JPH0615391Y2 - Charged particle moderator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Technical Field The present invention relates to an improvement in the structure of a deceleration pipe of a surface analysis device.

The surface analysis device is to measure the energy spectrum of the proton beam by irradiating the sample with the accelerated proton beam in a high vacuum, decelerating the proton beam scattered by the collision with the sample atoms. A device that analyzes what elements are present on the surface of a sample and what their proportion is.

Use a speed reducer for this device. This is a condition slightly different from the normal deceleration of charged particles. This is due to the structure of the surface analysis device.

The crystal structure inside the sample is known by X-ray diffraction.

The crystal structure near the surface of the sample can be examined by electron diffraction or the like.

However, in any of the methods, the distribution of elements on the monoatomic layer or diatomic layer of the sample and the surface thereof cannot be known.

The present inventors have used a proton energy loss spectrum analysis method (Proton Entropy method) as a method for measuring the amount of elements existing on the surface.
Energy Loss Spectroscopy (PELS).

(B) Principle of PELS Since this is a new measurement technology, the principle will be briefly explained.

In FIG. 3, the atom of mass M is stationary. On the other hand, it is assumed that a proton of mass m approaches at velocity U, is scattered by atom M, and flies at velocity V in the direction of angle Θ. It is assumed that the atom M is jumped at the velocity W in the direction of the angle Φ.
From the law of conservation of momentum, mU = mV _cos Θ + MW _cos Φ (1) 0 = mV _sin Θ−MW _sin Φ (2).

Energy conservation law assuming complete elastic collision Is established.

When W is eliminated from (1) to (3), (Γ + 1) V ² −2U _cos ΘV− (Γ−1) U ² = 0
(4) from here, Becomes

However And In (5), the negative sign indicates the scattering toward (π−Θ). It is correct to take only the positive sign for the scattering toward Θ.

Protons lose energy due to scattering. If the scattering angles are the same, the lighter the other atom, the larger the energy loss. This tells what the other atom is.

If the initial kinetic energy of the proton is E ₀ , Is. The energy after the collision is E ₁ . This is always less than E ₀ . This ratio is called the damping coefficient K.

E ₁ = KE ₀ (8) Is.

Θ is not a variable, but Γ is a variable. Θ becomes constant depending on the experimental device.

The inventors initially developed a low scattering angle (Θ0) device. For example, JP-A-59-180945 (published on S59.10.15) and JP-A-61-151958 (published on S61.7.10) disclose PELS devices having a low scattering angle.

However, low scattering angles have a drawback that they are easily affected by surface irregularities, as shown in FIG. In addition to once scattering, twice scattering may occur.

Further, in the case of a low scattering angle, there is a drawback that the difference in K due to Γ is small and the resolution is low as can be seen from the equation (9).

The principle of PELS is shown directly in Eq. (9), which is the attenuation due to single scattering. It should not be multiply scattered.

The low scattering angle was chosen because of the high yield. Figure 13 shows the relationship between the scattering angle Θ from Au and Si and the yield.

The relative ratio of yield is determined by the geometrical relationship,
It does not depend on the physical properties of the atom.

From this figure, it can be seen that the yield reaches its maximum even at Θ = 180 ° (π).

Furthermore, if Θ = π, it can be seen from Equation (9) that the resolution can be highest with respect to Γ.

At this time, the attenuation coefficient K is set to Θ = π in (9). Becomes Γ is the mass ratio of atom M and proton m. The atomic mass unit 1u and the proton mass m are slightly different, but if they are regarded as almost the same, Γ can be called a mass number.

If a certain element is specified, Γ can be determined and K can be obtained. Here's an example.

In the case of Al Γ = 26.98 K = 0.8621 In the case of Ga Γ = 69.72 K = 0.94422 As In the case of Γ = 74.9 K = 0.94799. It is easy to calculate K for all elements.

That is, if the final kinetic energy E ₁ of the proton is measured, K can be found from the ratio with E ₀ . Then, Γ is known, and the atom with which the proton is scattered is known.

Then, the abundance of the atom can be known from the energy spectrum.

In most cases, E ₀ is about 100 keV.

The PELS principle is thus simple.

