Supported Lanthanum Borohydride Catalyzes CH Borylation Inside Zeolite Micropores.

The zeolite-supported lanthanide La(BH4 )x -HY30 catalyzes C-H borylation of benzene with pinacolborane (HBpin), providing a complementary approach to precious, late transition metal-catalyzed borylations. The reactive catalytic species are generated from La grafted at the Brønsted acid sites (BAS) in micropores of the zeolite, whereas silanoate- and aluminoate-grafted sites are inactive under the reaction conditions. During typical catalytic borylations, conversion to phenyl pinacolborane (PhBpin) is zero-order in HBpin concentration. A turnover number (TON) of 167 is accessed by capping external silanols, selectively grafting at BAS sites, and adding HBpin slowly to the reaction.

Hydrocarbon C−H borylation with pinacol diborane (B2pin2), efficiently catalyzed by Group 9 organometallics, is a single‐step functionalization that provides versatile organoboranes.[ , , , ] Catalysts that access new mechanisms or distinct selectivity for C−H borylation, utilize earth‐abundant metal centers such as lanthanum (an over‐produced co‐product of technologically essential rare earth elements), or use the pinacolborane (HBpin) reagent could complement late metal catalysts that currently dominate this transformation. In particular, catalysts employing d 0 or d 0f centers are appealing for new C−H borylations, given well‐established C−H bond activations mediated or catalyzed via elementary σ‐bond metathesis reactions. Despite early promise, the sole example of rare earth‐catalyzed C−H borylation involves ortho‐directed derivatization of aromatic ethers catalyzed by 5–10 mol % (C5Me4H)2LnR (Ln=Y, Lu). Reactants and ancillary ligands’ steric properties, coordination of the aryl ether, and catalyst deactivation by ring‐opening of HBpin are important factors in that catalysis.

Oxide‐supported, confined organolanthanide compounds could potentially tolerate elevated reaction temperatures and access low‐coordinate metal centers needed for C−H bond activations. Because [Ln]H⋅HBpin adducts are proposed to directly react with carbonyls or epoxides during catalytic hydroboration reactions or act as masked hydrides, we hypothesized that related species from lanthanide borohydrides could mediate borylation of hydrocarbons. Cationic molecular and surface‐supported lanthanide borohydride complexes also have shown promise in ring‐opening and addition polymerization of esters.[ , ] In addition, late metal complexes supported on metal–organic frameworks (MOFs) or periodic mesoporous organosilicas (PMOs) have provided efficient catalysts for C−H borylation, often showing the benefits of site isolation.[ , , , , ] Metal centers supported on zeolites are well known to catalyze other conversions of hydrocarbons with size and shape selectivity. Here, we identify lanthanide borohydride species grafted at the bridging Si−OH−Al Brønsted acid sites (BAS) in zeolite micropores as catalysts for C−H borylation and describe strategies for increasing turnover number (TON), guided by solid‐state nuclear magnetic resonance (SS NMR), diffuse reflectance infrared Fourier‐transform spectroscopy (DRIFTS), and kinetic analysis.

First, La(BH4)3(THF)3 grafted onto inorganic supports, including SiO2, γ‐Al2O3, SiO2/Al2O3, or the microporous faujasite zeolite HY30 (Si/Al=30) were compared as potential catalysts for C−H borylation of benzene. The latter two supports contain heteroatomic bridging hydroxy as BAS, which are mostly isolated in the highly siliceous HY30 (

Table 1

C−H borylation of benzene catalyzed by lanthanum borohydride complexes.[a]

Catalyst	La [mmol g−1][b]	La [mo %][c]	Conversion [%]	Yield [%]	Turnovers	Selectivity [%]
La(BH4)3(THF)3	n/a	0.18	≈0	0	n/a	n/a
La(BH4) x (THF) n ‐SiO2	0.14	0.36	≈0	0	n/a	n/a
La(BH4) x (THF) n ‐Al2O3 [d]	0.02	0.18	≈0	0	n/a	n/a
La(BH4) x (THF) n ‐SiO2/Al2O3	0.09	0.22	100	1.3	6	1.1
La(BH4) x (THF) n ‐HY30	0.07	0.18	100	1.8	10	1.8
La(BH4)2(THF)2.5‐TPS‐HY30	0.05	0.12	95	7.4	62	7.8

[a] Reaction conditions: 50 mg catalyst, 1 mL benzene and 0.3 mL HBpin (0.002 mol) at 120 °C for 12 h. [b] Loading of La [mmol g−1] in catalytic materials measured by ICP‐OES. [c] Molar percentage [mol %] of La to HBpin. [d] 150 mg.

