Dec. 5, 1958
IONIZATION AND DISSOCIATION OF OXYGEN BY ELECTRON IMPACT
(NH4)2Crz07(c)are consistent to within about one and one-half kcal./mole. The “best” value for the standard heat of formation of CrOc‘(aq) is taken to be the average of the values based on CrO2 and (NH4)~Cr207~ namely, - 208.6 kcal./mole. From the “best” heat of formation of Cr04-(aq) and heats of reactions 1 and 3, the “best” heats of formation of HCrOr-(aq) and Cr207-(aq) have been calculated to be -207.9 and -352.2 kcal./ mole, respectively. These heats of formation have been combined with heats of solution and reaction from this and the earlier paper2 to calculate “best” heats of formation of KzCrO4(c) and KK!r207(c) to be - 332.8 and -488.3 kcal./mole, respectively. From heats of formation and entropies given in this paper and others tabulated by the Bureau of Standards,?free energies of formation of Cr04-(aq), HCrOl-(aq) and C r 2 0 ~ ( a q )have been calculated t o be -171.1, -1180.0 and -305.9 kcal./mole, respectively.
[CONTRIBUTION FROM THE
6183
Conclusions News and Rieman,‘ Tong and King“ and Davies and Pmee are in fair agreement as t o the value of K 3 a t 25’. Their concordant results give us confidence in the reported values for K3 that is substantiated by the satisfactory agreement between our calorimetric A H 3 O and the A H 3 O calculated from the temperature dependence of K3. It also seems very likely that the reported values for KP and K4 are a t least approximately correct. The equilibrium constant for reaction 1 has been less adequately investigated than has K3. It is planned to reinvestigate in this Laboratory this equilibrium a t 25’ with the aim of checking the results of Neuss and Rieman.4 It is also intended to investigate this equilibrium a t several temperatures in order that an independent value of AHlo may be obtained to be compared with the calorimetric AHl0 reported in this paper. CHARLOTTESVILLE, VIRGINIA
DEPARTMENT OF CHEMISTRY,
UNIVERSITY OF
BRITISHCOLUMBIA]
The Ionization and Dissociation of Oxygen by Electron Impact’ BY D. C. FROST AND C. A. MCDOWELL RECEIVED JUNE 30, 1958 The ionization and dissociation of molecular oxygen has been studied using the retarding potential difference (r.p,d.) method of obtaining essentially mono-energetic electrons. All the spectroscopically known excited states of molecular *If,, *nu, 42g-,and one additional state, have been observed. The vertical ionization potentials here deteroxygen, ;.e., mined are in very good agreement with those calculated from available spectroscopic data. Six different dissociation processes leading to the production of O+ ions have been observed. These ions arise from either of these two types of processes: 0 2 f e = O f f 0 f 2e (a), 0 2 f e = O + f 0- f e (b). Each particular process has been identified by a careful study of the negative and positive ions formed. Ambiguities in earlier work, which arose largely from the difficulty of calibrating the negative ion energy scale, have been overcome by using sulfur hexafluoride as a calibrating gas.
Oxygen has been the object of many fairly detailed electron impact studies. The most detailed work is perhaps reported by Hagstrum.2 Besides studying the various ionization and dissociation processes, Hagstrum measured the appearance potentials of the positive and negative oxygen ions. Hagstrum also determined the kinetic energies with which the various ions are formed in the different dissociation processes observed. Later work by ThorburnShas, however, shown that some of Hagstrum’s observations are perhaps not quite as well founded as they formerly appeared. Examination of Thorburn’s results indicates that these more recent observations are still not quite satisfactory. The main difficulties in all the earlier work arise from the fact that the various workers used electron beams which had quite a large spread in their energy distribution. These experimental methods thus precluded the observation of fine details which are to be expected if the known spectroscopic data on oxygen are studied (see McDowel14). (1) This work was supported in part by the Geophysical Research Directorate of the U. S. Air Force Cambridge Research Centre, Air Research and Development Command, under Contract No. AF19(604)-2275, and by grants from the National Research Council of Canada. (2) H. D. Hagstrum, Rev. Mod. P h y s . , 23, 185 (1951). (3) R. Thorburn, “Report of Conference on Applied Mass Spectrometry,” Institute of Petroleum, London, 1955, p. 185. (4) C. A. McDowell, “Applied Mass Spectrometry,” Institute of Petroleum, London, 1954, p. 129.
Furthermore] in studying negative ion efficiency curves, these workers were unable to observe the true shape of the electron resonance capture peaks, and also they were unable to calibrate accurately the negative ion energy scale. Both these major difficulties recently have been shown to be resolved if one uses essentially mono-energetic electrons.6-7 It is now possible to calibrate the negative ion energy scale accurately by using the near-zero appearance potential of the SFs- ion from sulfur hexafluoride as a standard.6 Our own previous work68819 has shown that the retarding potential difference method of obtaining essentially mono-energetic electrons developed by Fox, et aL17yields accurate values for the excited states of molecular and atomic ions. It thus seems that the application of this method to oxygen would resolve many of the discrepancies noted in earlier electron impact studies of this molecule.
