United States Patent 3805068

Method and apparatus for energy analysis of a stream of moving electrons by effecting electrostatic segregation and counting of an electron portion having a preselected kinetic energy.

Application Number:
Publication Date:
Filing Date:
Primary Class:
Other Classes:
International Classes:
G01Q60/00; H01J49/48; (IPC1-7): H01J37/26
Field of Search:
View Patent Images:

Primary Examiner:
Borchelt, Archie R.
Assistant Examiner:
Church C. E.
Parent Case Data:


This application is a continuation-in-part of U.S. application Ser. No. 99,475 filed Dec. 18, 1970, in the name of the same applicant and now abandoned.
1. A method for determining the energy of a preselected fraction of electrons contained within a moving stream of charged particles having different kinetic energies emanating from a common source, comprising,

2. A cylindrical energy analyzer for a moving stream of charged particles having different kinetic energies, comprising in combination:

3. The energy analyzer of claim 3 wherein said charged particles are

4. The energy analyzer of claim 4 wherein said filter means comprises two

5. The energy analyzer of claim 5 further comprising a beam stop positioned on the axis of the analyzer between said inlet means and said detector to prevent high energy charged particles from flowing directly from said

6. The energy analyzer of claim 5 wherein said filter means comprises two pairs of spherical grids, the first pair being concave and the second pair

7. The energy analyzer of claim 7 further comprising a beam stop positioned on the axis of the analyzer between said inlet means and said detector to prevent high energy charged particles from flowing directly from said

8. A cylindrical energy analyzer for a moving stream of charged particles having different kinetic energies, comprising in combination:

9. The energy analyzer of claim 9 wherein said charged particles are

10. The energy analyzer of claim 10 further comprising enhancing means to enhance the radial outwardly directed component of the velocity of said

11. The energy analyzer of claim 11 wherein said enhancing means is a charged needle positioned coaxially with respect to said collection means.

12. The energy analyzer of claim 11 further comprising a trap for those electrons which pass through said filter means and are not collected by

13. The energy analyzer of claims 10 further comprising an electrostatically charged prefilter, located before said filter means, to remove from the stream of electrons those electrons having kinetic

14. The energy analyzer of claims 10 wherein said filter means comprises a multiplicity of coaxially disposed metal rings maintained in mutual electrical isolation one from another at predetermined electrical potential levels decreasing from positive to negative in the direction of

15. The energy analyzer of claim 15 wherein the electrons in said moving stream of charged particles have both axial and radial velocity components and wherein said filter means also functions as said enhancing means by retarding the axial component of the velocity of said charged particles without retarding their radial components, whereby said electrons drift towards said collection means through a substantially field free region.

16. The energy analyzer of claims 16 wherein said collection means comprises a metallic ring collector disposed coaxially with respect to said filter means and wherein said filter means and said means to enhance combine to cause said electrons to drift radially outward to said metallic

17. The energy analyzer of claim 17 wherein said detection means is an electron multiplier provided with a sensing connection to said ring collector.


Generally, this invention comprises method and apparatus for energy analysis of a moving stream of electrons having different kinetic energies comprising, in seriatim, constraining a stream of electrons having diverse kinetic energies to a given flow path, applying to the moving stream an electrostatic repulsion field barring forward longitudinal passage of substantially all of the charged particles having kinetic energies below a first preselected energy level while permitting forward passage of the remainder of the charged particles, segregating by transverse drift and counting the fraction of the electrons having a preselected range of kinetic energies which it is desired to analyze for from the remainder of the moving stream, and ejecting the remainder of the moving electron stream.


