Enhancing the catalytic activity of hydronium ions through constrained environments

The dehydration of alcohols is involved in many organic conversions but has to overcome high free-energy barriers in water. Here we demonstrate that hydronium ions confined in the nanopores of zeolite HBEA catalyse aqueous phase dehydration of cyclohexanol at a rate significantly higher than hydronium ions in water. This rate enhancement is not related to a shift in mechanism; for both cases, the dehydration of cyclohexanol occurs via an E1 mechanism with the cleavage of Cβ–H bond being rate determining. The higher activity of hydronium ions in zeolites is caused by the enhanced association between the hydronium ion and the alcohol, as well as a higher intrinsic rate constant in the constrained environments compared with water. The higher rate constant is caused by a greater entropy of activation rather than a lower enthalpy of activation. These insights should allow us to understand and predict similar processes in confined spaces.


Supplementary Figure 2 | (a) Proportionality of reaction rate to the concentration of hydronium ion and total acid and (b) the reaction order with respect to the concentration of H 3 PO 4 .
The dehydration of cyclohexanol to cyclohexene was carried out at 160 °C in aqueous solutions containing 0.32 M (at r.t.) cyclohexanol and various concentrations of H 3 PO 4 (0.02−0.09 M at r.t.). Rates and concentrations in the plots are corrected for solution density at 160 °C. Reaction order with respect to total acid concentration is approximately 0.6 (b). This supports the claim that hydronium ions dissociated from the H 3 PO 4 are responsible for the catalytic reaction.  Table 11).

1000* (1/T) (K -1 )
Olefin formation E a =158 kJ/mol HBEA150 has particles with rounded corners and average diameters of ~200−300 nm. Two distinct bands of free OH groups were detected; the band at 3740 cm -1 is attributed to terminal and internal Si-OH groups (non-acidic), while the band at 3605 cm -1 is attributed to the Brønsted-acidic bridging hydroxyl groups associated with Al T sites.   Table 11).

Supplementary Figure 17 | A representative 31 P NMR spectrum of a H 3 PO 4 solution with
NaH 2 PO 2 as the internal standard.

Supplementary Note 1: Estimation of adsorption capacity under reaction conditions
At room temperature, the saturation uptake of cyclohexanol was determined to be 1.6 ± 0.1 mmol g HBEA -1 (Figure 1). For gas phase adsorption of cyclohexanol, of 2.2 ± 0.2 mmol g HBEA -1 was determined ( Supplementary Figure 10(a)). Assuming a liquid density (0.962 g cm -3 ) for the adsorbed cyclohexanol, an uptake of 2.2 mmol g HBEA -1 in the absence of water corresponds to 0.23 cm 3 g HBEA -1 , comparable to the micropore volume of the sample (Supplementary Table 2). In the presence of water, the saturation uptake of cyclohexanol corresponds to an occupied volume of 0.16 cm 3 g HBEA -1 . Subtracting this value from the micropore volume, and if solely attributing this difference (0.04 cm 3 g HBEA -1 ) to water in the pores, the adsorbed amount of water would be ~2 mmol g HBEA -1 .
The adsorption isotherms of cyclohexanol from aqueous solutions onto zeolite HBEA150 have been measured at various temperatures (7-80 °C). Langmuir-type adsorption model, as discussed in the main text, has been applied to fit these measured isotherms to obtain adsorption constant (K ads ) and saturation uptake (q max ) at each temperature. Detailed results will be reported in a subsequent publication. Important to this work is what we show below regarding the estimation of adsorption capacity under reaction conditions.
It was found that the saturation uptake decreased as adsorption temperature increased (Supplementary Table 5). This decrease in the saturation uptake with increasing adsorption temperature stems from the decrease in density of the adsorbate phase in the micropore (like thermal expansion of a liquid) as a function of temperature. The temperature dependence takes the form: where δ is the temperature coefficient of expansion.
Plotting measured/regressed saturation adsorption capacity at different temperatures as a function of temperature yielded a slope (-δ) of -0.0037 K -1 . Having extrapolating these experimentally determined q max and K ads to reaction temperatures using the same temperature dependence as determined between 7 and 80 °C, we found that the saturation uptake of cyclohexanol would decrease from 1.05 to 0.92 mmol g HBEA -1 at 160-200 °C. Assuming that the remaining micropore volume (total V micro = 0.20 cm 3 g -1 ) is filled by adsorbed water, the uptake of water in the pore would increase from 3.9 to 4.2 mmol g HBEA -1 (compared with 1.8 mmo g HBEA -1 at room temperature) with temperature increasing from 160 to 200 °C.

(Supplementary Figure 11) and calculations of the related parameters
An illustration of a reversible first-order dehydration reaction in aqueous phase is shown as follows: At reaction temperature, however, a significant amount of cyclohexene (B) is distributed into the gas phase. The distribution of B between gas and aqueous phases is defined by Henry's law. It is reasonable to neglect the portion of A that is in the gas phase due to its high boiling point relative to the reaction temperature. The reaction is assumed to occur only in the aqueous solution which contains the catalyst (hydronium ions). The total moles (n A + n B ) of A and B is constant in the reactor at different times. Therefore, n A,aq,t=0 + n B,g,t=0 + n B,aq,t=0 = n A,aq,t + n B,g,t + n B,aq,t (2) n A,aq,t=0 − n A,aq,t = n B,g,t − n B,g,t=0 + n B,aq,t − n B,aq,t=0 Using Henry's law that applies to the phase distribution for B between gas and solution, we obtain: The reaction occurs in the aqueous phase, such that the rate of cyclohexanol consumption is The integration of the above differential equation gives: aq be y (t) and t be x, rearrange Supplementary Equation (14) into The moles of A in the aqueous solution at equilibrium, n A,aq,t=∞ , almost equals to the total moles of B (n B,total,t=∞ ) in both gas (n B,g,t=∞ ) and solution (n B,aq,t=∞ ) subtracted from the initial moles of A (no reaction, n A,aq,ini ). Based on Supplementary Equation (4), together with the above consideration: constants. Overall, the reverse reaction occurs at a much slower rate at conversions < 10 % than the forward reaction. Thus the measured initial rate should primarily reflect the forward dehydration reaction.

