Parahydrogen-Induced Polarization in Heterogeneous Hydrogenations over Silica-Immobilized Rh Complexes

Parahydrogen-Induced Polarization in HeterogeneousHydrogenations over Silica-Immobilized Rh Complexes

Ivan V. Skovpin • Vladimir V. Zhivonitko •

Igor V. Koptyug

Received: 15 June 2011 / Revised: 27 July 2011 / Published online: 25 August 2011

� Springer-Verlag 2011

Abstract The use of heterogeneous catalysts for parahydrogen-induced polari-

zation (PHIP) of nuclear spins opens new horizons for production of hyperpolar-

ized substances. Immobilization of homogeneous hydrogenation catalysts is a

promising approach for designing the efficient heterogeneous catalytic systems

capable of PHIP generation. Herein, we study the formation of PHIP in the gas-

phase and in the liquid-phase hydrogenations of propyne and propylene catalyzed

by silica-immobilized Rh complexes synthesized by the ligand-exchange anchoring

of the Wilkinson’s complex RhCl(PPh3)3, the binuclear complex Rh2Cl2(C8H12)2

and the cationic complex [Rh(C8H12)2]?[BF4]- to the phosphine-modified silica

gel. We consider the stability and the mechanistic aspects of the hydrogenation

over the immobilized Wilkinson’s catalyst in terms of PHIP observation. Using a

PASADENA (parahydrogen and synthesis allow dramatically enhanced nuclear

alignment) effect, it is found, in particular, that liquid-phase propyne hydrogena-

tion over the immobilized Wilkinsons’s catalyst at 70�C proceeds in a stable

regime with a stereoselective cis addition of a hydrogen molecule, while in the gas

phase at the same temperature the hydrogenation stereoselectivity is observed only

for a short time after the reaction is started, and then the catalyst rapidly loses its

activity. The reasons of the catalyst deactivation are discussed based on the lit-

erature data, the results of infrared spectroscopy study, and the comparison to the

behavior of the immobilized binuclear and cationic Rh complexes. In addition, it is

shown that the immobilized Wilkinson’s catalyst is reduced as temperature

increases in the range of 90–130�C, as confirmed by X-ray photoelectron

spectroscopy.

I. V. Skovpin (&) � V. V. Zhivonitko � I. V. Koptyug

International Tomography Center, Siberian Branch of the Russian Academy of Sciences,

Institutskaya st. 3a, 630090 Novosibirsk, Russia

e-mail: [email protected]

URL: http://www.tomo.nsc.ru/

123

Appl Magn Reson (2011) 41:393–410

DOI 10.1007/s00723-011-0255-z

Applied

Magnetic Resonance

1 Introduction

Nowadays, nuclear magnetic resonance (NMR) spectroscopy is widely applied in

science, medicine, and industry. In many cases, however, NMR methods suffer from

a poor inherent sensitivity limited by the extremely small population difference

between the nuclear spin energy levels at thermal equilibrium. In the late 1980s,

Bowers [1] demonstrated that molecular hydrogen with an excess of molecules in

the para spin state (parahydrogen) can be used to generate spin systems with the

population differences greatly exceeding those observed for thermal polarization.

As a result, the sensitivity can be enhanced by several orders of magnitude for these

hyperpolarized systems [1]. This phenomenon is known in the literature as

parahydrogen-induced polarization (PHIP). The theory and applications of this

phenomenon have been reviewed in the past [2–4].

PHIP can be observed in the NMR spectra of parahydrogen addition products. On

top of the significant signal enhancement, PHIP can serve as a spin-labeling

technique, since the signal intensities and shapes for the protons derived from

parahydrogen depend on the spin system under study. In addition, depending on the

experimental procedure, a multiplet or a net polarization can be observed in the 1H

NMR spectra. The multiplet polarization is achieved in the PASADENA-type

(parahydrogen and synthesis allow dramatically enhanced nuclear alignment)

experiment when hydrogenation is performed in a high magnetic field of an NMR

magnet. In this case, the polarized signals have an antiphase line shape (A/E or E/A),

with both absorptive (A) and emissive (E) components present within each multiplet

[2]. For generating net polarization, the magnetic field cycling should be involved

according to the ALTADENA-type (adiabatic longitudinal transport after dissoci-

ation engenders net alignment) experiment. In this case, two polarized multiplets of

opposite signs are observed in the spectrum [3]. It is worth to note that PHIP

applications are not restricted only to the 1H nucleus. There are reported examples

of the polarization transfer to 13C, 19F, 15N, 31P and other nuclei [5–13].

In general, addition of molecular hydrogen is a catalytic process, which

represents an important step of hydrogenation and hydroformylation reactions.

There are many publications wherein PHIP was successfully utilized for detection

and investigation of metal dihydride species as well as for homogeneous catalytic

hydrogenation and hydroformylation studies. Moreover, there are several works

devoted to the use of the signal enhancement provided by PHIP in magnetic

resonance imaging (MRI) [5, 6, 14, 15]. It is known that one of the pathways of H2

addition catalyzed by transition-metal complexes is the oxidative addition leading to

the formation of dihydride complexes, which often serve as important intermediates

of both hydrogenation and hydroformylation reactions. The necessary condition for

PHIP observation is a pairwise transfer of the two protons of a parahydrogen

molecule to a substrate molecule. For this reason, until recently PHIP has been

observed exclusively in the investigations concerning homogeneous catalytic

systems by utilizing dissolved transition-metal complexes as homogeneous

catalysts. Indeed, the pairwise addition is an inherent feature of many homogeneous

catalysts. However, the applications of homogeneous catalysts in NMR spectros-

copy and MRI for the production of hyperpolarized substances are restricted

394 I. V. Skovpin et al.

123

because these expensive compounds are difficult to rapidly separate from the

polarized products.