The following drawings depict three embodiments of the invention as applied to electron spectroscopy, in which:

FIG. 1A is a plot of transmission current v. energy for the photoelectron stream emitted when a sample is bombarded with X-rays,

FIG. 1B is a plot of transmission current v. energy characteristic of a narrow-band energy-pass filter at an arbitrary energy Eo,

FIG. 2A is a schematic side elevation cross-sectional view of a typical prior art hemispherical electrostatic analyzer,

FIG. 2B is a plot of transmission current v. energy characteristic obtained with energy analyzers of this invention,

FIG. 3A is a schematic representation of a first preferred embodiment of apparatus according to this invention,

FIG. 3B is a diagrammatic representation of the effective regions of equipotential distribution between paired sections of the apparatus of FIG. 3A,

FIG. 4A is a transmission v. energy characteristic curve for a narrow-band energy pass apparatus such as that illustrated in FIG. 3A,

FIG. 4B is the transmission v. energy characteristic curve of FIG. 4A showing, in broken-line representation, the effect of a prefilter on the narrow band-pass output of FIG. 2B in removing the broad wing tail of FIG. 4A,

FIG. 5A is a schematic side elevation sectional view of a preferred design of prefilter,

FIG. 5B is a schematic representation of trajectories of typical photoelectrons passed by the prefilter of FIG. 5A,

FIG. 6 is a partially schematic side-elevation sectional view, in detail, of a first preferred embodiment of energy analyzer according to this invention shown in operational association with the X-ray-bombarded sample of an electron spectrometer, not otherwise detailed,

FIGS. 7A and 7B is a typical spectrum records (A) obtained in the analysis of a gold sample using apparatus constructed according to the first embodiment of this invention in comparison with a spectrum record (B) obtained with typical apparatus of the prior art,

FIG. 8 is a schematic longitudinal section view of a second embodiment of apparatus according to this invention, and

FIG. 9 is a schematic longitudinal section view of a third embodiment of apparatus according to this invention.


Electron analysis according to this invention can be applied to a wide variety of situations; however, it is particularly useful in the conduct of electron spectroscopy and, accordingly, is hereinafter described in particular application to this technique.

Electron spectroscopy for chemical analysis is a comparatively new procedure which has been described extensively in the article entitled "Electron Spectroscopy for Chemical Analysis (ESCA)" by Kai Siegbahn et al., Uppsala University, Uppsala, Sweden (1968) October (Processed for the Defense Supply Agency by the Clearinghouse for Scientific and Technical Information with the identification number AD-844-315).

Briefly, electron spectroscopy is the study of the energy (velocity) distribution of secondary electrons (photoelectrons) emitted by a sample upon irradiation of the sample by a primary energy source, such as a beam of X-rays. The operation is conducted by an electron spectrometer having a radiation source for exciting a sample, means for analyzing the velocities (energies) of the secondary electrons released due to the excitation, and means for recording electron energy vs. the quantity (current) of electrons falling within small increments of energy. The apparatus utilizes high vacuum pumps, a high-voltage source, an X-ray or other emitter of exciting energy, a sample module or holder, an energy analyzer, and a readout device such as an X-Y recorder.

ESCA has broad application to the analysis of the full range of individual chemical elements, even in the presence of other elements, and is particularly effective in organic chemistry, since the chief constituent elements carbon, nitrogen, oxygen, etc., are relatively easy to study. In addition, electron spectroscopy is better suited than X-ray analysis for studying the atomic structure of surfaces, because the secondary electrons (as contrasted with "secondary" X-rays), are emitted only from a surface layer 100A or less in thickness. Thus, information on composition, bonding states and the like peculiar to the surface exclusively is readily obtainable using this tool.

When a sample under analysis is irradiated from a primary source, the sample emits photoelectrons is essentially random directions and at velocities (energies) unique to the specific electron-level structure of the atoms in the sample. To be of value in chemical analysis, these photoelectrons must be categorized with respect to their energies and the number of electrons emitted in each energy category determined over a given interval of time. This categorizing is effected by an energy analyzer, such as the designs provided by this invention.

A successful analyzer must (a) provide high resolution, i.e., separation of electron fractions of closely adjacent energies and (b) provide high sensitivity, i.e., measureable and representative readouts for each small energy increment. Also, the analyzer must accommodate a high electron throughput, or luminosity, so that electron energy categorization can be accomplished within a relatively short time interval.

The energy distribution spectra of photoelectrons produced by X-ray excitation is such that the electrons which characterize a specific element lie at a particular energy level and, in general, are manifested as discrete maxima resting upon a background having a broad distribution of energy. This background exists because electrons which would otherwise have discrete energies which characterize the element (or sample) have undergone collisions within the sample and thus have lost varying amounts of energy. Another source of background current is the exciting X-ray background (bremsstrahlung), which is superimposed upon the desired exciting X-rays of narrow energy distribution (characteristic X-rays). A typical energy distribution of electrons emitted by a sample, including the unavoidable background, is represented by FIG. 1A.

