I. SOFT X-RAY SPECTROSCOPY

advertisement
SOFT X-RAY SPECTROSCOPY
I.
Y. Laohavanich
J. B. Lastovka
Prof. K. W. Billman
Prof. G. G. Harvey
VOLUMETRIC PHOTOELECTRON INVESTIGATION OF NICKEL
A.
A system has
in
an
and
source,
of a soft
essentially
consists
and
of solids
properties
electric
beams
atomic
a
monochromator,
x-ray
analyzer,
electron-energy
electrostatic
the
in
of photo-
investigation
for the
and tested
been constructed
It
region.
x-ray
soft
glow-discharge
radiation
schematically
as shown
Fig. I-i.
1
The monochromator is the soft x-ray spectrometer described by Piore and others
No major modi-
which was removed from storage for the purpose of this investigation.
fications have been made in this apparatus.
The radiation source is a Schiller lamp or hollow-cathode discharge tube similar to
that described by Newburgh and others.
2
electron-energy analyzer is
3
If the angle
cussed by Harrower.3
of the plane
The
there
is
are
widths
slit
Harrower's
The
cent.
half-wave
tron
in
The
this
resolution
has
and biasing
and the
gives
an
off
were
energy resolution
during the
from
detected
by a
heating
the input
and exit
to 45* ± 4o.
a
In this way,
accelerating
voltages;
filament
Allen type
was
of Be-Cu
the
thus
By
5 per
filament
the elec -
distribution that was
across the
by use of an
o
to 45 ,
close
of approximately
half-cycle.
Maxwellian
drop
is
experimentally by heating
for the modest
voltage
analyzer
restricted
angle is
input
type dis-
capacitor
device
present
In the
been confirmed
resolution
resulting
electrons
it
of input into this
action.
characterized
were
analyzer
energies
inch,
0. 0625
analysis,
energies
than the
focussing
an angular
parallel
smaller
spread
avoided.
electron
mul-
tiplier.
The analyzer system was enclosed in a double-shielded mu-metal box to eliminate
magnetic fields.
The box traversed the Rowland circle under the action of a lathe lead
screw turned externally by means of a rotary vacuum seal.
In order to check out the system, a determination of the distribution of energies of
photoelectrons from a gassy nickel target was made.
used to excite the volumetric photoelectrons.
A single 304 A helium line was
The energies of the volumetric photo-
electrons emitted from the nickel were measured.
Because of the low transmission of the electron-energy analyzer at very low energies, accelerating voltages of approximately 10 volts were applied between the target
and the input of the analyzer.
Admittedly, this distorts the energy distribution some-
what, but no significant difference was observed between the distributions obtained at,
say, 5 and 20 volts.
QPR No. 67
(I.
SOFT X-RAY SPECTROSCOPY)
ELECTRON
ENERGY
ELECTRON
MULTIPLIER
TO
DIFFUSION .
PUMP
DOVE - TAIL
TO
MECHANICAL
PUMP
TRACK
KOVAR
ANODE
HOLLOW
GRAPHITE
CATHODE
Fig. I-1.
Photoelectron spectrometer mounted inside
the soft x-ray spectrometer.
A typical photoelectron energy distribution is shown in Fig. I-2. The important
features are that, despite a photon energy of 44. 4 ev, the peak of the photoelectron
energy distribution is at -2 ev, that is, at a total energy after acceleration of 11. 2 ev,
and that no electrons are detected which have energies greater than -10 ev above the
noise.
All such distributions, of which 19 were taken, have a sharp edge on the low-energy
30
304A RADIATI ON
Z
B9.2V
APPLIED, ACCELERATING VOLTAGE
2 O/o PROBABL E ERROR
t
20
z
10
0
I L.
8.0
I
9.0
10.0
i
11.0
I
12.0
I
13.0
I
14.0
I
15.0
I
16.0
ELECTRON ENERGY (e v )
Fig. 1-2.
QPR No. 67
Photoelectron energy spectrum
of nickel photocathode.
I
I
17.0
Il
l
I
18.0
19.0
(I.
side,
presumably at zero initial electron energy.