Heterogeneous catalysts have several advantages over their homogeneous

counterparts as they are easier to remove from the reaction medium and can be

used in subsequent catalytic cycles with little loss in their activity. For a long time,

however, heterogeneous catalysts were not considered for PHIP applications. As

common heterogeneous catalysts are dispersed metals, the two hydrogen atoms

from a parahydrogen molecule were expected to rapidly loose each other upon

chemisorption on a metal surface, leading to the loss of original correlation between

the two nuclear spins of parahydrogen. Therefore, no polarization effects were

expected for such catalysts. Recently, however, Kovtunov et al. [16] reported

observation of PHIP in a hydrogenation reaction over supported platinum and

palladium metal catalysts. They observed both PASADENA and ALTADENA

effects in the 1H NMR spectra of propane generated in the heterogeneous gas-phase

hydrogenation of propylene. Later, the PHIP effects were also observed in the liquid

phase hydrogenation reactions catalyzed by supported metal nanoparticles [17, 18].

These observations indicate that the pairwise transfer of hydrogen in heterogeneous

hydrogenation over supported metal catalysts is possible, but estimated selectivities

toward the pairwise addition did not exceed 3%.

Immobilization of homogeneous hydrogenation catalysts on a solid support is an

alternative strategy for the use of the advantages of heterogeneous catalysis in PHIP

applications. In an ideal case, the immobilization of transition-metal complexes can

be expected to preserve their chemical properties, i.e., one could expect that the

reaction mechanism of homogeneous hydrogenation would remain unchanged [19].

Indeed, the Wilkinson’s complex [RhCl(PPh3)3] immobilized on either styrene–

divinylbenzene copolymer or phosphine-modified silica gel as well as a tridentate

cationic rhodium complex immobilized on hydrophilic silica gel were successfully

utilized for generation of both PASADENA and ALTADENA polarization patterns

in heterogeneous hydrogenations of styrene and propylene [19]. However, the

stability of PHIP generation and the actual mechanism of hydrogenations over these

catalysts were not addressed in detail.

In this work, we consider the mechanistic aspects of liquid-phase and gas-phase

hydrogenations over several immobilized Rh complexes. The immobilized Wilkin-

son’s catalyst is mostly addressed, although immobilized binuclear and cationic

Rh complexes are also examined. In particular, it is shown that liquid-phase

hydrogenation over the immobilized Wilkinson’s catalyst exhibits a stereoselective

hydrogen molecule addition to propyne, similar to the homogeneous hydrogenation

with the Wilkinson’s catalyst dissolved in benzene. In contrast to the liquid-phase

hydrogenation, the stereoselectivity in the gas-phase hydrogenation of propyne over

the immobilized Wilkinson’s catalyst is observed only for a short time after the

reaction is started. The catalyst is unstable under the gas-phase conditions and

deactivates rapidly. Possible reasons for this deactivation process are discussed

along with the experimental findings obtained in propyne and propylene hydroge-

nations over the other types of immobilized Rh catalysts. It is also shown that at

higher temperatures (90–130�C) the immobilized rhodium catalysts are being

partially reduced under gas-phase reaction conditions, which leads to a significant

Parahydrogen-Induced Polarization in Heterogenous Hydrogenations 395

123

increase in the catalyst activity and the observation of PHIP. The results obtained

for the reduced immobilized Rh catalysts are compared to those obtained for a

supported Rh metal catalyst.

2 Experimental

2.1 General Information

All 1H NMR spectra were recorded on a Bruker AV300 SB spectrometer operating

at 300 MHz proton resonance frequency. The temperature control unit of the

instrument was used to vary the sample temperature in the range from room

temperature to 130�C. To maximize the signals in 1H NMR spectra in the

PASEDENA and ALTADENA experiments, the p/4 and p/2 radio-frequency (rf)

pulses were used, respectively. The experimental setup was equipped with a tubular

resistive oven, which was used to heat the reaction mixture and the catalyst bed

outside the NMR spectrometer (see Fig. 1).

2.2 Parahydrogen Enrichment

Parahydrogen-enriched H2 can be produced by cooling normal hydrogen in the

presence of an appropriate ortho–para interconversion catalyst [3]. FeO(OH)

(Sigma-Aldrich, 30–50 mesh) was used for this purpose in our experiments. This

catalyst was loaded in a �00 copper tube, which then was placed in a Dewar vessel

with liquid nitrogen (see Fig. 1), and normal hydrogen gas was passed through the

tube at a low flow rate. The interaction of cold H2 gas with FeO(OH) at 77 K led to

Fig. 1 Scheme of the experimental setup. 1 Dewar vessel with liquid nitrogen, 2 ortho–parainterconversion catalyst, 3 gas cylinder for mixing the reagents (C3H4 and H2 or C3H6 and H2 in the1:4 ratio), 4 gas flow meter, 5 vacuum meter, 6 pressure gauge, 7 tubular reactor. Hydrogenation is carriedout (a) in the strong magnetic field (7 T) inside a NMR spectrometer (PASADENA experiments) or (b) inthe low magnetic field (Earth’s field, 50 9 10-6 T) followed by an adiabatic transfer of the reactionmixture into the NMR spectrometer for analysis (ALTADENA experiments)


123

a constant flow of parahydrogen-enriched H2 with a 1:1 ortho–para ratio of spin

isomers, which from here on is referred to as parahydrogen.

