- Email: [email protected]
Field evaporation experiments with a magnetic sector atom-probe FIM
Surface Science 50 (1975) 38-44 0 North-Holland Publishing Company
FIELD EVAPORATION
EXPERIMENTS
WITH A MAGNETIC SECTOR
ATOM-PROBE FIM TOSHIO SAKURAI * and Erwin W. MULLER Physics Department, Pennsylvania State University, University Park, Pennsylvania 16802, US.A. Received 14 February 1975
A magnetic sector atom-probe FIM has been successfully operated for dc field evaporation of tip materials such as Rh, W, Ir, MO and Ti. A limited number of evaporated metal ions were clearly identified forming a line spectrum. Field evaporation of Rh in the presence of ‘He and 4He gases showed that the formation of the helium compound (RhHej2+ is quite sensitive to He gas pressure; no helium compound were observed below 5 X lo- Torr and all ions detected as helium compdund above 5 X lo-* Torr at 78 K.
1. Introduction The atom-probe FIM [l] is an advanced microanalytical tool combining the atomic resolution of a FIM with the ultimately single-ion identification capability of a sensitive mass spectrometer. A desired surface area or even a single atom can be selected by manipulating the tip orientation and placing the desired imaged surface of the specimen over a probe hole on the FIM screen. The surface atoms are field desorbed by temporary increase of surface field and are projected towards the screen. Only those ions passing through the probe hole enter the mass spectrometer and are chemically identified. There are two versions of atom-probes depending upon the mass analyzer employed, one is a time-of-flight (ToF) atom-probe [2] and the other a magnetic sector (MS) atom-probe [3]. The former has been used successfully with pulse field technique for more than five years, achieving many significant discoveries [4]. Realizing the disadvantages associated with pulse field technique, such as premature evaporation [S] , we have developed a MS atom-probe [6] in which dc field evaporation technique can be employed as well as pulse technique. So far our MS atom-probe has been mainly employed for quantitative investigation of the energy distributions of ions produced by field ionization [7,8] because of its unique high energy resolution equivalent to a fractional electron volt, and its capability to analyze the field desorbed surface atom. * Present address: Bell Laboratories, Murray Hill, New Jersey 07974, U.S.A.
T. Sakurai, E. W.MiiIler/Magneticsector atom-probe HM
39
Previously we have briefly reported on palladium field evaporation [3]. We now present the results of evaporation experiments with rhodium in connection with field adsorption of noble gases such as He or Ne. In addition, field evaporation of tungsten, iridium, molybdenum and titanium are discussed.
2. Experimental details 2.1. A MS atom probe The details of our MS atom probe were previously published [ 1,3]. The specific features of this instrument are the following. (1) Using the additional negative electrode independent of the tip voltage, ion energy can be maintained as low as 1OOOeV which yields the very high energy or mass resolution of better than l/ 1000. A conventional ToF atom probe has a mass resolution of no better than l/l00 N l/SO and an energy resolution of about lOeV, although the most recent energy focused ToF [2] has yielded the same high resolution as our MS atom probe. (2) Since the mass or energy identification is carried out by spatial correlation, no trigger pulse is needed. Thus rate-dependent processes can be studied. In order to investigate the field ionization or field desorption of gas atoms, it is ideal to use the dc field technique. An individual ion which hits the active detector surface produces a bright flash on the Chevron channel plate-phosphor screen and thus is easily photographed with an ordinary camera. A long exposure time records many detected ions forming a line spectrum. However, a short exposure time, such as l/8 set, forms the line but enables us to count the number of ions which form the line. 2.2. Rh field evaporation Rhodium is monoisotopic of mass number 103 and is easily identified by slow field evaporation. Rhodium tips were field evaporated in the temperature range of 21 K to 700 K in the presence of helium. Using the dc evaporation technique, the evaporation rate was chosen to be as slow as possible, about 0.2 layers/set in this case. Thus the tip does not change its geometry rapidly and maintains its shape to produce quantitative data. At the early stage of evaporation the probe hole was always set by manipulating a cold finger to cover the edge of the (011) plane. This was found to be an important procedure to locate the exact position of the expected spectrum and adjust the evaporation rate. The (011) plane is the optimum plane for dc field evaporation because each layer of evaporation yields a few tens of metal ions consecutively in a short time interval and forms a series of bright spots forming a clear spectral line, while other planes produce fewer ions randomly. With exposure times of 1 set to
40
T. Sakurai, E. W.MiillerlMagneticsector atom-probe FIM
