Leed and thermal desorption studies of small molecules (H2, O2, CO,CO2, NO,C2H4, C2H2 AND C) chemisorbed on the rhodium (111) and (100) surfaces

The chemisorption of H₂, O₂, CO, CO₂, NO, C₂H₄, C₂H₂ and C has been studied on the clean Rh(111) and (100) surfaces. LEED, AES and thermal desorption were used to determine the surface structures, disordering and desorption temperatures, displacement and decomposition characteristics for each species. All of the molecules studied readily chemisorbed on both surfaces. A large variety of ordered structures was observed, especially on the (111) surface. The disordering temperatures of most ordered surface structures on the (111) surface were below 100°C. It was necessary to adsorb the gases at 25° C or below in order to obtain well-ordered surface structures. Chemisorbed oxygen was readily removed from the surface by H₂ or CO gas at crystal temperatures above 50°C. CO₂ appears to dissociate to CO upon adsorption on both rhodium surfaces as indicated by the identical ordering and desorption characteristics of these two molecules. C₂H₄ and C₂H₂ also had very similar ordering and desorption characteristics and it is likely that the adsorbed species formed by both molecules is the same. Decomposition of ethylene produced a sequence of ordered carbon surface structures on the (111) face as a result of a bulk-surface carbon equilibrium. The chemisorption properties of rhodium appear to be generally similar to those of iridium, nickel and palladium.     

Interaction of O2, CO2, CO,C2H4 and C2H4O with Ag(110)

The interaction of O₂, CO₂, CO, C₂H₄ AND C₂H₄O with Ag(110) has been studied by low energy electron diffraction (LEED), temperature programmed desorption (TPD) and electron energy loss spectroscopy (EELS). For adsorbed oxygen the EELS and TPD signals are measured as a function of coverage (θ). Up to θ = 0.25 the EELS signal is proportional to coverage; above 0.25 evidence is found for dipole-dipole interaction as the EELS signal is no longer proportional to coverage. The TPD signal is not directly proportional to the oxygen coverage, which is explained by diffusion of part of the adsorbed oxygen into the bulk. Oxygen has been adsorbed both at pressures of less than 10^-4 Pa in an ultrahigh vacuum chamber and at pressures up to 10³ Pa in a preparation chamber. After desorption at 10³ Pa a new type of weakly bound subsurface oxygen is identified, which can be transferred to the surface by heating the crystal to 470 K. CO₂ is not adsorbed as such on clean silver at 300 K. However, it is adsorbed in the form of a carbonate ion if the surface is first exposed to oxygen. If the crystal is heated this complex decomposes into O_ad and CO₂ with an activation energy of 27 kcal/mol(1 kcal = 4.187 kJ). Up to an oxygen coverage of 0.25 one CO₂ molecule is adsorbed per two oxygen atoms on the surface. At higher oxygen coverages the amount of CO₂ adsorbed becomes smaller. CO readily reacts with O_ad at room temperature to form CO₂. This reaction has been used to measure the number of O atoms present on the surface at 300 K relative to the amount of CO₂ that is adsorbed at 300 K by the formation of a carbonate ion. Weakly bound subsurface oxygen does not react with CO at 300 K. Adsorption of C₂H₄O at 110 K is promoted by the presence of atomic oxygen. The activation energy for desorption of C₂H₄O from clean silver is ～ 9 kcal/mol, whereas on the oxygen-precovered surface two states are found with activation energies of 8.5 and 12.5 kcal/mol. The results are discussed in terms of the mechanism of ethylene epoxidation over unpromoted and unmoderated silver.     

Nitric oxide reduction by CO ON Rh(111): Temperature programmed reaction

The reaction of NO with CO on Rh(111) has been studied with temperature programmed reaction (TPR). Comparisons are made with the reaction of O₂ with CO and the reaction of NO with H₂. The rate-determining step for both CO oxidation reactions is CO(a) + O(a) → CO₂(g). Repulsive interactions between adsorbed CO and adsorbed nitrogen atoms lead to desorption of CO in a peak at 415 K which is in the temperature range where the reaction between CO(a) and O(a) produces CO₂(g). Thus the extent of reaction of CO(a) with NO(a) is less than that between CO(a) and O(a) due to the lower coverage of CO caused by adsorbed N atoms and NO. A similar repulsive interaction between NO(a) and H(a) suppresses the NO + H₂ reaction. CO + NO reaction behavior on Rh(111) is compared to that observed on Pt(111).     