SOFT X-RAY SPECTROSCOPY)
This edge was used to calibrate the
analyzer and evaluate the effect of contact potential that was, incidentally, zero within
±0. 1 ev.
The only previous measurements of extreme ultraviolet photoelectron energy dis4,5
tribution from nickel were made with a planar geometry, stopping-potential apparatus.
Such a device,
the kine-
at best, only gives the "normal" energy distribution, that is,
tic enegry associated with motion normal to the surface.
Furthermore, these measure-
ments were made with an electron multiplier placed behind a coarse retarding grid.
This obviously allows field distortion between the grid wires, causing a parabolic potential distribution that may dip as much as 25 volts, although the report does not give
sufficient data on which to base a precise calculation.
In any event, it is not surprising
that they get identical photoelectron distributions with nickel and tungsten.
However,
their results do agree with ours in that they get very few electrons above 5 ev.
PLASMA LOSS, 19.5 ev
PLASMA LOSS, 19.5 ev
REMAINING ENERGY,
- lev
hv= 44.4 ev
WORK FUNCTION,- 4.5 ev
FERMI LEVEL
OF VALENCE
SWIDTH
5.0 ev
BAND,
HIGHEST DISCRETE LEVEL,
66 ev BELOW
-,
M
FERMI LEVEL
Fig. 1-3.
Level diagram of nickel
(not drawn to scale).
The observed results can be explained on the basis of "collective" electron-electron
The width of the
interactions, which give rise to the "plasma" oscillations in metals.
valence band in nickel is 5. 0 ev and the work function is approximately 4.5 ev. (See
6
Fig. 1-3.) From the energy of the M2,3 absorption edge, it is seen that the 3p levels
are approximately 66 ev below the Fermi level.
QPR No. 67
Thus, all photoemission from the
(I.
SOFT X-RAY SPECTROSCOPY)
304 A (44.4 ev) photons must come from the valence band.
The mean-free path for excitation of plasma oscillations is considerably shorter
than the photon penetration depth.
Thus, a photoelectron has many opportunities to
excite plasma oscillations on its way out of the metal.
Robins and Swan 7 list the char-
acteristic electron-energy losses of interest to the present discussion as occurring at
4.3,
8.3,
and 19.5 ev.
These are due to M4,
5
shell ionization,
a lowered plasma loss,
and a normal plasma loss, respectively.
The lowered plasma loss is an effect of the fact that the specimen boundary permits
another mode of oscillation distinct from the normal mode.
Of course, this is very sen-
sitive to surface conditions and contamination and hence cannot be discussed intelligently in the present case,
since very little target preparation was attempted.
such preparation would have greatly reduced the quantum yield.
Indeed,
4
It is seen that an electron liberated from the Fermi level by a 304 A photon and
escaping the surface after exciting two normal plasma oscillations will have approximately 1 ev of energy remaining.
This is in good accord with the peak of the distri-
butions observed here, and hence that mechanism is postulated as being primarily
responsible for the low-energy photoelectron distribution observed.
H.
H. Barrett, K. W.
Billman
References
1. E. R. Piore, G.
Instr. 23, 8 (1952).
G. Harvey,
E. M. Gyorgy,
and R.
2. R. Newburgh, L. Heroux, and H. E. Hinteregger,
the extreme ultraviolet (to be published in Appl. Opt.).
3.
G. A. Harrower, Rev. Sci. Instr. 26,
4.
H. E.
Hinteregger and K. Watanabe,
H. Kingston,
Rev.
Sci.
Two light sources for use in
850-854 (1955).
J.
Opt. Soc. Am.
43,
604 (1953).
5. L. Heroux and H. E. Hinteregger, Photoelectron energy distribution measurements for 10 to 50 ev photons with a planar analyzer and electron multiplier (to be
published in Appl. Opt.).
6. D. H. Tomboulian, Handbuch der Physik, Vol.
1957), pp. 243-304.
7.
J.
L. Robins and J.
QPR No. 67
30 (Springer
B. Swan, Proc. Phys. Soc. (London) 16,
Verlag, Berlin,
857 (1960).
Download