Document [original]

This version is available at https://doi.org/10.14279/depositonce-8330

right to use is granted. This document is intended solely for

personal, non-commercial use.

Trunk, M., Teichert, J. F., & Thomas, A. (2017). Room-Temperature Activation of Hydrogen by Semi-

immobilized Frustrated Lewis Pairs in Microporous Polymer Networks. Journal of the American Chemical

Society, 139(10), 3615–3618. https://doi.org/10.1021/jacs.6b13147

This document is the Accepted Manuscript version of a Published Work that appeared in final form in the

technical editing by the publisher. To access the final edited and published work see

https://pubs.acs.org/doi/10.1021/jacs.6b13147.

Trunk, M., Teichert, J. F.; Thomas, A.

Room-Temperature Activation of

Hydrogen by Semi-immobilized Frustrated

Lewis Pairs in Microporous Polymer

Networks

Accepted manuscript (Postprint)Journal article |

ROOM-TEMPERATURE ACTIVATION OF HYDROGEN BY SEMI-

IMMOBILIZED FRUSTRATED LEWIS PAIRS IN MICROPOROUS

POLYMER NETWORKS

Matthias Trunk,† Johannes F. Teichert,‡ and Arne Thomas*,†

†Department of Chemistry, Functional Materials, Technische Universitat Berlin, Hardenbergstraße 40, 10623 Berlin, Germany

‡Department of Chemistry, Organic Chemistry/Sustainable Synthetic Methods, Technische Universitat Berlin, Straße des 17.

Juni 115, 10623 Berlin, Germany

ABSTRACT: Porous polymer networks based on sterically encumbered triphenylphosphine motifs, mimicking the basic

sites employed in frustrated Lewis pair (FLP) chemistry, were synthesized via Yamamoto polymerization and their

interactions with the strong Lewis acid B(C6F5)3 probed. The combinations yield semi-immobilized FLPs, which are able

to cleave dihydrogen heterolytically at ambient temperature and low hydrogen pressure.

Over the past decade, microporous polymer networks have garnered attention as materials for a wide variety of applications.1,2

Emanating from rigid organic building blocks, materials with very small pores between 0.5 and 2 nm and accessible surface

areas of up to 6400 m2 g−1, as well as high thermal and chemical stability,3 have been synthesized by a number of polymerization

methods, such as metal-mediated homocoupling reactions, metal-catalyzed cross-coupling reactions, condensation reactions,

and many others.2,4 In contrast to metal−organic frameworks and covalent organic frameworks, these materials do not exhibit

crystallinity; emerging from C−C bond-forming reactions, they can forego the use of heteroatoms entirely, allowing for a pure

hydrocarbon skeleton. This feature is especially advantageous when designing catalysts with highly sensitive active sites, as it

enables the exclusion of nonorthogonal functional groups.

Such types of functional groups can be found as catalytic sites in frustrated Lewis pairs (FLPs).7 The members of this

emerging class of molecular catalysts are solely based on maingroup elements and in most cases consist of a boron-based Lewis

acid (LA) and a Lewis base (LB), usually comprising phosphorus or nitrogen as basic sites.8 Both active sites are sterically

encumbered by bulky moieties in order to prevent them from forming a strong covalent bond and consequently quenching their

opposing reactivities. Small molecules (H2, CO2, etc.) fitting in between these bulky groups can be activated through double

polarization by the acidic and basic sites, leading to (zwitter)ionic products (see Scheme 1).6,7,9 For this activation to take place,

the steric hindrance must not be sufficient to prohibit any interaction; in solution, a weak but reversible coordination must be

possible, lest the polarization be precluded entirely.10 Subsequently, the activated intermediate(s) can be used for reactions with

a number of substrates.