Protons need to be scattered only once to be able to make such measurements.

Instead of measuring E ₁ directly, the energy loss ΔE
= (1-K) E ₀ can also be measured. PELS is named because it measures the distribution of proton energy loss.

(C) Depth analysis by shear dough cone As shown in FIG. 4, when the proton m collides with and is scattered by the heavy atom M, a part where the proton cannot reach is generated. When this is called Shed corn.

Protons passing far away from the atom M are hardly scattered, and protons passing near are easily scattered. Because of this, shed corn is produced.

The traveling direction of the proton is the x-axis, and the direction perpendicular to this is the y-axis.

If the repulsive force acting between the atom and the proton is the Coulomb force, the Sheadow-Corn equation becomes Given by. e is the elementary charge, z is the charge of the proton,
Z is the charge of the atom (in units of e).

Which atoms make up the first layer of the sample when the sheared corn is cleverly used? I understand that.

In FIG. 5, it is assumed that the proton beam is vertically incident on the surface of the GaAs sample. Since it is scattered by both Ga and As, there are two peaks in the energy loss distribution as shown in FIG.

Next, as shown in FIG. 7, a proton beam is obliquely incident. Then, the atoms of the second layer are allowed to enter the shear cone formed by the atoms of the first layer. In this case, only those corresponding to the atoms of the first layer appear in the proton energy loss distribution. As shown in FIG. 8, only the Ga peak appears. From these results, it can be seen that the first layer of the sample is Ga.

(D) Depth analysis using electron collision When the proton beam enters the sample, the proton loses energy even by collision with the electron. Energy loss due to electron collision is proportional to the distance traveled in the sample.

This will be described with reference to FIG.

There is a proton beam that makes an angle θ / 2 with the surface. There is a difference in energy loss between the case where they are scattered at point J of the first layer and scattered in the direction of θ / 2 and the case where they are scattered at point G of the second layer. Energy loss difference Is. d is the interlayer distance, and S is the electron stopping power. The smaller θ is, the larger ΔF is. Even if θ = π, for example, d
= 5.6Å and S = 10eV / Å, ΔF = 112eV.

Even at normal incidence (θ = π), the energies of the scattered protons in the first and second layers differ by about 100 eV. This means that even if the atom M, which is the main body of scattering, is the same, the energy loss differs between the first layer and the second layer.

If θ is sufficiently small and ΔF is large, it is possible to identify the proton scattering from the different atoms in the first layer and the different atoms in the second layer.

(E) Outline of measuring device Figure 10 shows the conventional PELS when the scattering angle Θ is 180 °.
The outline of a measuring device is shown.

The proton ion beam extracted from the ion source A is mass-analyzed by the magnet B. Only monovalent proton ions are selected and enter the accelerating tube C for acceleration.

It has a kinetic energy E ₀ that is the sum of the drawing energy Eex and the acceleration energy Eacc in the acceleration tube C,
This energy hits the sample Σ.

Protons are scattered by the surface atoms of the sample Σ. Only those with a scattering angle of Θ = π run backwards through the accelerating tube. Those with Θ ≠ π collide with the chamber wall, return from the ion to the neutral molecule H ₂ , and are exhausted.

Those with Θ = π run through the accelerating tube in the opposite direction, so they are decelerated. In other words, this time it works as a speed reducer.

The deceleration energy Edec is equal to the acceleration energy Eacc.

Eacc = Edec (13) It is an excellent advantage when Θ = π that the proton beam can be accelerated and decelerated by the same accelerating tube C and decelerating tube D.

The decelerated proton beam is bent at a right angle by the magnet E and further bent at a right angle by the magnet F. And
The protons are obliquely incident on the analyzer G.

The two magnets of the magnets E and F are necessary because the proton beam energy has a variation due to the scattering loss ΔE and is converged to one point.

The energy of the proton beam focused on one point is detected by the analyzer G.

There is a voltage V ₀ across the plates of the analyzer. The protons entering from the slit draw a parabola and enter one of the channels of the microchannel plate H. The distance L from the slit to the position where you descended is known depending on which channel you entered. From this, the kinetic energy of the proton when it is incident is known.