Since it is desired to measure only the intensity and shape of the photoelectron peak representing a discrete energy difference, both the sensitivity and accuracy of measurement are reduced proportionately if any of the background other than that directly under the peak is measured. This occurs because the statistical variation in the number of electrons arising from the broad background is large. For this reason, it is preferred to measure the current (representative of the number of electrons in that narrow energy band) within a photoelectron peak with a narrow-band energy pass apparatus. FIG. 1B shows the desired transmission characteristics of such a narrow-band energy pass device at an arbitrary energy Eo. One such narrow-band pass device of the prior art which accomplishes this type of discrimination is a hemispherical electrostatic analyzer, illustrated schematically in side elevation cross-section in FIG. 2A. Only those electrons with energy Eo = eERo /2 can pass both slits, where E is the electric field produced by the potential difference V, e is the charge on an electron and Ro is the radius of the hemisphere including the slits.


Conceptually, it is the purpose of this invention to provide means which affords a transmission v. energy characteristic which can be visualized as a conjoined filter and cutoff means preselected so that the energy passed by the filter is precisely at the trailing edge of the maxima region corresponding to Eo, whereas the energy level preselected by the cutoff means is precisely at the leading edge of the maxima region corresponding to Eo, all as shown in FIG. 2B.

Referring to FIG. 3A, there is shown, for purposes of explanation, an over-simplified schematic representation of a first embodiment of apparatus according to this invention comprising two cylindrical co-axially aligned metallic tube sections 10 and 11, which collectively constitute an electrostatic high-pass filter, followed by a collector ring 11a, i.e., the cutoff means, which demonstrates the principle of the invention. The individual sections are electrically isolated one from another and are maintained at preselected electrostatic potentials through taps from the voltage source 14. In addition, the negative potential side of source 14 is connected through microammeter 16 to ring 11a.

The course of the moving photoelectron stream is shown as proceeding from left to right along the line Z in FIG. 3A, corresponding to the common longitudinal axis of the sections 10 and 11, and ring 11a.

The equipotential distribution in terms of the percent difference in potential between paired tubular sections, such as sections 10 and 11 of FIG. 3A, is shown diagrammatically in FIG. 3B.

Now, if the electrostatic potential of section 11 is preselected with respect to section 10 at a negative level repelling all photoelectrons having energies below a given level, a first separation corresponding to the action of the filter, i.e., the high energy pass filter of FIG. 2B, is obtained. Good operation is obtained if ring 11a is maintained at the same potential as section 11, or within ±1/2 percent thereof. Under these circumstances photoelectrons with energies lying in the maxima corresponding substantially to Eo are deflected into contact with the metal wall of ring 11a, and the fraction of photoelectrons of analytical interest is segregated. The number of photoelectrons in this segregated fraction is measured as the electrical current passed through microammeter 16.

The photoelectrons completely escaping ring 11a are ejected from the apparatus unanalyzed.

From the foregoing, it will be apparent that the width of ring 11a, i.e., the dimension coparallel with the longitudinal axes of sections 10 and 11, determines, together with the ring potential, the energy rnage of the fraction of photoelectrons measured. As will become more clear as this description proceeds, the incoming stream of photoelectrons passing from left to right through sections 10 and 11, as seen in FIG. 3A, is divergent, and the transverse velocity component thereof, denoted vt, is the effective force radially segregating the photoelectron fraction to be measured. Two other embodiments of apparatus hereinafter described operate by effectively counterbalancing the divergent transverse velocity component vt and substituting therefore a convergent component vt ', thereby effecting axial segregation of the electron fraction of interest.

The transmission v. energy characteristic of a narrow-band energy pass device such as that of FIG. 3A is shown in FIG. 4A, and it is seen that there is a relatively broad wing extending forward from the leading edge. This wing contributes to inaccuracy in the photoelectron peak measurement, unless it is eliminated.

