Chen The electron capture detector

Qan ratio of unity. This was one of the first determinations of a Qan ratio for electron molecule reactions. The ECD values of the excited-state electron affinity, Qan, and A1 and E1 agree with the swarm values. The low-temperature response of the ECD can be calculated from parameters measured in swarm experiments.

In surface ionization experiments the electron source is a heated filament where reactions take place. The reactions are studied at high temperatures and relatively low pressures. The filament has been heated by a flame and resistance heating. The experimental data are the electron and ion currents. These have been measured using an electric field, a magnetic field, or combinations of both [19–22]. The magnetron method is a surface ionization technique in which a magnetic field is used to measure the ion and electron currents. Both ions and electrons carry the current from the filament to a concentrically mounted electrode in the absence of an intense magnetic field. In the presence of a sufficiently intense magnetic field, the electrons are driven into concentric paths between the filament and the anode so that only the ions are collected. The ratio of the electrons and ions can be measured by applying a magnetic field. In the 1960s the magnetron method was used to determine the electron affinity of several organic molecules. This direct capture process is only one of four possible mechanisms observed in the magnetron. Another reaction involves gas phase dissociation prior to electron capture. This yields electron affinities of radicals. Many of these values are equal to the current Ea within the experimental error. Two other mechanisms involve the adsorption of a radical on the filament. These give values that are higher than the current evaluated values for the phenyl and alkyl radicals. These high values cannot be currently explained. When the equilibrium mechanism can be established, the magnetron Ea is the AEa or an excited state Ea.

In Figure 6.2 global plots of the magnetron data as ln KT3/2 versus 1,000/T for tetracyanoethylene, tetracyanobenzene, and dicyanoethylene studied in the direct capture method are presented. These data were taken from the review on the magnetron method [2]. The similarity of these data to ECD and NIMS data is clear. The magnetron method has also been used to determine Qan values. For tetracyanoethylene there is a change in slope at the lower temperatures so that other ion losses compete with detachment. Shown in Figure 6.3 are the electron swarm data for O2, the ECD and magnetron data for NO and C6F6, and NIMS and magnetron data for SF6. The calculated equilibrium method data for an Ea of 0.02 eV for NO are also shown. These illustrate different temperature ranges and the similarity of data from the equilibrium methods. An ideal equilibrium method would extend the temperature range from the ECD/NIMS range ð1;000=T ¼ 1:5–3:0Þ to the magnetron range ð1;000=T < 1:0Þ. The magnetron values for SF6 and C6F6 are higher than for the evaluated values by about 0.3 eV. Two states are observed in the ECD data for C6F6, as indicated by the structure at lower temperatures [3, 23].

The electron affinities of molecules determined from the direct capture magnetron method are plotted in a P and A graph in Figure 6.4. The zero intercept slope is 1.01(5). The standard deviation is 0.13 eV. This gives a nominal precision of0.15 eV, including the low value for C4H2N2 and high values for SF6 and C6F6. The magnetron values for anthraquinone, benzoquinone, fluoranil, and chloranil are

108 COMPLEMENTARY EXPERIMENTAL AND THEORETICAL PROCEDURES

Figure 6.2 Plots of magnetron data as ln KT3/2 versus 1,000/T. The similarity of these data compared to the ECD data should be noted. The data were taken from [27] and original references cited therein. This shows the range of the electron affinities that can be obtained using the magnetron which is 0.75 to 3.3 eV for hexacyanobutadiene (not shown).

lower (0.42 eV) than the EvV. Since this is more than twice the nominal uncertainty, these are assigned to excited states.

The equilibrium methods yield electron affinities directly from experimental data and fundamental constants. The ECD is the most extensively applied equilibrium method. It is limited to molecules with electron affinities less than about 1.5 eV (C6F5NO2) by the upper temperature limit. Figure 6.5 is a plot of data that illustrates the range of electron affinities which have been measured in the ECD. If the temperature range can be extended, higher electron affinities can be measured. The NIMS method has the potential of being applied to molecules with electron affinities higher than 1.5 eV. For molecules, such as acetophenone, with a large enough linear temperature range and only one negative-ion state, the ECD or NIMS has attained a precision of 0.01 eV for Ea. By repeating the determinations, this precision can be improved. The swarm technique used for the determination of an electron affinity of oxygen and the detailed balancing procedures have the same potential precision as the ECD and NIMS procedures. The possibility of anionic excited states must be considered in all cases. The magnetron method