The greater the kinetic energy of the proton, the greater the range L in the analyzer G.

Let Ea be the kinetic energy of the proton when it enters the slit of the analyzer G, Ψ is the angle between the proton beam and the parallel electrode when it is incident, h is the parallel electrode distance, and V ₀ is the electrostatic voltage between the electrodes. Is. From the distribution of range L, the distribution of proton energy Ea is known. That is, the energy distribution of protons can be measured.

The device of Figure 10 is all in a high vacuum. Especially, a part of the sample Σ must be in an ultrahigh vacuum. If not vacuum
This is because the protons collide with the gas molecules, which causes energy loss, and the scattering loss ΔE due to the sample Σ cannot be known.

For simplicity, the vacuum chamber and the vacuum exhaust device are not shown in FIG.

(F) Proton energy change Figure 11 explains the proton energy change.

The position of the proton moves from left to right. This represents potent energy.

The proton extraction voltage of the ion source is Vex. If the proton charge is q, the potential energy due to this is qVex. When the ion source comes out, this much energy becomes kinetic energy.

The acceleration voltage at the acceleration tube C is Vacc. The kinetic energy just before leaving the accelerating tube and hitting the sample Σ becomes q (Vex + Vacc). This is E ₀ . It is about 100 keV.

The proton beam hits the sample and is scattered. The energy loss due to scattering is ΔE.

Next, it goes backwards through the accelerating tube and loses kinetic energy equal to qVacc.

The kinetic energy Ea when entering the analyzer G is Ea = qV ₀ −ΔE (15). Here, V ₀ is equal to the extraction voltage Vex. It can be said that the energy is incident on the analyzer G when the energy loss is assumed to be zero.

Ea is set to about 0.5 keV, and scattered proton energy Es Es = q (Vex + Vacc) −ΔE (16) is set to about 100 keV.

Ea should originally be a variable. However, the target atom is determined in advance, and Vacc is determined so that Ea for this atom is about 0.5 keV.

(G) Entire PELS device FIG. 14 is a schematic perspective view of the entire PELS device.

The proton beam emitted from the ion source is focused by the Einzel lens, bent by the magnet, and accelerated by the acceleration / deceleration tube.

The sample is set in the ultra-high vacuum chamber. The sample can be manipulated by a manipulator.

The accelerated proton beam is focused by the Q lens and hits the sample in the ultra-high vacuum chamber.

Of the protons scattered on the sample surface, those with Θ = 180 ° emerge from the ultra-high vacuum chamber and are decelerated. This is bent 180 ° by the magnets 1 and 2 and enters the analyzer. Here, the energy loss ΔE is measured.

(H) Conventional Technology The outline of the PELS device has been described above.

The present invention relates to the structure of a speed reducer of a PELS device.

The acceleration / deceleration tube has the function as an acceleration tube and the function as a deceleration tube. When the scattering angle Θ is 180 °, the incident proton and
This is because the paths of scattered protons are the same and opposite.

Although the present invention relates to an improvement as a deceleration pipe, it can also be used as an acceleration / deceleration pipe because it can be used for acceleration.

Decelerating means decelerating a proton having a kinetic energy of 100 keV to a proton having a kinetic energy of 0.5 keV.

In this case, it is important that the beam emitted from the moderator tube does not spread too much. It is desirable that the divergence of the incident beam and the outgoing beam be approximately the same.

The emitted beam is bent by a magnet or the like and enters the electrostatic analyzer, and the kinetic energy is measured. The analyzer correlates the range of protons with the energy.

If the beam divergence is large, there will be many errors in the measurement of proton range. Inaccurate measurements.

If the divergence of the beam is large, it will collide with an orifice, magnet, etc.
I can't earn d. (In the detector, H ₊ is 10 cps
It is indispensable to improve the transport efficiency because it seems that only a certain degree can be obtained.) Furthermore, the beam divergence in the deceleration tube is more difficult to handle than the beam divergence in the accelerator tube. Further, the beam in the deceleration tube is more likely to diverge.

Therefore, in the present invention, the divergence of the beam in the speed reducer is a problem.

However, conventionally, there is no document in which the beam divergence has been a problem in the moderator tube. This is because the deceleration tube is not frequently used like the acceleration tube.

