crystals belong

to the orthorhombic crystalline system, and the asymmetric unit

contains a Pd{(PC₈H₁₄)CH₂}(MeOH) moiety, one triflate anion,

and one methanol of crystallization. The rest of the atoms

(denoted by “_2” or #) can therefore be generated by appropriate

symmetry operations. In this species, the cation can be regarded

as a solvento complex where two methanols have displaced the

triflate counteranions from the palladium center.

The P(1)-Pd-P(2) angle is significantly smaller in this

species (84.36(3)°) than in both [Pd(^tbucope)(Cl)₂] and

[Pd(^tbucope)(OH₂)₂](OTf)₂‚2(H₂O) (90.497(19) and 90.049-

(16)°, respectively). In contrast, the P-Pd distance of 2.2446-

(5) Å here is slightly longer than that in [Pd(^tbucope)(OH₂)₂]-

(OTf)₂‚2(H₂O), 2.2343(4) Å.

(39) Anandhi, U.; Holbert, Y.; Lueng, D.; Sharp, P. R. Inorg. Chem. 2003, 42,

1282.

(40) Song, L. C.; Jin, G. X.; Zhan, W. X.; Hu, Q. M. Organometallics 2005,

24, 700-706.

(41) Rochon, F. D.; Melanson, R. Inorg. Chem. 1987, 26, 989.

(42) Braga, D.; Grepioni, F. Acc. Chem. Res. 1994, 27, 51.

(43) Braga, D.; Grepioni, F.; Sabatino, P.; Desiraju, G. R. Organometallics 1994,

13, 3532.

(44) Quek, G. H.; Leung, P.-H.; Mok, K. F. Inorg. Chim. Acta 1995, 239, 185-

188.

(45) Aime, S.; Gobetto, R.; Canet, D. J. Am. Chem. Soc. 1998, 120, 6770-

6773.

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6517

A R T I C L E S

Lo´pez-Serrano et al.

Table 1. Selected Crystallographic Data for Pd(^tbucope)Cl₂, [Pd(bcope)(MeOH)₂](OTf)₂‚2MeOH, and [Pd(^tbucope)(OH₂)₂](OTf)₂‚2(H₂O)

Pd(^tbucope)Cl₂

[1a(MeOH)₂](MeOH)

‚

2(OTf)₂

[1b(OH₂)₂](OTf)₂‚2(H₂O)

formula

M_r

C₂₃H₃₈Cl₂P₂Pd

553.77

C₂₄H₄₈F₆O₁₀P₂PdS₂

843.08

C₂₅H₄₆F₆O₁₀P₂PdS₂

853.08

cryst size (mm)

cryst syst

space group

0.24 × 0.16 × 0.08

monoclinic

P2₁/c

0.15 × 0.13 × 0.11

orthorhombic

P2₁2₁2

0.24 × 0.23 × 0.20

monoclinic

P2₁/n

cell constants

a, Å

b, Å

c, Å

11.057(2)

15.655(3)

14.870(3)

90

109.911

90

10.1425(6)

12.0361(8)

14.1564(9)

90

14.8608(15)

16.6018(16)

14.8694(15)

90

108.403(2)

90

R, deg

â, deg

γ, deg

V, Å³

2420.1(8)

4

1728.16(19)

2

3480.9(6)

4

Z

λ, Å

0.71073

1.520

1.128

0.71073

1.620

0.831

0.71073

1.628

0.827

F(calcd), Mg m^-3

µ, mm^-1

F(000)

1144

868

1752

T, K

θ range, deg

no. of rflns measd

no. of indep reflns

max and min transmissn

R_int

115(2)

1.95-28.30

24424

110(2)

2.22-30.03

19608

100(2)

1.69-30.04

27146

10013

1.000 and 0.624

0.0222

0/453

0.0646

5994

4996

1.000 and 0.873

0.0224

0/290

0.0514

0.0229

1.054

0.913 and 0.845

0.0247

0/214

0.0658

0.0272

1.056

no. of restraints/params

R_W(F²) (all rflns)

R(I) (>2σ(I))

S

0.0260

1.029

0.607 and -0.419

largest diff peak and hole, e Å^-3

0.578 and -0.545

0.679 and -0.466

Scheme 1

tion of a cationic hydride complex in the formation of methyl

propanoate from CO and ethane. This work was completed in

deuterated methanol, and evidence was presented that Pd-H/

CH₃OD exchange was slow relative to propagation.⁴⁶We will

return to this point later in the manuscript.

The mechanistic routes taken to form these species are now

discussed explicitly.

Inorganic Complex Speciation 1. Reactions of Pd(bcope)-

(OTf)₂(1a) with p-H₂and Diphenylacetylene in CD₂Cl₂. We

have previously described the finding that, when a 1 µΜ solution

of 1a in CD₂Cl₂containing a 50-fold excess of diphenylacety-

lene-d₁₀is placed under 3 atm of 100% p-H₂at 213 K and

rapidly introduced into a 400 MHz NMR spectrometer, a new

palladium complex is detected.²⁴The studies now described

demonstrate that this corresponds to the cation [Pd(bcope)-

(CHPhCH₂Ph)](OTf) (2a) and not Pd(bcope)(CHPhCH₂Ph)(H)

as originally suggested.

The starting ³¹P{¹H} NMR spectrum contains a singlet at δ

73.11 due to 1a, but the most notable feature of the correspond-

ing ¹H NMR spectrum of 2a is the observation of substantially

enhanced proton signals at δ 4.94 and δ 3.13, and a further

weakly enhanced signal at δ 2.92 (Table 2). These signals

contained characteristic antiphase components due to their origin

as protons in p-H₂, proved to be coupled in a COSY spectrum,

and simplified on ³¹P decoupling. ¹H-³¹P heteronuclear multi-

quantum coherence (HMQC) spectra then revealed that this

though this signal does not show p-H₂enhancement is consistent

with it being the thermodynamically preferred product.

When such reactions are examined using ¹³C-enriched diphen-

ylacetylene-d₁₀, the δ 6.66 signal is strongly enhanced, while

the δ 7.18 signal shows limited PHIP enhancement. These

effects arise because both of the two alkenic protons of cis-

and trans-PhCHd¹³CHPh are magnetically inequivalent due to

the presence of the ¹³C label, and efficiently enhanced NMR

transitions are now observed.

The % conversion obtained after 1 h was determined by GC/

MS when 2 mg of catalyst was added to a 50-fold excess of

substrate in 1 mL of solvent and the resultant reaction with 3

atm of H₂was initiated at 312 K. These data (see Supporting

Information) confirm that (i) the hydrogenation activity is greater

in methanol than CD₂Cl₂, (ii) 1b favors the formation of trans-

stilbene, and (iii) 1b is more active than 1a. It also reveals that

1a shows a higher degree of C-C coupling products. It should

also be noted that, when these reactions were completed in

methanol-d₄, essentially no deuteration of the products was

observed according to the GC/MS fragmentation patterns. In a

related study, Cole-Hamilton et al. demonstrated the participa-

species yielded two ³¹P doublets at δ 32.3 and δ 42.9 (J_PP

)

47 Hz). A ¹H-¹³C HMQC spectrum produced correlations from

the proton resonance at δ 4.94 to a ¹³C signal at δ 63.0, with

the remaining proton resonances connecting to a signal at δ

37.1. We therefore concluded that these resonances arise from

a ligand that is attached to palladium. When this reaction was

(46) Eastham, G. R.; Tooze, R. P.; Kilner, M.; Foster, D. F.; Cole-Hamilton,

D. J. Dalton Trans. 2002, 1613-1617.

9

6518 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

Table 2. Selected ¹H, ³¹P, and ¹³C NMR Data for Complexes 2-7 at 298 K as a Function of Solvent