Referring to FIG. 4B, it is seen that the wing can be largely eliminated by use of a low-energy pass prefilter having the transmission characteristic denoted in broken line representation, which substantially "cuts off" the wing in close adjacency with the maxima of interest without, however, affecting the latter in any way.

A prefilter of novel design is detailed in FIG. 5A and consists of a metal plate 19, maintained at a positive potential (typically, +300 volts) and apertured via metal screens 19a and 19b to, respectively, admit and eject the electron stream, such as the gross photoelectron output from an X-ray sample bombardment chamber (not shown). This is followed by a coparallel metal screen 20 maintained at a negative potential with respect to plate 19 (typically, +90 volts), spaced at a distance h (typically, 1/4 in.) from plate 19. Screens 19a, 19b and 20 are all of the same mesh size (approximately USS No. 28) wherein the nickel wire diameter is 0.00054 inch and the openings 0.04946 inch on a side, providing 97 percent open area. The prefilter is disposed across the beam of electrons at a bias angle of θ, typically 45°, under which conditions the fraction of electrons having energies less than a cutoff value Eo entering this prefilter is "brought to a focus" by the field between screen 20 and plate 19 at a distance Xo (which is the center-to-center distance between apertures 19a and 19b) along plate 19 from the point of electron beam entry, as shown in FIG. 5A. At θ = 45°, Xo has a maximum value.

From FIG. 5B it can be seen that the height to which the focused beam rises is Xo /4. Thus, screen 20 is spaced from plate 19 by an amount h = Xo /4, in which case the trajectories of electrons having energies E≤Eo pass under or just graze the surface of screen 20. In actuality, the beam entering aperture 19a will always have some divergence dθ, in which case, the cutoff will no longer occur at sharply Eo. At θ = 45°, it can be shown that the cut-off-energy spread dE, from Eo, that occurs from a beam divergence dθ is given by dE = 2 Eo dθ. To accommodate beam sizes passing greater flux, apertures 19a and 19b can be opened to a width of some value less than Xo, typically √2/2Xo.

The bulk of the high energy (E>Eo) photoelectrons pass through the meshes of screen 20 and thus are discarded to avoid disturbance of the subsequent analysis. However, the photoelectron stream of analytical interest, inclusive of a few high energy photoelectrons, is deflected to the right and is passed via exit aperture 19b to the inlet end of the analyzer, i.e., the inlet end of section 10, FIG. 3A.

Referring to FIG. 6, there is shown, in partial representation only, an electron spectrometer, evacuated throughout to about 10-6 mm. absolute, wherein the sample 25 irradiated is mounted within a cylinder 37, which can be thin-walled aluminum of, typically, 3 micron thickness maintained at a positive potential of 0 to 1,500 volts. The exit end of cylinder 37 is covered by a metal screen 36, which can be the same mesh size as screens 19a, 19b and 20, through which the photoelectrons pass along line Z1 to the prefilter, hereinbefore described, denoted generally at 26. In this design, inlet aperture 19a is covered by metal screen 27, whereas exit aperture 19b is similarly covered by screen 28.

With prefilter 26 inclined at an angle of 45° towards the energy analyzer, denoted generally at 30, and with a +90 volt potential carried on screen 20, the photoelectron stream of analytical interest is deflected approximately 90° clockwise to take the horizontal course Z2, at which it proceeds generally along the analyzer longitudinal axis. Actually, the photoelectrons of analytical interest, as well as the residue of higher energies finally ejected, invariably possess a substantial transverse velocity component which, over the relatively short length of tube 34, does not bring them into contact with the tube sidewalls but does, ultimately, effect segregation of the photoelectron fraction of analytical interest, all as hereinafter described. High-energy rejected photoelectrons pass vertically through screen 20 and are preferably trapped in a metal cup, not shown.

An electrical field preferably carried between cylinder 37 and the screen 36, as well as a field between the screens 36 and 27, adjusts the photoelectron velocities and focuses the photoelectrons prior to their being channeled through screen 27. Typical operating potentials are as follows: Sample 25 and cylinder 37 adjustable between 0 and +1,500 volts, screen 36 potential adjustable between ground and +1500 volts, support block 31 and screens 27 and 28, typically at a fixed potential of +300 volts, and screen 20 +90 volts (thereby preselecting the cutoff point of the prefilter at around 300 electron volts).