110 COMPLEMENTARY EXPERIMENTAL AND THEORETICAL PROCEDURES

Figure 6.5 ECD data plotted as ln KT3/2 versus 1,000/T illustrating the range of electron affinities measured in the ECD. This is 0.05 eV to 1.5 eV. The data have been published, but not in this specific combination [31, 34, 40].

can be used to obtain accurate electron affinities of molecules with a demonstrated precision of 0.15 eV. The major problems with the magnetron method are the establishment of the mechanism, the identification of the state of the anion, and the correction of the results from the high temperature of the measurement to absolute zero. The magnetron procedure can be improved by obtaining a more accurate measurement of the temperature and concentration of the reactant. The magnetron electron affinities have been generally disregarded because of the lack of mass analysis. The major ion formed for TCNQ and TCNE at a heated filament is the parent negative ion. The use of mass spectrometry will not eliminate the problem of identifying the state of the anion.

6.3PHOTON TECHNIQUES

The reaction ABð Þ þ hn ! AB þ e is the basis of photoelectron spectroscopy and photodetachment methods. Many precise and accurate ionization potentials of molecules have been obtained by studying the photoionization of neutral molecules. The same principles apply to the photon methods for determining electron affinities, except that negative ions are studied. The electron affinities of over 1,000 atoms, radicals, clusters, and small molecules have been determined using

Figure 6.6 Plots of ion intensities versus photon energy for O( ) and O2( ) from [26]. The atomic threshold is very sharp but the molecular threshold is gradual.

these techniques. Only about a dozen electron affinities of large organic molecules have been determined by photon methods. The combination of photodetachment with an ECD has been used to estimate the electron affinities of molecules [24]. These experiments require a monochromatic light source, source of negative ions, and way to measure the intensity and energy of the electrons. In photodetachment experiments the threshold for electron formation is related to adiabatic electron affinity. In photoelectron spectroscopy the intensity and energy of the electrons formed when photons of a fixed energy interact with the anions are measured. The absorption and emission spectra of atoms or their anions have also been used to obtain the electron affinities of atoms [7]. Bound excited states of anions have been studied using two-photon photodetachment spectroscopy [25].

The first accurate and precise determinations of the electron affinity of atoms other than the halogens were carried out using the photodetachment procedure in the late 1950s. With a high-intensity carbon arc lamp and interference filters with a glow discharge source, the photodetachment spectrum of O( ) was measured and the electron affinity determined from the threshold [26]. Figure 6.6 is a graph of the data for O( ) and a graph for O2( ) [27]. The difference in the shape of the threshold for the molecule and atom is important. The onset for the photodetachment of the atomic anion is very sharp and gives an accurate and precise estimate of electron affinity. However, the onset for the molecule is gradual and extrapolates to well below adiabatic electron affinity. This was believed to result from the population of excited states of the anions [28]. In the 1960s photoabsorption and photoemission

112 COMPLEMENTARY EXPERIMENTAL AND THEORETICAL PROCEDURES

Figure 6.7 High-resolution photoelectron spectra of O2( ). The large peaks form a progression from one negative-ion state. The small peaks were unexplained in the original article, but coincide with electron affinities measured by other techniques. The PES spectrum was taken from [33].

spectra of the halogen atoms or anions were used to obtain electron affinities. Some of these values remain the most accurate and precise Ea [7, 29].

With the development of laser light sources the PES and PD methods have become the preferred method of measuring atomic electron affinities. The electron affinities of most of the atoms have been determined to precisions of 0.1% using laser PES. The use of laser photodetachment spectroscopy has improved the precision to parts per million. The evaluated values of the electron affinities of the atoms and their relationship to the Periodic Table will be discussed in Chapter 9. The problems in establishing the threshold and in assigning the onset to a given state make the determination of molecular electron affinities by photon techniques more difficult. The ground-state negative ions of many homonuclear diatomic molecules have been characterized by PES [7]. In the photoelectron spectra of diatomic molecules, hot bands due to excited vibrational levels of the negative ion have been observed. The highest resolution, 0.02 eV, photoelectron spectra for oxygen, is shown in Figure 6.7 [30]. The internuclear distance and fundamental frequency for the excited-state anion are obtained from a detailed analysis of this spectrum. The most intense progression is for the excited state, with an electron affinity of 0.430(2) eV and 0.450(2) eV. The doublets result from spin orbital coupling. However, other peaks in the spectrum were unexplained. We have interpreted these peaks as arising from the higher-energy, lower-population anion states observed in other experiments [31].