2.3 Synthesis of Immobilized Rhodium Catalysts

In the present study, we used 2-diphenylphosphinoethyl-functionalized silica gel

(Sigma-Aldrich, product #538019, 5 g, extent of labeling: 0.7 mmol/g loading,

200–400 mesh) for immobilization of the following rhodium complexes: 1tris(triphenylphosphine)rhodium(I) chloride (Wilkinson’s catalyst) (Sigma-Aldrich,

product #199982, 5 g); 2 bis(1,5-cyclooctadiene)dirhodium(I) dichloride (Sigma-

Aldrich, product #14692, 1 g); 3 bis(1,5-cyclooctadiene)rhodium(I) tetrafluoroborate

(Sigma-Aldrich, product #14694, 100 mg). We denote the immobilized complexes

as 1/PPh2-SiO2, 2/PPh2-SiO2 and 3/PPh2-SiO2 in the text below. All immobilized

catalysts were synthesized according to the procedure described by Shyu et al. [19,

20], wherein the supported analogue of a Wilkinson’s complex was prepared by

interaction of RhCl(PPh3)3 with a phosphine-modified support. An approximately

stoichiometric amount of the modified silica gel was added to the solution of 1 in

degassed benzene in the case of the Wilkinson’s complex immobilization. For the

complexes 2 and 3, the modified silica gel was taken so as to have 1:1 and 1:2 Rh:P

ratios, respectively. The resulting suspension was stirred overnight at room

temperature. Finally, the corresponding immobilized catalyst was filtered out,

washed three times with degassed benzene and dried under vacuum at room

temperature. All the procedures described above were usually carried out under

argon. For some selected experiments, the preparation procedures were carried out

under the atmosphere of hydrogen gas or hydrogen-propylene (1:1) gas mixture, as

described in Sect. 3.

2.4 Liquid-Phase Hydrogenation

Hydrogenation catalyst (1/PPh2-SiO2 or 2/PPh2-SiO2) was placed in a 10 mm NMR

sample tube charged with 4 ml of benzene-d6 solvent, which was preliminarily

degassed by three freeze–pump–thaw cycles under vacuum. The sample tube was

then placed inside the magnet of the NMR spectrometer and equilibrated at a

required temperature for some time. Thereafter, hydrogen–propyne mixture (4:1)

supplied at a flow rate of 0.1–1 ml/s was bubbled through the solution for ca. 10 s

using a Teflon capillary. Then, the 1H NMR spectra of the reaction solution were

recorded just after the bubbling was stopped with the use of a shut-off valve.

2.5 Gas-Phase Hydrogenation Procedure

In PASADENA experiments, the immobilized catalysts (1/PPh2-SiO2, 15 mg;

2/PPh2-SiO2, 20 mg; or 3/ PPh2-SiO2, 15 mg) were placed in a 10 mm NMR

sample tube, which then was transferred into the magnet of the NMR spectrometer.

Thereafter, the reagents mixture containing propylene and parahydrogen or propyne

and parahydrogen in a 1:4 ratio supplied via a Teflon capillary at a flow rate of

1–5 ml/s was flowing through the catalyst layer heated up to a certain temperature.


123

The acquisition of 1H NMR spectra of the reaction mixture was started immediately

after the flow of the mixture was enabled.

In the ALTADENA experiments, 1/PPh2-SiO2 catalyst (15 mg) was packed

between the two plugs of a glass wool in a tubular quartz reactor with an inner

diameter of 1.5 mm, and the mixture of gaseous reagents was passed through the

resulting catalyst bed kept at a certain temperature (see Fig. 1). The hydrogenation

step was performed outside the magnet of the NMR spectrometer. The hydroge-

nation reaction mixture was continuously flowing from the reactor to an empty

10 mm sample tube inside the NMR magnet through a system of capillaries. The 1H

NMR spectra of the reaction mixture were recorded at various flow rates (1–5 ml/s)

in order to find conditions providing the maximum signal intensities.

3 Results and Discussion

3.1 Stereoselectivity of Dihydrogen Addition in Liquid Solution

The hydrogenation over Wilkinson’s catalysts is known to be stereoselective toward

the cis addition of the two hydrogen atoms to the substrate [21]. Taking this into

account, we investigated the stereoselectivity of the hydrogenation over 1/PPh2-

SiO2, which is a heterogenized analog of a Wilkinson’s catalyst. The heterogeneous

hydrogenation of propyne was carried out at 7 T (the PASADENA experiment), and

the spectrum of the reaction mixture is shown in Fig. 2.

As one can see, the antiphase PASADENA signals at 4.74 (‘Hc) and 5.54 ppm

(‘Hb) are observed in the spectrum. These signals correspond to the vinyl fragment

of the reaction product propylene, namely, to the vicinal protons which are in a cisposition relative to each other in the propylene molecule. Trans addition of two

hydrogens would lead to the polarization of the signal at 4.87 ppm (Hd). As this

signal is not polarized, we can conclude that the transfer of dihydrogen to the triple

Fig. 2 PASADENA 1H NMRspectrum acquired during theheterogeneous hydrogenation ofpropyne catalyzed by theimmobilized Wilkinson’scatalyst (1/PPh2-SiO2) inbenzene-d6 solution atT = 70�C. The polarized signalscorresponding to protons of theCH and CH2 groups ofpropylene are labeled as ‘Hb and‘Hc, respectively. Signals fromthe CH and CH3 groups ofpropane are labeled as Hb andHa, respectively. The signal ofthe hydrogen atom of the vinylfragment of propylene markedas Hd is not polarized


123

bond of propyne proceeds stereoselectively over the 1/PPh2-SiO2 catalyst, and cisdihydrogen addition takes place. As it was noted above, the homogeneous

hydrogenation of triple carbon–carbon bonds catalyzed by a Wilkinson’s complex

almost exclusively occurs via cis addition of dihydrogen. However, in the earlier

PHIP studies the stereoselectivity issues were not explicitly addressed for the

conventional homogeneous hydrogenation with the use of a Wilkinson’s complex as

a catalyst. At the same time, there are several publications wherein PHIP is applied

as a spin-labeling technique in order to study the pathways of hydrogen transfer in the

hydrogenation reactions of alkynes on other homogeneous catalysts [2, 3, 8, 22–24].