δ

¹H

δ

³¹P

δ

¹³C

in CD₂Cl₂

in methanol-d₄

in CD₂Cl₂

in methanol-d₄

in CD₂Cl₂

in methanol-d₄

Pd(bcope)²⁺System

2a

4.94 (CHPh);

3.13, 2.92 (CH₂Ph)

4.98 (CHPh); 3.06,

2.89 (CH₂Ph)

32.30, 42.90 (d,

J_PP) 47 Hz)

32.20, 43.40 (d,

J_PP) 40 Hz)

63.0 (CH), 37.1

(J_CP) 42, 14 Hz;

CH₂)

62.30 (dd, J_CP

54, 16 Hz; CH),

35.60 (dd, J_CP

)

16, 5 Hz; CH₂)

161.2 (CPhd),

130.6 (dCH₂Ph)

3a

4a

6.75 (dd, J_HP

13.6, 6.8 Hz)

)

21.79 (m, trans to

vinyl), 39.04 (cis

to vinyl)

20.17 (m, trans to

vinyl), 33.41 (cis

to vinyl)

6.72 (dd, J_HP

13.2, 5.6 Hz)

)

6.70 (dd, J_HP

12.7, 5.2 Hz)

19.70 (trans to

hydride), 33.51

(cis to hydride)

129.9 (d, J_CP

18 Hz; CPhd),

)

163.3 (dd, J_CP

)

109, 10 Hz; dCH₂Ph)

5a

-7.04 (dd, J_HP

)

-6.96 (dd, J_HP

)

22.33 (d, J_PP)

22 Hz, cis to

223.2, 19.0 Hz)

232.3, 17.2 Hz)

hydride), 43.40

(trans to hydride)

28.1, 57.4

6

-6.40 (ddd, J_HP

190.7, 25.3, 11.6 Hz)

)

-6.30 (ddd, J_HP)

191.1, 23.8, 12.9 Hz)

Pd(^tbcope)²⁺System

2b

4.92 (CHPh); 3.02,

2.95 (CH₂Ph)

4.93 (CHPh); 3.03,

2.91 (CH₂Ph)

70.70 (P^tBu₂trans

to alkyl), 20.40

70.43 (P^tBu₂trans

to alkyl), 19.81

(J_PP) 60 Hz)

63.94 (CHPh)

35.98 (CH₂Ph)

64.0 (d, J_CP)

55 Hz; CH), 34.8

(CH₂)

3b

4b

6.72 (dd, J_HP

12.0, 8.9 Hz)

6.62 (dd, J_HP

11.7, 8.8 Hz)

)

6.61 (dd, J_HP

11.4, 8.4 Hz)

)

15.11 (d, J_PP

42 Hz), 57.62 (d,

)

128.6 (CPhd),

163.9 (d, J_CP

)

J_PP) 42 Hz)

101 Hz; dCH₂Ph)

5b

-7.77 (dd, J_HP

)

-7.69 (major) (dd,

J_HP) 209.4, 8.3 Hz)

-2.20 (d, J_HP

)

-2.54 (d, J_HP)

208.5, 9.7 Hz)

24 Hz, trans to

27 Hz, trans to

hydride), 100.96

hydride), 101.92

(d, J_HP) 27 Hz,

trans to pyridine)

5b′

-7.91 (dd, J_HP

205.2, 13.5 Hz)

-4.21 (ddd, J_HH)

-7.84 (minor) (dd,

J_HP) 205.5, 13.3 Hz)

7

18.0 Hz; J_HP) 138,

9 Hz; trans to P^tBu₂),

-4.71 (ddd, J_HH

)

18.0 Hz; J_HP) 146, 9

Hz; trans to PC₈H₁₁)

re-examined using ¹³CtC-enriched diphenylacetylene-d₁₀, strong

signals were observed in the one-dimensional, fully coupled ¹³

Scheme 2. Proposed Mechanism for Alkene Isomerization Based

on Neutral Intermediates

C

NMR experiment at δ 63.0 and δ 37.1; the former showed two

³¹P splittings of 42 and 14 Hz in addition to a single proton

splitting of 147 Hz. The δ 63.0 signal therefore corresponds to

a CH group that is bound directly to palladium. In this spectrum,

the δ 37.1 signal appears as a pseudo-triplet as a consequence

of the PHIP effect, with lines of relative intensities 1, 0, and

-1, and therefore it corresponds to a CH₂moiety (J_CH) 130

Hz). This would be consistent with the fact that cis-PhsCHd

CHsPh has been converted into a CHPhCH₂Ph group which

is bound to the metal.

In order to explore the reactivity of 2a, a series of modified

one-dimensional exchange spectroscopy (1D EXSY) experi-

ments were recorded where a single alkyl proton resonance was

selected and magnetization transfer from this site monitored as

a function of mixing time. For the δ 4.94 peak, strong

magnetization transfer into the trans-stilbene signal at δ 7.18

was observed at 295 K. When the sample was warmed to 313

K, even greater magnetization transfer from the δ 4.94 site into

trans-stilbene was seen in conjunction with weaker transfer into

cis-stilbene and into both of the previously described CH₂Ph

proton sites at δ 3.13 and δ 2.92. When the peak at δ 3.13 was

selected in an analogous EXSY experiment at 313 K, exchange

into free H₂, and weaker exchange into free cis- and trans-

stilbene as well as into the δ 4.94 position of 2a, was observed.

However, when the resonance at δ 2.92 was selected, only

transfer into trans-stilbene was seen. These observations confirm

that the proton yielding the resonance at δ 3.13 transfers directly

into free H₂, while those at δ 4.94 and δ 2.92 move into trans-

stilbene. In our earlier report,²⁴these data were taken to indicate

that 2a was an alkyl hydride, even though the hydride resonance

itself was not observed (Scheme 2). This deduction was

supported by a number of theoretical calculations.²⁵

Inorganic Complex Speciation 2. Reactions of Pd(bcope)-

(OTf)₂(1a) with p-H₂and Diphenylacetylene in Methanol-

d₄. (a) Identification of 2a as the Cation Pd(bcope)-

(CHPhCH₂Ph)⁺. A sample containing 1a, diphenylacetylene-

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6519

A R T I C L E S

Lo´pez-Serrano et al.

1

Figure 7. (a) H NMR spectra for the reaction of 1a with diphenylacetylene-d₁₀and p-H₂in methanol-d₄, showing H resonances for trans-stilbene, cis-

stilbene (in emission), antiphase PHIP signals for the alkyl protons of 2a, and an emission doublet of doublets for the vinyl proton of 3a. In the expanded

inset traces, the effect of broadband ³¹P decoupling (c) on the emission resonances for cis-stilbene and the vinyl proton of 3a is illustrated (b).

d₁₀, and p-H₂in methanol-d₄was prepared and monitored by

NMR spectroscopy. The starting ³¹P{¹H} NMR spectrum

contained a singlet at δ 72.12 due to the methanol solvate of

of 2a at δ 32.20 showed substantial broadening, which we

originally took to indicate the presence of such a ligand. The

earlier demonstration of the concomitant elimination of H₂and

trans-stilbene from 2a, and the fact that the hydrogenation

1

1a. The corresponding H NMR spectrum yielded signals for

1

the same organic products as observed in CD₂Cl₂and enhanced

proton signals due to the palladium-based species 2a at δ 4.98,

3.06, and 2.89 that were now of much greater intensity, in

accordance with a much higher hydrogenation activity, but there

were still no signals seen in the hydride region of the ¹H NMR

spectrum (Figure 7). These three proton NMR signals still

contained the characteristic antiphase components seen in CD₂-

Cl₂and were all mutually coupled. Now, the PHIP-enhanced δ

4.98 signal clearly possesses two J_HHcouplings of 11.5 and

4.0 Hz and two ³¹P splittings of 10.6 and 4.3 Hz, which can be

removed by appropriate ³¹P decoupling. The corresponding ¹H-

³¹P HMQC spectrum, which utilizes magnetization transfer from

the three enhanced proton resonances, showed two ³¹P doublets

at δ 43.40 and δ 32.20 (d, J_PP) 40 Hz). Two ¹³C signals were

also located for this species in methanol-d₄, at δ 62.30 (dd, J_CP

) 54.0 and 16.0 Hz, CH) and δ 35.60 (dd, J_CP) 16.0 and 5.0

Hz, CH₂). The chemical shifts of the enhanced proton reso-

nances, and the associated ³¹P and ¹³C signals of 2a seen in

methanol-d₄, are therefore very similar to those seen in

CD₂Cl₂(Table 2). This suggests that 2a does not contain bound

solvent.