A molecular-beam investigation of the reaction H2 + 12O2 → H2O on Pd(111)

Using molecular-beam relaxation techniques and isotopic exchange experiments, the water-formation reaction on Pd(111) has been shown to proceed via a Langmuir-Hinshelwood mechanism. The reaction product H₂O is emitted from the surface with a cosine distribution. The rate-determining step is the formation of OH_ad in the reaction O_ad + H_ad → OH_ad. The activation energy for this step is 7 kcal/mole with a pre-exponential factor, v, of 4 × 10^?8 cm² atom^?1 sec^?1. This value for v lies well below that observed for simple second-order desorption of dissociatively adsorbed diatomic gases, but is roughly of the order of that obtained for the oxidation of CO on Pd(111). The formation of H₂O proceeds differently under conditions of excess O₂ or H₂. In an excess of H₂, the kinetics is dominated by the transport of atomic hydrogen between the bulk and the surface as was found for the H?D exchange reaction on Pd(111). In an excess of O₂, diffusion of hydrogen into the bulk is blocked by adsorbed oxygen and the hydrogen reservoir available for reaction at the surface is decreased by several orders of magnitude. This results in a drastic reduction of the reaction rate which can be reversed by increasing the partial pressure of H₂.     

Adsorption and coadsorption of carbon monoxide and hydrogen on Pd(111)

The adsorption and coadsorption of CO and H₂ have been studied by means of thermal desorption (TD) and electron stimulated desorption (ESD) at temperatures ranging from 250 to 400 K. Three CO TD states, labelled as β₂, β₁, and β₀ were detected after adsorption at 250 K. The population of β₂ and β₁ states which are the only ones observed upon adsorption at temperatures higher than 300 K was found to depend on adsorption temperature. The correlation between the binding states in the TD spectra and the ESD O⁺ and CO⁺ ions observed was discussed. Hydrogen is dissociatively adsorbed on Pd(111) and no ESD H⁺ signal was recorded following H₂ adsorption on a clean Pd surface. The presence of CO was found to cause an appearance of a H⁺ ESD signal, a decrease of hydrogen surface population and an arisement of a broad H₂ TD peak at about 450 K. An apparent influence of hydrogen on CO adsorption was detected at high hydrogen precoverages alone, leading to a decrease in the CO sticking coefficient and the relative population of CO β₂ state. The coadsorption results were interpreted assuming mutual interaction between CO and H at low and medium CO coverages, the “cooperative” species being responsible for the H⁺ ESD signal. Besides, the presence of CO was proved to favour hydrogen penetration into the bulk even at high CO coverage when H atoms were completely displaced from the surface.     

The adsorption of CO,H2, H2, CO2 and H2O on carburized and graphitized Ni(110)

A study of the adsorption/desorption behavior of CO, H₂O, CO₂ and H₂ on Ni(110)(4 × 5)-C and Ni(110)-graphite was made in order to assess the importance of desorption as a rate-limiting step for the decomposition of formic acid and to identify available reaction channels for the decomposition. The carbide surface adsorbed CO and H₂O in amounts comparable to the clean surface, whereas this surface, unlike clean Ni(110), did not appreciably adsorb H₂. The binding energy of CO on the carbide was coverage sensitive, decreasing from 21 to 12 $\text{kcal}\text{mol}$ as the CO coverage approached 1.1 × 10¹⁵ molecules cm^?2 at 200K. The initial sticking probability and maximum coverage of CO on the carbide surface were close to that observed for clean Ni(110). The amount of H₂, CO, CO₂ and H₂O adsorbed on the graphitized surface was insignificant relative to the clean surface. The kinetics of adsorption/desorption of the states observed are discussed.     