Therefore, in order to compare the results obtained in the hydrogenation over the

immobilized 1/PPh2-SiO2 catalyst with the conventional homogeneous hydrogenation

of propyne catalyzed by the dissolved Wilkinson’s complex, we acquired the NMR

spectrum of the reaction mixture under homogeneous conditions (Fig. 3).

As expected, in the PASADENA patterns observed in this case the signals at 4.74

(‘Hc) and 5.54 ppm (‘Hb) are polarized while the one at 4.87 ppm (Hd) is not.

Therefore, the heterogeneous hydrogenation over the immobilized 1/PPh2-SiO2

catalyst and the homogeneous hydrogenation with the dissolved Wilkinson’s

catalyst both proceed as a cis addition. It has been demonstrated recently for the

1/PPh2-SiO2 catalyst that the leaching of the rhodium complex from the support

surface is negligible under the applied conditions [20]. Therefore, we believe that in

our case, the reaction proceeds on the immobilized complex. However, it is worth to

note that homogeneous Wilkinson’s catalyst is more active in the hydrogenation

reaction of propyne in C6D6 solution than its heterogenized analog, evidenced in

particular by the fact that the NMR signal of propylene CH3 group (labeled Ha) is

clearly seen in Fig. 3 but not in Fig. 2. Indeed, for the same number of rhodium

centers the homogeneous catalyst shows a notable activity already at 50�C, whereas

for the immobilized complex the reaction mixture must be heated to 70�C to have a

comparable activity. Since the addition of dihydrogen to propyne occurs in both

Fig. 3 PASADENA 1H NMRspectrum of the reaction mixturedetected for the homogeneoushydrogenation of propynecatalyzed by the dissolvedWilkinson’s catalystin benzene-d6 solution atT = 50�C


123

cases in a pairwise manner and stereoselectively as a cis addition, we can conclude

that the mechanism of the heterogeneous hydrogenation over the immobilized

1/PPh2-SiO2 catalyst is very similar to the classical mechanism of the homogeneous

hydrogenation on the Wilkinson’s catalyst, i.e., most likely the nature of the

catalytic centers in the case of the immobilized complex is not changed

significantly.

3.2 Stereoselectivity of Dihydrogen Addition in the Gas Phase

Previously, it was shown that PHIP can be observed in both the PASADENA and

the ALTADENA experiments in the gas-phase propylene hydrogenation performed

using the immobilized Wilkinson’s catalyst (1/PPh2-SiO2) [19]. Moreover, signal

enhancement in the ALTADENA experiment was utilized for boosting the NMR

sensitivity in the MRI and remote-detection NMR experiments [14, 15, 25]. Despite

the significant success achieved in this field, little is known about the mechanism of

the gas-phase heterogeneous hydrogenation on such catalysts. Since a solvent is

absent under these conditions, one cannot simply assume that homogeneous

hydrogenation mechanism is preserved.

In the present work, we tested the immobilized Wilkinson’s catalyst in the

heterogeneous hydrogenation of propyne in the temperature range of 70–130�C. It

was found that initially, when the gas mixture starts to flow through the layer of the

immobilized Wilkinson’s catalyst, the PASADENA patterns in the detected 1H

NMR spectrum are the same as for the liquid-phase hydrogenation described above

(Fig. 4).

The observable polarized signals in this case correspond to ‘Hc and ‘Hb protons

located in the cis position with respect to each other. However, the polarized signals

are observed during only a short time period (*1 min) after enabling the mixture

flow through the catalyst bed. Thereafter, the intensity of the polarized signals

Fig. 4 PASADENA 1H NMRspectrum recorded during theheterogeneous hydrogenationreaction of propyne catalyzed bythe immobilized Wilkinson’scatalyst under the gas-phaseconditions at T = 70�C


123

quickly diminishes with the simultaneous loss of catalyst activity. This result is an

indication that there are significant differences in the reaction mechanisms in the

liquid-phase and in the gas-phase hydrogenation processes. In spite of the

similarities observed at the initial stage, a rapid deactivation is observed in the

gas-phase hydrogenation, whereas in the liquid-phase hydrogenation the polariza-

tion is observed during a long period of time without catalyst deactivation.

It is known that the reaction mechanism of olefin hydrogenation reaction on the

Wilkinson’s catalyst involves four key intermediates, [RhCl(PPh3)2], [RhCl(H)2

(PPh3)2], [RhCl(H)2(alkene)(PPh3)2] and [RhCl(alkyl)(H)(PPh3)2] [26–29]. It is also

known that in solution under reactive conditions, [RhCl(PPh3)2] and [RhCl(H)2

(PPh3)2] complexes can interact with other rhodium intermediates to form binuclear

compounds [28, 29]. Importantly, these compounds are less active as hydrogenation

catalysts in comparison to their mononuclear precursors. Tolman et al. [29] have

shown that for binuclear rhodium complexes [RhCl(P(p-tolyl)3)2]2, the equilibrium

constant of the [RhCl(H)2(P(p-tolyl)3)2(l-Cl)2RhCl(P(p-tolyl)3)2] dihydride inter-

mediate formation is 3.6 times lower than that for the formation of [RhCl(H)2

(PPh3)3] from [RhCl(PPh3)3] and H2. On the other hand, it was shown that H2

addition to [RhCl(PPh3)2] is 104 times faster in comparison to [RhCl(PPh3)3] [28].