Species 2a is therefore minimally described as Pd(bcope)-

(CHPhCH₂Ph). This moiety would, however, be paramagnetic,

and 2a must therefore either be cationic or, as suggested in our

previous report,²⁴contain a hydride ligand, which is not

observed. We thought that this situation was not too surprising

since the successful observations of palladium hydride reso-

nances are relatively rare and their appearance is very temper-

ature-sensitive. Furthermore, we note that, for the PHIP effect

to operate in the traditional way for two inequivalent protons,

we need a resolvable H*-H* coupling, which in this case would

be vanishingly small for the necessary four-bond Pd(H*)-

(CHPhCHH*Ph) connection. We also recorded a ³¹P insensitive

nuclei enhanced by polarization transfer (INEPT) spectrum,

where a hydride coupling should be retained. The ³¹P resonance

products contain only additional H labels, even in methanol-

d₄, were taken as additional indirect proof, but surprisingly, an

alternative, more complex explanation for these observations

was revealed when we continued to explore the reactivity of

the methanol system.

In the related cationic bis-phosphine complex Pd(1,2-(CH₂P-

^tBu₂)₂C₆H₄)(CH₂CH₃)⁺, the ethyl ligand interacts with the metal

center by an agostic C-H interaction from the â carbon.²An

alternative situation has been reported in related cationic

benzylpalladium systems featuring coordination of the pendant

arene to the metal center.^8,47

Since the chemical shifts of the alkyl ligand resonances in

2a are remarkably similar in CD₂Cl₂and methanol-d₄, coordina-

tion of the solvent or water is unlikely.⁴⁸Values of J_HChave

been used extensively to assess such agostic interactions,^49-51

and the corresponding proton-carbon coupling constants for

the CH and CH₂carbons of the alkyl ligand in 2a are 151 and

130 Hz, respectively. While these values are similar to those

reported by Heaton et al. for the palladium ethyl system,¹⁰they

are relatively large for an agostic interaction.

In order to establish the exact nature of 2a, full characteriza-

tion of the proton and carbon resonances for the alkyl ligand is

necessary. An NMR sample of 1a (3 mg) and protio-cis-stilbene

(12 µL) in methanol-d₄was therefore prepared, cooled to 258

K, and reacted with H₂(3 atm). The slow conversion of 1a

1

into 2a was followed by H and ³¹P NMR spectroscopy. The

(47) Hii, K. K. M.; Claridge, T. D. W.; Giernoth, R.; Brown, J. M. AdV. Synth.

Catal. 2004, 346, 983-988.

(48) CD₂Cl₂was carefully dried from calcium hydride and stored with 4 Å

molecular sieves in the absence of light. Methanol-d₄was similarly dried

with iodine and magnesium turnings. When samples of 1a and dipheny-

lacetylene were prepared using these solvents, no changes in the chemical

shift of the detected intermediates were observed.

(49) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250, 395-

408.

(50) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789-805.

(51) Shultz, L. H.; Tempel, D. J.; Brookhart, M. J. Am. Chem. Soc. 2001, 123,

11539-11555.

9

6520 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

two inequivalent ³¹P centers (Figure 7). When such a sample is

cooled to 275 K, this emission signal grows in strength and

can ultimately be seen as a normal NMR signal from a thermally

equilibrated spin system. In contrast, the associated resonances

Table 3. Selected NMR Data Obtained for 2a in Methanol-d₄at

258 K

assignment

δ

¹H

assignment

δ

¹³C

a

3.13, 2.94

a

35.5

1

(J_CP) 5 Hz)

62.7

(J_CP) 12, 44 Hz)

for 2a were only visible in the associated H NMR spectra

b

d

5.11

b

because of the PHIP effect. This observation agrees with the

fact that, when a ³¹P{¹H} NMR spectrum was recorded, in

addition to the dominant signal for 1a, two doublets appeared

at δ 39.04 and δ 21.75, with J_PP) 40.0 Hz. In the corresponding

(J_HH) 5 Hz,

J_HP) 5 Hz,

J_HC) 144 Hz)

c

119.0

(t, J_CP) 6 Hz)

105.4

1

¹H-³¹P HMQC experiment, these signals coupled to the H

6.98

d

signal at δ 6.75, with the former resonance providing the 13.6

Hz splitting.

(J_HH) 5 Hz,

J_HP) 5 Hz,

J_HC) 143 Hz)

8.02

(J_HC) 162 Hz)

7.54

(d, J_CP) 14 Hz)

Before we identify this species (3a), we must first comment

on the fact that the δ 6.75 signal appears in emission. Such a

situation also occurs for the cis isomer of stilbene, where the

two alkene protons are chemically and magnetically equivalent

and appear at δ 6.66. In this case, the involvement of an

intermediate in the formation of cis-stilbeneis indicated, where

two hydrogen atoms of the original p-H₂molecule become

inequivalent but are then transported into the final, symmetrical

environment;⁴⁵in other words, “para-cis-stilbene” is formed.

The binding of cis-CHPhdCHPh to palladium would meet this

requirement, but the corresponding proton resonances of Pd-

(bcope)(η²-CHPhdCHPh) would be expected to appear around

δ 5, and only single ³¹P and alkenenic ¹³C signals should be

observed for this species.^7,52,53

e

f

e

134.2

f, h

129.2/132.2

(J_CP) 3 Hz)

129.1

g

h

7.05

7.54

g

i

141.0

(t, J_CP) 10 Hz)

j, j′

k, k′

l

7.12

7.21

7.18

j, j′

k, k′

l

128.4

127.8

126.0

Chart 1. Proposed Structure for 2a

We therefore proceeded to record a ¹H-¹³C HMQC spectrum

using ¹³C-enriched diphenylacetylene-d₁₀to confirm the exact

nature of 3a. These spectra were recorded at 313 K and showed

two ¹³C resonances, at δ 130.6 and δ 161.2, coupled to the

proton resonance of 3a at δ 6.75. The carbon resonance at δ

161.2 is indicative of a vinyl carbon signal, but the poor

resolution prevented the detection of diagnostic carbon-

phosphorus couplings that would confirm this. Nonetheless, this

suggests that 3a has the form [Pd(bcope)(cis-CPhdCHPh)], and

the remaining coordination site must correspond to either a

hydride ligand if the complex is neutral or methanol if it is

cationic. Furthermore, the emission signal arises from a single

proton that was previously found in para-hydrogen. This effect

has been seen before and has been referred to as a one-proton

PHIP.⁵⁴On the basis of the fact that we have previously

identified cis-Pd(PEt₃)₂(CPhdCHPh)(H) through enhanced and

coupled hydride and vinyl signals, and that these are not seen

this time, we believe that 3a is the cation [Pd(bcope)(cis-CPhd

CHPh)(MeOD)](OTf). This species can be observed in CD₂-

Cl₂if 10 µL of methanol-d₄is added to the reacting solution. A

similar complex can be detected when ethanol rather than

methanol is added to the solution (δ 6.76 at 313 K, J_HP) 14.6,

9.2 Hz).

additional NMR data obtained for 2a are presented in Table 3.