The photoelectron spectroscopy data for CS2 led to the first accurate and precise value for the electron affinity for the ground-state bent anion, 0.895 0.02 eV [32].

Figure 6.8 Photoelectron spectra of coronene( ), from [38]. The electron affinity was determined from the first peak in the first series of peaks. If the second series of peaks is from the ground state of the anion, the electron affinity is 0.8 eV, in agreement with the reduction potential and CURES-EC value.

A second higher-resolution PES study gave an upper limit of 0.8 eV for the Ea. The high-resolution PES spectrum is shown in Chapter 9 [33]. In neither case was an electron affinity for the excited state reported, although both spectra have peaks at low energy attributed to ‘‘hot’’ bands. If these bands are assigned to the excited linear state, the excited-state Ea can be estimated from the initial onset, while the ground-state value can be obtained from the identification of the 0-0 band of the second series [31–33]. The examples of O2 and CS2 illustrate the problems encountered with excited states in PES. Both the ground-state electron affinity and linear excited-state electron affinity have been determined using the ECD [34] (see Figure 4.3). The excited-state electron affinity has been measured in TCT and AMB studies, whereas the higher Ea has been measured in a separate AMB and TCT study [3, 35–37].

The data obtained from the ECD and reduction potentials can be used to interpret PES data. Three examples for molecules are the ‘‘lower’’ values for the electron affinities of nitromethane, anthracene, and coronene. Based on the observation of excited states in anthracene and tetracene in the ECD data, it is reasonable to assume that the lower value for coronene derives from the population of an excited state. In Figure 6.8 the PES of coronene is shown with two sets of peaks. If the ground-state Ea for coronene is taken from the initial onset, it is much lower than the value obtained from reduction potential or electronegativity data. In addition, the second onset must be explained [38–42].

The electron affinity of nitromethane has been determined by ECD, PES, AMB, and TCT. The ground-state electron affinity for nitromethane has been measured by

114 COMPLEMENTARY EXPERIMENTAL AND THEORETICAL PROCEDURES

Figure 6.9 Photodetachment spectrum of tetracyanoquinodimethane( ), from [25]. The two peaks represent detachment from an excited state by way of a two-photon process.

ECD, TCT, and AMB studies. In the case of the PES data there are low-energy bands that can be attributed to excited valence states, one of which is a dipole bound state. The new question becomes, ‘‘Why are the excited states not stabilized to the ground state?’’ The answer could be that the pressure is not sufficiently high [43–46].

Photon experiments have produced convincing evidence that bound excited anionic states of atoms, small molecules, and large organic molecules such as tetracyanoethylene (TCNE) and tetracyanoquinodimethane (TCNQ) exist. One of the most convincing studies made use of electron photodetachment and one-photon and two-photon intensity dependence studies to establish at least one excited state of TCNQ( ) that is lower in energy than electron detachment. The most direct evidence lies in the photodetachment spectra shown in Figure 6.9 [25]. The similarity between this spectrum and that in Figure 6.8 is striking. Here the low-energy process is attributed to a two-photon absorption in which the anion is excited and then the electron is photodetached with a second photon. The existence of a bound excited state was also used to explain the difference between the photodetachment value for the Ea of TCNE (1.7–2.3 eV) and the TCT and magnetron values of 2.9 eV [47–49]. It was further concluded that this appears to be a general phenomenon which should apply to a whole class of radical anions [25].

In the case of the photoelectron spectra of cyclooctatetraene (COT) the ECD and TCT values for the ground-state electron affinity can be used to assign the onset to the formation of a transition state. Thus, the onset occurs at an energy higher than the adiabatic electron affinity. In the case of the photodetachment data for COT the onset occurs between the adiabatic electron affinity and the onset for photoelectron spectroscopy [50]. These will be considered more extensively

Figure 6.10 Precision and accuracy plot for photon data. The nitromethane, CS2, and anthracene values are lower than the current ‘‘best’’ value by 0.3 eV. The dipole bound Ea are all less than 0.25 eV. The random uncertainty in the other values is 0.01 eV.

in Chapter 10. In another important case the PES data give the dipole bound Ea of the purines and pyrimidines. Only the dipole bound state is observed in rare gases, but in the presence of water, the ground state can be accessed. In addition, twophoton absorption of the hydrates leads to the dipole bound anion. These will be described in Chapter 12 [51].

