n-alkanes on MgO(100). I. Coverage-dependent desorption kinetics of n-butane

High-quality temperature-programmed desorption (TPD) measurements of n-butane from MgO(100) have been made for a large number of initial butane coverages (0-3.70 ML, ML-monolayers) and a wide range of heating ramp rates (0.3-10 K/s). We present a TPD analysis technique which allows the coverage-dependent desorption energy to be accurately determined by mathematical inversion of a TPD spectrum, assuming only that the preexponential factor (prefactor) is coverage independent. A variational method is used to determine the prefactor that minimizes the difference between a set of simulated TPD spectra and corresponding experimental data. The best fit for butane desorption from MgO is obtained with a prefactor of 10(15.7+/-1.6) s(-1). The desorption energy is 34.9+/-3.4 kJ/mol at 0.5-ML coverage, and varies with coverage approximately as Ed(theta)=34.5+0.566theta+8.37 exp(-theta/0.101). Simulations based on these results can accurately reproduce TPD experiments for submonolayer initial coverages over a wide range of heating ramp rates (0.3-10 K/s). Advantages and limitations of this method are discussed.     

The adsorption and desorption of ethanol ices from a model grain surface

Reflection absorption infrared spectroscopy (RAIRS) and temperature programed desorption (TPD) have been used to probe the adsorption and desorption of ethanol on highly ordered pyrolytic graphite (HOPG) at 98 K. RAIR spectra for ethanol show that it forms physisorbed multilayers on the surface at 98 K. Annealing multilayer ethanol ices (exposures >50 L) beyond 120 K gives rise to a change in morphology before crystallization within the ice occurs. TPD shows that ethanol adsorbs and desorbs molecularly on the HOPG surface and shows four different species in desorption. At low coverage, desorption of monolayer ethanol is observed and is described by first-order kinetics. With increasing coverage, a second TPD peak is observed at a lower temperature, which is assigned to an ethanol bilayer. When the coverage is further increased, a second multilayer, less strongly bound to the underlying ethanol ice film, is observed. This peak dominates the TPD spectra with increasing coverage and is characterized by fractional-order kinetics and a desorption energy of 56.3+/-1.7 kJ mol(-1). At exposures exceeding 50 L, formation of crystalline ethanol is also observed as a high temperature shoulder on the TPD spectrum at 160 K.     

Adsorption energies, inter-adsorbate interactions, and the two binding sites within monolayer benzene on Ag(111)

The adsorption of monolayer and multilayer benzene on the Ag(111) surface was characterized using temperature programmed desorption (TPD). TPD spectra revealed two broad peaks at approximately 205 and approximately 150 K at submonolayer coverage and a sharper, multilayer peak at 140 K. Analysis of the coverage-dependent shape and shift of the two submonolayer peaks has resulted in their assignment to desorption from two different binding geometries on threefold-hollow sites with symmetries C(3v)(sigma d) and C(3v)(sigma v). The TPD peak analysis incorporated inter-adsorbate repulsive interaction that resulted from the local dipole moment at the adsorption site induced by the adsorbate-surface charge transfer bonding. The analysis has yielded desorption energies of 54.9 +/- 0.8 and 50.4 +/- 0.4 kJ/mol for the C(3v)(sigma d) and C(3v)(sigma v) configurations, respectively. The interface dipole and polarizability of the benzene-silver complex have been determined to be 5.4 +/- 1.8 D and 14 +/- 10 A3, respectively. Repulsive interactions in the monolayer were found to lower the desorption energy from the zero-coverage value by 14.8 kJ/mol. Leading edge analysis of the multilayer peak yielded a desorption energy of 40.9 +/- 0.7 kJ/mol.     