The most significant feature of the H NMR spectrum cor-

1

responded to the observation of five distinct resonances

for one of the phenyl rings of the alkyl ligand, which indicates

that there is restricted rotation about the corresponding C-C

bond. Four of these protons coupled to the δ 43.5 ³¹P signal

arising from the phosphorus center, which is cis to the alkyl

1

ligand in the corresponding H-³¹P spectrum. The size of the

corresponding ortho-phenyl proton-phosphorus splitting was

5 Hz. In the corresponding ¹H-¹³C spectrum, the ortho proton

connected with a ¹³C center at δ 105.4 which possesses a ³¹P

splitting of 14 Hz; in addition, it showed NOE connections to

the δ 5.11 proton of the alkyl, thereby confirming that the

phenyl group was on the R-carbon of the alkyl. The correspond

J_HCvalue was determined to be 143 Hz. The ipso carbon of the

phenyl ring showed a triplet ³¹P splitting of 6 Hz, in accord

with this deduction. These data provide unambiguous

evidence that the palladium center interacts with the π-system

of the arene rather than the C-H bond. We therefore conclude

that 2a is stabilized by metal-ring association, as illustrated in

Chart 1.⁴⁷

A series of 1D EXSY experiments were then used to probe

magnetization transfer within 2a in methanol-d₄. Selective

irradiation of the ¹H signals at δ 4.98 and δ 2.98 led to transfer

into trans-stilbene. Interestingly, excitation of the remaining

resonance at δ 3.06 revealed exchange into the δ 6.75 signal

due to 3a and not into free H₂, as seen in CD₂Cl₂.

(b) Detection of the Vinyl Complex [Pd(bcope)(CPhd

CHPh)(MeOH)](OTf) (3a). In the early stages of this reaction,

when the catalytic activity of the system is optimal, a weak

emission signal appears as a doublet of doublets at δ 6.75 (13.5

and 6.8 Hz) in methanol-d₄; this signal is not seen in CD₂Cl₂.

This new signal becomes a singlet on ³¹P decoupling, and

consequently the two splittings must arise from interactions with

(52) Liu, W.; Brookhart, M. Organometallics 2004, 23, 6099-6107.

(53) Brunker, T. J.; Blank, N. F.; Moncarz, J. R.; Scriban, C.; Anderson, B. J.;

Glueck, D. S.; Zakharov, L. N.; Golen, J. A.; Sommer, R. D.; Incarvito, C.

D.; Rheingold, A. L. Organometallics 2005, 24, 2730-2746.

(54) Permin, A. B.; Eisenberg, R. J. Am. Chem. Soc. 2002, 124, 12406-1407.

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6521

A R T I C L E S

Lo´pez-Serrano et al.

1

Figure 9. Hydride region of H (upper trace) and H{³¹P} (lower trace)

NMR spectra obtained during the reaction of 1a with diphenylacetylene-

d₁₀, pyridine, and p-H₂in CD₂Cl₂at 313 K. The emission signals for the

hydride resonances of 5a (right) and 6 (left) are illustrated.

Figure 8. ¹H-¹³C HMQC NMR spectra, optimized for long-range

correlations, illustrating the connection between the vinylic protons of 4a

(left) and 4b (right) and the two vinyl carbon signals.

and 6 in methanol. However, when ¹⁵N-labeled pyridine was

employed, the signal at δ -6.96 showed some evidence of

broadening, although a discrete ¹⁵N splitting was not observed.

Nonetheless, species 5a is proposed to be the cation [Pd(H)-

(bcope)(pyridine)](OTf).

The corresponding ¹H-¹³C HMQC spectrum of 4a in

methanol-d₄, obtained using the ¹³C-enriched diphenylacetylene-

d₁₀, showed two resonances at δ 129.9 (d, J_CP) 18 Hz) and δ

163.3 (dd, J_CP) 109, 10 Hz). The improved resolution confirms

that the δ 163.3 signal corresponds to a vinyl carbon with

phosphorus couplings that indicate it is bound to palladium

(Figure 8).

In order to test for the presence of bound solvent, in this

case methanol in 3a, we decided to add pyridine to the reaction

mixture. We now describe those results.

(c) Reactions of Pd(bcope)(OTf)₂(1a) with p-H₂and

Diphenylacetylene in Methanol-d₄Containing 2 µL of

Pyridine. A methanol-d₄sample of 1a containing p-H₂, diphe-

nylacetylene, and 2 µL of pyridine was prepared. At the start

of the experiment, the ³¹P{¹H} NMR spectrum contained a

singlet at δ 57.65, which suggests that the Pd(bcope)(OTf)₂is

converted into the pyridine solvate [Pd(bcope)(pyridine)₂](OTf)₂.

When the corresponding reaction was monitored 295 K, much

weaker polarization was observed, although when the sample

was warmed to 305 K, 2a could be readily characterized. Since

Modified 1D EXSY experiments were then used to probe

magnetization transfer within 2a, and selective irradiation of

1

the H signals at δ 4.98 and δ 2.89 led to transfer into trans-

stilbene. Interestingly, excitation of the remaining resonance at

δ 3.06 revealed exchange into the δ 6.70 signal due to 3a and

the δ -6.96 signal due to 5a. These data are consistent with a

cationic reaction cycle based on 2a being Pd(bcope)(CHPhCH₂-

Ph)(OTf) rather than a neutral cycle based on Pd(bcope)-

(CHPhCH₂Ph)(H).

1

the corresponding H and ³¹P resonances for 2a are almost

identical to those seen in methanol-d₄or CD₂Cl₂, we believe

that 2a does not contain pyridine, methanol, or CD₂Cl₂within

its coordination sphere as discussed before. However, we now

see two enhanced emission signals, one at δ 6.70 (dd, J_HP

12.7, 5.2 Hz) due to 4a and a second at δ -6.96 (dd, J_HP

)

(d) Reactions of Pd(bcope)(OTf)₂(1a) with p-H₂and

Diphenylacetylene in CD₂Cl₂Containing 2 µL of Pyridine.

When the reaction of 1a with p-H₂and diphenylacetylene is

monitored in CD₂Cl₂with 2 µL of pyridine, no polarization

was visible until the temperature reached 313 K. The initial

³¹P{¹H} NMR spectrum contained a singlet at δ 54.30, which

suggests that the Pd(bcope)(OTf)₂is converted into the pyridine

solvate [Pd(bcope)(pyridine)₂](OTf)₂prior to the start of the

reaction. This species then reacts relatively slowly at 313 K,

and ultimately the PHIP-enhanced signals of 2a become visible.

In addition, ¹H NMR signals for cis- and trans-stilbene are again

observed, but now a doublet of doublets appears at δ 6.72 (J_HP

) 13.2, 5.6 Hz) due to 4a and a hydride signal appears at δ

-7.04 (dd, J_HP) 223.2, 19.0 Hz) due to 5a. Both of these

resonances appear as enhanced emission signals, and as the

reaction proceeds a second hydride resonance appears at δ

-6.40 due to 6. These spectral features are illustrated in Figure

9. The addition of pyridine to the CD₂Cl₂solvent system

therefore facilitates the detection of the cationic vinyl and

hydride products. The lower activity of the CD₂Cl₂-based system

232.3, 17.2 Hz) due to a new species, 5a. Both of these proton

signals collapse into singlets upon broadband ³¹P decoupling.