Figure 6.10 is a precision and accuracy plot for the Ea of several molecules determined using photon methods. The excited-valence-state values (shown in the squares) are clearly below the unit slope zero intercept line. The dipole bound states (shown in the triangles) are all less than 0.25 eV. The ground-state values are within 0.01 eV of the unit slope zero intercept line. Thus, the photon methods are capable of measuring ground-state values within this uncertainty. When a value is much lower than the largest accurate value, excited states must be considered. In Chapters 9 to 12 the use of this concept to interpret photon data for other molecules in the same manner as for oxygen and coronene will be discussed.

The absorption, emission, photodetachment, and photoelectron spectroscopy experiments are capable of providing accurate and precise values for the electron affinities of atoms. The best precision is about 1 part per million, more than precise enough for chemical purposes. The state of the ion must be identified and some excited-state electron affinities of atoms have been reported. The photoelectron spectroscopy and photodetachment procedure can give the accurate and precise electron affinities of molecules and radicals when the state of the anion is assigned.

116 COMPLEMENTARY EXPERIMENTAL AND THEORETICAL PROCEDURES

In the case of tetracyanoquinodimethane, carbon disulfide, nitromethane, and the purines and pyrimidines, two or more negative-ion states have been observed. In some cases the photoelectron spectrum can be assigned to an excited state and reveal an electron affinity lower than adiabatic electron affinity. In the case of cyclooctatetraene the onset in the PES spectrum is higher than adiabatic electron affinity because of the significant change in the geometry of the anion.

6.4THERMAL CHARGE TRANSFER METHODS

In the thermal charge transfer methods the electron affinity of a molecule is determined by bracketing the electron affinity of a test species between that of two species with known electron affinities. When the reaction studied is ABð Þ þ CD , AB þ CDð Þ direct charge transfer, the relative electron affinities are obtained. A variation on this method is the measurement of the intensity of ions formed by the collisional ionization of an electron bound dimer. This has been applied to a number of aromatic and halogen-substituted aromatic hydrocarbons, but can give Ea for excited states. When the reaction studied is proton transfer, Að Þ þ BH ! AH þ Bð Þ, the relative gas phase acidities are obtained. With experimental values of the AH and BH bond energies the relative electron affinities of the radicals may be determined [52]. In order to carry out these experiments, there must be a source of the test material and a source of the reactant anion. There must also be a method for measuring the concentration of the test anion and the product anion. The electron affinities or gas phase acidities of the reference compounds and the accuracy and precision must be known. The general procedure has been to develop a ladder of bracketing reactions and to reference that ladder to an accurate and precise acidity or electron affinity. These reactions can be studied by observing the direction of charge transfer from kinetic experiments or by measuring the equilibrium concentrations of the anions and neutrals. In either case the relative values of the electron affinity or gas phase acidity are obtained. The error in the measured electron affinity is no smaller than the errors in the electron affinity of the bracketing species. The determination of the equilibrium concentrations can be completed at one temperature or multiple temperatures. When the reaction is carried out at a single temperature, it is necessary to assume that the entropy changes for the two reactions are equal so as to eliminate the need for temperature corrections. If the entropy and energy changes are determined, the gas phase acidity or electron affinity must be corrected to absolute zero.

The first thermal electron transfer measurements to give molecular electron affinities were observations of the direction of charge transfer using negative-ion mass spectrometry. The electron affinity of nitrobenzene was bracketed between that of SO2 and NH2 in 1959 [53]. In 1961 the electron affinities of CS2 and SO2 were bracketed between that of the O atom and NH2 [54]. In the 1970s the use of nega- tive-ion mass spectrometry to study equilibrium reactions began to flourish. This coincided with the development of the flowing afterglow [55], ion cyclotron resonance (ICR) mass spectrometer, and high-pressure mass spectrometer (HPMS)