5 set the spectra were photographed. Then the helium compound of rhodium was studied at 78 K with 3He and 4He. Although the existence of the helium compound ion RhHe 2+ has been established [5,6], its confirmation by dc field evaporation proves that the lifet-time of these compound ions is much longer than that required by ToF atom-probe experiments, namely 3 to 5 nsec. In our system, the time required for ions to travel through the magnetic sector lens, will be more than 2 to 41.tsec, indicating that helide ions are a definitely stable compound. Field evaporation were carried out in vacuum and 4He was introduced gradually. Up to the helium pressure of 5 X 10e7 Torr, no helium compound was formed and doubly charged rhodium Rh2+ ions were detected. With increasing helium gas pressure, the percentage of helium compound becomes appreciable and above 5 X 10h5 Torr all rhodium ions are desorbed as a helide RhHe2+ as shown in fig. 1. The following experiments were performed adding the lighter helium isotope 3He. With 4He gas only, the field evaporation spectrum of rhodium at 78 K consists of two lines. As soon as 3He gas is introduced to the system, an additional line appeared between the original two lines as seen in fig. 2. By mass calibration, the new line was found to be (Rh3He)2+ while the others are Rh2+ and (Rh4He)2+. Increasing the helium gas eliminates the pure rhodium ions Rh2+ completely, leaving the two helium compound lines. In field evaporation in the presence of He-Ne mixture gas, no neon compound was observed, instead a strong spectral line at mass 57.5 amu indicating RhC2+ ap pears with its intensity depending upon the neon pressure. This may be explained by the fact that the evaporation field of rhodium 4.5 V/A may be higher than the desorption field of neon so that no neon exists on a metal surface during the dc field evaporation. Instead, neon atoms release slow electrons as the result of free-space PROBABILITY
oi
Fwd+
_-_____--____--_____-~~-~~-~~-~-
1 t
C&h, /
Fig. 1. Dependence of the field evaporation product RhHe2+ on helium gas pressure. Above 5 X lo-’ Torr of helium there is no pure rhodium evaporated.
T. Sakurai, E. W. MiilIer/Magneticsector atom-probe FIM
a
b
lib’+
51
52
RhH:+
53
54 amu
Fig. 2. Field evaporation spectrum of rhodium at 78 K: (a) in vacuum; (b) in the presence of 4He; (c) in the presence of ‘He and 4He; with increasing He gas pressure, pure Rh2+ disappears (Cl
-+ c2 +
c3).
field ionization which excite surface metal complexes and promote their desorption. In addition to the observation of RhC2+, similar phenomena were noticed in some
42
T. Sakurai, E. W. Miiller/Magnetic sector atom-probe FIM
other occasions, such as the appearance of PdNe+ or Mo02+ compounds in the presence of hydrogen or neon gas. Rhodium field evaporation [l] with ToF atom-probe yields Rh3+ and RhNe2+ in addition to Rh2+ and RhHe2+. The observation of the neon compound RhNe2+ with the pulsed field technique can be attributed to field desorption of the apexadsorbed neon atom at the dc holding field lower than the evaporation field of rhodium. Occurrence of highly charged ions Rh3+ or RhHe3+ which have not been observed [7] in dc field evaporations using MS atom-probe is suspected to be fielddependent, that is, the higher field temporarily produced during the pulse favors the highly charged ions. 2.3. Some other materials Tungsten, iridium, molybdenum and titanium were field evaporated by dc field in order to identify evaporation products. In the case of tungsten which has five isotopes of mass 180, 182, 183, 184 and 186, it was found that slow dc evaporation with a rate of less than 1 layerlsec is extremely difficult to detect. The most abundant isotope, lwW, is only 27.7% and rarely forms a straight line spectrum with this evaporation rate. Thus continuous evaporation with a rate of the order of 10 atomic layers/set was required, whereupon a satisfactorily intense line spectrum was observed at masses 182, 183, 184 and 186 amu, while the rare isotope l*OW with the natural abundance of 0.16% could not be discriminated from noise (fig. 3a). In our experiment, no fourfold charged ion W4+ was detected and the triply charged ion W3+ was abundant with occasional observation of the doubly charged ion W2+ at the temperature 78 K. So far tungsten helide ions have not been identified in our work, though apex-adsorption of helium [9] is well established by the observation of field ionization within the forbidden zone [8, lo]. Using the same procedure as above, iridium, molybdenum and titanium were field evaporated and the spectra shown in figs. 3b, c and d were obtained. Identified ions of field evaporation are listed in table 1, in which the results [ 1 l] of pulsed field evaporation with ToF atom-probe are also cited. Comparison may indicate that field evaporation at fields higher than the evaporation field produces a greater variety of ions than are produced at the evaporation field, and these ions are more highly charged. In this aspect, the MS atom-probe is a more flexible and suitable instrument than the ToF atom-probe in order to investigate field evaporation quantitatively since both the techniques of dc and pulsed field evaporation can be employed. However, it was realized some time during the field evaporation experiments that identification of a very small number of desorbed atoms is not so easy. Because the 75 mm Chevron channel plate detector cannot scan the entire mass range, but scans a limited mass range, occurrence of unexpected compound ions may often be overlooked. Furthermore, it is essential to detect more than a few ions with the same mass to form a “spectral line” since no time correlation is employed and only spatial correlation is used to discriminate signals from noise spots.