Most likely, in the case of the immobilized catalyst, the catalytic cycle is mainly

driven by the formation of the anchored 14-electron rhodium intermediates, namely,

[RhCl(PPh3)(PPh2)-(CH2)2-(SiO2)n]. These intermediates are extremely active,

similar to [RhCl(PPh3)2] intermediates in the case of homogeneous hydrogenation.

These immobilized intermediates can react with H2 and C3H4 to produce the

reaction product (C3H6). However, these intermediates can also interact with each

other to form much less active dimers, and with surface functional groups such as

siloxane group (:Si–O–Si:) or several forms of silanol groups. These interactions

lead to the formation of Rh species that are much less active in the hydrogenation

reaction [30, 31]. Since it was shown that in the case of rhodium complex

[HRh(CO)(PPh3)3], the number of Rh–O–Si bonds is very small, we can assume

that the dominating contribution to the deactivation of the immobilized Wilkinson’s

catalyst is made by the dimerization reactions [32].

To verify this, we have purposely synthesized the immobilized catalyst starting

directly from the rhodium dimer complex [Rh(PPh3)2(l-Cl)]2 via its interaction with

the modified silica gel. The immobilized complex was found to be much less active,

and PHIP was not observed at all. Most probably, the dimer complex immobilizes as

a mixture of mono- and binuclear complexes on the support during its synthesis, and

the number of the anchored binuclear complexes is dominant. These results are in

agreement with the assumption that the deactivation of the immobilized Wilkinson’s

catalysts is mainly caused by the dimerization of the active mononuclear

intermediate complexes.

In practice, the anchoring of binuclear complexes can also occur in solution during

immobilization of a Wilkinson’s complex at the stage of 1/PPh2-SiO2 catalyst

synthesis. Indeed, Tolman et al. [29] have established that the Wilkinson’s complex is

prone to dimerization in solution. Moreover, recently, in the studies of hydrogenation

reaction on the Wilkinson’s catalyst it was shown that under H2 atmosphere in

addition to [Rh(PPh3)2(l-Cl)]2 the dihydride [Rh(H)2(PPh3)2(l-Cl)2Rh(PPh3)2] and


123

the tetrahydride [Rh(H)2(PPh3)2(l-Cl)]2 complexes are formed in solution [28, 29].

Indeed, when we synthesized the immobilized catalysts in the presence of H2 (under

the H2 or H2 ? C3H6 atmosphere), the activity of the catalysts was much lower as

compared to the immobilized catalysts prepared under Ar, and PHIP was not observed

even at the initial moment after starting the reagent mixture flow. Infrared (IR)

spectroscopy analysis revealed that when the immobilized catalyst was synthesized in

the presence of H2, the relative intensities of the signals at m = 1,440 cm-1,

m = 743 cm-1, m = 719 cm-1 and m = 694 cm-1 were lower in comparison with the

catalyst prepared under Ar (Figs. 5, 6).

Since these signals correspond to the aromatic rings of PPh3 ligands, we can

assume that the synthesis in the presence of H2 leads to the decrease in the number

of phosphine ligands in the immobilized catalyst. This indicates that binuclear

complexes, which are present during the homogeneous hydrogenation catalyzed by

the dissolved Wilkinson’s catalyst, are likely to get immobilized when the synthesis

of the immobilized catalyst is carried out in the presence of H2.

In order to suppress the dimerization reaction, the synthesis of the immobilized

Wilkinson’s complex was carried out in solutions containing various concentrations

of PPh3 (0.1, 0.05 and 0.01 M). It was found that catalysts synthesized in 0.1 and

0.05 M solutions of PPh3 were deactivating more slowly as compared to the

catalysts prepared without PPh3 addition. However, after 5–7 min of propylene

hydrogenation the activity of these catalysts also starts to decrease significantly. The

addition of PPh3 to solution during the catalyst synthesis should suppress the

dimerization reaction by shifting the equilibrium of the dimerization reaction

toward the formation of [RhCl(PPh3)3]. Therefore, only the mononuclear Wilkin-

son’s complex should be immobilized and the catalyst deactivation associated with

the low-activity binuclear complexes should be suppressed.

Fig. 5 IR spectrum of the immobilized Wilkinson’s catalyst synthesized under the Ar atmosphere


123

The results described above indicate that synthesis in the presence of extra PPh3

in solution under inert atmosphere is the most effective approach for preparation of

the immobilized 1/PPh2-SiO2 catalyst. However, this catalyst is not very stable.

Another way to suppress the dimerization reaction is to use the cationic rhodium

complexes as homogeneous precursors in the synthesis of immobilized catalysts, as

electrostatic interactions between the ions should prevent their dimerization.

Moreover, the observation of polarization in homogeneous conditions has been

described for several cationic complexes of Rh, Ir, Ru and Pd [2, 3, 23, 33, 34].

Herein, we used [Rh(COD)2]?[BF4]- complex as a precursor of an immobilized

cationic catalyst. It was found that at room temperature in the initial moments after

starting the propyne–parahydrogen gas mixture flow through the catalyst bed, the

signals corresponding to vicinal protons (‘Hc and ‘Hb in Fig. 7) of the vinyl

fragment of the propylene molecule are polarized. Interestingly, the polarization of

the reagent molecule is observed, which presumably points to the presence of an

exchange process that leads to a mutual pairwise exchange between the two

hydrogen atoms of propylene and the parahydrogen molecule.

In earlier studies, proton exchange was observed in the interaction of a

parahydrogen molecule with cationic rhodium catalysts for the geminal protons of

double bonds [33]. In our case, polarization is observed for the vicinal protons of

propylene, and only for those in the cis position with respect to each other (‘Hc and

‘Hb). Most probably, the p-bond between the two carbon atoms is not broken,

otherwise the polarization would be observed on all three protons of the vinyl

fragment of propylene, as described for the cationic rhodium and palladium

catalysts [3, 24, 33].