Furthermore, as the reaction progresses, a second hydride

resonance, also emissive, appears at δ -6.30 due to 6, which

has a doublet of doublets of doublets multiplicity due to coupling

to three ³¹P nuclei, where the corresponding splittings are 191.1,

23.8, and 12.9 Hz.

The δ 6.70 signal of 4a couples to two ³¹P centers which

1

appear at δ 20.17 and δ 33.41 in the corresponding H-³¹P

HMQC experiment. These differ from those seen for 3a in

methanol-d₄, which appear at δ 21.79 and δ 39.04, and suggest

that 4a is the cation [Pd(bcope)(cis-CPhdCHPh)(pyridine)]-

(OTf). This was confirmed when ¹⁵N-labeled pyridine was

employed, since in that case the ³¹P resonance at δ 33.41 showed

an extra 17 Hz splitting. This also provides further support that

3a is indeed [Pd(bcope)(cis-CPhdCHPh)(MeOD)](OTf).

Unfortunately, because of low signal strengths and short

lifetimes, it was not possible to determine the chemical shift of

the phosphorus resonances for the bound bcope ligand in 5a

9

6522 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

Scheme 3. Schematic Proposed Mechanism for trans-Stilbene

Formation Based on Cationic Species As Deduced from the New

Observations

Figure 10. NMR traces from a series of 1D EXSY experiments acquired

during the reaction of 1a with diphenylacetylene-d₁₀, pyridine, and p-H₂in

CD₂Cl₂. Magnetization transfer is seen in traces (a) and (b) from the δ

4.97 and δ 2.87 resonances of 2a into trans-stilbene; in (c) the resonance

of 2a, appearing at δ 3.10, moves into that of 4a at δ 6.70 and the hydride

signals of 5a and 6, and in (d) transfer from the δ -7.04 signal of 5a moves

into the vinyl proton resonance of 4a and the hydride signal of 6.

allowed the PHIP effect to be observed for a longer time, and

hence better-quality NMR spectra could be obtained. The

characterization of 5a was further secured by the fact that its δ

-7.04 hydride signal couples to two phosphorus resonances that

appear at δ 22.33 (d, J_PP) 22 Hz) and δ 43.4, with the former

methanol-d₄or CD₂Cl₂solution containing either cis- or trans-

stilbene no reaction was observed.

1

providing the large H-³¹P splitting and the latter the 19 Hz

(f) Reactions of Pd(bcope)(OTf)₂(1a) with p-H₂and cis-

Stilbene or trans-Stilbene in Methanol-d₄. When a methanol-

d₄solution of 1a containing either cis-stilbene or trans-stilbene

coupling.

The weaker hydride signal due to 6 at δ -6.40 was shown

to correlate via the large J_HPcoupling to a phosphorus center

which resonated at δ 28.1 in the corresponding spectrum,

although in this case we were unable to determine the size of

any of its phosphorus-phosphorus couplings, but we did locate

a single cis ³¹P signal at δ 57.4. Under these conditions, when

the 1D EXSY approach was used to probe magnetization

1

was cooled to 273 K and p-H₂added, the associated H NMR

spectra showed no polarization effects. The formation and

depletion of the proton resonances associated with 2a at δ

4.98, δ 3.06, and δ 2.89 could be observed to occur over the

period of an hour during these experiments. It should be noted

that there was no evidence in any of these spectra for the

formation of 3a, and even though ¹H NMR spectra were

recorded down to 220 K, no hydride signal could be located

for 2a. These two observations make it clear that H₂plays a

role in activating 1a.

Inorganic Complex Speciation 3. Pd(^tbucope)(OTf)₂(1b).

We have completed the analogous experiments with

Pd(^tbucope)(OTf)₂(1b) in both CD₂Cl₂and methanol-d₄. The

organic components of the reactions mixtures show behaviors

similar to those found for 1a, and the reactivity again proves to

be higher in methanol than inCD₂Cl₂. The situation is, however,

more complex than was found with 1a, because of the reduced

symmetry of the phosphine ligand.

1

transfer within 2a, selective irradiation of the H signals at δ

4.97 and δ 2.87 again led to transfer into trans-stilbene.

However, excitation of the resonance at δ 3.10 revealed

exchange of this proton into the site of 4a that gave rise to the

δ 6.72 signal and the two hydride resonances at δ -7.04 and δ

-6.40 of 5a and 6 rather than free H₂. When the δ -7.04

resonance was excited, magnetization transfer into trans-stilbene

and the hydride signal of 6 was observed. This information is

portrayed in Figure 10 and a mechanism shown in Scheme 3.

The sample was then removed from the spectrometer, and both

free bcope and p-H₂were added. The hydride resonances due

to 4a and 5a were quenched in the resulting ¹H NMR spectrum,

while those of 6 were found to grow in intensity. These results

suggest that 4a is formed from 5a via incorporation of free

diphenylacetylene. It also indicates that 6 corresponds to [Pd-

(H)(bcope)(η¹-bcope)](OTf), although the remote ³¹P center of

the η¹-bcope ligand could be bound to palladium.

The addition of free bcope also rapidly stopped catalysis, and

ultimately, a new hydride resonance appeared at δ -10.77, a

phosphorus-coupled quintet, which coupled to a single ³¹P center

that was located at δ 1.4. This signal appeared as a doublet in

the corresponding ³¹P{¹H} spectrum when the hydride coupling

was retained. The new species could therefore be the monohy-

dride Pd(bcope)₂H⁺, although formation of a higher order cluster

cannot be excluded.

In methanol-d₄, the species which shows the largest polariza-

tion, 2b, yields PHIP-enhanced ¹H signals at δ 4.93 (m, J_HH

11.0, 6.0 Hz, J_HP) 8.0 Hz, J_HC) 156 Hz), δ 3.03 (m, J_HH

)

6.4 Hz), and δ 2.91 (m, J_HH) 10 Hz, J_HP) 5.7, 3.0 Hz, J_HC

) 130 Hz) which are very similar to those of 2a.

The characterization of species 2b was completed by record-

ing a ¹H-³¹P HMQC spectrum which showed, by virtue of the

location of two ³¹P resonances at δ 70.43 (d, J_PP) 60 Hz) and

δ 19.81 (d, J_PP) 60 Hz) for the ^tBu- and cope-substituted ³¹

P

centers, respectively, that 2b contains two inequivalent phos-

phorus centers. On the basis of the fact that decoupling at δ 72

in the phosphorus spectrum (i.e., the ^tBu₂P center) removes the

large ³¹P splitting seen for the proton signal at δ 4.93 we

(e) Reactions of Pd(bcope)(OTf)₂(1a) with cis-Stilbene or

trans-Stilbene Alone. When a sample of 1a was dissolved in a

t

conclude that the alkyl ligand lies trans to the Bu-substi-

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6523

A R T I C L E S

Chart 2

Lo´pez-Serrano et al.

tuted phosphorus center. ¹³C-¹H HMQC experiments, using

¹³CtC-enriched diphenylacetylene-d₁₀as the substrate, enabled

the ¹H resonance of 2b, which appears at δ 4.93, to be connected

to a ¹³C signal at δ 64.0 (d, J_CP) 55 Hz), while the proton at

δ 2.91 for the CH₂group connected to a signal at δ 34.8. The

structure of 2b shown in Chart 2 reflects an arene π coordination

motif similar to that deduced for 2a.

