(167c) Mechanism of C-C Hydrogenolysis On Ir Catalysts

Mechanism of C-C Hydrogenolysis on Ir Catalysts

David D. Hibbitts¹, David W. Flaherty^1,2
and Enrique Iglesia^1*

¹Department
of Chemical and Biomolecular Engineering, University of California at Berkeley,
Berkeley, CA 94720, United States

²Department
of Chemical and Biomolecular Engineering, University of Illinois at
Urbana-Champaign, Urbana, IL 61801, United States

*iglesia@berkeley.edu

Catalytic hydrogenolysis of alkanes has
been widely studied due to the necessity of selectively removing sulfur and
oxygen through C-S and C-O hydrogenolysis as well as optimizing chain length of
hydrocarbon mixtures through C-C hydrogenolysis. Hydrogenolysis is considered a
structure-sensitive reaction in which the composition of the metal and the
particle size have large effects on the turnover rates and activation energies
of the reaction. Substituted alkanes contain distinguishable C-C bonds, which
have distinct mechanisms that result in different dependencies on H₂
pressure and temperature, allowing for limited control over product
selectivity. Kinetic analysis and density functional theory (DFT) of
hydrogenolysis on well-defined catalyst particles are used to determine the
mechanisms for C-C bond activation and how they differ with the substitution of
the involved C-atoms. Ethane, the simplest alkane with a C-C bond, is
investigated here to determine general trends in hydrogenolysis on Ir.

Ethane hydrogenolysis was examined on 3.0
wt% Ir/SiO₂ catalysts with mean particle diameters of 7.1 nm at 593
K at high H₂:C₂H₆ ratios (0.6-1.8 MPa H₂,
10-80 kPa C₂H₆). Atomic hydrogen (H*) is the most
abundant surface intermediate (MASI) at these high H₂:C₂H₆
ratios and the rate of hydrogenolysis follows:

                                                          (1)

where (H₂) and (C₂H₆)
are partial pressures, α is the product of rate and equilibrium constants
and λ was found to be 3.3 ± 0.2. Chemical interpretation of this rate
expression suggests that the kinetically-relevant transition state is derived
from C₂H₆ and λ represents rate-averaged number of H₂
molecules evolved from dehydrogenation of C₂H₆ and from
desorption of H* to accommodate the kinetically-relevant transition state on
the catalyst surface.

            Previous studies have proposed that
dehydrogenation of C₂H₆ is quasi-equilibrated on Ir
surfaces and that the kinetically-relevant step for hydrogenolysis is the C-C
bond cleavage, followed by quasi-equilibrated hydrogenation to form CH₄,
Scheme 1.

H₂ + 2*  2H*                                                                   (1.1)

C₂H₆ + *  C₂H₆*                                                              (1.2)

C₂H₆* + (6-x)*  C₂H_6-x*
+ xH*                                       (1.3)

C₂H_x*  CH_y*
+ CH_z*                                                       (1.4)

CH_x* + (4-x)H*
 CH₄                                                                        (1.5)

For Scheme 1, a L-H expression can be derived which
matches the functional form of Eq. 1:

                 (2)

where k_CC,j is the rate constant for
the kinetically-relevant C-C bond cleavage reaction, K_j is
the equilibrium constant for a quasi-equilibrated C-H activation and j
represents the number of H atoms removed from the ethane. Taking into account
isomers of C₂H₄, C₂H₃ and C₂H₂,
there are ten possible dehydrogenated C₂ intermediates each with
their own kinetically-relevant C-C bond activation, representing ten distinct
possible mechanisms of the form shown in Scheme 1.

Periodic, plane-wave based DFT calculations
were modeled on a four-layer Ir(111) slab using VASP with RPBE gradient
corrections and PAW potentials. Reaction energetics confirmed that all
dehydrogenation reactions were quasi-equilibrated from C₂H₆
to *C₂H₂*; further dehydrogenations to form *CHC* and *CC*,
however, were not quasi-equilibrated. As such, there exists an equilibrium
between C₂H₆ in the gas phase and eight C₂H_x*
species, each of which can undergo a kinetically-relevant C-C bond activation with
a rate exponentially proportional to the Gibbs free energy difference between
the transition state with λ H₂ molecules in the gas phase and a
H*-covered surface with C₂H₆ in the gas phase:

                                  (3)

DH҂  and DS҂  were determined by measuring the
dependence of the rate on temperature and are 213 ± 2 kJ mol^-1 and 134
± 5 J mol^-1 K^-1, respectively at 20 kPa C₂H₆
and 1.8 MPa H₂.

In order to compute Gibbs free energy
differences through DFT, vibrational frequency calculations were performed to
determine zero-point vibrational energy corrections (ZPVE) as well as
vibrational enthalpy and entropy of adsorbed species. For gas-phase species,
additional rotational and translational enthalpy and entropy terms were
computed from statistical mechanics. Results from theory suggest that C-C bond
rupture occurs in *CHCH* with rates that would be 10⁷ times greater
than rates for any other potential intermediates. The predicted values of DH҂
and DS҂ are 218 kJ mol^-1 and 126 J mol^-1
K^-1, respectively, which are in excellent agreement with values from
experiments.

These results on C₂H₆
hydrogenolysis with near-quantitative matching between experiment and theory
establish a set of principles which will guide the investigation of larger n-alkanes
and branched alkanes to determine how substitution across the C-C bond affects
the mechanism for hydrogenolysis, ultimately affecting rates and selectivity.