La(BH4) (THF) ‐SiO2

La(BH4) (THF) ‐Al2O3 [d]

La(BH4) (THF) ‐SiO2/Al2O3

La(BH4) (THF) ‐HY30

The HBpin starting material was the only borane species observed in attempted CH borylations of benzene using ≈0.2 mol % of either unsupported La(BH4)3(THF)3, La(BH4) (THF) ‐SiO2, or La(BH4) (THF) ‐Al2O3, as precatalysts, and phenyl pinacolborane (PhBpin) was not detected in the mixture. We infer that neither (≡SiO)3−La(BH4) (THF) ‐type sites grafted by reaction with surface OH in silica or alumina nor cationic [La(BH4) (THF) ]+ resulting from activation at Lewis acid sites (LAS) in γ‐Al2O3 are catalytically active for C−H borylation of benzene under these conditions. In contrast, HBpin was completely consumed in neat benzene and in the presence of La(BH4) (THF) ‐SiO2/Al2O3 or La(BH4) (THF) ‐HY30 as precatalysts, giving 1.1 % and 1.8 % yield of PhBpin, which corresponds to turnovers of 6 and 10 (moles PhBpin/mols La). Although the initial precatalyst formulations and reaction conditions give low yields and poor selectivity for PhBpin, the experiments importantly provided new insight that La(BH4) species bonded at BAS, the functionality that is common to SiO2/Al2O3 and zeolite HY30 but not SiO2 or γ‐Al2O3, are catalytically active for CH borylation of benzene.

We then identified the surface species on the bare La‐free supports capable of decomposing HBpin, by examining reactions of HBpin with the calcined oxides, to develop strategies for increasing benzene borylation yields by limiting the undesired pathway. SiO2 or γ‐Al2O3 gave constant HBpin concentration after heating with HBpin and benzene for 2 h at 120 °C (Table S1), ruling out silanol, aluminol, and LAS sites as catalysts for HBpin decomposition. In contrast, HBpin was quantitatively consumed within 2 h in the presence of BAS‐containing HY30 and SiO2/Al2O3, while PhBpin was not detected in these experiments. Considering these observations and the hypothesis that lanthanum species grafted on BAS are the active sites, strategies to improve catalytic performance could involve selective placement of lanthanum borohydrides at all bridging Si−OH−Al moieties. DRIFTS analysis (Figure S4), however, showed that nearly all silanols in the SiO2/Al2O3 and HY30 had been consumed during grafting, giving a high proportion of inactive La species.

We designed an approach for grafting La selectively at the BAS in HY30 to improve the performance of this catalyst. The Si−OH−Al BAS are primarily located in microporous cages (Figure S1) while silanols are mainly on the external surface after thermal treatment at 500 °C. This spatial segregation suggested that silanol capping with Ph3SiCl (TPSCl) could block unwanted grafting of La(BH4)3(THF)3 on the external surface. TPSCl is too large (≈10 Å diameter) to enter the 7.35 Å micropores and react with BAS in HY30. La(BH4) (THF) ‐HY30 also provided slightly higher turnovers, yield, and selectivity for PhBpin than the SiO2/Al2O3‐grafted catalyst, which cannot be selectively capped. In addition, selective grafting of La(BH4)3(THF)3 on the BAS adjacent to crystallographically equivalent Al‐based T4 positions in HY30 could provide a single‐site catalyst with structurally uniform sites.