T. Sakurai, E. W.h4iiller/Magneticsector atom-probe FIM
W
43
182 186 183 184
3+
C 94 96 96 97 98
n#lo2+
92
1i+
46 -4r
48- 49
10U
66
Fig. 3. Field evaporation spectra of multiple isotope metals: (a) tungsten as W3+; (b) tridium as I:+; (c) molybdenum as MO’+; (d) titanium as Ti+.
In order to improve this situation, the detector should be activated by a pulse synchronized with field evaporation events. This will cut the noise level considerably
Table 1 Identified ions of field evaporation of tip materials (data by ToF atom-probe are in parenthesis) -~ _--_ Helium Neon Vacuum Metal Rhodium
Rhs+ (Rh’+ * Rh3+ , Rh+)
Tungsten
w2+ (w2;.
w3+ w3+,
Rh’+, RhHea+ Rh2+ RhC2+ (Rhs+ t RhHe’+, Rh’+, RhHe3+) (Rh”, Rhs*, RhNe’+) w3+
w4+)
wi^te3+
(9)
w2+
fJv3/ I w4+ 9 W&3+)
w3;,
w3+ iv++,
Iridium
Ir2+ (HZ+, I?+)
HZ+, IrHe2+ (Hz+, Ir3+, IrHe2+, IrHe3+)
I?+ (Ir2+, lr3+, IrNe2+)
~oly~enum
MO+, MO’+
(Mo3+, MoHe’+)
Mo02+
Titanium
Ti+, TiZf
Ti+ 9Ti2+
and enable us to identify rare species of ions in the case of pulsed field evaporation.
3. Conclusion A magnetic sector atom-probe was successfully used for the study of field evaporation of Rh, W, Ir, MOand Ti and each isotope was clearly separated and identified by forming a line spectrum. In the case of Rh, the helium compound RhHe2+ was investigated quantitatively and was found to be sensitive to the helium gas pressure. No helium compound ions were observed below 5 X 10e7 Torr, and ah observed ions were helium compound ions above 5 X 10w5 Torr. By comparison with the results of the pulsed field evaporation with a ToF atom-probe, it was concluded that the occurrence of highly charged ions is field dependent [ 121 and is e~anced as the applied field increases.
References [ 11 E.W. MtiBer and T.T. Tsong, Progr. Surface Sci. 4 (1973) 1. [2] E.W. Miilier and S.V. ~s~aswamy, Rev. Sci. Irish. 45 (1974) 1053. [3] E.W. Mtiiler and T. Sakurai, J. Vacuum Sci. Technd. 11 (1974) $78. [4] E.W. MiiBer, Quart. Rep~(~ondon) 23 (1969) 177; see also ref. [I]. [S] E.W. Mtiher, Ber. Bunsenges Physik. Chem. 75 (1971) 971. [ 61 E.W. Mtiller and T. Sakurai, 19th Field Emission Symposium, Urbana, Ill. (1972). [7] T. Sakurai and E.W. Miiller, whys. Rev. Letters 30 (1973) 532. (81 T. Sakurai and E.W. Mtiller, Surface Sci 49 (1975) 497. [9] T.T. Tsong and E.W. hitiller, J. Chem. whys. 55 (1971) 2884. [ 101 E.W. Mtiller and S.V. Krirhnaswamy, Surface Sci. 36 (1973) 29. [ll] Ref. [l],p.46. [ 121 J.A. Panitz, J. Vacuum Sci. Technol. 11 (1974) 206.