The most intensely researched application for this type of catalyst is the heterolytic splitting of dihydrogen and its

Scheme 1. (Top) Equilibrium State of a Generic Frustrated Lewis Pair (LA = Lewis Acid, LB = Lewis Base) and Activation

of Carbon Dioxide and Hydrogen; (Bottom) (Non)reversible Activation of Hydrogen by Reported FLPs5,6

subsequent use in the reduction of C=C,11 C=N,12 and C=O13 double bonds as well as cleavage of silylenol ethers,14 among

others. LA and LB can be part of the same molecule; i.e., both intra- and intermolecular FLPs are capable of activating small

molecules.9 In many cases, though, the mechanistic properties of an activation process are deceptively simple. Even in cases of

both active sites within the same molecule, theoretical studies rather suggest intermolecular activation via a two-fold head-to-

tail interaction between two intramolecular FLP molecules.15

Whereas a great number of Lewis basic polymers has been published, since these are easy to obtain and often display

interesting CO2 adsorption properties, only a handful of Lewis acidic moieties in porous polymers have been reported,16,17 none

of which are used for catalytic applications. Due to the use of very strong LAs such as tris(pentafluorophenyl)borane (commonly

abbreviated to BCF) or related compounds, these catalysts are often highly sensitive to external factors such as humidity or

oxygen, which makes their handling challenging. To overcome this issue, extensive studies toward electronic fine-tuning have

produced robust LAs which can be handled without the use of inert gas techniques.18 Recent advances have furthermore afforded

so-called “inverse” FLPs based on a weak and more robust LA but which are used in conjunction with a very strong and

sensitive LB on the other hand.19

In order to be able to enhance stability and recyclability and to further increase the appeal of FLPs, the incorporation of FLPs

into solid materials, e.g., a microporous polymer network, seems an advantageous endeavor. In this work we focused on the

generation of FLPs emerging from porous polymers comprising sterically encumbered phosphine moieties, which are

impregnated with a readily available strong LA. We demonstrate the ability of these semi-immobilized FLP systems to activate

dihydrogen at ambient temperature and low hydrogen pressure.

The syntheses of the phosphine polymers were carried out according to the synthesis procedure for porous polymer networks

via Yamamoto polymerization20 starting from tris(pbromophenyl)phosphine derivatives (see Scheme 2) with

Scheme 2. Syntheses of Triarylphosphine Polymersa

aReagents and conditions: (a) i-PrMgCl, THF, 0 °C, 1 h, then PCl3, THF, −78 °C to rt, 12 h; (b) Ni(COD)2, DMF/THF, rt,

15 h; (c) nBuLi, Et2O, −78 °C, 2 h, then PCl3, Et2O, −78 °C to rt, 12 h; (d) Ni(COD)2, DMF/THF, rt, 22 h.

increasing steric demand around the phosphorus atoms. The structure of unmethylated triphenylphosphine polymer has been

described previously.21, 22 Due to the use of the less reactive aryl chloride starting material, harsher synthesis conditions were

required, whereas in this work a milder synthesis protocol23 was applied due to the exclusive use of the more reactive brominated

species.

Purification via Soxhlet extraction with methanol and drying in vacuo at 80 °C yields the product networks (Figure S1).

Tris(2-methylphenyl)phosphine polymer P1 was collected as white and light solid, whereas tris(2,6-dimethylphenyl)phosphine

polymer P2 has an intense yellow color and fluffy texture. P1 and P2 feature surface areas of 1045 and 947 m2 g−1, respectively,

with P2 featuring a remarkable hysteresis (see Figure 1). These results are in line with the reported surface area of unmethylated

triphenylphosphine polymer of 1284 m2 g−1.21 The decrease in surface area can thus be interpreted as a result of the increased

monomer weight and pore blocking by increasing amounts of methyl groups.

Organic solvents such as toluene, bromobenzene, dichloromethane, and THF are readily absorbed by the polymers, which

exhibit strong swelling properties and form gel-like mixtures with these solvents. To probe the properties of the synthesized

materials as LBs, their behavior toward LAs such as commercially available BCF has to be examined. When a substance is pre-

dissolved in the absorbed solvent, it can be introduced into the pores of the polymer. Choosing a volatile

Figure 1. Nitrogen sorption isotherms of P1 (blue) and P2 (red).

solvent facilitates drying of the obtained material. BCF is well soluble in aromatic and chlorinated solvents. To gain information

on the interaction between BCF as a LA and the bulky LB, no other Lewis acidic or basic functions may be present. Therefore,

inside the glovebox, BCF and porous phosphine polymer were suspended in DCM, which is volatile, non-coordinating, and

neither acidic nor basic, readily dissolves BCF, and leads to rapid swelling of the polymer. When impregnated with BCF

solutions in non-coordinating solvents, e.g., DCM, the LA is readily absorbed into the polymer. The impregnation of the yellow

polymer network P2 was accompanied by darkening of the color, while colorless P1 remained colorless. Washing with

coordinating solvents such as diethyl ether or THF restores the original network. Even weakly coordinating solvents such as

methanol or acetone are coordinated preferentially over the bulky phosphine moieties.