There are many devices that accelerate charged particles. These often accelerate and hit the target. There are many cases of investigating scattering from the target, but it is (PELS
Except)

For this reason, it seems that divergence in the speed reducer has not been a problem.

Japanese Unexamined Patent Publication (Kokai) No. 61-124099 (S61.6.11) discloses an ion implantation accelerating tube in which the shape of the electrode is devised so as to suppress the divergence of the accelerating beam and obtain a focused beam.

Ion implantation is performed on a Si wafer by using n-type impurities and p-type impurities (B
Etc.) is injected at an energy of several tens keV to several hundred keV. The accelerated one is driven into the wafer. You have to accelerate because you hit.

The accelerating electrode is composed of a plate-shaped electrode curved in a concave shape with respect to the traveling direction of ions, and an electrode having a thin center and a thick side circumference like a concave lens.

In this case, the equipotential surface becomes concave toward the traveling direction, and the accelerating beam tends to be perpendicular to the equipotential surface, so that the beam is converged.

This is the prior art regarding convergence in an accelerating tube. Since the ions are implanted, they do not return from the wafer and there is no need to decelerate.

The beam focusing problem in the deceleration tube is more difficult than in the acceleration tube.

Particles flying at high speed do not bend easily. Since the accelerating tube uses a high speed, charged particles can move straight and are difficult to bend.

Since it is difficult to bend, it can be converged simply by moving the gyro inward.

However, in the moderator tube, kinetic energy is lost by climbing the electric field. And it becomes a slow particle. Slow particles have a small momentum, so they bend easily even if a slight lateral force is applied.

Susceptible to distortion of electric field.

Therefore, in the deceleration tube, the divergence is more important than the convergence in order to converge the beam.

It can be said that this paradox is unique to the speed reducer.

Suppose that the beam converges inside the deceleration tube. Let Q be the convergence point. The beam converged at point Q becomes a divergent beam when passing through this point. It will diverge by this recurrence. If the convergence is rapid, the divergence after passing the convergence point Q is also rapid.

Therefore, it is best not to create the convergence point Q in the deceleration pipe.

However, there is always a convergence point. It is impossible to eliminate the convergence point.

If the occurrence of the convergence point cannot be avoided, it is desirable to set the convergence point Q as rearward as possible. Also, it is required that the convergence is not rapid.

Figure 2 shows the speed reducer used in conventional PELS.

There are flanges F and G at the start and end of the speed reducer. Flange F, between G, and eight flat disc-shaped electrodes P ₁ to P _8, the insulating ring H partitioning them, H, ... are provided.

Electrodes P ₁ to P ₈ is a simple circular plate having a hole in the center.

The deceleration voltage Vdec was equally distributed and applied to this electrode. That is, the number of electrodes is n, the P ₁ O (V), the P ₂ Vdec / n
The voltage was applied to (V), P ₃ at 2Vdec / n (V), and so on. A voltage of Vdec / n will be applied between the adjacent plates.

(K) Problems with the prior art When the beam is decelerated with such a simple electrode structure, it occurs almost uniformly in the electrode having the equipotential surface in the middle. That is,
It occurs at right angles to the beam axis.

However, at both ends, since the electrodes have holes, the equipotential surface bulges outward at the center. That is, the equipotential surface has a shape similar to that of a convex lens.

Since the distortion of the equipotential surface occurs symmetrically like a convex lens,
The divergence and convergence effects seem to be canceled out for the proton beam passing through here.

But not so.

Since the beam is decelerated, when the beam obliquely crosses the equipotential surface, it is refracted so as to come close to the equipotential surface. That is, the proton beam is refracted in the same direction as when the light is incident on the substance having the high refractive index and the substance having the low refractive index.

The exact opposite of acceleration. You have to be careful.

In the vicinity of the first electrode P _1, the equipotential surface bulges outward at the center. Therefore, the proton beam is refracted in the direction of divergence.

It is desirable to be able to bend in the direction of divergence. This point has already been mentioned, but I will reiterate it. It is desirable not to converge. I have already explained why we should not converge.