Chemisorption of H2 on W(100): Absolute sticking probability,coverage, and electron stimulated desorption

Measurements of both the absolute sticking probability near normal incidence and the coverage of H₂ adsorbed on W(100) at ~ 300K have been made using a precision gas dosing system; a known fraction of the molecules entering the vacuum chamber struck the sample crystal before reaching a mass spectrometer detector. The initial sticking probability S₀ for H₂/W(100) is 0.51 ± 0.03; the hydrogen coverage extrapolated to S = 0 is 2.0 × 10¹⁵ atoms cm^?2. The initial sticking probability S₀ for D₂/W(100) is 0.57 ± 0.03; the isotope effect for sticking probability is smaller than previously reported. Electron stimulated desorption (ESD) studies reveal that the low coverage β₂ hydrogen state on W(100) yields H⁺ ions upon bombardment by 100 eV electrons; the ion desorption cross section is ~ 1.8 × 10^?23 cm². The H⁺ ion cross section at saturation hydrogen coverage when the β₁ state is fully populated is ? 10^?25 cm². An isotope effect in electron stimulated desorption of H⁺ and D⁺ has been found. The H⁺ ion yield is ? 100 × greater than the D⁺ ion yield, in agreement with theory.     

Femtosecond electron dynamics on solid surfaces probed by low energy ion scattering and stimulated desorption of secondary ions

The cations emission from condensed matter surfaces has been investigated on the basis of localization and delocalization of valence hole(s) in the femtosecond timescale. The yield of scattered H⁺ (E₀=100 eV), though negligibly small from the Pt(1 1 1) substrate, increases markedly when Ar is adsorbed on it, indicating the localization of a valence (H⁺ 1s) hole on the physisorbed Ar layer. However, the yield of H⁺ scattered from a thick H₂O layer is considerably small relative to that from Ar and CO layers. The delocalized nature of a valence hole in water ice is caused by some covalency in hydrogen bonds. Hydrated protons, H⁺(H₂O)_n, n=1,2,…,10, are emitted efficiently in electron stimulated desorption from water molecules adsorbed on the Ar layer; the ion yields are highest at the initial adsorption stage and decay steeply with increasing coverage. Coulombic repulsion between the hydrated protons confined in physisorbed nanoclusters is responsible for the explosive ion emission.     

Electron stimulated desorption of H3O from 316L stainless steel

Surface ions generated by electron stimulated desorption from mass spectrometer ion source grids are frequently observed, but often misidentified. For example, in the case of mass 19, the source is often assumed to be surface fluorine, but since the metal oxide on grid surfaces has been shown to form water and hydroxides, a more compelling case can be made for the formation of hydronium. Further, fluorine is strongly electronegative, so it is rarely generated as a positive ion. A commonly used metal for ion source grids is 316L stainless steel. Thermal vacuum processing by bakeout or radiation heating from the filament typically alters the surface composition to predominantly Cr₂O₃. X-ray photoelectron spectral shoulders on the O 1s and Cr 2p_3/2 peaks can be attributed to adsorbed water and hydroxides, the intensity of which can be substantially increased by hydrogen dosing. On the other hand, the sub-peak intensities are substantially reduced by heating and/or by electron bombardment. Electron bombardment diode measurements show an initial work function increase corresponding to predominant hydrogen desorption (H₂) and a subsequent work function decrease corresponding to predominant oxygen desorption (CO). The fraction of hydroxide concentration on the surface was determined from X-ray photoelectron spectroscopy and from the deconvolution of temperature desorption spectra. Electron stimulated desorption yields from the surface show unambiguous H₃O⁺ peaks that can be significantly increased by hydrogen dosing. Time of flight secondary ion mass spectrometry sputter yields show small signals of H₃O⁺, as well as its constituents (H⁺, O⁺ and OH⁺) and a small amount of fluorine as F⁻, but no F⁺ or F⁺ complexes (HF⁺, etc.). An electron stimulated desorption cross-section of σ⁺ ∼ 1.4 × 10⁻²⁰ cm² was determined for H₃O⁺ from 316L stainless steel for hydrogen residing in surface chromium hydroxide.     

Adsorption and reaction of nitric oxide and oxygen on Rh(111)