When the reaction of Pd(^tbucope)(OTf)₂(1b) with diphenyl-

acetylene-d₁₀and p-H₂was followed in methanol-d₄at 298 K,

in addition to the signals of 2b already described, an emission

signal was seen at δ 6.72 in the proton spectrum which

possessed two phosphorus splittings of 8.9 and 12.0 Hz. The

formation of [Pd(^tbucope)(CPhdCHPh)(MeOH)](OTf) (3b) is

therefore indicated. NMR data for 3b are presented in Table 2

(below). The analogous emission signal was not seen in

CD₂Cl₂.

Inorganic Complex Speciation 4. Pd(^tbucope)(OTf)₂

(1b): Effect of Pyridine. A series of analogous reactions in

methanol-d₄and CD₂Cl₂with 1b, diphenylacetylene, and para-

hydrogen in the presence of 2 equiv of pyridine were carried

out. In these experiments, polarized resonances were always

seen for the alkyl ligand of 2b. Two emission signals were also

detected in these reactions, one at δ 6.62 (or δ 6.61 in CD₂Cl₂)

for 4b (T ) 303 K using CD₂Cl₂and 295 K with methanol-d₄)

and one at δ -7.77 for 5b (in CD₂Cl₂, Figure 11). As expected,

the hydride signal possessed a large trans phosphorus coupling

of 208.5 Hz and a small cis phosphorus coupling of 9.7 Hz.

When the signal-to-noise ratio observed for this resonance

became substantial, a much weaker second hydride was seen

at δ -7.91 (5b′) with similar multiplicity and J_PHvalues of

205.2 and 13.5 Hz.

Figure 11. Hydride region of ¹H (upper trace) NMR spectra of the reaction

of 1b with diphenylacetylene-d₁₀, pyridine, and p-H₂in CD₂Cl₂at 313 K.

The dominant emission signals are for 5b (right), but those for 5b′ and 7

(expanded inset) can also be seen.

centered at δ -4.21 (ddad, J_HH) -18.0 Hz , J_HP) 138, 9 Hz)

and δ -4.71 (ddad, J_HH) -18.0 Hz, J_HP) 146, 9 Hz) (where

in the multiplicities, “a” ) antiphase) (Figure 11). These two

proton resonances proved to be coupled to each other and,

therefore, must correspond to a dihydride species, which is most

likely [Pd(^tbucope)(H)₂] (7). This information indicates that

reduction of palladium(II) to palladium(0) is possible under the

reaction conditions and provides evidence that a neutral cycle

could be entered in CD₂Cl₂, where we have demonstrated the

involvement of species 2-5. It should be noted that complex 7

was only observed in CD₂Cl₂.

NMR data for complexes 2-7 as a function of solvent are

summarized in Table 2.

Modified 1D EXSY experiments were then recorded in both

solvents to probe magnetization transfer within 2b. Selective

irradiation of the ¹H signals at δ 4.93 and δ 2.91 led to transfer

into trans-stilbene, while excitation of the δ 3.03 signal revealed

exchange into the δ -7.69 signal due to 5b and the δ 6.63

signal of 4b. When the δ -7.69 signal was irradiated, transfer

into both 4b and trans-stilbene was indicated, although when

the signal for 4b was examined, no chemical activity was

evident.

Detection of [Pd(^tbucope)(D)(pyridine)](OTf). When a

sample of 1b was examined that had been left in methanol-d₄

with ca. 5 equiv of pyridine for 24 h at room temperature, in

the absence of hydrogen or the alkyne, a roughly 1:1 mixture

of 1b and a new species, with ³¹P resonances at positions almost

identical to those seen for 5b in methanol-d₄, was obtained. The

signal at δ -2.45 in the ³¹P{¹H} spectrum exhibited an extra

1:1:1 triplet splitting of 30 Hz due to an additional coupling to

a single trans deuterium, which must come from the methanol-

d₄solvent. These two ³¹P resonances therefore correspond to

the deuterated analogue of 5b, [Pd(^tbucope)(D)(pyridine)](OTf)

(5b_²H), as shown in Figure 12. As expected, the many-scan

1D proton NMR spectrum showed only a very weak reso-

nance at δ -7.70 for 5b due to trace amounts of protons in the

system. Further proof was obtained when ¹⁵N-labeled pyridine

was used, since the ³¹P resonance at δ 101.92 showed the

The spectroscopic data for these species in both solvents are

listed in Table 2. The characterization of the vinyl- and hydride-

based species is also complicated by the symmetry of the

phosphine, and, not surprisingly, when a ¹H-³¹P HMQC

experiment was recorded in CD₂Cl₂, two mutually coupled

doublets were located for 5b at δ -2.2 (J_PP) 24 Hz) and δ

100.96. The former signal arises from the cyclooctane-sub-

t

stituted phosphorus center, while the latter is due to the Bu-

substituted group. Unfortunately, the weakness of the hydride

signal for 5b′ prevented the analysis of the corresponding ³¹P

information, but when a ¹H NMR spectrum was collected with

selective phosphorus decoupling around δ -2.2, the large

hydride-phosphorus coupling was removed for the hydride

resonance of 5b and the small one for 5b′. When this experiment

was repeated with irradiation around δ 100 in the ³¹P spectrum,

the reverse happened: the small coupling was removed for the

hydride resonance of 5b and the large coupling for 5b′. We

therefore conclude that 5b and 5b′ are simply isomers that differ

according to the relative orientations of the hydride, pyridine,

and phosphorus ligands.

expected extra coupling due to the trans pyridine ligand (J_PN

)

26 Hz).

A remarkable feature of these experiments is the observation

in CD₂Cl₂of a set of weakly enhanced hydride resonances

When this sample was removed from the spectrometer and

an excess of diphenylacetylene was added to the solution, the

9

6524 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

Table 4. Rate Constants, Activation Parameters ∆H^q(kJ mol^-1

and ∆S^q(J mol^-1K^-1), and Free Energy of Activation ∆G^q(kJ

)

mol^-1) for the Elimination of trans-Stilbene from 2a and 2b at 300

K

1

complex

k (s^-

)

∆

H^q

∆

S^q

∆

G^q

300

2a in CD₂Cl₂

0.18 79 ( 7

5 ( 24 77 ( 4

3 ( 16 77 ( 1

77^a

2a in CD₂Cl₂with 10% methanol 0.21 78 ( 4

2a in CD₂Cl₂+ 2 equiv of py

0.27^a

2a in methanol-d₄

0.53 42 ( 9 -107 ( 31 75 ( 8

2a in methanol-d₄+ py

2b in CD₂Cl₂

0.61 76 ( 1

5 ( 3

75 ( 1

1.04^b94 ( 6

69 ( 22 73 ( 1

^aPoor data. ^bExtrapolated from Eyring plot.

but when methanol-d₄was used instead, the enhanced signals

decayed too fast to record any data. The corresponding observed

rates (s^-1) for trans-stilbene elimination at 300 K, and ∆H^q,

∆S^q, and ∆G^q

values, where obtainable, are presented in

(300)

Table 4. These data will be discussed in the Conclusions section.

Conclusions

The symmetrical chelating phosphine complex [Pd(bcope)-

(OTf)₂] (1a), where bcope is (C₈H₁₄)PCH₂-CH₂P(C₈H₁₄), and

the lower symmetry complex [Pd(^tbucope)(OTf)₂] (1b), where

^tbucope is (C₈H₁₄)PC₆H₄CH₂P(^tBu)₂, were prepared and have

been shown to catalyze the hydrogenation of diphenylacetylene.

In this reaction, cis- and trans-stilbene and 1,2-diphenylethane

are the dominant products.