Notably, at room temperature there are no signals in the spectrum corresponding

to the hydrogen addition product propane (Fig. 7), whereas at 70�C the activity of

Fig. 6 The IR spectrum of the immobilized Wilkinson’s catalyst synthesized under the atmosphere ofH2 ? C3H6


123

the immobilized catalyst toward hydrogenation increases and the polarized signals

of propane are also observed (Fig. 8). However, the immobilized cationic catalyst is

unstable at this temperature and rapidly changes its color. The hydrogenation

continues with a high efficiency, but the intensity of the polarized signals goes

down. Most likely, at the elevated temperatures the nature of the catalytic sites is

changing. This issue is discussed below.

Fig. 7 PASADENA 1H NMRspectrum recorded during theheterogeneous reaction betweenparahydrogen and propylenecatalyzed by the 3/PPh2-SiO2

catalyst under the gas-phaseconditions at 25�C. Thepolarized signals correspondingto protons of the CH and CH2

groups of propylene are labeledas ‘Hb and ‘Hc, respectively.Signals from the CH3 groupprotons of propylene are markedas Ha. The signal of thehydrogen atom of the vinylfragment of propylene markedas Hd is not polarized

Fig. 8 PASADENA 1HNMR spectrum recordedduring the heterogeneoushydrogenation of propylenecatalyzed by the 3/PPh2-SiO2

catalyst under the gas-phaseconditions at T = 70�C


123

3.3 Partial Reduction of Immobilized Catalysts

As it was noted above, the catalytic activity of the immobilized Wilkinson’s catalyst

quickly declines in the gas-phase heterogeneous hydrogenation at 70�C. Since

increasing the temperature should increase the hydrogenation reaction rate, after the

deactivation of the immobilized catalyst it was also tested at 110�C in the gas-phase

hydrogenation of propyne. Interestingly, during several minutes after elevating the

temperature the increase in the catalyst activity is observed, and polarized signals of

propylene protons are seen in the 1H NMR spectra of the reaction mixture (Fig. 9).

It can be seen that in this spectrum the intensity of the signal at 4.87 ppm is larger as

compared to that in the spectrum recorded at 70�C (see Fig. 4), i.e., in addition to the

polarized signals of the (‘Hc) and (‘Hb) protons, the (‘Hd) proton is also polarized.

Moreover, when the temperature is increased further from 110 to 130�C, the increase

in both the activity of the catalyst and the intensity of the polarized signals is observed.

It should be noted that the catalyst changes its color from orange to dark grey after the

temperature increase. The observation of PHIP on all protons of propylene vinyl

fragment indicates that the H2 addition stereoselectivity is lost. Most probably, at a

higher temperature (110�C) the nature of the reaction center changes. Previously, in

the studies by Allum et al. [35], it has been shown that immobilized Wilkinson-type

catalysts under certain reaction conditions (15 bar of H2) at a relatively high

temperature (100�C) in n-heptane tend toward reduction of the Rh metal center. In our

case, most likely, the observed changes of the catalyst activity are also associated with

the reduction processes. X-ray photoelectron spectroscopy (XPS) analysis of the

catalysts after the reaction at 130�C revealed that 3d5/2 and 3d3/2 electron binding

energies of Rh were equal to 309.1 and 313.8 eV, respectively (Fig. 10).

These values are different from those found for the immobilized catalyst (310.1

and 315.1 eV) before the reaction. The values of 310 and 315 eV correspond to 3d5/2

and 3d3/2 electron binding energies of Rh for the Wilkinson’s complex in the solid

Fig. 9 PASADENA 1H NMRspectrum recorded in theheterogeneous hydrogenation ofpropyne catalyzed by theimmobilized Wilkinson’scatalyst under the gas-phaseconditions at T = 110�C


123

state [36]. Therefore, the obtained values of 310.1 and 315.1 eV for the fresh

immobilized catalyst agree well with those reported in the literature for the

Wilkinson’s complex. However, the values of 309.1 and 313.8 eV are different from

those of metallic rhodium particles (307.2 and 312.1 eV). At the same time, it was

reported previously that 3d5/2 and 3d5/2 Rh binding energies for the Wilkinson’s

complex reduced at 100 kPa H2 and 100�C for 18 h are equal to 308.2 and 312.7 eV,

respectively. As in our case, these values differ from those of the metallic rhodium

particles. On the other hand, in the study by Yan et al. [31], it has been reported that

for metallic rhodium catalyst Rh/SBA-15, which was prepared by the reduction of

RhCl3 under H2 at 637 K and subsequent treatment with a solution of PPh3, the

surface Rh atoms are in a Rh?d state (0 \ d\ 1). In that study, the binding energies

for particles have been also shown to be in the range of 307.5–308.0 eV. Since in our

case the reaction proceeds under milder conditions (1 bar H2, T = 130�C), we can

assume that the observed values of the binding energies, 309.1 and 313.8 eV, are the

consequence of the presence of the surface Rh?d species.