Reactions of ammonia on stoichiometric and reduced TiO(2)(001) single crystal surfaces

The reaction of NH(3) on the surface of the 011-faceted structure of the TiO(2)(001) single crystal is studied and compared to that on the O-defected surface. Temperature-programmed desorption (TPD) conducted after NH(3) adsorption at 300 K shows only molecular desorption at 340 K. Modeling of TPD signals as a function of surface coverage indicated that the activation energy, E(d), and pre-exponential factor, v(eff), decrease with increasing coverage. Near zero surface coverage, E(d) was found to be equal to 92 kJ/mol and v(eff) to be close to 10(13) /s. Both parameters decreased to approximately 52 kJ/mol and approximately 10(7) /s at saturation coverage. The decrease is due to a repulsive interaction of adsorbed NH(3) molecules on the surface. Computing of the TPD results show that saturation is obtained at 1/2 monolayer coverage (referred to Ti atoms). Both the amount and shape of NH(3) peak change on the reduced (Ar(+)-sputtered) surfaces. The desorption peak at 340 K is considerably attenuated on mildly reduced surfaces (TiO( approximately )(1.9)) and has totally disappeared on the heavily reduced surfaces (TiO(1.6)(-)(1.7)), where the main desorption peak is found at 440 K. This 440-K desorption is most likely due to NH(x) + H recombination resulting from ammonia dissociation upon adsorption on Ti atoms in low oxidation states.     

Benchmarking the accuracy of coverage-dependent models: adsorption and desorption of benzene on Pt (1 1 1) and Pt3Sn (1 1 1) from first principles

Bimetallic catalysts have demonstrated properties favorable for upgrading biofuel through catalytic hydrodeoxygenation. However, the design and optimization of such bimetallic catalysts requires the ability to construct accurate, predictive models of these systems. To generate a model that predicts the kinetic behavior of benzene adsorbed on Pt (1 1 1) and a Pt₃Sn (1 1 1) surface alloy (Pt₃Sn (1 1 1)), the adsorption of benzene was studied for a wide range of benzene coverages on both surfaces using density functional theory (DFT) calculations. The adsorption energy of benzene was found to correlate linearly with benzene coverage on Pt (1 1 1) and Pt₃Sn (1 1 1); both surfaces exhibited net repulsive lateral interactions. Through an analysis of the d-band properties of the metal surface, it was determined that the coverage dependence is a consequence of the electronic interactions between benzene and the surface. The linear coverage dependence of the adsorption energy allowed us to quantify the influence of the lateral interactions on the heat of adsorption and temperature programmed desorption (TPD) spectra using a mean-field model. A comparison of our simulated TPD to experiment showed that this mean-field model adequately reproduces the desorption behavior of benzene on Pt (1 1 1) and Pt₃Sn (1 1 1). In particular, the TPD correctly exhibits a broadening desorption peak as the initial coverage of benzene increases on Pt (1 1 1) and a low temperature desorption peak on Pt₃Sn (1 1 1). However, due to the sensitivity of the TPD peak temperature to the desorption energy, precise alignment of experimental and theoretical TPD spectra demands an accurate calculation of the adsorption energy. Therefore, an analysis of the effect of the exchange-correlation functional on TPD modeling is presented. Through this work, we show the necessity of incorporating lateral interactions into theoretical models in order to correctly predict experimental behavior.     

Energetic mapping of Ni catalysts by detailed kinetic modeling

Temperature-programmed desorption (TPD) of CO has been performed on supported and unsupported nickel catalysts. The unsupported Ni catalyst consists of a Ni(14 13 13) single crystal which has been studied under ultrahigh vacuum conditions. The desorption energy for CO at low CO surface coverage was found to be 119 kJ/mol, and the binding energy of C to the Ni(111) surface of the crystal was 703 kJ/mol. The supported catalysts consist of nickel supported on hydrotalcite-like compounds with three different Mg2+/Al3+ ratios. The experimental results show that for the supported Ni catalysts TPD of CO leads to desorption of both CO and CO2, with the latter being dominant. Dissociation of CO takes place, and considerable amounts of residue C are left on the surface. The residue C is removed by temperature-programmed oxidation (TPO). The results show that a low Mg2+/Al3+ ratio in the hydrotalcite precursor seems to result in more steplike sites, kinks, and defects for carbon monoxide dissociation. A detailed kinetic modeling of the TPO results based on elementary reaction steps has been conducted to give an energetic map of supported Ni catalysts. Experimental results from the ideal Ni surface fit nicely with literature values, providing useful information for identifying active sites on supported Ni catalysts.     