At the entrance of analyzer tube 34 there is provided an electrostatic lens 35, maintained at a typical potential of +100 volts, which can be, if desired, a quadrupole lens of standard design, but is preferably a simple so-called Einzel lens as shown (i.e., one wherein a single value of potential determines its optical characteristics). The concentrated photoelectron stream is thence passed rightward as seen in FIG. 6 through longitudinally adjusted centrally apertured plate 38.

The third quarter right-hand length of the analyzer, constituting the electrostatic filter denoted generally at 66, is made up of a plurality of coaxially arranged electrically isolated 1-5/16 inches O.D. × 1/2 inch long copper rings, each carried at the following typical potentials: ring 40 +270 v., ring 41 +210 v., ring 42 +150 v., ring 43 +90 v., and ring 44 +30 v. These rings are disposed between screened apertures 39 on the left and 46 on the right, so that there is maintained a substantially uniform potential gradient between the two screens.

The described segmented construction is preferred because there is thereby obtained enhanced planarity over that of a single field maintained across the total potential difference, resulting in improved resolution and sensitivity.

The magnitudes of the resistors 50 and 55 are, typically, 0.5 megohm, whereas resistors 51, 52, 53 and 54 typically have magnitudes of one megohm each.

As seen in FIG. 6, the opening of aperture 39 is smaller than that of aperture 46. Both apertures can be conveniently spring-mounted within their associated tube sections 34 and 45, respectively. Typically, the cylinder 34-aperture 39 voltage is held at +300 volts, whereas the subassembly 45-46 is held at 0 volts.

The next section in sequence is the cutoff means, denoted generally at 67, comprising tubular copper skeleton ring 59, which is cut away peripherally, as indicated at 59a and 59b, over most of its circumference to permit unimpeded radial passage of the electron fraction of analytical interest into the coaxially disposed surrounding shroud casing 60 maintained at -0.8 volts potential.

The left-hand wall of casing 60 is cut away as indicated at 60a to present an opening in opposition to the conical flared entrances 63 (maintained at, typically, +90 volts) os the continuous dynode surface of a conventional electron multiplier detection means, indicated generally at 64. The voltage along the axis of the dynode surface increases to, typically, +2,500 volts to accelerate the segregated electron fraction. Thus, each electron entering the flared openings 63 collides with the dynode wall surface, producing a secondary emission of more than one electron. These secondary electrons, in turn, are accelerated and undergo wall collisions until the output of a single segregated incoming photoelectron produces as many as one hundred million electrons, constituting a gain of 108. The output of electron multiplier 64, in turn, is routed to an amplifier, a counter and an X-Y recorder (not shown).

In the operation of the analyzer apparatus, the output of electron multiplier 64 is fed through a rate meter into the Y-axis of the recorder, while the output of a motor-driven potentiometer, which is the retarding voltage at cylinder 37, is fed into the X-axis of the recorder. The resulting display is the electron transmission versus electron energy.

The right-hand end of the analyzer is provided with trap 68 which accommodates ejected high energy electrons and comprises co-axially disposed rings 70, maintained at, typically, -3 volts, 71 maintained at +90 volts, and 72 maintained at 0 volts. There is also preferably provided a coaxially mounted leftward-extending 1/16 inch dia. needle electrode 72a carried at a retarding potential of -3 volts to provide assisting radial deflection for analyzed photoelectrons.

In operation, photoelectrons entering tube 34 at, typically, 303 electron volts (±0.5 volt) can pass radially through the opening of skeleton ring 59 at ground level. The reason for this is that high pass filter 66 applies a force leftwards in the direction Z2, FIG. 6, which substantially counteracts the longitudinal force component acting on the photoelectrons of analytical interest, while leaving the radially outward transverse velocity components vt unaffected. This remaining transverse drift, being substantially the sole remaining force acting on the photoelectron fraction of interest, propels these photoelectrons through the opening of skeleton ring 59, so that they can be measured by electron multiplier 64. Needle electrode 72a assists in the radial deflection and thus facilitates the segregation. Photoelectrons of less than 303 volts fall on the succession of rings 40-45, inclusive, of the high-pass filter. On the other hand, photoelectrons of energies exceeding 303 electron volts are ejected and neutralized on the rings 70-72, inclusive, of the trap 68 disposed to the right of ring 59.