Previously, observation of PHIP in the gas-phase propylene hydrogenation

reaction catalyzed by the immobilized Wilkinson’s catalyst at 80–150�C was

reported in Refs. [14, 15, 19]. Moreover, Bouchard et al. [15] explicitly stated that

activation of the catalyst required several hours at 145�C under 15 psig of hydrogen–

propylene gas mixture. The results of the current study indicate that, in fact, the gas-

phase hydrogenation at such temperatures proceeds on the partially reduced

immobilized catalyst. In the gas-phase hydrogenation of propylene in the ALTA-

DENA experiment, the enhancement factor of up to 300 was reported [14, 15]. Here,

we utilized the same type of the catalyst (1/PPh2-SiO2) in the hydrogenation in the

temperature range of 100–300�C, i.e., under the conditions when hydrogenation

occurs on the reduced catalyst. At 150�C, the maximum conversion of propylene was

Fig. 10 Rh 3d5/2 and 3d3/2 photoelectron spectra of the immobilized Wilkinson’s catalyst (a) before and(b) after the gas-phase hydrogenation at 130�C


123

24% and the signal enhancement factor in an ALTADENA experiment was 180. This

enhancement factor is comparable to the value reported earlier. The small difference

can be caused by the slightly different experimental conditions. In particular, the

transfer time of the polarized product produced outside the spectrometer to the NMR

magnet was slightly longer in our case, thus more polarization was lost due to

relaxation during the transport from the hydrogenation reactor to the NMR probe.

For comparison, we examined the PHIP effects in the heterogeneous hydroge-

nation reaction using the purposely reduced metallic catalyst Rh/SiO2. It was found

that in a PASADENA experiment, both propylene and propyne hydrogenations gave

rise to PHIP (Figs. 11, 12). Interestingly, if supported metal catalyst Rh/SiO2 is used

in the hydrogenation of propylene, then in addition to the polarization of propane

protons (‘H1 and ‘H2) the polarized signals of propylene vinyl fragment protons are

also observed (‘Hc, ‘Hd and ‘Hb).

The observation of polarization on the methyl and methylene groups of propane

in propylene hydrogenation suggests that the hydrogenation mechanism over Rh/

SiO2 catalysts involves the pairwise route of the reaction, which, however, is the

minor one on supported metal particles [16]. In addition, it should be noted that

the polarization of –CH=CH2 group protons in propylene hydrogenation indicates

the presence of side reaction processes, which lead to a pairwise hydrogen exchange

in the propylene molecule. In the case of propyne hydrogenation over Rh/SiO2, the

polarized signals are only observed for protons of the vinyl fragment of the

propylene molecule while propane is not polarized. Moreover, polarization is

observed on all protons of the vinyl fragment of the produced propylene, i.e., the

transfer of hydrogen atoms occurs without stereoselectivity. This is in accord with

the observed loss of stereoselectivity in propyne hydrogenation over the immobi-

lized Wilkinson’s catalyst at elevated temperatures described above. Therefore, this

Fig. 11 PASADENA 1H NMRspectrum recorded in theheterogeneous hydrogenationreaction of propylene catalyzedby Rh/SiO2 catalyst under thegas-phase conditions atT = 120�C


123

observation is an additional evidence of the possible reduction of immobilized Rh

catalysts at elevated temperatures in the presence of H2.

4 Conclusions

The use of heterogeneous catalysts can offer much more flexibility in developing

various implementations of the PHIP-based hyperpolarization techniques than the

widely utilized homogeneous catalysts. In spite of this, the first observation of PHIP

in heterogeneous hydrogenations was reported only recently. At present, the major

challenge is to find a heterogeneous catalyst for which hydrogenation predominantly

occurs via a pairwise transfer of two hydrogen atoms of a dihydrogen molecule to a

substrate. In this respect, immobilization of homogeneous catalysts is a promising

approach since in this case the pairwise addition of dihydrogen should be an

inherent feature. Herein, we focused mainly on the study of the immobilized

Wilkinson’s catalyst, while the immobilized binuclear and cationic Rh complexes

were also examined for comparison. It is found that in the liquid-phase and in the

gas-phase propyne hydrogenation over the immobilized Wilkinson’s catalyst, the

stereoselectivity toward cis addition in the reaction is observed, indicating that

the mechanism of the homogeneous hydrogenation is most likely preserved. In the

gas-phase hydrogenation, however, the catalyst is not stable, PHIP is observed only

during the initial moments after the reaction is started and deactivation of the

catalyst occurs. It is likely that deactivation takes place because the active Rh

species tend to dimerize and interact with the surface of the support. The use of the

immobilized cationic Rh complex, for which the dimerization is suppressed,

revealed that proton exchange was taking place to a significant extent in the

propylene hydrogenation reaction over this catalyst. The immobilized binuclear Rh

complex had a very poor activity in the hydrogenations, and no PHIP was observed

for this catalyst. Elevating the temperature above 70�C leads to the reduction of all

examined immobilized Rh catalysts under reaction conditions in the gas-phase

hydrogenations. At the same time, particularly for the immobilized Wilkinson’s

Fig. 12 PASADENA 1H NMRspectrum recorded in theheterogeneous hydrogenation ofpropyne catalyzed by the Rh/SiO2 catalyst under the gas-phase conditions at T = 120�C


123

complex, hydrogenation activity of the catalysts significantly increases upon

reduction, which leads to the observation of PHIP in the 1H NMR spectra.

Comparison with the experiments performed using fully reduced supported Rh

metal catalysts also demonstrated PHIP generation in heterogeneous gas-phase

hydrogenations of propyne and propylene.

Acknowledgments We are grateful for the support received from the Russian Academy of Sciences

(RAS, 5.1.1), Russian Foundation for Basic Research (11-03-93995-CSIC_a, 11-03-00248-a), Siberian

Branch of RAS (67, 88), (NSh-7643.2010.3), Russian Ministry of Science and Education

(02.740.11.0262) and the Council on Grants of the President of the Russian Federation (MK-

1284.2010.3). We acknowledge Dr. Elena V. Karpova for IR-spectroscopy analysis, Dr. Igor P. Prosvirin

for XPS studies and I. E. Beck for providing the Rh/SiO2 catalyst.