Silanol capping with TPSCl affords the silylated zeolite TPS‐HY30 on a 2 g scale (see Supporting Information). The normalized intensity of the νSiOH signal at 3746 cm−1 in the DRIFTS of TPS‐HY30 was decreased by 54 % compared to that of the HY30 starting material (Figure 1), while BAS‐assigned signals at 3630 and 3566 cm−1 were constant in the two samples. In addition, the similar pattern of peaks from ca. 3100–2850 cm−1 associated with νCH in TPSCl (Figure S4) and in the modified zeolite indicate that organosilyl groups were grafted on HY30. These grafted Ph3 Si‐OSi surface species were unambiguously assigned to a signal at −20 ppm in the 29Si{1H} cross‐polarization magic angle spinning (CPMAS) SS NMR spectrum of TPS‐HY30 (Figure S5). Finally, TPS‐HY30 and HY30 mediated HBpin decomposition in comparable times (2 h), reinforcing the notion that BAS are similarly available inside the pores of TPS‐HY30 and HY30, as well as that capping primarily sequesters the external silanols.

Figure 1

DRIFTS spectra of HY30 (black, top), TPS‐HY30 (blue, middle) and La(BH4)2(THF)2.5‐TPS‐HY30 (red, bottom), normalized to the zeolite framework signal at 1860 cm−1.

Reaction of La(BH4)3(THF)3 and TPS‐HY30 affords La(BH4)2(THF)2.5‐TPS‐HY30 (composition established below). The 0.75 La wt % in the TPS‐capped zeolite is expectedly lower than 1.0 La wt % in unprotected La(BH4) (THF) ‐HY30. DRIFTS revealed new bands at ≈2500 and 2200–2300 cm−1 (Figure S4) assigned to terminal and bridging νB‐H from grafted borohydride species. Moreover, the normalized intensity of peaks associated with BAS diminished while the silanol signal retained similar intensity, suggesting that the La complex was selectively grafted at BAS and residual silanols are not accessible. 11B direct‐polarization (DP)MAS and multiple‐quantum (MQ)MAS SS NMR spectra (Figures 2 and S6, respectively) of HY30‐ and TPS‐HY30‐supported catalysts showed one broad signal at 5 ppm assigned to BO Hy and two signals at −15 and −22 ppm assigned to BH4 − groups in La(BH4) species.[ , ] BO H and BH4 − species are present in a ≈2 : 1 ratio. We also tentatively assign the 11B NMR signals at −22 and −15 ppm to silanol‐grafted ≡SiO−La(BH4)2 and BAS‐grafted ≡Si(≡Al)O−La(BH4)2 sites, respectively. The more shielded signal matches the chemical shift of the material obtained by grafting La(BH4)3(THF)3 on silica. The −15 ppm signal is assigned as ≡Si(≡Al)O−La(BH4)2 because this peak is present in the aluminum‐containing HY30‐grafted materials and not silica‐grafted materials. In addition, its intensity increases relative to the ≡SiO−La(BH4)2 signal after selective capping of silanols, which increased the proportion of BAS‐grafted sites.

Figure 2

11B DPMAS NMR spectra of a) La(BH4) (THF) ‐HY30 and b) La(BH4)2(THF)2.5‐TPS‐HY30 before (black) and after (red) reaction with BHT.

In addition, elemental analysis (EA) revealed a ca. 3.2 : 1 ratio of B : La in the TPS‐HY30‐supported catalyst (Table S2). Thus, NMR and EA data suggest that the B2H6 byproduct from borohydride protonolysis reacts with the zeolite. The final composition of the active site is assigned as La(BH4)2(THF)2.5‐TPS‐HY30 from these data and the loss of one THF per two La after grafting revealed by quantitative solution‐phase 1H NMR (Figure S3).

The spatial and site selectivity of La grafting was further established by a poisoning spectrochemical analysis (see Supporting Information). Reaction of La(BH4) (THF) ‐HY30 with 3,5‐di‐t‐butyl‐4‐hydroxytoluene (BHT) resulted in almost complete disappearance of [BH4]− signals in the 11B SS NMR spectrum (Figure 2a). In contrast, an experiment involving La(BH4)2(THF)2.5‐TPS‐HY30 under the same conditions gave a similar 11B SS NMR spectrum to that of the TPS‐decorated starting materials (Figure 2b). The ≈7.0×9.3 Å estimated molecular dimensions of BHT impede its easy diffusion through the HY30 micropores. These experiments are wholly consistent with La(BH4)3(THF)3 grafting primarily inside the micropores in the TPS‐capped zeolite. In contrast, most of the lanthanum borohydride grafts on the external surface of uncapped HY30, with a minority of active species grafted at BAS.