Supplementary Note 3: Derivation of rate expression for cyclohexanol dehydration in aqueous phase
A proposed sequence of steps within an E1-type mechanistic framework (with the C-H bond cleavage being kinetically relevant) for aqueous phase dehydration of cyclohexanol catalyzed by H 3 PO 4 is shown below: Association of the alcohol with hydronium ion and the subsequent protonation is proposed to be sufficiently fast and quasi-equilibrated (a circle on top of a two-way arrow). The hydronium ion is represented as H + (H 2 O) 4 (aq), the association complex as H + (H 2 O) 3 ROH(aq), the olefin product as R(-H).
It has been demonstrated from isotope experiments (see main text) that the prevalent dehydration mechanism is of E1-type with the C β -H bond cleavage as the kinetically relevant step, for aqueous phase dehydration of cyclohexanol both in dilute H 3 PO 4 and in HBEA. A classical sequence of steps for homogeneous acid catalyzed dehydration is proposed above. A similar sequence should apply to HBEA-catalyzed dehydration in aqueous phase, yet with additional adsorption (from aqueous phase to intrazeolite voids where active sites reside) and desorption steps (from intrazeolite sites to aqueous/gas phases).
Next, we derive the kinetic expression for this mechanistic sequence. We use concentration terms instead of activities for solution species in dilute systems, assuming activity coefficients for the solution species are unity.
For the first step shown above , i.e., association of cyclohexanol with hydronium ion, letting the initial proton concentration be [H 3 O + ] 0 , we have: For the second step, proton transfer from water cluster to ROH, we have: Thus, we have: For the third step, C-O bond cleavage, applying steady-state assumption to the solvated carbenium-ion intermediate, , we have: The expression for TOF is (equal to that of the fourth step):

Supplementary Note 4: Mathematical approach for the determination of hydronium ion concentration, association equilibrium constant and intrinsic rate constant for H 3 PO 4catalyzed dehydration
Since H 3 PO 4 is a weak acid with incomplete dissociation of even its first proton (the other two are hardly dissociated) in water at all practical temperatures, the extent of H 3 PO 4 dissociation is affected by temperature, total acid concentration, as well as additional equilibria that involve (i.e., consume or produce) any of the species (e.g., H 3 PO 4 , hydronium ion, anions) that is present in the acid dissociation equilibrium: While solving K L,a directly from the above functional relations hips seems quite challenging, an alternative approach is: 1) give initial guess for K L,a and obtain θ L,a (recall that θ L,a =  Figure 14).
From equation (3) in the main text, the rate constants (k L,d ) were determined from TOF ratios ( Table 1 in the main text) and θ L,a (Supplementary Table 8). Then, the intrinsic activation enthalpy and entropy (reported in Table 4 in the main text) were determined from the Eyring plot of ln(k L,d /T) as a function of 1/T (Supplementary Figure 15).
For HBEA-catalyzed cyclohexanol dehydration, is likely close to 1, as a result of the [ROH]/[H 2 O] in the pore being 0.25 (5 and 20/u.c. for cyclohexanol and water, respective ly, at reaction conditions); almost every hydronium ion in the pore is associated with cyclohexanol. In this case, TOF z ≈ k z,d , and the ratio of k z,d /k L,d was determined to be 2.7 ± 0.2, indicating that the intrinsic rate constants for cyclohexanol dehydration in HBEA are substantially higher than in aqueous solution.

Supplementary Note 6: Dehydration experiments using mixtures of H 3 PO 4 and zeolite
The observed rate with a mixture of H 3 PO 4 and HBEA was higher than the sum of rates obtained with the individual acids (see Supplementary Table 4), presumably as a result of phosphoric acid being adsorbed in the pore. On the contrary, no increase in dehydration rate was observed using the mixture of siliceous BEA (Si-BEA) and H 3 PO 4 . We attribute this to the well-known high hydrophobicity of the all-siliceous BEA that prevents an appreciable amount of water and H 3 PO 4 from entering the pore.
To confirm our speculations, H 3 PO 4 uptake on HBEA150 and siliceous BEA (Si-BEA) was measured at 25 °C by 31 P NMR spectroscopy. We observed significant H 3 PO 4 adsorption on zeolite HBEA150 (56 μmol g -1 ) but found no measurable uptake by Si-BEA (Supplementary dissociation in the zeolite pore is not known, it is currently not possible to establish a quantitative relation between the increase in the number of acidic species in the pore and the activity enhancement with the combination of H 3 PO 4 and HBEA150.

Supplementary Note 7: Calculation of activation enthalpies and entropies based on transition state theory formalism
Transition state theory (TST) assumes that a hypothetical transition state (activated complex) exists between reactants and products during a chemical reaction and that a quasi-equilibrium is established between the reactant and the TS. According to the Erying equation, if the rate constant has been experimentally determined, the theory can be used to calculate the Gibbs free energy, activation enthalpy and entropy. The results are compiled in Table 4 of the main text. The approach is briefly summarized below: Rearrange the Supplementary Equation (31) into the logarithmic form: Thus, the enthalpy required (∆H ‡ ) and entropy gained/lost (∆S ‡ ) to reach the transition state complex can be determined using Eyring plots (ln(k/T)-(1/T)), see Supplementary Equation (32).