Both polymers (see Figure 2) as well as their composites with BCF exhibit strong fluorescence under UV irradiation (bottom

row) at 254 (weak) and 366 nm (strong). BCF is readily absorbed by the polymer and quenches the fluorescence (see Supporting

Information), especially for polymer P1,

Figure 2. P1 (left block) and P2 (right block) in DCM suspension before and after addition of BCF. Top row, UV lamp off;

bottom row, under irradiation of UV light (366 nm).

pointing to a stronger interaction with the LA (Figure S1). Furthermore, after addition of BCF, the finely dispersed polymer

particles contract and sediment within a matter of minutes, whereas the empty polymer particles tend to collect at the surface

of DCM.

Solid-state magic angle spinning NMR spectroscopy of the dried solids confirms the assumed structures (see Figure 3 and

Figure 3. Impregnation procedure of triphenylphosphine derivative polymers (exemplified for P2) followed by 31P NMR

spectroscopy: P1 (blue) and P2 (red) before (upper spectra) and after (lower spectra) impregnation with BCF. Asterisks denote

spinning sidebands.

Figure S2). The 31P NMR spectra each exhibit one main peak in the region of −34 and −36 ppm alongside a small impurity at

approximately +32 ppm, indicative of small amounts of phosphine oxide present in each network. These traces of phosphine

oxide have been reported previously21 and seem to be the result of a side reaction of the Yamamoto polymerization, since the

polymers were synthesized under rigorously inert conditions. Additionally, P1 and P2 are prone to phosphine oxide formation

in air at elevated temperatures over elongated periods of time and should therefore be stored under inert gas.

Upon impregnation of the polymers with BCF and subsequent solvent removal, the 31P solid-state NMR signals are shifted

to lower field, as expected when coordinating to an acidic species. It is noteworthy that nearly the entire signal of the phosphine

within the polymers is shifted downfield, showing that almost all phosphine groups in the polymer networks P1 and P2 are

accessible to the molecular LA. Notably, the extent of the shift strongly correlates with the degree of methylation of the

triphenylphosphine derivative. For P1 a shift of 14.7 ppm is observed, whereas for P2 a shift of only 6.5 ppm is observed (see

Figure 3). This shift difference indicates that the bond strength decreases with increasing steric hindrance around the central

phosphorus atom, which means that the higher the degree of frustration is, the weaker the emerging bond is rendered. This

observation is quite remarkable.

According to Heine et al.,10 the formation of a FLP is usually not observed at ambient temperature in solution; i.e., the solution

NMR spectra of the individual components do not differ from the spectra of stoichiometric mixtures. Due to the only weakly

coordinating nature of the FLP, the favorable interactions gained from association of LA and LB are dominated by the entropic

contribution of the dissociation process. In the solid state, on the other hand, after solvent removal, LA and LB are forced into

and kept in a “not-sofrustrated” state, so that the interaction can actually be observed, and its strength is determined by the

difference in the chemical shifts, Δδ.

To verify this hypothesis, we compared the chemical shifts of the solid-state NMR spectrum of a “dry” FLP, consisting of

the molecular counterpart of P2, tris(2,6-dimethylphenyl)phosphine, and BCF to the solid-state NMR spectrum of the phosphine

molecule on its own. As expected, in the “dry” FLP, the phosphorus signal is shifted downfield by ca. 10 ppm compared to the

phosphine (see Figure S3).