FIELD EVAPORATION
EXPERIMENTS
WITH A MAGNETIC SECTOR
ATOM-PROBE FIM TOSHIO SAKURAI * and Erwin W. MULLER Physics Department, Pennsylvania State University, University Park, Pennsylvania 16802, US.A. Received 14 February 1975
A magnetic sector atom-probe FIM has been successfully operated for dc field evaporation of tip materials such as Rh, W, Ir, MO and Ti. A limited number of evaporated metal ions were clearly identified forming a line spectrum. Field evaporation of Rh in the presence of ‘He and 4He gases showed that the formation of the helium compound (RhHej2+ is quite sensitive to He gas pressure; no helium compound were observed below 5 X lo- Torr and all ions detected as helium compdund above 5 X lo-* Torr at 78 K.
1. Introduction The atom-probe FIM [l] is an advanced microanalytical tool combining the atomic resolution of a FIM with the ultimately single-ion identification capability of a sensitive mass spectrometer. A desired surface area or even a single atom can be selected by manipulating the tip orientation and placing the desired imaged surface of the specimen over a probe hole on the FIM screen. The surface atoms are field desorbed by temporary increase of surface field and are projected towards the screen. Only those ions passing through the probe hole enter the mass spectrometer and are chemically identified. There are two versions of atom-probes depending upon the mass analyzer employed, one is a time-of-flight (ToF) atom-probe [2] and the other a magnetic sector (MS) atom-probe [3]. The former has been used successfully with pulse field technique for more than five years, achieving many significant discoveries [4]. Realizing the disadvantages associated with pulse field technique, such as premature evaporation [S] , we have developed a MS atom-probe [6] in which dc field evaporation technique can be employed as well as pulse technique. So far our MS atom-probe has been mainly employed for quantitative investigation of the energy distributions of ions produced by field ionization [7,8] because of its unique high energy resolution equivalent to a fractional electron volt, and its capability to analyze the field desorbed surface atom. * Present address: Bell Laboratories, Murray Hill, New Jersey 07974, U.S.A.
T. Sakurai, E. W.MiiIler/Magneticsector atom-probe HM
39
Previously we have briefly reported on palladium field evaporation [3]. We now present the results of evaporation experiments with rhodium in connection with field adsorption of noble gases such as He or Ne. In addition, field evaporation of tungsten, iridium, molybdenum and titanium are discussed.
2. Experimental details 2.1. A MS atom probe The details of our MS atom probe were previously published [ 1,3]. The specific features of this instrument are the following. (1) Using the additional negative electrode independent of the tip voltage, ion energy can be maintained as low as 1OOOeV which yields the very high energy or mass resolution of better than l/ 1000. A conventional ToF atom probe has a mass resolution of no better than l/l00 N l/SO and an energy resolution of about lOeV, although the most recent energy focused ToF [2] has yielded the same high resolution as our MS atom probe. (2) Since the mass or energy identification is carried out by spatial correlation, no trigger pulse is needed. Thus rate-dependent processes can be studied. In order to investigate the field ionization or field desorption of gas atoms, it is ideal to use the dc field technique. An individual ion which hits the active detector surface produces a bright flash on the Chevron channel plate-phosphor screen and thus is easily photographed with an ordinary camera. A long exposure time records many detected ions forming a line spectrum. However, a short exposure time, such as l/8 set, forms the line but enables us to count the number of ions which form the line. 2.2. Rh field evaporation Rhodium is monoisotopic of mass number 103 and is easily identified by slow field evaporation. Rhodium tips were field evaporated in the temperature range of 21 K to 700 K in the presence of helium. Using the dc evaporation technique, the evaporation rate was chosen to be as slow as possible, about 0.2 layers/set in this case. Thus the tip does not change its geometry rapidly and maintains its shape to produce quantitative data. At the early stage of evaporation the probe hole was always set by manipulating a cold finger to cover the edge of the (011) plane. This was found to be an important procedure to locate the exact position of the expected spectrum and adjust the evaporation rate. The (011) plane is the optimum plane for dc field evaporation because each layer of evaporation yields a few tens of metal ions consecutively in a short time interval and forms a series of bright spots forming a clear spectral line, while other planes produce fewer ions randomly. With exposure times of 1 set to
40
T. Sakurai, E. W.MiillerlMagneticsector atom-probe FIM