Experimental The mass spectrometer used was that described in our earlier publications.810 Krypton served as a calibrating gas for the measurements on the positive ions. Sulfur (5) D. C. Frost and C. A. McDowell, Proc. Roy. SOC.(London), 8282, 227 (1955). (6) W.M.Hickam and R. E. Fox, J . Chem. Phys., 26, 642 (1956). (7) R. E. Fox, W. M. Hicknm, T. Kjeldaas. Jr., and D. J. Grove, Phys. Rev., 84, 859 (1951). (8) D. C. Frost and C. A. McDowell, Proc. Ray. SOC.(London),
A236, 278 (1956). (9) D. C.Frost and C. A. McDowell, C a n . J . Chem., 3 6 , 39 (1958).
D. C. FROST AND C. A . MCDOWELL
12 Fig. 1.-The
Vol. 80
15 16 17 18 19 20 21 Electron energy, e.v. uncorrected. ionization efficiency curve for the formation of the Oz+ ion from oxygen by electron impact. 13
14
hexafluoride (Matheson Co. Inc.) was used to calibrate the negative ion energy scale. The ionization potential of krypton is 112914.5 cm.-’ lo or 13.997 e.v. if 8066.83 cm.-l equals 1 e.v.I1 The appearance potential of the SFsion from sulfur hexafluoride was taken as 0.03 e.v.6 Oxygen pressures of 5 X 10-6 mm. and 4 X mm. were used for positive and negative ion studies, respectively, and 3 X 10-6 mm. pressure of each calibrating gas was found to give adequate K r + and SF6- abundances. The retarding potential difference (r.p.d.) method could not be used for 0 - ions a t high electron energies because of the small ion current. 0 - appearance potentials under these conditions were obtained in the usual way from the initial portion of the 0 - resonance capture process (of established onset energy). For the results quoted for the 0 2 + ion, the 0” ion and the 0 - ion a t low energies (resonance capture process) the theoretical energy band width of the electrons was 0.1 e.v.
Experimental Results The ionization efficiency curve for the 0 2 + ion is shown in Fig. 1. Here there are clearly five ionization processes observed. The energies a t which (10) R . E. Becker and S. Goudsmit, “Atomic Energy States,” &ICGraw-Hill Book Co., New York, N. Y., 1952. (11) J. W.hf. D u M o n d and E . R.Cohen, P h y s . Rev., 82, 555 (19.51).
these occur, with the standard deviations, are 12.21 h 0.04, 16.30 h 0.03, 17.18 =t 0.02, 18.42 0.02, and 21.34 0.02 e.v. These energies will later be shown to refer to the formation of the 02+ ion in its ground and first four excited states. The O f ionization efficiency curve is shown in Fig. 2. This graph shows evidence for five distinct dissociation processes leading to 0 + ions. The energies a t which these five dissociation processes occur are 17.30 =t 0.10, 1S.99 + 0.05, 20.42 i 0.04, 21.30 h 0.03 and 22.03 h 0.03 e.v. Figure 3 shows the ionization efficiency curve for the formation of 0 ions a t higher energies. These results were obtained using electrons with half the width of the normal energy distribution. Since the 0 - ion here appears a t 17.36 i 0.04 and 21.22 h 0.05 e.v., i t is apparent that the first and fourth dissociation reactions leading to the formation of the O+ positive ions in Fig. 2 must be ion-pair processes. Figure 4 shows the resonance capture peak for the 0- ion a t low electron energies. The SFs- ion resonance capture peak is also depicted as this is the
*
Dec. 5, 1958
t
IONIZATION AND DISSOCIATION OF OXYGEN BY ELECTRON IMPACT
17 18 19 20 21 22 23 Electron energy, e.v. uncorrected.
17
Fig. 2.-The ionization efficiency curve for the formation oE the O + ion from oxygen by electron impact.