Adsorption of NO and O₂ on Rh(111) has been studied by TPD and XPS. Both gases adsorb molecularly at 120 K. At low coverages (θ_NO < 0.3) NO dissociates completely upon heating to form N₂ and O₂ which have peak desorption temperatures at 710 and 1310 K., respectively. At higher NO coverages NO desorbs at 455 K and a new N₂ state obeying first order kinetics appears at 470 K. At saturation, 55% of the adsorbed NO decomposes. Preadsorbed oxygen inhibits NO decomposition and produces new N₂ and NO desorption states, both at 400 K. The saturation coverage of NO on Rh(111) is approximately 0.67 of the surface atom density. Oxygen on Rh(111) has two strongly bound states with peak temperatures of 840 and 1125 K with a saturation coverage ratio of 1:2. Desorption parameters for the 1125 peak vary strongly with coverage and, assuming second-order kinetics, yield an activation energy of $\text{85 ± 5}\text{kcal}\text{mol}$ and a pre-exponential factor of 2.0 cm² s^?1 in the limit of zero coverage. A molecular state desorbing at 150 K and the 840 K state fill concurrently. The saturation coverage of atomic oxygen on Rh(111) is approximately 0.83 times the surface atom density. The behavior of NO on Rh and Pt low index planes is compared.     

binding states of CO and H2 on clean and oxidized (111)Pt

The binding states and sticking coefficients of CO and H₂ on clean and oxide covered (111)Pt are examined using flash desorption mass spectrometry and Auger electron spectroscopy (AES). On the clean surface at 78 K there is one major binding state of CO with a desorption activation energy which decreases with coverage plus a second smaller state, while H₂ exhibits three binding states with peak temperatures of 140, 230 and 310 K and saturation density ratios of 0.5 : 1 : 1. Desorption kinetics of CO are consistent with a first order state with a normal pre-exponential factor of 10^{13 ± 1} sec^?1, while all three peaks of H₂ are broader than expected. Interpretations in terms of anomalous pre-exponential factors, coverage dependent desorption activation energies, and desorption orders are considered. On the oxidized surface saturation densities of both gases are nearly identical to those on the clean surface, but desorption temperatures are increased significantly and the initial sticking coefficient on the oxide decreases slightly for CO and increases slightly for H₂.     

The temperature dependence of the interaction of NO + CO on Pt{1 0 0}

Infrared reflection absorption spectroscopy together with mass spectrometry has been used to investigate the interaction of NO and CO on Pt{1 0 0}, initially prepared in the reconstructed ‘hex’ phase, under ambient pressures of these gases, in the temperature range 300-500 K. The results allow the local and total coverages of adsorbed CO and NO to be related to the rate of reaction to produce gas phase CO₂, and provide insight into the species present on the surface during the so-called low temperature oscillatory reaction regime of this process. At temperatures below that at which NO dissociation occurs (approximately 390-400 K) adsorption is controlled by the non-reactive displacement of NO by CO and results in a CO-poisoned surface. Above 400 K when significant CO₂ production occurs, the NO coverage increases to produce a surface with NO and CO fully intermixed; the increase in NO coverage is attributed to the higher rate of NO arrival from the gas phase (with a partial pressure ratio of P_NO:P_CO>1) at free surface sites created by NO dissociation and subsequent reaction with CO. The competition between these two processes of non-reactive NO displacement by CO and reactive displacement of CO by NO is proposed to determine the parameter space of the low temperature oscillatory regime. Rapid equilibration between bridged and atop CO species leads to them appearing to exhibit identical reaction behaviour. Particularly at the lowest reaction temperatures (around 400 K), islands of pure CO may coexist on the surface but not participate in the reaction. Under conditions corresponding to the high temperature oscillatory regime, small quantities of absorbed CO, but no NO, are seen on the surface.     

Hydrogen-induced low temperature CO displacement from the Pt(111) surface

A new low temperature displacement mechanism for CO on the Pt(111) surface has been observed in the presence of high pressures of hydrogen (0.001 to 0.1 Torr H₂). Temperature-programmed fluorescence yield near-edge spectroscopy (TP FYNES) was used to continuously monitor the CO coverage as a function of temperature both with and without hydrogen. For hydrogen pressures above 0.01 Torr, removal of CO begins at 130 K (E_d = 10.6 kcal/mol) instead of near the desorption temperature of 400 K (E_d = 26 kcal/mol). The large decrease in CO desorption energy appears to be caused by substantial repulsive interactions in the compressed monolayer induced by coadsorbed hydrogen. The new low temperature CO desorption channel appears to be caused by displacement of the compressed CO adlayer by coadsorbed hydrogen. In addition, the desorption activation energy for the main desorption channel of CO near 400 K is lowered by ~ 1 kcal/mol for hydrogen pressures in the 0.001 to 0.1 Torr range. These new results clearly emphasize the importance of in-situ methods capable of performing kinetic experiments at high pressures on well characterized adsorbed monolayers on single crystal surfaces. High coverages of coadsorbed hydrogen resulting from substantial overpressures may substantially modify desorption activation energies and thus coverages and kinetic pathways available even for strongly chemisorbed species. These phenomena may play an important role in surface reactions which occur at high pressure.     