The La(BH4)2(THF)2.5‐TPS‐HY30‐catalyzed reaction of benzene and HBpin at 120 °C gives PhBpin in substantially improved yield and turnovers (7.4±0.3 % and 62±3, respectively). The four‐fold increase in PhBpin yield and six‐fold increase in turnovers validates our silanol‐capping approach, indicates that a larger fraction of BAS reacts with the La precursor after TPS capping and matches the 11B NMR data. This result also affirms the hypothesis that ≡Si(≡Al)O−La(BH4)2‐type sites lead to catalytically active benzene borylation.

A time‐resolved study of the La(BH4)2(THF)2.5‐TPS‐HY30‐catalyzed benzene borylation revealed a ≈6 h half‐life for HBpin and a linear increase in turnovers over that time (Figure 3). The much longer half‐life for HBpin during benzene borylation than during its TPS‐HY30‐mediated decomposition (<2 h) indicates that this modified catalyst contains few accessible BAS to mediate the unwanted pathway. The rate of PhBpin formation has zero‐order dependence in [HBpin]. Comparison of the slopes of turnovers vs time for borylation reactions of benzene and benzene‐d 6 revealed a kinetic isotope effect (k H/k D=2.9), which is in the range of metalations of sp 2 hybridized C−H bonds via σ‐bond metathesis steps.

Figure 3

Benzene borylation, catalyzed by La(BH4)2(THF)2.5‐TPS‐HY30, showing zero‐order concentration dependence for PhBpin production and a small, normal kinetic isotope effect for C−H/C−D bond activation.

Because the rate of HBpin decomposition depends on its concentration (Figure S7), reactions employing low [HBpin] could lead to increased yields and improved selectivity. This idea was tested through experiments in which HBpin was added in smaller portions (0.69 mol every 6 h) to a heated mixture of benzene and the La(BH4)2(THF)2.5‐TPS‐HY30 catalyst (Figure 4). Addition of HBpin in 3 portions (totaling 2.1 mmol) gave full conversion of HBpin, 101 turnovers, and 12 % yield of PhBpin after 18 h, which favorably compares to the 62 turnovers and 7.4 % yield from the experiment involving an addition of HBpin in a single portion. Clearly, the inequivalent [HBpin] dependences for benzene borylation and HBpin decomposition may be leveraged to access higher catalytic performance.

Figure 4

Portion‐wise addition of HBpin (0.10 mL, 0.69 mmol every 6 h) in benzene borylation.

Addition of 2.8 mmol of HBpin in four portions gave 143 turnovers, corresponding to a linear increase in product over 24 h. Smaller increases, however, were observed in subsequent additions of HBpin. For example, addition of 4.1 mmol divided over six portions gave the maximum TON of 167. Separation of the catalytic material from the PhBpin product and HBpin decomposition product after addition of four portions of HBpin, washing the material with pentane and drying, and performing additional catalytic experiments provide approximately equivalent conversion as addition of HBpin in situ. Thus, under these conditions, the catalyst begins to deactivate for PhBpin production after ≈140 turnovers. We suspect that the side reactions which decompose HBpin are also related to catalyst deactivation, since these conditions that lead to greater selectivity also provide a higher absolute TON. Thus, improved yields and TON are likely achievable by avoiding HBpin decomposition, giving comparable performance to iridium catalysts (Table S3).

In conclusion, the identification of BAS both as detrimental, leading to HBpin decomposition, and as essential for creating the reactive lanthanum sites for benzene borylation enables future systematic and rational investigations to design catalysts with greater selectivity and efficiency. In this context, the porous and crystalline zeolite framework is particularly enticing, giving specific loadings of BAS which are either accessible or inaccessible based on the size and shape of the reactant. We are currently investigating these catalysts for such molecular sieving properties in reactive separations and to access enhanced selectivity. In addition, the behavior of C−H borylation versus HBpin decomposition at lower HBpin concentrations, suggested by the observed rate law, further implies improved conversions will develop by influencing concentrations of HBpin and reactant in microporous environment of the catalytic sites.

Conflict of interest

The authors declare no conflict of interest.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

Click here for additional data file.