calibrating ion. The appearance potential for the 0-ion is 4.53 f 0.03 e.v. Discussion The Molecular Oxygen Ion 02+.-The ground state of the oxygen molecule has the electronic structure KK( ~ 3 s' ()~ u 2 ~ ) '~( g 2 ~ ) ' (
T,~P)',3&-
6185
( 1)
18 19 20 21 22 23 Electron energy, e.v. uncorrected. Fig. 3.-The ionization efficiency curve for the formation of the 0 - ion from oxygen by ion pair processes as the result Of impact. I
t
4I
cc-
0-
The first ionization potential will therefore refer to the removal of an electron from the (rg2p) orbital to form an Oz+ ion in its zIIgground state. Earlier electron impact studies by Tate and SmithI2 and 0 1 2 3 4 5 6 7 Hagstrum2 yielded the values 12.5 f 0.1 and 12.1 Electron energy, e.v. uncorrected. f 0.2 v. Our value is 12.21 f 0.04 v. Recent photo-ionization studies by Watanabe13 give the Fig. 4.-The resonance capture peak for the formation of 0 value 12.075 f 0.01 v. for the first ionization potenions from oxygen by electron impact. tial. Careful study of the initial portion of the ioniza- These go to limits a t 18.16 e.v.,I5 18.216and 16.11 tion efficiency curve for this 0 2 + ion by our method e.v.16 These limits refer to the upper 42g-and lower shows that the curve is non-linear over the first 0.5 411, components of the first negative band sysseparae.v. Since about 80% of the electron energy dis- tem of molecular oxygen, the 42g--41Tu tribution we use lies within 0.15 e.v., the non-linear- tion being 2.05 e.v.17 The second ionization potenity of the true ionization efficiency curve is not tial we observed to be 16.30 e.v. This obviously reion in its 411ustate. likely to extend over more than 0.3 e.v. This curv- fers to the formation of the 0 2 + The third ionization potential we observed for the ature may be attributed to vertical transitions from ion occurs a t 17.18 e.v. As it is known that the the ground 32,state of the oxygen molecule to the 02+ first two vibration levels of the 211gground state of second negative bands of oxygen involve transitions the Oz+ ion, since these levels are separated by only between the 211gground state of Os+ and an upper state, 4.81 e.v. above the ground state, i t is 0.25 e.v. The 02+ 211gions have an equilibrium interapparent that this third ionization potential which nuclear distance 0.085 A. less than that of the ground state of the oxygen molecule, so i t is quite we observed must refer to the formation of the 02+ likely that the Oz+ ion will be formed by transitions ion in its 2TI, excited state. We observed the fourth ionization potential of to other than the v' = 0 vibrational level of the the 02+ion to occur a t 18.42 e.v. Since the 4Zgground state. is known to lie a t the limit of the RydEquation 1 indicates that the second ionization state of 02+ berg series observed by Price and Collins a t 18.16 potential will involve the removal of an electron from the (nu2p) orbital with the formation of an 02+ e.v., i t is evident that this fourth ionization potential which we found must be due to the formation of ion with the configuration ( T ~ ) ~ ( T ~ This ) ~ . configuration can give rise to the electron states 411u,2@u the Oz+ion in its 4Bg- state. This is most easily enand 2112,but Mulliken14 has shown that only the 411, visaged to be formed by the removal of an electron and states of the 02+ ion need here be considered. from the (a,2p) orbital of equation 1 leading to an (15) Y Tanaka and T. Takamine, P h y s . Rev., 6'9, 771 (1941). Two Rydberg series have been observed for oxygen. (12) J. T . Tate and P. S. Smith, P h y s . Rev., 39, 270 (1932). (13) K. Watanabe, J. Chem. P h y s , 2 6 , 542 (1957). (14) R S. Mulliken, Res. M o d . P h y s . , 4 , 3 (1932).
(16) W. C.Price and G. Collins, ibid., 48, 714 (1935). (17) T. E.Nevin, Phil. Trans., A237, 481 (1938); Proc. Roy. SOC. (London), A174, 371 (1940);.T. E. Nevin and T. Murphy, Proc. R o y . Irish A c a d . , 46, 169 (1941).
D. C. FROST AND C. A. MCDOWELL
CilSO
ion with the electronic configuration (cp)(nu)4 ( ~ and so in a 42,-state. Between 18.5 and 23 e.v. our work shows evidence for only one further ionization potential, i.e., that found a t 21.34 0.02 e.v. This could be the ion formed by the removal of an electron from a (uu2s) or a (ag2p) orbital with or without simultaneous excitation of another electron in one of the other orbitals. It is not possible a t the present time to say what the particular electronic structure of this ion may be, as i t has not been detected spectroscopically. The various electron impact and spectroscopic ionization potentials found for oxygen are collected in Table I.