Surface chemistry of CN bond formation from carbon and nitrogen atoms on Pt(111)

The mechanism of CN bond formation from CH3 and NH3 fragments adsorbed on Pt(111) was investigated with reflection absorption infrared spectroscopy (RAIRS), temperature-programmed desorption (TPD), and X-ray photoelectron spectroscopy (XPS). The surface chemistry of carbon-nitrogen coupling is of fundamental importance to catalytic processes such as the industrial-scale synthesis of HCN from CH4 and NH3 over Pt. Since neither CH4 nor NH3 thermally dissociate on Pt(111) under ultrahigh vacuum (UHV) conditions, the relevant surface intermediates were generated through the thermal decomposition of CH3I and the electron-induced dissociation of NH3. The presence of surface CN is detected with TPD through HCN desorption as well as with RAIRS through the appearance of the vibrational features characteristic of the aminocarbyne (CNH2) species, which is formed upon hydrogenation of surface CN at 300 K. The RAIRS results show that HCN desorption at approximately 500 K is kinetically limited by the formation of the CN bond at this temperature. High coverages of Cads suppress CN formation, but the results are not influenced by the coadsorbed I atoms. Cyanide formation is also observed from the reaction of adsorbed N atoms and carbon produced from the dissociation of ethylene.     

Identification and hydrogenation of C2 on Pt(111)

Reflection absorption infrared spectroscopy (RAIRS) and temperature-programmed desorption (TPD) were used to identify the molecular species formed upon the reaction of hydrogen with surface carbon that is deposited by exposing acetylene to a Pt(111) surface held at 750 K. At this temperature, the acetylene is completely dehydrogenated and all hydrogen is desorbed from the surface. Upon subsequent hydrogen exposure at 85 K followed by sequential annealing to higher temperatures, ethylidyne (CCH3), ethynyl (CCH), and methylidyne (CH) are formed. The observation of these species indicates that carbon atoms and C2 molecules exist as stable species on the surface over a wide range of temperatures. Through a combination of RAIRS intensities, hydrogen TPD peak areas, and Auger electron spectroscopy, quantitative estimates of the coverages of the various species were obtained. It was found that 79% of the acetylene-derived carbon was in the form of C2 molecules, with the remainder in the form of carbon atoms. Essentially all of the acetylene-derived carbon could be hydrogenated. In contrast, 85% of an equivalent coverage of carbon deposited by ethylene exposure at 750 K was found to be inert toward hydrogenation.     

Isotope tracer study of hydrogen spillover on carbon-based adsorbents for hydrogen storage

A composite material comprising platinum nanoparticles supported on molecular sieve templated carbon was synthesized and found to adsorb 1.35 wt % hydrogen at 298 K and 100 atm. The isosteric heat of adsorption for the material at low coverage was approximately 14 kJ/mol, and it approached a value of 10.6 kJ/mol as coverage increased for pressures at and above 1 atm. The increase in capacity is attributed to spillover, which is observed with the use of isotopic tracer TPD. IRMOF-8 bridged to Pt/C, a material known to exhibit hydrogen spillover at room temperature, was also studied with the hydrogen-deuterium scrambling reaction for comparison. The isotherms were reversible. For desorption, sequential doses of H2 and D2 at room temperature and subsequent TPD yield product distributions that are strong indicators of the surface diffusion controlled reverse spillover process.     

Physisorption of N2, O2, and CO on fully oxidized TiO2(110)

Physisorption of N(2), O(2), and CO was studied on fully oxidized TiO(2)(110) using beam reflection and temperature-programmed desorption (TPD) techniques. Sticking coefficients for all three molecules are nearly equal (0.75 +/- 0.05) and approximately independent of coverage suggesting that adsorption occurs via a precursor-mediated mechanism. Excluding multilayer coverages, the TPD spectra for all three adsorbates exhibit three distinct coverage regimes that can be interpreted in accord with previous theoretical studies of N(2) adsorption. At low coverages (0-0.5 N(2)/Ti(4+)), N(2) molecules bind head-on to five-coordinated Ti(4+) ions. The adsorption occurs preferentially on the Ti(4+) sites that do not have neighboring adsorbates. This arrangement minimizes the repulsive interactions between the adsorbed molecules along the Ti(4+) rows resulting in a relatively small shift of the TPD peak (105 --> 90 K) with increasing coverage. At higher N(2) coverages (0.5-1.0 N(2)/Ti(4+)) the nearest-neighbor Ti(4+) sites become occupied. The close proximity of the adsorbates results in strong repulsion thus giving rise to a significant shift of the TPD leading edges (90 --> 45 K) with increasing coverage. For N(2)/Ti(4+) > 1, an additional low-temperature peak (approximately 43 K) is present and is ascribed to N(2) adsorption on bridge-bonded oxygen rows. The results for O(2) and CO are qualitatively similar. The repulsive adsorbate-adsorbate interactions are largest for CO, most likely due to alignment of CO dipole moments. The coverage-dependent binding energies of O(2), N(2), and CO are determined by inverting TPD profiles.     