5 set the spectra were photographed. Then the helium compound of rhodium was studied at 78 K with 3He and 4He. Although the existence of the helium compound ion RhHe 2+ has been established [5,6], its confirmation by dc field evaporation proves that the lifet-time of these compound ions is much longer than that required by ToF atom-probe experiments, namely 3 to 5 nsec. In our system, the time required for ions to travel through the magnetic sector lens, will be more than 2 to 41.tsec, indicating that helide ions are a definitely stable compound. Field evaporation were carried out in vacuum and 4He was introduced gradually. Up to the helium pressure of 5 X 10e7 Torr, no helium compound was formed and doubly charged rhodium Rh2+ ions were detected. With increasing helium gas pressure, the percentage of helium compound becomes appreciable and above 5 X 10h5 Torr all rhodium ions are desorbed as a helide RhHe2+ as shown in fig. 1. The following experiments were performed adding the lighter helium isotope 3He. With 4He gas only, the field evaporation spectrum of rhodium at 78 K consists of two lines. As soon as 3He gas is introduced to the system, an additional line appeared between the original two lines as seen in fig. 2. By mass calibration, the new line was found to be (Rh3He)2+ while the others are Rh2+ and (Rh4He)2+. Increasing the helium gas eliminates the pure rhodium ions Rh2+ completely, leaving the two helium compound lines. In field evaporation in the presence of He-Ne mixture gas, no neon compound was observed, instead a strong spectral line at mass 57.5 amu indicating RhC2+ ap pears with its intensity depending upon the neon pressure. This may be explained by the fact that the evaporation field of rhodium 4.5 V/A may be higher than the desorption field of neon so that no neon exists on a metal surface during the dc field evaporation. Instead, neon atoms release slow electrons as the result of free-space PROBABILITY
oi
Fwd+
_-_____--____--_____-~~-~~-~~-~-
1 t
C&h, /
Fig. 1. Dependence of the field evaporation product RhHe2+ on helium gas pressure. Above 5 X lo-’ Torr of helium there is no pure rhodium evaporated.
T. Sakurai, E. W. MiilIer/Magneticsector atom-probe FIM
a
b
lib’+
51
52
RhH:+
53
54 amu
Fig. 2. Field evaporation spectrum of rhodium at 78 K: (a) in vacuum; (b) in the presence of 4He; (c) in the presence of ‘He and 4He; with increasing He gas pressure, pure Rh2+ disappears (Cl
-+ c2 +
c3).
field ionization which excite surface metal complexes and promote their desorption. In addition to the observation of RhC2+, similar phenomena were noticed in some
42
T. Sakurai, E. W. Miiller/Magnetic sector atom-probe FIM
other occasions, such as the appearance of PdNe+ or Mo02+ compounds in the presence of hydrogen or neon gas. Rhodium field evaporation [l] with ToF atom-probe yields Rh3+ and RhNe2+ in addition to Rh2+ and RhHe2+. The observation of the neon compound RhNe2+ with the pulsed field technique can be attributed to field desorption of the apexadsorbed neon atom at the dc holding field lower than the evaporation field of rhodium. Occurrence of highly charged ions Rh3+ or RhHe3+ which have not been observed [7] in dc field evaporations using MS atom-probe is suspected to be fielddependent, that is, the higher field temporarily produced during the pulse favors the highly charged ions. 2.3. Some other materials Tungsten, iridium, molybdenum and titanium were field evaporated by dc field in order to identify evaporation products. In the case of tungsten which has five isotopes of mass 180, 182, 183, 184 and 186, it was found that slow dc evaporation with a rate of less than 1 layerlsec is extremely difficult to detect. The most abundant isotope, lwW, is only 27.7% and rarely forms a straight line spectrum with this evaporation rate. Thus continuous evaporation with a rate of the order of 10 atomic layers/set was required, whereupon a satisfactorily intense line spectrum was observed at masses 182, 183, 184 and 186 amu, while the rare isotope l*OW with the natural abundance of 0.16% could not be discriminated from noise (fig. 3a). In our experiment, no fourfold charged ion W4+ was detected and the triply charged ion W3+ was abundant with occasional observation of the doubly charged ion W2+ at the temperature 78 K. So far tungsten helide ions have not been identified in our work, though apex-adsorption of helium [9] is well established by the observation of field ionization within the forbidden zone [8, lo]. Using the same procedure as above, iridium, molybdenum and titanium were field evaporated and the spectra shown in figs. 3b, c and d were obtained. Identified ions of field evaporation are listed in table 1, in which the results [ 1 l] of pulsed field evaporation with ToF atom-probe are also cited. Comparison may indicate that field evaporation at fields higher than the evaporation field produces a greater variety of ions than are produced at the evaporation field, and these ions are more highly charged. In this aspect, the MS atom-probe is a more flexible and suitable instrument than the ToF atom-probe in order to investigate field evaporation quantitatively since both the techniques of dc and pulsed field evaporation can be employed. However, it was realized some time during the field evaporation experiments that identification of a very small number of desorbed atoms is not so easy. Because the 75 mm Chevron channel plate detector cannot scan the entire mass range, but scans a limited mass range, occurrence of unexpected compound ions may often be overlooked. Furthermore, it is essential to detect more than a few ions with the same mass to form a “spectral line” since no time correlation is employed and only spatial correlation is used to discriminate signals from noise spots.