Figure 12. ³¹P{¹H} NMR spectra of a methanol-d₄solution of 1b after 24

h at room temperature (upper trace), showing a ca. 1:1 mixture of the

methanol adduct of 1b [δ_P97.53 (P^tBu₂), 17.96 (PC₈H₁₄)] and [Pd(^tbucope)-

(D)(pyridine)]⁺(the inset shows the extra trans deuterium splitting seen

for the ³¹P-C₈H₁₄center); and the same solution after addition of

diphenylacetylene (lower trace), indicating the quantitative transformation

of [Pd(^tbucope)(D)(pyridine)]⁺into 4b.

A number of reaction intermediates have been observed and

some kinetic data collected on these two systems that rationalize

this behavior. In the first instance, neither 1a nor 1b reacts with

diphenylacetylene directly in these solvents to produce the three

organic products in the absence of hydrogen. However, when

methanol-d₄or CD₂Cl₂solutions of 1a or 1b containing pyridine

are examined, the slow formation of the monodeuteride-

containing cations [Pd(P₂)(D)(pyridine)](OTf) is indicated.⁴⁶In

contrast, when H₂is added to similar solutions, the formation

of the monohydrides [Pd(P₂)(H)(pyridine)](OTf) (5a and 5b)

occurs, even though methanol-d₄is present. Examination of the

proton signal for free hydrogen reveals that no HD is present

after 1 h. This indicates that, under H₂, 1a and 1b can access

catalytically active hydride cations of the type that have been

proposed for numerous palladium-catalyzed reactions.^1,51,52

new ³¹P{¹H} NMR spectrum indicated that 5b_²H had been

1

converted into 4b_²H. Not surprisingly, the corresponding H

NMR spectrum did not contain a resonance at δ 6.62 due to

the vinylic proton of 4b because the insertion reaction now

involved a PdD bond. When this sample was then degassed

and charged with 3 atm of para-hydrogen, emission signals were

seen for the hydride resonance of 5b, and as well as the vinylic

proton of 4b at δ 6.62 and the three protons of the alkyl cation

2b showed the PHIP effect. This indicates that [Pd(^tbucope)-

(D)(pyridine)](OTf) is able to initiate the hydrogenation reaction.

Furthermore, GC/MS analysis was also undertaken for the

organic components produced from the reaction of preformed

[Pd(^tbucope)(D)(pyridine)](OTf) with H₂and a 3-fold excess

of protio-diphenylacetylene (based on the original amount of

1b). Evidence for complete conversion into trans-stilbene was

obtained after a reaction time of 5 min, and the ratio of D₀to

D₁isotopomers of trans-stilbene proved to be approximately

2:1. This information confirms that 5b_²H is catalytically active

and that H₂addition leads to the generation of protio-5b rather

than 5b_²H after 5b_²H is turned over.

Quantitative Magnetization Transfer Studies on 2a and

2b. In order to explore the reactivity of 2a and 2b in more detail,

a series of 1D EXSY experiments were recorded where the alkyl

proton R to the metal was selected, and magnetization transfer

from this site was monitored as a function of reaction time.

We have already indicated that, when 2a was examined in this

way in CD₂Cl₂, strong magnetization transfer into the trans-

stilbene signal at δ 7.18 was observed. This process has been

followed as a function of reaction temperature for 2a in several

solvent systems: CD₂Cl₂, methanol-d₄, CD₂Cl₂containing 10%

methanol, and both CD₂Cl₂and methanol-d₄containing 2 µL

of pyridine. 2b was examined in a similar way using CD₂Cl₂,

When the reactions of 1a and 1b with diphenylacetylene and

para-hydrogen in methanol-d₄or CD₂Cl₂are examined,

the detection of [Pd(bcope)(CHPhCH₂Ph)](OTf) (2a) and

[Pd(^tbucope)(CHPhCH₂Ph)](OTf) (2b) is achieved. In 2b, the

alkyl ligand lies trans to the ^tBu-substituted phosphorus center.

Crystal structures for the di-aqua complex, [Pd(^tbucope)(OH₂)₂]-

(OTf)₂‚2(H₂O), and the methanol solvate, [Pd(bcope)(MeOH)₂]-

(OTf)₂‚2MeOH, illustrate the fact that these cations coordinate

effectively to relatively weak donor ligands, and it might be

expected that such a situation would be found during catalysis.

The spectroscopic signatures of the two alkyl cations do not,

however, change with solvent, even when pyridine is added,

and to achieve an acceptable 16-electron count, a stabilizing

interaction with the arene π system of the alkyl ligand is

suggested from our experimental data. The NMR data support

an η³-benzyl interaction, in accordance with DFT calculations

that have revealed a preference for η³-benzyl interactions over

â-agostic interactions in a related complex.⁵⁵In addition, a role

for related η³-benzyl complexes has been demonstrated by

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6525

A R T I C L E S

Lo´pez-Serrano et al.

Scheme 4. Schematic Mechanism Showing Alkene Formation

and Isomerization

In CD₂Cl₂, transfer is observed from the δ 3.13 site of 2a

into both cis-stilbene and H₂. In contrast, when the analogous

process is monitored in methanol-d₄, magnetization proceeds

from the δ 3.13 site into the vinyl proton of 3a, [Pd(bcope)-

(CPhdCHPh)(methanol)](OTf). The process by which this

interconversion occurs was identified by the addition of pyridine,

since magnetization transfer from this site proceeds into [Pd-

(bcope)(CPhdCHPh)(pyridine)](OTf) (4a), the pyridine-sol-

vated monohydride [Pd(bcope)(H)(pyridine)](OTf) (5a), and the

related species [Pd(H)(bcope)(η¹-bcope))](OTF) (6); when free

bcope is added, magnetization transfer from this site proceeds

only into 6. This confirms that the â-H of the CHPhCH₂Ph

group of 2a, which resonates at δ 3.13, actually transfers first

into [Pd(bcope)(H)(L)](OTf), where L ) methanol or pyridine,

and then into [Pd(bcope)(CPhdCHPh)(L)](OTf) via reaction

with a fresh molecule of diphenylacetylene. At the same time

that this occurs, trans-stilbene is eliminated from 2a. It proved

possible to demonstrate a role for 5a in this process directly by

showing that magnetization transfer from the hydride site of

5a proceeds directly into both the vinyl proton site of 4a and

trans-stilbene.

When the corresponding process was examined for 2b,

magnetization transfer into trans-stilbene was again observed

from two sites of the alkyl, while transfer into either hydride

species 5b or 5b′, or the vinyl complexes 3b and 4b, was seen,

depending on whether the pyridine was present or not. These

deductions are summarized in Scheme 4.

It is worth noting at this stage that, remarkably, the single

Hartwig et al. in the hydroamination of vinylarenes^56,57and by

Brown et al. and Brookhart in the Heck reaction.^8,47,58

Evidence for â-hydrogen transfer in related palladium bis-

phosphine cations containing an alkyl ligand has also been

presented by Brown et al. during a study of the Heck reaction.⁵⁹

They suggest that an undetected palladium alkene hydride

complex is involved in the liberation of the corresponding

alkene.

In the case of 2a and 2b, the interaction with the arene π

system would need to be broken in order for this to happen.

Scheme 4 shows an alternative formulation for species 2,

featuring a â-agostic interaction between one of the â-CH

protons of the alkyl ligand and the metal center. The resulting

â-hydrogen transfer is likely to involve an intermediate with a

â-agostic interaction, and intermediates of this type have been

shown to play a direct role in related reactions (i.e., ethylene

polymerization).⁵¹

Both 2a and 2b have, in fact, been shown by EXSY to play

a direct role in the formation of the two semihydrogenation

products. When magnetization transfer from the R-H of the

CHPhCH₂Ph is monitored, the formation of trans-stilbene is

indicated as the major pathway, although limited exchange into

the other two sites is also seen. The two â-H’s of the alkyl

cations are distinct, with magnetization transfer from one

proceeding into trans-stilbene and from the other into a number

of distinct sites that depend explicitly on the solvent.

vinylic CH protons in 3 and 4 and the single hydride ligand in

1

5 appear as strong emission signals in the corresponding H

NMR experiments. This is an example of one-proton PHIP and

illustrates the care with which p-H₂experiments need to be

examined if the correct deductions are to be made, since species

such as [Pd(P₂)(cis-stilbene)] or [Pd(P₂)(H)₂(pyridine)] might

otherwise be invoked. Magnetization transfer experiments

indicated that the kinetic fate of 3 and 4 could not be determined

because they were unreactive on the NMR time scale.