The 303 e.v. photoelectrons, now reduced to 3 e.v., bounce around elastically and inelastically within the shroud casing 60 until they come within escape distance of the annular opening 60a and are drawn off therethrough into dynode 64 through confronting openings 63.

FIG. 7A is illustrative of a typical electron energy spectrum recorded with the first embodiment of the analyzer of this invention, in which an X-ray source was used with a gold sample, the irradiation being with aluminum Kα radiation. The X-ray source input power was 250 ma at 10 KV. The time of scan for this spectrum was 1.6 min.

The ordinate values of this spectrum are a measure of transmitted electron current expressed in terms of counts/-second, as registered by the photoelectron detection means, whereas the abscissa values are a measure of photoelectron energy expressed in electron volts. The spectrum obtained is typical of that expected for a gold sample, showing two characteristic peaks associated with the NVI and NVII energy levels in gold, as well as certain of the satellite peaks (not labeled) which appear at higher energies. Complete resolution of the NVI and NVII spin doublet into two lines was obtained.

For comparison, there is shown, in FIG. 7B, a spectrum of gold obtained with a typical prior-art electrostatic analyzer. It will be seen that there exists general agreement as regards appearance and positioning of the photoelectron peaks. The time required to obtain the spectrum of FIG. 7B is not known. However, the significant difference is the superiority in count rate obtained with the analyzer of this invention, i.e., 156,000 counts/sec. as compared with 9,000 counts/sec. for the NVII peak.

Thus, at a source current of 250 ma for which my instrument is designed, the counting rate is better than seventeen times that of the apparatus producing the spectrum of FIG. 7B.

Referring to FIG. 8 the second embodiment of this invention utilizes a converging electron lens which is schematically denoted by the two limiting potential leads Vo, typically +100 volts, connecting with metal screen 39', and V1, typically approximately +50 volts. The electron lens sets up focusing equipotential lines across the tube 34' .

The sample 25' to be examined is supported within a cylinder 37' coaxially disposed with respect to tube 34', within which it is maintained at a potential V2, typically, +1500 volts. The sample is irradiated by a beam of X-rays 77 introduced through port 25a ' covered by window 25b ' to emit photoelectrons having characteristic energies, the path of one of which is denoted by the broken line trace B.

As in the first embodiment, the top end of the analyzer is made up of a plurality of coaxially arranged electrically isolated copper rings, each carried at the following typical potentials, ring 40' 75 v, ring 41' 50 v and ring 44' 25 v. These rings are disposed between screened apertures 39' and 46' (which latter is maintained at ground potential) so that a substantially uniform potential gradient is maintained between the two screens, which thus constitute the equivalent of the electrostatic filter 66 (FIG. 6) of the first embodiment.

Detector 64' can be of an electron multiplier type (typically, a Model 4219 marketed by Bendix Corporation) the collector opening of which is maintained at potential V3, typically +1,000 v. To prevent front face bridging of the field created by the +1,000 v potential applied to detector 64', it is good practice to house the detector within a grounded metal housing, not shown in FIG. 8. In this design the detector is disposed substantially coaxially of tube 34', directly back of a beam stop 75 which can be a circular metal plate 0.001-0.01 inch thick maintained at potential Vo (i.e., +100 volts) by attachment to screen 39'. Beam stop 75 extends outward past the periphery of detector 64' approximately 0.1 inch to bar ingress of any high energy photoelectrons which follow a substantially axial course from sample 25'.

Cylindrical screen 76, constituting, in this design the cutoff means 67', is coaxially mounted with respect to tube 34', and can be 10,000 openings/sq,in. nickel wire metal approximately 1 inch diameter by 1 inch long, maintained at ground potential by attachment to screen 46', thereby effecting photoelectron collection by detector 64'.