According to the molecular compounds,5,6 the combinations P1/BCF and P2/BCF should form FLPs and be active toward the

cleavage of dihydrogen. When used in combination with BCF, the molecular equivalent of P1, tris(o-tolyl)phosphine, has been

reported to activate hydrogen reversibly at ambient conditions to form [(o-C6H4Me)3PH][HB(C6F5)3], whereas the molecular

counterpart of P2, trimesitylphosphine, forms the hydrogen adduct [(1,3,5-C6H2Me3)3PH][HB(C6F5)3], which remains stable

even when exposed to temperatures of up to 150 °C. To prove the hydrogen activation capability of our polymeric materials,

we performed H/D isotope scrambling experiments typically employed to demonstrate FLP behavior in a new material24 for

composites of BCF with P1, P2, and the molecular counterpart of P1, tris(o-tolyl)phosphine. To that end, inside a glovebox, a

Young-type NMR tube was charged with a stoichiometric mixture of P1/BCF or P2/BCF and cyclohexane-d12, and the

suspension was subjected to ultrasonication for 15 min to ensure sufficient intermixing of the poorly soluble BCF and the

polymer. Subsequently, the tube was pressurized with an H2/D2 mixture (ca. 1:1 v/v) to 6 bar and analyzed via 1H NMR

spectroscopy beginning several hours after pressurization of the tube. In both cases, an emerging triplet can be observed at 4.54

ppm shortly after pressurization, indicating the formation of HD (see Figure 4). The control reaction comprising only porous

phosphine polymer P2 and solvent under H2/D2 pressure did not result in any conversion to HD. Although the control reaction

with BCF as the sole reactant gave the scrambling product, albeit at much lower rate than the BCF/P2 composite, it must be

pointed out that it has been proven theoretically25 and experimentally under low hydrogen pressure5 that BCF alone is not active

in the heterolytic cleavage of H2. In our case, residues of protic solvents attached to BCF or trace water on the glass surface

might be responsible for providing a proton shuttle in the activation of hydrogen. Furthermore, if a composite of BCF and

another highly microporous polymer network devoid of any functional groups (PPN-6, 4172 m2 g−1, also prepared by Yamamoto

polymerization) is used, no hydrogen activation product can be detected. It seems, therefore, that the impregnation procedure

of a polymer has a negating effect on the activity of BCF itself. This could be caused by the highly hydrophobic and

nonpolarizable surface of the networks. It can

Figure 4. (A) Kinetic measurement of H/D exchange catalyzed be P2/BCF determined via 1H NMR. (B) Integral ratio of HD/H2

plotted as a function of time.

therefore be assumed that the observed reactivity of the composites is not only independent of but inherently much higher than

the activity of “pure” BCF. The composite tris(otolyl)phosphine/BCF gives the expected statistical H2/HD ratio after merely a

few hours. This shows that the incorporation of the active sites into the polymer imposes a strong kinetic restriction on the H/D

exchange process.

In summary, we hereby report the first synthesis of two semiimmobilized frustrated Lewis pairs with the basic components

embedded into the backbones of microporous polymer networks. Addition of the commercially available Lewis acid,

tris(pentafluorophenyl)borane, to polymers structurally akin to trimesitylphosphine yielded two functional heterogeneous

hydrogen activation catalysts. Their ability to cleave dihydrogen at ambient temperature and low pressure was demonstrated

via isotope scrambling experiments.

Further combinations of Lewis acidic and basic species need to be created to obtain heterogeneous catalyst systems capable

of splitting dihydrogen at low pressures and to enable catalytic hydrogenation. Furthermore, full heterogenization of FLPs could

even further enhance the reactivity and applicability of such systems.

ASSOCIATED CONTENT

*S Supporting Information

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/jacs.6b13147.

Experimental details on synthesis of monomers and polymers and hydrogen activation; 13C NMR of polymers and

molecular model compounds; fluorescence spectra and TGA of the polymers, including Figures S1− S4 (PDF)

AUTHOR INFORMATION

Corresponding Author

*arne.thomas@tu-berlin.de

ORCID

Johannes F. Teichert: 0000-0003-1043-8092

Arne Thomas: 0000-0002-2130-4930

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS

The authors thank Christina Eichenauer for surface area and TGA measurements. We also thank Douglas Stephan for his

insights into the subject of FLP materials. This work was funded by the ERC Project ORGZEO (Grant No. 278593) and the

DFG (Cluster of Excellence UniCat).