T. Sakurai, E. W.h4iiller/Magneticsector atom-probe FIM
W
43
182 186 183 184
3+
C 94 96 96 97 98
n#lo2+
92
1i+
46 -4r
48- 49
10U
66
Fig. 3. Field evaporation spectra of multiple isotope metals: (a) tungsten as W3+; (b) tridium as I:+; (c) molybdenum as MO’+; (d) titanium as Ti+.
In order to improve this situation, the detector should be activated by a pulse synchronized with field evaporation events. This will cut the noise level considerably
Table 1 Identified ions of field evaporation of tip materials (data by ToF atom-probe are in parenthesis) -~ _--_ Helium Neon Vacuum Metal Rhodium
Rhs+ (Rh’+ * Rh3+ , Rh+)
Tungsten
w2+ (w2;.
w3+ w3+,
Rh’+, RhHea+ Rh2+ RhC2+ (Rhs+ t RhHe’+, Rh’+, RhHe3+) (Rh”, Rhs*, RhNe’+) w3+
w4+)
wi^te3+
(9)
w2+
fJv3/ I w4+ 9 W&3+)
w3;,
w3+ iv++,
Iridium
Ir2+ (HZ+, I?+)
HZ+, IrHe2+ (Hz+, Ir3+, IrHe2+, IrHe3+)
I?+ (Ir2+, lr3+, IrNe2+)
~oly~enum
MO+, MO’+
(Mo3+, MoHe’+)
Mo02+
Titanium
Ti+, TiZf
Ti+ 9Ti2+
and enable us to identify rare species of ions in the case of pulsed field evaporation.
3. Conclusion A magnetic sector atom-probe was successfully used for the study of field evaporation of Rh, W, Ir, MOand Ti and each isotope was clearly separated and identified by forming a line spectrum. In the case of Rh, the helium compound RhHe2+ was investigated quantitatively and was found to be sensitive to the helium gas pressure. No helium compound ions were observed below 5 X 10e7 Torr, and ah observed ions were helium compound ions above 5 X 10w5 Torr. By comparison with the results of the pulsed field evaporation with a ToF atom-probe, it was concluded that the occurrence of highly charged ions is field dependent [ 121 and is e~anced as the applied field increases.
References [ 11 E.W. MtiBer and T.T. Tsong, Progr. Surface Sci. 4 (1973) 1. [2] E.W. Miilier and S.V. ~s~aswamy, Rev. Sci. Irish. 45 (1974) 1053. [3] E.W. Mtiiler and T. Sakurai, J. Vacuum Sci. Technd. 11 (1974) $78. [4] E.W. MiiBer, Quart. Rep~(~ondon) 23 (1969) 177; see also ref. [I]. [S] E.W. Mtiher, Ber. Bunsenges Physik. Chem. 75 (1971) 971. [ 61 E.W. Mtiller and T. Sakurai, 19th Field Emission Symposium, Urbana, Ill. (1972). [7] T. Sakurai and E.W. Miiller, whys. Rev. Letters 30 (1973) 532. (81 T. Sakurai and E.W. Mtiller, Surface Sci 49 (1975) 497. [9] T.T. Tsong and E.W. hitiller, J. Chem. whys. 55 (1971) 2884. [ 101 E.W. Mtiller and S.V. Krirhnaswamy, Surface Sci. 36 (1973) 29. [ll] Ref. [l],p.46. [ 121 J.A. Panitz, J. Vacuum Sci. Technol. 11 (1974) 206.