This is because if the beam converges at a certain point Q, it becomes a divergent beam after the point Q. If the point Q is in front of the deceleration pipe, the traveling distance of the diverging beam will be long, and the divergence of the beam will finally be remarkable.

But somewhere it converges. There is no choice but to converge.

Then, the best point is to push the convergence point Q to the rear of the speed reducer and to reduce the convergence angle.

The equipotential surface of the intermediate portion is substantially orthogonal to the beam traveling axis. Therefore, the beam does not bend much in the middle part.

However, the last electrode has an equipotential surface that bulges outward. Therefore, the proton beam is focused.

Since they diverge first and finally converge, it may be thought that they cancel each other out.

However, that is not the case.

The energy of the proton beam is large when it enters the moderator tube. There is about 100 keV. The speed is also great. It is difficult to bend the trajectory of a high-speed object even if force is applied from the side. This is because the speed is high, and the strain passes through the equipotential surface in a short time. Since the impulse received is small, the orbit is hardly bent.

However, the energy is extremely small near the last electrode of the speed reducer. For example, it is about 0.5 keV.
Pass through here at a leisurely speed. Therefore, it takes a long time to receive the force in the converging direction. In other words, it receives more impulse in the direction of convergence.

Then, eventually, the converging action due to the distortion of the equipotential surface formed behind the last electrode is superior. The convergence point Q can be just behind the last electrode. The convergence angle is large. This is naturally large because the divergence angle is equal to the convergence angle.

(C) Structure FIG. 1 shows the electrode structure of the speed reducer of the present invention.

This acts as a deceleration tube as the proton beam travels from left to right. Since the present invention is concerned with the function of the speed reducer, only the case where the proton beam travels from left to right will be described.

Of course, in an actual PELS device, if Θ = 180 ° scattering, it also functions as an accelerating tube. However, its function as an accelerator tube will not be described here.

There are flanges F and G on both sides. During this time, appropriate number of electrodes P ₁ , P
₂ and insulating rings H, H, ... Are provided alternately.

Here, 7 electrodes are shown between the flanges, but if the flange G is regarded as the last electrode, 8 electrodes (P ₈
= G). The number of electrodes is arbitrary.

The electrodes P ₅ , P ₆ , and P ₇ are simple disk-shaped plates. There is no feature in the shape.

The front electrode has a shape feature.

The electrode P ₁ is composed of a disc portion 1 and a cylindrical portion 2 continuous with the disc portion 1 at its inner edge.
Have and. The cylindrical portion 2 is arranged to continue to the rear of the disc portion 1 as seen from the flow of the beam. The mounting portion 21 on the outer periphery of the disc portion 1 is sandwiched and fixed by an insulating ring H.
Further, the mounting portion 21 is connected to the power source.

As compared with the conventional electrode, the disc portion 1 projects inward, and the hole through which the beam passes is narrowed.

Another characteristic of P ₁ is the cylindrical portion 2 extending to the rear of the flow, which is described above.

These are for bending the equipotential surface so that a divergent effect is produced.

The equipotential surfaces Υ _{1 to} Υ ₂₀ are shown, which will explain the meaning of the electrode shape.

The equipotential surface shown here is an equipotential surface when the flange F is set to 0 kV and each electrode is set to P ₁ -2 kV P ₂ 21.5 kV P ₃ 45.5 kV P ₄ 69.5 kV P _{5 to} G 99.5 kV. When the distribution of voltage is changed, the equipotential surface also changes, but since the electrode shapes are the same, it can be said that the tendency does not change.

At P ₁ , the central opening diameter is small, and since the cylindrical portion 2 is provided, the equipotential surfaces Υ ₁ and Υ ₂ have a shape protruding forward. Especially, the curvature of Υ ₁ is large. This causes the beam to diverge strongly.

Since there is the cylindrical portion 2, the intermediate portion 3 between P ₁ and P ₂ becomes narrow. Here, equipotential surfaces Υ _{1 to} Υ ₅ are strongly narrowed.

After being squeezed, since it expands in the center, Υ ₁ and Υ ₂ will bend strongly forward.

Since Υ ₄ and Υ ₅ are concave toward the front, they focus the beam. However, since the curvature is small, the force for converging is weak.