Nonetheless, these species are not stable under these

conditions and must be involved in the formation of the

hydrogenation products on a longer time scale. The addition of

H₂in a slow step to 2, the alkyl cation, is necessary if the

formation of H₄-labeled 1,2-diphenylethane in methanol-d₄is

to be explained.

For the ^tbucope system, single isomers of both [Pd(^tbucope)-

(CHPhCH₂Ph)](OTf) (2b) and [Pd(^tbucope)(CPhdCHPh)-

(pyridine)](OTf) (4b) were seen, while two isomers of

[Pd(^tbucope)(H)(pyridine)](OTf) (5b and 5b′) were detected. In

2b, the alkyl arm is trans to the ^tBu-bearing phosphorus center,

and the vinyl group is arranged in a similar way in 4b, but the

opposite is true for the hydride orientation in 5b, where it is

trans to the cope-bearing phosphorus center. This information

is summarized in Chart 2 and Scheme 4. â-H transfer in 2b

locates the new hydride ligand trans to the cope-bearing

phosphorus arm, and loss of alkene and binding of pyridine

then lead directly to 5b, while binding of the alkyne leads to

4b if the hydride ligand moves onto one of the unsaturated

alkyne carbons.

(55) Ludwig, M.; Stro¨mberg, S.; Svensson, M.; Akermark, B. Organometallics

1999, 18, 970-975.

(56) Nettekoven, U.; Hartwig, J. F. J. Am. Chem. Soc. 2002, 124, 1166-1167.

(57) Johns, A. M.; Utsunomiya, M.; Incarvito, C. D.; Hartwig, J. F. J. Am. Chem.

Soc. 2006, 128, 1828-1839.

It should also be noted that, when the symmetry-breaking

(58) Rix, F. C.; Brookhart, M.; White, P. S. J. Am. Chem. Soc. 1996, 118, 2436-

2448.

t

phosphine bucope is employed, enhanced hydride signals are

(59) Brown, J. M.; Hii, K. K. M. Angew. Chem., Int. Ed. Engl. 1996, 35, 657-

659.

also seen for the dihydride complex [Pd(^tbucope)(H)₂] (7). This

9

6526 J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007

Palladium-Catalyzed Alkyne Hydrogenation

A R T I C L E S

means that, when CD₂Cl₂is the reaction medium, the formation

of a Pd(0), Pd(II) reaction cycle involving neutral species is

possible.

The activation entropy for alkene liberation for 2a in

CD₂Cl₂is -5 ( 24 J mol^-1K^-1, which suggests that solvation

effects are relatively limited. As a consequence, the enthalpy

of activation is relatively high, at 79 ( 7 kJ mol^-1. Upon

changing the solvent to methanol, the activation entropy for

alkene liberation for 2a falls to -107 ( 31 J mol^-1K^-1, which

suggests that solvation effects become much more important,

in accordance with the fact that methanol-d₄is a better donor

than CD₂Cl₂. This deduction is matched by the lowering of the

enthalpy of activation to 42 ( 9 kJ mol^-1in methanol-d₄.

It is also important to note that GC/MS analysis of the organic

products produced in these reactions revealed proton rather than

deuterium incorporation for reactions involving H₂in methanol-

d₄until 5b_²H was used as the catalyst precursor. This situation

changed when 5b_²H was used as the catalyst precursor because

evidence for ²H incorporation into the cis-stilbene was observed.

We therefore conclude that the active palladium hydrides are

recycled, as shown in Scheme 4, via reactions involving H₂

rather than methanol-d₄. The initial formation of 5 via oxidation

of the solvent rather than H₂addition cannot, however, be ruled

out.^2,10

This reaction proved to be solvent-dependent, with methanol

leading to higher activities than CD₂Cl₂for both 1a and 1b,

although 1b shows a higher overall activity than 1a for the

consumption of diphenylacetylene. This is reflected in the higher

first-order rate constants for the production of trans-stilbene that

were measured for the bcope system in methanol-d₄, as

illustrated in Table 4. The product distribution is critically

dependent on the phosphine, with the bcope system forming

substantial amounts of the double hydrogenation product 1,2-

t

diphenylethane when compared to the bucope system (25%

rather than <1% when the complete conversion of diphenyl-

acetylene is achieved). Furthermore, the bcope system also

produces more of the dimerization product E,E-1,2,3,4-tetraphen-

t

ylbuta-1,3-diene than the bucope-based catalyst, although the

level of conversion, at 1%, means that it is not formed in a

significant reaction pathway.

We also know that 1a and 1b are stable indefinitely in

CD₂Cl₂, and in methanol we were successful at crystallizing

1a‚MeOH. In contrast, while 1b was stable in methanol, it

reacted in methanol-d₄containing pyridine to form the mono-

deuteride 5b_²H. We therefore believe that the greater activity

seen in methanol results from the improved efficiency of catalyst

conversion into active species in this solvent.

The product distribution seems to be related to the lifetime

of the alkyl cation, which is longest for 2a, where substantial

amounts of 1,2-diphenylethane are obtained in a reaction which

involves H₂addition to 2. The binding of diphenylacetylene to

either 3 or 4, followed by C-C bond formation and the addition

of H₂, is necessary to account for the formation of E,E-1,2,3,4-

tetraphenylbuta-1,3-diene.

For 2a, the elimination of trans-stilbene proceeds in methanol-

d₄at a rate of 0.53 s^-1at 300 K, where ∆H^q) 42 ( 9 kJ

mol^-1and ∆S^q) -107 ( 31 J mol^-1K^-1. When the corre-

sponding reaction is completed in CD₂Cl₂, the corresponding

rate constant falls to 0.18 s^-1, while ∆H^qbecomes 79 ( 7 kJ

mol^-1and ∆S^qbecomes 5 ( 24 J mol^-1K^-1. The analogous

process for 2b proved to be too fast to monitor accurately in

methanol, but when CD₂Cl₂was employed, the corresponding

rate constant for trans-stilbene formation is 1.04 s^-1at 300 K,

Acknowledgment. We are grateful for EU funding under the

HYDROCHEM network (contract HPRN-CT-2002-00176). We

are also grateful to Prof. Paul Pringle (University of Bristol)

for the bcope and ^tbucope ligands and to Prof. Robin N. Perutz

(University of York) for helpful discussions.

Supporting Information Available: X-ray tables and some

synthetic details (PDF, CIF). This material is available free of

charge via the Internet at http://pubs.acs.org.

with ∆H^q) 94 ( 6 kJ mol^-1and ∆S^q) 69 ( 22 J mol^-1K^-1

.

These data match those obtained from the product studies, where

2b proved faster and produced the largest amounts of trans-

stilbene.

JA070331C

9

J. AM. CHEM. SOC. VOL. 129, NO. 20, 2007 6527

Article Doi

A para-hydrogen investigation of palladium-catalyzed alkyne hydrogenation

DOI: 10.1021/ja070331c

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja070331c

Full text of DOI:10.1021/ja070331c

Products guided by the article

R&D Labs maybe for 942628-33-3

Relevant to this article

Hot Product