Mechanism of the decomposition of the surface oxide film on polycrystalline palladium

The decomposition of thin surface oxide films on polycrystalline palladium Pd(poly) at 500–1300 K was investigated by mathematical modeling. This process was analyzed in terms of a model including O₂ desorption from the chemisorbed oxygen layer (O_ads) and the passage of oxygen inserted under the surface layer of the metal (O_abs) and oxygen dissolved in metal subsurface layers (O_dis) to the surface. O₂ desorption was modeled on a surface with a square lattice of adsorption sites, with account taken of the energy of the lateral repulsive interactions between adjacent O_ads atoms (ε_aa). At ε_aa = 10 kJ/mol and when the activation energy of O₂ desorption for a chemisorbed-oxygen surface coverage of θ ≈ 0 is E_des⁰ = 230 kJ/mol, the calculated spectra are in agreement with the oxygen temperature-programmed desorption (TPD) spectra obtained for Pd(poly) at θ ≤ 0.5. The passage of O_abs and O_dis atoms to the surface was calculated using a first-order equation, with account taken of the activation energy for these atoms coming out to the surface (E₂ and E₃, respectively). As the oxide film is heated, O₂ desorption is accompanied by the passage of O_abs and then O_dis to the surface, which leads to an increase in the O_ads surface coverage and, accordingly, to a buildup of lateral surroundings in the adsorbed layer. Owing to this fact and to the repulsive interactions between O_ads atoms, the bonds between O_ads and the surface weaken and E_des decreases. As a consequence, the O₂ desorption rate increases and a low-temperature peak with T_max ≈ 710 K, which is due to the passage of O_abs atoms to the surface, and then a high-temperature peak with T_max ≈ 770 K, which is due to the passage of O_dis atoms to the surface, appear in the TPD spectrum. At ε_aa = 10 kJ/mol, E_des⁰ = 230 kJ/mol, E₂ = 145 kJ/mol, and E₃ = 160 kJ/mol and when the number of inserted oxygen monolayers is θ_abs ≤ 0.3 and the number of oxygen monolayers dissolved in subsurface layers is θ_dis ≤ 10, the TPD spectra calculated for the given model are in agreement with the O₂ TPD spectra that are observed for Pd(poly) and are due to the decomposition of surface oxide films.     

Acetone and water on TiO2(110): competition for sites

The competitive interaction between acetone and water for surface sites on TiO2(110) was examined using temperature programmed desorption (TPD). Two surface pretreatment methods were employed, one involving vacuum reduction of the surface by annealing at 850 K in ultrahigh vacuum (UHV) and another involving surface oxidation with molecular oxygen. In the former case, the surface possessed about 7% oxygen vacancy sites, and in the latter, reactive oxygen species (adatoms and molecules) were deposited on the surface as a result of oxidative filling of vacancy sites. On the 7% oxygen vacancy surface, excess water displaced all but about 20% of a saturated d6-acetone first layer to physisorbed desorption states, whereas about 40% of the first layer d6-acetone was stabilized on the oxidized surface against displacement by water through a reaction between oxygen and d6-acetone. The displacement of acetone on both surfaces is explained in terms of the relative desorption energies of each molecule on the clean surface and the role of intermolecular repulsions in shifting the respective desorption features to lower temperatures with increasing coverage. Although first layer water desorbs from TiO2(110) at slightly lower temperature (275 K) than submonolayer coverages of d6-acetone (340 K), intermolecular repulsions between d6-acetone molecules shift its leading edge for desorption to 170 K as the first layer is saturated. In contrast, the desorption leading edge for first layer water (with or without coadsorbed d6-acetone) shifted to no lower than 210 K as a function of increasing coverage. This small difference in the onsets for d6-acetone and water desorption resulted in the majority of d6-acetone being compressed into islands by water and displaced from the first layer at a lower temperature than that observed in the absence of coadsorbed water. On the oxidized surface, the species resulting from reaction of d6-acetone and oxygen was not influence by increasing water coverages. This species was stable up to 375 K (well past the first layer water TPD feature) where it decomposed mostly back to d6-acetone and atomic oxygen. These results are discussed in terms of the influence of water in inhibiting acetone photo-oxidation on TiO2 surfaces.     