*
Vol. 80
3 ~ close to the value found by Thorburn,a fairly namely, 4.7 e.v., and less than the value of (3.3 e.v. observed by H a g s t ” n 2 The forrnation of the 0ion from 0,by process 2 can be represented energetically by the equation V(O-) = D(0z)
- E(0)+ K
+E
(3)
where V(0-) is the appearance potential of the 0ion, D(02) is the dissociation energy of the oxygen molecule, E ( 0 ) is the electron affinity of the oxygen atom, K is the kinetic energy with which the fragment may be endowed, and E is the excitation energy of either the 0 atom, or the 0.negative ion, or both. If D(O2) = 5.115 e.v.,21and E ( 0 ) = 1.45 f. 0.15,22then
V ( O - ) = 3.67 e.v. + K + E (4) If the fragments are formed in their ground states, the kinetic energy K is 0.86 e.v. Earlier workers have used retarding potential techniques to Spectroscopic determine V(O-)for ions a t zero kinetic energy. ionization Electron impact lllectronic state potential of 0 2 ionization potenInteratomic Values which have been reported are 2.gIz32.g2 and of the 02-ion (e.v.) iial of 0 2 (e.v.) distance (A,) 3.2.24 The discrepancies between these values and *& 12.16“ 12.21 i 0 . 0 4 1.1227 ours probably are due to uncertainty in the cali411~ 16.1116 16.30 i .03 1.3813 bration of the electron energy scale in earlier X, 16. 97b 17.18 i .02 1.4089 work. Here it is to be noted that the unambigu42,18.1615,’6 18.42 zt .02 1.2795 ous SFg method we have used was only recently ? . ..... 21.34 .02 .... introduced.6 The 0-ion resonance capture peak 1.207 ( 0 2 ) which we observed using electrons with about 0.2 a The value given for the first spectroscopic ionization potential depends on the value chosen for D(O2+) in its e.v. energy band width extended over about 3 e.v. state. A short extrapolation of the TIg state gives and was nearly symmetrical, about a maximum of D(Ot+)211, as 6.57 e.v.18p1g Hence from the cycle 5.98 e.v. The shape of the resonance capture peak should D(0d I ( 0 ) = D(Oz+) ~ ( O Z ) we get I(0.J = 12.16 e.v. T h e true value for ~ ( O Zis) be a reflection of the 02 molecular ground state eilikely to be nearer the value of 12.075 e.v. obtained by genfunction in the potential energy curve for the Assum- 02- molecular ion. This is illustrated in Fig. 5 Watanabe.I3 This gives D(02+)211gas 6.65 e.v. ing Watanabe’s value for I(O2) to be correct, then the ioniza- where the 0-ion peak is drawn to scale and the tion potential of oxygen referring to the formation of the upper curve is constructed so that it crosses the ap0 2 + ion in its state is 16.89 e.v. proximate boundaries of the Franck-Condon reThe interatomic distance of the 321,- ground gion a t points corresponding to the upper and state of oxygen is 1.2074 A., so if the “width” of the lower threshold energies a t which the 0-ion peak lowest vibrational ejgen-function in this state is is observed. The upper 02-state dissociation limit assumed to be 0.1 A., then electronic transitions must lie below the 0-ion appearance potential a t obeying the Franck-Condon principle will reach 4.53 e v., and must involve an 0 atom in the ground the lowest vibrational energy levels of those states of or first excited state. No excited states of the 0the 02+ion haying interatomic distances between ion are known so the dissociation limit has been put 1.16 and 1.26 A. The interatomic distances for a t an energy of D(O2) - E ( 0 ) = 3.67 e.v. Kethe various known states20of the 0 2 + ion are given cently Dibeler, Reese and Mann25have observed in Table I where that for the 32g-ground state of 0 2 the 02- ion in the mass spectrum of perchloryl is included for comparison. The interatomic dis- fluoride. This proved that the 02- must be tances of the 411u,WT,and 42g-states of the OZ+ion stable enough to have a life-time greater than 10lie outside the limits of 1.16 and 1.26 A., SO on sec. This 02- ion must have a shallow minimuni these grounds the electron impact ionization poten- in its potential energy curve a t a distance greater tials leading the formation of the 0 2 + ion in these than the internuclear distance of oxygen ; otherwise states, would be expected to be greater than the 02could be observed in the mass spectrum of 0 2 . corresponding spectroscopic or adiabatic ionization (ii) Of and 0- Ions Formed at Higher Enerpotentials. The data in Table I show that this ex- gies.-There are several processes by which atomic pectation is fulfilled. po2itive and negative ions may arise through the The Formation of Atomic Oxygen Ions. (i) electronic bombardment of oxygen molecules. 0 - Ions at Low Energies.-The appearance poten- Table I1 gives the minimum energies which such tial of the 0-ion formed by the resonance capture processes may be expected between zero and 23 e.v. process Above the 0- resonance capture peak, the O f Op+ e- = 0 + 0(2) and 0- ions were first formed a t 17.30 and 17.36 was found to be 4.63 f.0.03 e.v. (see Fig. 4). This is (21) P. Brix and G. Herzberg, J. Chem. P h y s . , 21, 2240 (1953). TABLE I AL)IABATICA N D VERTICALIONIZATION POTENTIALS AND IKTERATOMIC DISTASCESF O R 0 2 AND 0 2 ’
*
+
+
(18) I