In operation, assuming that the apparatus is to be employed for the analysis of 1,500 eV photoelectrons emanating from sample 25', the typical potentials hereinbefore reported for Vo -V3 inclusive are adequate. Then the equipotentials maintained between V0 and V1 constitute a converging lens which imparts a convergent transverse velocity component, v't, to each photoelectron traversing the lens. But v't = (2eVo /m)1/2 sin α, where m = the mass of the electron and α is the angle between a line drawn parallel to the longitudinal axis of tube 34' and the electron final trajectory A'. Filter 66', in this instance, applies a homogeneous retarding field which counterbalances, and thus nullifies, most of the axial velocity component for photoelectrons at energy .about. eVo to be analyzed. The high energy photoelectrons, not being appreciably deflected by the electric fields, travel essentially straight paths and, therefore, go past detector 64' without being counted. Low energy photoelectrons, on the other hand, are repelled by the homogeneous retarding field. However, detector 64', being under a positive potential V3, attracts all photoelectrons which have entered the cutoff means 67' comprised of collection cylinder 76.

It will be understood that the potential at V1 must be experimentally adjusted to yield the proper path direction α for any given as-received photoelectron energy input to be analyzed. Moreover, it is essential that a field-free electron drift region of substantial magnitude be maintained between the retarding field of filter 66' and cutoff means 67' in order to obtain the sharp segregation characteristic of FIG. 1B. The extent of this region is denoted by line length Aa ' in FIG. 8, which should be a minimum of one cut-off element diameter (i.e., one diameter d of cylindrical screen element 76). The lens action obtained by the potential difference maintained between screen 39' and tube 34' can have other geometries, such as, for example, a spherical screen lens.

Thus, referring to FIG. 9, wherein all corresponding elements are denoted by the same reference numerals as for FIGS. 6 and 8, except double-primed, the axial velocity retarding field is provided by an electrostatically charged filter constituting the two-part spherical metal screen means denoted generally at 80a, 80b having radii preselected to present the screen wires generally normal to the paths of incoming electrons. Screens 80a and 80b are concave on the electron input side and are in electrical connection with potential taps V1 and V2, respectively, thereby maintaining an equipotential line pattern (not shown) of substantially uniform repulsion strength. Typically, screens 80a and 80b can be formed on 10 cm. radii drawn from centers on the longitudinal axis of tube 34". Electrically isolated rings 40", 41" and 44" are disposed between screens 80a and 80b and are separated one from another by resistors 50", 51" and 52", with final connection to V2 ground through resistor 55". This sub-assembly brings the equipotential lines into parallelism with screens 80a and 80b while, at the same time, reducing field distortion.

In this third embodiment, segregation of the photoelectron fraction to be analyzed is by convergence, and this is effected by the electron lens made up of spherically curved screens 81a, 81b and the auxiliary ring 85, 86, 87 - resistor 90, 91, 92, 93 sub-assembly. Spherical screens 81a, 81b can be of identical construction with screens 80a, 80b, except reversed in disposition, so that their convex sides confront the incoming electrons. As in the second embodiment, beam stop 75' is interposed between the retarding field means and the converging means to block direct electron impingement on detection means 64" which is, in this design, disposed coaxially of tube 34".

In operation, a typical analyzed photoelectron trajectory is from sample 25", irradiated as hereinbefore described but not detailed again here, via line A", during which rejection of lower energy electrons is effected, thence via A'", which is at only a small divergent angle with respect to the axis of analyzer tube 34", and thence sharply convergent along path B' to the collection face of detector 64".

The cutoff means 67" in this design constitutes aperture plate 95, covered by metal screen 95a assisting field-free drift maintenance, disposed coaxially of tube 34". Again, a field-free electron drift region greater than about one diameter of aperture opening 95 must be provided, the length of line section B' denoting this construction feature. Any electrons not segregated by the convergence cutoff described are removed by the electron trap denoted generally at 68". This comprises an apertured cylindrical metal wire screen section 98 disposed concentrically within a solid metal companionate outer enclosure 99, section 98 being typically maintained at V4 (+100v) whereas enclosure 99 can be V5 (+1,000v).