Reflection-absorption infrared and thermal desorption studies of carbon monoxide and hydrogen adsorbed on rhodium

Reflection-absorption infrared spectroscopic and thermal desorption techniques have been used to study the interaction of mixtures of carbon monoxide and hydrogen with evaporated rhodium films. For equimolar mixtures near 10^?9 Torr, hydrogen adsorbed much more rapidly, but long exposure times or increases in CO pressures to 10^?6 Torr led to its partial, but never complete, displacement by adsorbed carbon monoxide. Hydrogen desorption spectra taken during the displacement process showed two peaks which was consistent with a cooperative interaction between adsorbed CO and H species. In contrast to previous transmission studies of CO adsorption on small rhodium particles, the present reflection—absorption infrared study of the film system showed a single absorption band at 2075 ±10 cm^?1. While explanations for the discrepancy in terms of particle size effects are possible it is considered more likely that all CO molecules are linearly bound to individual Rh atoms in the present situation. In our work, increases in CO pressure (especially above 10^?6 Torr) were accompanied by an upward frequency shift (from 2065 cm^?1 to 2085 cm^?1) and a narrowing in half width (from 25 to 17 cm^?1). Several possible explanations for the latter unusual effect are discussed.     

The coadsorption of CO and H2 on Fe(100)

The coadsorption of CO and hydrogen on an Fe(100) surface was studied by temperature programmed desorption and X-ray photoelectron spectroscopy. It was found that CO adsorption blocked the subsequent dissociative adsorption of H₂, although it did not seem to affect the hydrogen binding energy. Preadsorption of hydrogen was observed to reduce the binding energy of CO subsequently adsorbed and to inhibit the dissociation of CO. A new surface species was identified in a coadsorbed layer of CO and hydrogen. This species was evidenced by the formation of a desorption peak for H₂ at 475 K when CO was adsorbed subsequent to H₂ adsorption.     

The behaviour of carbon species produced by CO disproportionation on Ru(1, 1, 10) and Ru(001) surfaces

The behaviour of adsorbed CO on Ru(001) flat and Ru(l,1,10) stepped surfaces in the CO pressure range between 10^?6 and 10¹ Pa has been investigated by TDS, AES, LEED and UPS. The disproportionation of CO proceeds rapidly on the stepped surface and its apparent activation energy was obtained to be 20 kJ mol^?1 at nearly zero coverage. The carbon species produced by CO disproportionation show non-uniform reactivity with ¹⁸O₂ and provide four CO desorption peaks in TPR spectra, which are assigned to α-C¹⁸O,ß-C¹⁸O and those derived from carbidic and graphitic carbons. At smaller carbon coverage, only α-CO and β-CO were observed, but with increasing coverage the amount of ß-CO reaches a maximum and carbidic carbon is newly formed. Further increase of carbon deposition gives graphitic carbon. The conversion from carbidic to graphitic carbon and the dissolution into the bulk took place upon heating to 1000 K. It is remarkable that very active carbon species are converted to molecular CO through the reaction with O₂ even at low temperature such as 200 K. It was also confirmed that active carbon species are formed on Ru surface during COH₂ reaction.     

Low-energy electron diffraction,Auger electron spectroscopy,and thermal desorption studies of chemisorbed CO and O2 on the (111) and stepped [6(111) × (100)] iridium surfaces

The adsorption of CO, O₂, and H₂O was studied on both the (111) and [6(111) × (100)] crystal faces of iridium. The techniques used were LEED, AES, and thermal desorption. Marked differences were found in surface structures and heats of adsorption on these crystal faces. Oxygen is adsorbed in a single bonding state on the (111) face. On the stepped iridium surface an additional bonding state with a higher heat of adsorption was detected which can be attributed to oxygen adsorbed at steps. On both (111) and stepped iridium crystal faces the adsorption of oxygen at room temperature produced a (2 × 1) surface structure. Two surface structures were found for CO adsorbed on Ir(111); a (√3 × √3)R30° at an exposure of 1.5–2.5 L and a (2√3 × 2√3)R30° at higher coverage. No indication for ordering of adsorbed CO was found on the Ir(S)-[6(111) × (100)] surface. No significant differences in thermal desorption spectra of CO were found on these two faces. H₂O is not adsorbed at 300 K on either iridium crystal face. The reaction of CO with O₂ was studied on Ir(111) and the results are discussed. The influence of steps on the adsorption behaviour of CO and O₂ on iridium and the correlation with the results found previously on the same platinum crystal faces are discussed.     