Dewetting growth of crystalline water ice on a hydrogen saturated Rh(111) surface at 135 K

We investigated the water (D(2)O) adsorption at 135?K on a hydrogen pre-adsorbed Rh(111) surface using temperature programmed desorption and infrared reflection absorption spectroscopy (IRAS) in ultrahigh vacuum. With increasing the hydrogen coverage, the desorption temperature of water decreases. At the saturation coverage of hydrogen, dewetting growth of water ice was observed: large three-dimensional ice grains are formed. The activation energy of water desorption from the hydrogen-saturated Rh(111) surface is estimated to be 51 kJ/mol. The initial sticking probability of water decreases from 0.46 on the clean surface to 0.35 on the hydrogen-saturated surface. In IRAS measurements, D-down species were not observed on the hydrogen saturated surface. The present experimental results clearly show that a hydrophilic Rh(111) clean surface changes into a hydrophobic surface as a result of hydrogen adsorption.     

1-Phenyl-1-propyne on Cu(111): TOFMS TPD, XPS, UPS, and 2PPE Studies

The bonding properties of 1-phenyl-1-propyne (PP, C6H5CCCH3) on Cu(111) at 100 K have been studied using temperature-programmed desorption (TPD), and X-ray, ultraviolet, and two-photon photoemission spectroscopies (XPS, UPS, and 2PPE). In TPD, there is no evidence for dissociation. Multilayer desorption occurs at 187 K, and monolayer desorption occurs at 320 (83.5 kJ/mol) and 390 K (102.4 kJ/mol), with the latter dominating. Based on the calibrated C(1s) XPS, the saturation monolayer coverage is one PP per four surface Cu atoms. The broad and asymmetric C(1s) intensity profile of the monolayer can be resolved into three symmetric components, with peaks at 283.6, 284.5, and 285.2 eV and intensities of 2:6:1, respectively. These are attributed, respectively, to acetylenic carbons bound to Cu, phenyl, and methyl carbons. The monolayer valence band ultraviolet photoemission spectrum profile contains four resonances attributable to PP perturbed by interactions with the Cu(111) substrate. With the exception of the highest occupied molecular orbital (HOMO) that is shifted by 0.4 eV, these are uniformly shifted by 1 eV further from the Fermi level for the multilayer. Calculated electron density plots of the occupied orbitals coupled with UPS profiles suggest a spectator role for the phenyl group and bonding to Cu via the acetylenic carbons. The adsorption of 1.0 monolayer (ML) of PP on Cu(111) lowers the work function by 0.85 eV. Using 2PPE, two unoccupied orbitals were identified at 1.0 (U1*-LUMO) and 0.6 eV (U2*-image state) below the vacuum level. A chemisorption model consistent with these spectroscopic results and the major chemisorption peak in TPD involve di-sigma-bonding of the acetylenic carbons to a pair of second-nearest neighbor surface Cu atoms (cross-bridge).     

Microkinetic modeling of CO TPD spectra using coverage dependent microcalorimetric heats of adsorption

CO adsorption on the ternary methanol synthesis Cu/ZnO/Al2O3 catalyst was studied in detail by means of adsorption microcalorimetry and flow temperature-programmed desorption (TPD). Based on these experimental data, we established a microkinetic analysis method, which provides information about the adsorption kinetics of CO on the catalyst surface. Experimentally derived microcalorimetric heats of adsorption were applied in a microkinetic model to simulate TPD curves with varying initial coverage. Two approaches were used: an integral approach based on evaluation of the integral heats of adsorption which predicts the experimental TPD curves roughly and provides first approximations for the preexponential factors. The second, more detailed approach was based on the simulation of the adsorption isotherm taking the experimentally determined coverage-dependence of the heat of adsorption into account. This approach led to a significantly improved agreement between experimental and simulated TPD curves. Moreover, it was possible to derive the standard entropy of adsorption. The general applicability of our approaches is demonstrated by analyzing the CO TPD and microcalorimetry data obtained with a binary ZnO-free Cu/Al2O3 catalyst.     