REFERENCES

(1) Thomas, A. Angew. Chem., Int. Ed. 2010, 49, 8328.

(2) Dawson, R.; Cooper, A. I.; Adams, D. J. Prog. Polym. Sci. 2012, 37, 530.

(3) Yuan, D.; Lu, W.; Zhao, D.; Zhou, H.-C. Adv. Mater. 2011, 23, 3723.

(4) Rose, M. ChemCatChem 2014, 6, 1166.

(5) Welch, G. C.; Stephan, D. W. J. Am. Chem. Soc. 2007, 129, 1880. (6) Ullrich, M.; Lough, A. J.; Stephan, D. W. J.

Am. Chem. Soc. 2009, 131, 52.

(7) Welch, G. C.; Juan, R. R. S.; Masuda, J. D.; Stephan, D. W. Science 2006, 314, 1124.

(8) Stephan, D. W. J. Am. Chem. Soc. 2015, 137, 10018.

(9) Spies, P.; Erker, G.; Kehr, G.; Bergander, K.; Fröhlich, R.; Grimme, S.; Stephan, D. W. Chem. Commun. 2007, 2,

5072.

(10) Zeonjuk, L. L.; Vankova, N.; Mavrandonakis, A.; Heine, T.; Röschenthaler, G.-V.; Eicher, J. Chem. - Eur. J. 2013,

19, 17413.

(11) Greb, L.; Oña-Burgos, P.; Schirmer, B.; Grimme, S.; Stephan, D.

W.; Paradies, J. Angew. Chem., Int. Ed. 2012, 51, 10164.

(12) Spies, P.; Schwendemann, S.; Lange, S.; Kehr, G.; Fröhlich, R.; Erker, G. Angew. Chem., Int. Ed. 2008, 47, 7543.

(13) Lindqvist, M.; Sarnela, N.; Sumerin, V.; Chernichenko, K.; Leskela, M.; Repo, T. Dalton Trans. 2012, 41, 4310.

(14) Wang, H.; Fröhlich, R.; Kehr, G.; Erker, G. Chem. Commun. 2008, 45, 5966.

(15) Liu Zeonjuk, L.; St. Petkov, P.; Heine, T.; Röschenthaler, G.-V.; Eicher, J.; Vankova, N. Phys. Chem. Chem. Phys.

2015, 17, 10687.

(16) Suresh, V. M.; Bandyopadhyay, A.; Roy, S.; Pati, S. K.; Maji, T.

K. Chem. - Eur. J. 2015, 21, 10799.

(17) Li, Z.; Li, H.; Xia, H.; Ding, X.; Luo, X.; Liu, X.; Mu, Y. Chem. Eur. J. 2015, 21, 17355.

(18) Scott, D. J.; Fuchter, M. J.; Ashley, A. E. Angew. Chem., Int. Ed. 2014, 53, 10218.

(19) Mummadi, S.; Unruh, D. K.; Zhao, J.; Li, S.; Krempner, C. J. Am. Chem. Soc. 2016, 138, 3286.

(20) Schmidt, J.; Werner, M.; Thomas, A. Macromolecules 2009, 42, 4426.

(21) Zhang, Q.; Yang, Y.; Zhang, S. Chem. - Eur. J. 2013, 19, 10024. (22) Pei, C.; Ben, T.; Guo, H.; Xu, J.; Deng, F.;

Xiang, Z.; Cao, D.; Qiu, S. Philos. Trans. R. Soc., A 2013, 371, 20120312.

(23) Lu, W.; Yuan, D.; Sculley, J. P.; Zhao, D.; Krishna, R.; Zhou, H.-

C. J. Am. Chem. Soc. 2011, 133, 18126.

(24) vom Stein, T.; Perez, M.; Dobrovetsky, R.; Winkelhaus, D.; Caputo, C. B.; Stephan, D. W. Angew. Chem., Int. Ed. 2015,

54, 10178. (25) Nikonov, G. I.; Vyboishchikov, S. F.; Shirobokov, O. G. J. Am. Chem. Soc. 2012, 134, 5488.