References

1. C.R. Bowers, in Encyclopedia of Nuclear Magnetic Resonance, ed. by D.M. Gant, R.K. Harris, vol. 9

(Wiley, Chichester, 2002), p.750

2. S.B. Duckett, C.J. Sleigh, Prog. Nucl. Magn. Reson. Spectrosc. 34, 71 (1999)

3. S.B. Duckett, N.J. Wood, Coord. Chem. Rev. 252, 2278 (2008)

4. J. Natterer, J. Bargon, Prog. Nucl. Magn. Reson. Spectrosc. 31, 293 (1997)

5. K. Golman, O. Axelsson, H. Johannesson, S. Mansson, C. Olofsson, J.S. Petersson, Magn. Reson.

Med. 46, 1 (2001)

6. U. Bommerich, T. Trantzschel, S. Mulla-Osman, G. Buntkowsky, J. Bargon, J. Bernarding, Phys.

Chem. Chem. Phys. 12, 10309 (2010)

7. F. Reineri, A. Viale, G. Giovenzana, D. Santelia, W. Dastru, R. Gobetto, S. Aime, J. Am. Chem. Soc.

130, 15047 (2008)

8. L.T. Kuhn, U. Bommerich, J. Bargon, J. Phys. Chem. A 110, 3521 (2006)

9. R. Eisenberg, T.C. Eisenschmid, M.S. Chinn, R.U. Kirss, Adv. Chem. Ser. 230, 47 (1992)

10. M. Haake, J. Natterer, J. Bargon, J. Am. Chem. Soc. 118, 8688 (1996)

11. M. Roth, P. Kindervater, H.P. Raich, J. Bargon, H.W. Spiess, K. Munnemann, Angew. Chem. Int. Ed.

49, 8358 (2010)

12. T.C. Eisenschmid, J. Mcdonald, R. Eisenberg, R.G. Lawler, J. Am. Chem. Soc. 111, 7267 (1989)

13. L.T. Kuhn, J. Bargon, Top. Curr. Chem. 276, 25 (2007)

14. L.S. Bouchard, K.V. Kovtunov, S.R. Burt, M.S. Anwar, I.V. Koptyug, R.Z. Sagdeev, A. Pines,

Angew. Chem. Int. Ed. 46, 4064 (2007)

15. L.S. Bouchard, S.R. Burt, M.S. Anwar, K.V. Kovtunov, I.V. Koptyug, A. Pines, Science 319, 442

(2008)

16. K.V. Kovtunov, I.E. Beck, V.I. Bukhtiyarov, I.V. Koptyug, Angew. Chem. Int. Ed. 47, 1492 (2008)

17. A.M. Balu, S.B. Duckett, R. Luque, Dalton Trans., 5074 (2009)

18. I.V. Koptyug, V.V. Zhivonitko, K.V. Kovtunov, ChemPhysChem 11, 3086 (2010)

19. I.V. Koptyug, K.V. Kovtunov, S.R. Burt, M.S. Anwar, C. Hilty, S.I. Han, A. Pines, R.Z. Sagdeev, J.

Am. Chem. Soc. 129, 5580 (2007)

20. S.G. Shyu, S.W. Cheng, D.L. Tzou, Chem. Commun., 2337 (1999)

21. J.A. Osborn, F.H. Jardine, J.F. Young, G. Wilkinson, J. Chem. Soc. A, 1711 (1966)

22. R.U. Kirss, R. Eisenberg, Inorg. Chem. 28, 3372 (1989)

23. D. Schleyer, H.G. Niessen, J. Bargon, New J. Chem. 25, 423 (2001)

24. J. Lopez-Serrano, S.B. Duckett, S. Aiken, K.Q.A. Lenero, E. Drent, J.P. Dunne, D. Konya,

A.C. Whitwood, J. Am. Chem. Soc. 129, 6513 (2007)

25. V.V. Telkki, V.V. Zhivonitko, S. Ahola, K.V. Kovtunov, J. Jokisaari, I.V. Koptyug, Angew. Chem.

Int. Ed. 49, 8363 (2010)

26. D.J. Nelson, R.B. Li, C. Brammer, J. Org. Chem. 70, 761 (2005)

27. S.A. Colebrooke, S.B. Duckett, J.A.B. Lohman, R. Eisenberg, Chem. Eur. J. 10, 2459 (2004)

28. S.B. Duckett, C.L. Newell, R. Eisenberg, J. Am. Chem. Soc. 116, 10548 (1994)

29. C.A. Tolman, P.Z. Meakin, D.I. Lindner, J.P. Jesson, J. Am. Chem. Soc. 96, 2762 (1974)


123

30. R. Duchateau, Chem. Rev. 102, 3525 (2002)

31. A. Grunberg, Y.P. Xu, H. Breitzke, G. Buntkowsky, Chem. Eur. J. 16, 6993 (2010)

32. X.J. Lan, W.P. Zhang, L. Yan, Y.J. Ding, X.W. Han, L.W. Lin, X.H. Bao, J. Phys. Chem. C 113,

6589 (2009)

33. A. Harthun, R. Giernoth, C.J. Elsevier, J. Bargon, Chem. Commun., 2483 (1996)

34. H.G. Niessen, D. Schleyer, S. Wiemann, J. Bargon, S. Steines, B. Driessen-Hoelscher, Magn. Reson.

Chem. 38, 747 (2000)

35. K.G. Allum, R.D. Hancock, I.V. Howell, T.E. Lester, S. McKenzie, R.C. Pitkethly, P.J. Robinson,

J. Organomet. Chem. 107, 393 (1976)

36. C. Furlani, G. Mattogno, G. Polzonetti, G. Braca, G. Valentini, Inorg. Chim. Acta 69, 199 (1983)


123

Parahydrogen-Induced Polarization in Heterogeneous Hydrogenations over Silica-Immobilized Rh Complexes

Documents