The electrode P ₂ also has a disc portion 4 and a cylindrical portion 5 that follows the inner end thereof. The cylindrical portion 5 is for narrowing the space between the cylindrical portion 5 and the next electrode P ₃ . At the intermediate portion 7, the equipotential surfaces Υ _{6 to} Υ ₉ are strongly narrowed.

P _{2 also} has a front ridge 6. The mounting portion 22 fixes P ₂ to the insulating ring H. Further, it is connected to a power source at the mounting portion 22.

The same applies to the electrode P ₃ . A cylindrical portion 9 is provided at the center of the flat plate portion 8 having a hole. The front protrusion 10 is provided on the side opposite to the extending direction of the cylindrical portion 9.

The front ridges 6 and 10 further bend the equipotential surface by increasing the electric field here and expanding it in the central portion.

Since Υ ₆ and Υ ₇ are convex toward the front, they diverge the beam. However, since Υ ₄ and Υ ₅ extend backward,
The forward curve of Υ ₆ and Υ ₇ is suppressed.

Υ ₉ and Υ ₁₀ are equipotential surfaces that act to focus the beam. Since Υ ₁₀ bends along the cylindrical portion 9 of P _3, the curvature at the central portion is small. Therefore, the convergence effect of Υ ₁₀ is small.

The electrode P ₄ has a more peculiar shape. It is designed to cover the next electrode P ₅ halfway. The front surface 11 is a disc surface perpendicular to the axis. Following this, the inner cylindrical surface 12 is formed. This is longer than the cylindrical surfaces 2, 5 and 9 of P _{1 to} P ₃ .

A rear disc surface 13 is formed following the inner cylindrical surface 12.
Further, a cylindrical surface 14 and a conical surface 15 are formed toward the outer edge.

It is fixed by an insulating ring H at the mounting portion 24.

Equipotential surfaces Υ _{16 to} Υ ₂₀ are strongly pressed into the intermediate portion 18 between the electrodes P ₄ and P ₅ . Since the intermediate portion 18 opens rearward, the equipotential surface bends gently.

Υ _{17 to} Υ ₂₀ have a function of converging the beam. However, since Υ ₂₀ is an equipotential surface that is gently curved with both ends constrained by P ₅ , it has a small curvature.

The same applies to Υ ₁₈ and Υ ₁₉ .

Υ _{16 to} Υ ₂₀ have a binding region between the conical surface 15 and P ₅ of P ₄ and have a distance. Therefore, the constraint is loose. Therefore, the curvature becomes small, and the beam focusing effect of Υ _{17 to} Υ ₂₀ is small.

Of course, there are diverging effects of Υ ₁₅ and Υ ₁₆ . These have a divergent effect because they are convex forward. It is pushed forward by Υ _{17 to} Υ ₂₀ . When it is pushed forward, it is obstructed by the rear disc surface 13 of the electrode P ₄ , and a new inflection point is generated at a point near the central axis.

Therefore, the curvature becomes considerably large. Therefore, Υ ₁₅ , Υ
₁₆ great divergence effect.

As mentioned at the beginning, it is only for relatively slow beams that the proton beam can be focused and diverged by the force of the electric field. A fast beam exits the electric field rapidly and therefore experiences a small impulse.

At P ₁ , the proton has an energy of almost 100 keV. It is a fast beam. However, at P ₄ , there is only about 30 keV of energy. By crossing Υ _{15 to} Υ ₂₀ , the speed is reduced to about 0.5 keV.

Then, it can be seen that the strongest influence on the convergence and divergence of the proton beam is the gradient of the equipotential surface of Υ ₁₅ to Υ ₂₀ .

Of these, Υ ₁₅ and Υ ₁₆ contribute to divergence, and Υ ₁₈ to Υ ₂₀ contribute to convergence. Υ _{18 to} Υ ₂₀ draw a gentle curve and have a small curvature at the beam axis.

The points of inflection Υ ₁₅ and Υ ₁₆ have inflection points protruding toward the beam axis due to the structure of the electrode P ₄ protruding downstream, and the curvature increases on the beam axis. The divergent effect can be increased.

This is because the intermediate portion 18 of the electrodes P ₄ and P ₅ is open rearward.