The role of co-adsorbed O2 on the catalytic reduction of NO with C2H5OH on the surface of Pt(3 3 2)

The influence of pre-dosed O₂ on the catalytic reduction of NO with ¹³C₂H₅OH on the surface of stepped Pt(3 3 2) was investigated using Fourier transform infra red reflection-absorption spectroscopy (FTIR-RAS) and thermal desorption spectroscopy (TDS). We show that the oxidation of ¹³C₂H₅OH with O₂ is a very effective reaction, occurring at 150 K and giving rise to acetate. The presence of NO does not lead to any evident oxidation of ¹³C₂H₅OH irrespective of the annealing temperature. For the case of O₂ + ¹³C₂H₅OH + NO co-adlayers, oxidation of ¹³C₂H₅OH also takes place at 150 K. However, no new surface species that are supposed to be an intermediate for the production of N₂ are detected.The influence of O₂ on the production and desorption of N₂ is intimately related to both O₂ and ¹³C₂H₅OH coverage. The presence of pre-dosed O₂ does not greatly promote N₂ desorption. In fact, N₂ desorption is suppressed quantitatively with increasing O₂ coverage, after which unreacted, or left-over O atoms appear and remain on steps. It is concluded that the presence of pre-dosed O₂ does not play a role of activating reactants in the catalytic reduction of NO with ¹³C₂H₅OH on the surface of Pt(3 3 2).     

DRIFTS study of different gas adsorption for CO selective oxidation on Cu-Zr-Ce-O catalysts

The adsorptions of different gases (CO, H₂ and O₂) in the hydrogen-rich gas on the co-precipitated Cu-Zr-Ce-O catalyst were discussed and the active sites were ascertained with infrared spectroscopy technique. It was shown that the adsorption strength of CO was stronger than that of O₂ or H₂. Hydrogen and CO were competitive adsorption and the coexistence H₂ and CO on the surface accelerated the rate of CO desorption. Adsorbed H₂ could convert into geminal OH groups on the ceria surface at high temperatures in the absence of oxygen, while it was easy to form surface hydroxyl groups at low temperatures and condensed to physical water with increasing desorption temperature in the existence of oxygen. The adsorption of CO₂ was strong and it could transform into thermal stable carbonate species even in the reaction conditions. The active sites of the Cu-Zr-Ce-O catalyst were Cu²⁺ and Cu⁺, mainly the latter. The oxygen defect sites could be formed on the Cu-Zr-Ce-O catalyst surface through dehydration and decarboxylation.     

LEED and thermal desorption studies of small molecules (H2, O2, CO,CO2, NO,C2H4, C2H2 and C) chemisorbed on the stepped rhodium (755) and (331) surfaces

The chemisorption of H₂, O₂, CO, CO₂, NO, C₂H₂, C₂H₄ and C has been studied on the clean stepped Rh(755) and (331) surfaces. Low energy electron diffraction (LEED), Auger electron spectroscopy (AES) and thermal desorption spectroscopy (TDS) were used to determine the size and orientation of the unit cells, desorption temperatures and decomposition characteristics for each adsorbate. All of the molecules studied readily chemisorbed on both stepped surfaces and several ordered surface structures were observed. The LEED patterns seen on the (755) surface were due to the formation of surface structures on the (111) terraces, while on the (331) surface the step periodicity played an important role in the determination of the unit cells of the observed structures. When heated in O₂ or C₂H₄ the (331) surface was more stable than the (755) surface which readily formed (111) and (100) facets. In the CO and CO₂ TDS spectra a peak due to dissociated CO was observed on both surfaces. NO adsorption was dissociative at low exposures and associative at high exposures. C₂H₄ and C₂H₂ had similar adsorption and desorption properties and it is likely that the same adsorbed species was formed by both molecules.