Probing the interaction of hydrogen chloride with low-temperature water ice surfaces using thermal and electron-stimulated desorption

The interaction and autoionization of HCl on low-temperature (80-140 K) water ice surfaces has been studied using low-energy (5-250 eV) electron-stimulated desorption (ESD) and temperature programmed desorption (TPD). There is a reduction of H(+) and H(2)(+) and a concomitant increase in H(+)(H(2)O)(n=1-7) ESD yields due to the presence of submonolayer quantities of HCl. These changes are consistent with HCl induced reduction of dangling bonds required for H(+) and H(2)(+) ESD and increased hole localization necessary for H(+)(H(2)O)(n=1-7) ESD. For low coverages, this can involve nonactivated autoionization of HCl, even at temperatures as low as 80 K; well below those typical of polar stratospheric cloud particles. The uptake and autoionization of HCl is supported by TPD studies which show that for HCl doses ≤0.5 ± 0.2 ML (ML = monolayer) at 110 K, desorption of HCl begins at 115 K and peaks at 180 K. The former is associated with adsorption of a small amount of molecular HCl and is strongly dependent on the annealing history of the ice. The latter peak at 180 K is commensurate with desorption of HCl via recombinative desorption of solvated separated ion pairs. The activation energy for second-order desorption of HCl initially in the ionized state is 43 ± 2 kJ/mol. This is close to the zero-order activation energy for ice desorption.     

A DFT study of the adsorption and dissociation of CO on Fe(100): influence of surface coverage on the nature of accessible adsorption states.

In the present article, we report adsorption energies, structures, and vibrational frequencies of CO on Fe(100) for several adsorption states and at three surface coverages. We have performed a full analysis of the vibrational frequencies of CO, thus determining what structures are stable adsorption states and characterizing the transition-state structure for CO dissociation. We have calculated the activation energy of dissociation of CO at 0.25 ML (ML = monolayers) as well as at 0.5 ML; we have studied the dissociation at 0.5 ML to quantify the destabilization effect on the CO(alpha3) molecules when a neighboring CO molecule dissociates. In addition, it is shown that the number and nature of likely adsorption states is coverage dependent. Evidence is presented that shows that the CO molecule adsorbs on Fe(100) at fourfold hollow sites with the molecular axis tilted away from the surface normal by 51.0 degrees. The asorprton energy of the CO molecule is -2.54 eV and the C-O stretching frequency is 1156 cm(-1). This adsorption state corresponds to the alpha3 molecular desorption state reported in temperature programmed desorption (TPD) experiments. However, the activation energy of dissociation of CO(alpha3) molecules at 0.25 ML is only 1.11 eV (approximately 25.60 kcal mol(-1)) and the gain in energy is -1.17 eV; thus, the dissociation of CO is largely favored at low coverages. The activation energy of dissociation of CO at 0.5 ML is 1.18 eV (approximately 27.21 kcal mol(-1)), very similar to that calculated at 0.25 ML. However, the dissociation reaction at 0.5 ML is slightly endothermic, with a total change in energy of 0.10 eV Consequently, molecular adsorption is stabilized with respect to CO dissociation when the CO coverage is increased from 0.25 to 0.5 ML.     

Heat of adsorption of naphthalene on Pt(111) measured by adsorption calorimetry

The heat of adsorption of naphthalene on Pt(111) at 300 K was measured with single-crystal adsorption calorimetry. The heat of adsorption on the ideal, defect-free surface is estimated to be (300 - 34 - 199(2)) kJ/mol. From this, a C-Pt bond energy for aromatic hydrocarbons on Pt(111) of approximately 30 kJ/mol is estimated, consistent with earlier results for benzene on Pt(111). There is higher heat of adsorption at very low coverage, attributed to step sites where the adsorption heat is >/=330 kJ/mol. Saturation coverage, = 1 ML, corresponds to 1.55 x 10(14) molecules/cm(2). Sticking probability measurements of naphthalene on Pt(111) give a high initial value of 1.0 and a Kisliuk-type coverage dependence that implies precursor-mediated sticking. The ratio of the hopping rate to the desorption rate of this precursor is approximately 51. Naphthalene adsorbs transiently on top of chemisorbed naphthalene molecules with a heat of adsorption of 83-87 kJ/mol.