Of course, since there is a difference in speed, the action of Υ _{15 to} Υ ₂₀ is
After all, the converging action becomes superior. 0.5 keV for Υ ₂₀
This is because it becomes extremely easy to feel the electric field.

It has the same potential as the P ₅ ~P ₇ and the flange G.

Therefore until P ₅ ~ flange G are all equal potential. This space is called a drift space 30.

No equipotential surface exists in the drift space 30. Therefore, there are no lines of electric force that are orthogonal curves of equipotential surfaces. In the drift space 30, there is no electric field effect. The proton beam moves at a constant linear velocity.

The effects of Υ _{16 to} Υ ₂₀ are more converging in total.
However, since the difference is slight, the drift space 30
Then, the proton beam converges smoothly.

Since the convergence point Q is extremely backward and the convergence angle is narrow,
The divergence will be gradual.

The beam divergence at the front opening 19 and the beam divergence at the rear opening are substantially the same.

(B) Effect It is possible to suppress the proton beam from converging in the deceleration tube and move the converging point to the rear. When it exceeds the convergence point Q, it becomes a divergent beam. Since the convergence angle is small, the divergence angle is extremely small.

The beam divergence upon entering the front opening 19 and the beam divergence upon exiting the rear opening 20 are substantially the same.

After that, it is bent by a magnet and the energy is measured by an analyzer, but since the beam divergence is narrow, the energy measurement accuracy becomes high. Since the measurement can be performed with high precision, the element existing amount on the sample surface can be obtained with high precision.

Among the PELS devices, the one with a scattering angle of Θ = 180 ° was mainly explained. However, in the present invention, PELS with any value of Θ
It can also be applied to the speed reducer of the device.

If Θ ≠ 180 °, the acceleration tube and the deceleration tube are different. In this case, the invention applies only to the speed reducer.

[Brief description of drawings]

FIG. 1 is a longitudinal sectional view of a speed reducer showing an example of the present invention. FIG. 2 is a vertical cross-sectional view of a conventional reduction gear. FIG. 3 is an explanatory diagram of the velocity relationship before and after the collision of the atom m with the proton m. FIG. 4 is an explanatory view of a shed cone produced in the forward direction when protons are scattered by atoms. Figure 5 shows GaAs for explaining normal incidence scattering in PELS.
Crystal surface diagram. Fig. 6 is a proton energy loss diagram of PELS in the incident direction of Fig. 5. FIG. 7 is a GaAs surface layer diagram for explaining oblique incidence scattering in PELS. Fig. 8 is a proton energy loss diagram of PELS in the incident direction of Fig. 7. FIG. 9 is a scattering cross-sectional view explaining that the first layer and the second layer can be distinguished by the increase in the proton energy loss due to electron collision. Figure 10 shows the principle configuration of PELS. Figure 11 shows the proton potential energy coordination diagram. Figure 12 is an illustration of scattered beams at low scattering angles. Figure 13 is a graph showing the difference in yield due to the difference in scattering angle for Au and Si. FIG. 14 is a schematic perspective view showing the actual PELS device. F, G …… Flange P _{1 to} P ₇ …… Electrode H …… Insulation ring Υ _{1 to} Υ ₂₀ …… Equipotential surface

Claims (1)

[Scope of utility model registration request] 1. A proton beam is accelerated to irradiate a sample placed in a high vacuum, and the proton beam scattered in a certain direction is decelerated to measure energy, and the kind and distribution of elements on the sample surface are measured. A deceleration tube used in a surface analysis device which is determined from energy loss, and includes a suitable number of electrodes P ₁ , P ₂ ... Separated from each other and insulating rings H, H, ...
, And electrodes P ₁ ... and flanges G, H provided at the front and rear ends of the insulating rings H, H, .... Among the electrodes, one or several electrodes at the front are a disk part and a disk part. In contrast to this, one electrode following these electrodes has a disk portion and a cylinder portion continuing to the rear, and is thick and between the next electrode and the next electrode. In the deceleration tube for charged particles, a narrow opening facing obliquely rearward is formed, the rear electrode and the rear flange G have the same potential, and the front electrode has a potential difference.