DE2628368A1 - METHOD FOR EFFECTING FUSION REACTIONS - Google Patents

Description

Method for causing fusion reactions

The invention relates to a method for causing fusion reactions between particles of the same polarity.

U.S. Applications 247,473 dated April 25, 1972 and 490,693 dated July 22 1974 are methods and facilities for achieving nuclear mergers or fusion reactions through mutual influencing of self-colliding accelerated ions of the same charge in one organized mixture ("migma") of ions orbits under the influence of a magnetic field. Owned here as a "Migma cell" designated device an evacuated space in which a magnetic field is formed, the field strength of which with increasing radial distance from a vertical central or z-axis decreases and in this central axis with increasing axial distance from one perpendicular to the central axis running median plane (x- and y-axes) increases. Accelerated ions are fed into the middle area of the evacuated room, its path of entry into the migma cell through the orientation of a magnetic channel on the circumference of the room in relation to the central plane is determined. As a result of a considerable introduction angle of about 60 degrees and a magnetic field formed in the manner mentioned vertical oscillations are triggered, so that the injected ions follow paths in the center of the evacuated space in the form of a figure eight. These vibrations, together with the characteristic precession of each injected ion in an orbit, produce Radius is determined by the respective feed-in energy, a metastable cohesive restraint of the ions in the migma cell,

609883/0347

590204 - / - 26283 * 88

at which the ions fail for a variety of precession periods return to their entry point. This and a high instantaneous ion introduction current self-colliding ion bundles are formed in the evacuated space. The ones with dissociations and splits associated collisions create an atomic migma and cause nuclear fusion reaction products to be radiated from the migma cell, for example, of charged particles that allow electricity to be generated directly if they are slowed down by external electrical fields will.

The proposed Migma cell works under operating conditions that limit their usefulness in certain applications. Only then does this migma cell cause a metastable bondage over many Periods of precession when such a large size lead-in angle is used. Because the magnetic field is generated by coil assemblies which are arranged in the central axis at a distance from one another on opposite sides of the central plane, and there is the lead-in path the ions is at such a large angle to the center plane, the design and the spacing of the coil arrangements must therefore be chosen so that they do not physically hinder the introduction of ions.

It has also been suggested that the restrictions on to mitigate the design and arrangement of the magnetic field generating components that a substantially in the center plane of the Migma cell fed ion beam to achieve metastable bondage is electrostatically deflected. So effective is this measure also, the device required for the electrostatic deflection of the beam means an additional effort.

In contrast, the invention is based on the object of a method of the aforementioned type to the effect that a metastable bondage of the molecular ions fed in over many Precession periods and thus such a bulky feeding of these ions is no longer necessary at all.

This is achieved according to the invention in that the following process steps are carried out: (a) creating a vacuum, (b) forming a magnetic field in this vacuum, whose field strength is with

609883/0347

0FUG5NAL INSPECTED

Formation of a magnetic field in this vacuum, its field strength with increasing radial distance from a central axis of the vacuum decreases and in the central axis with increasing axial distance from a increasing central plane extending perpendicular to the central axis, (c) generating particles of a predetermined polarity with a quantity distribution, in which the light of the particles of higher vibrational state exceeds that of such particles in a normal distribution, and (d) Feeding these particles in a plane with the median plane from the circumference of the vacuum radially inward into the central part of the Magnetic field.

By this measure, the injected molecular ions are only subjected to a single pass through the reactor space. The ion beam to be fed in is prepared in such a way that its quantity distribution the molecules are in higher vibrational states, than is the case with normal quantity distributions. As a result, the injected molecules are already experienced in one pass in the migma cell under magnetic influence a collision-free dissociation, the so-called Lorentz dissociation.

In the following, the invention is explained in more detail with reference to some embodiments shown in the accompanying drawings, wherein further features, advantages and embodiments of the invention result. Show it

1 shows a front view of a migma cell for carrying out the method according to the invention, which has been rotated ninety degrees from its operating position, ''

FIG. 2 shows a schematic cross section through the Migma cell according to FIG. along the line H-II,

Fig. 3 (a) and (b) the course of the magnetic field, Fig. 4 the paths traversed by the ions,

5 shows the dependence of the D ₁ ⁺ * Dp ⁺ and D, ⁺ currents on the pressure in the ion source and

6 shows the yield of D ₂ ⁺ in the highest oscillation state as a function of the ion source pressure and

FIG. 7 shows a modification of the arrangement according to FIG. 1.

609883/0347

In Fig. 1 and 2, a Migma cell 10 is shown, which is cylindrical is and a 'fusion reactor chamber 12 with side walls I4 as well has upper and lower terminations 16 and 18. The reactor chamber 12 is an evacuated space with a center plane ζ = 0 and one Central or z-axis that extends perpendicular to the central plane and passes through its center at the intersection of the x and y axes (Fig. 2). Upper and lower coil assemblies 20 and 22, respectively, are aligned with the central axis.

The coil arrangements 20 and 22 constitute a superconducting magnet system They are located in chambers 24a and 24b for liquid helium, the is supplied via pipes 26a and 26b. The whole arrangement is surrounded by chambers 28a and 28b for liquid nitrogen, which is supplied via pipes 30a and 30b is supplied. The coil assemblies 20 and 22 are electrically excited via lines 32a and 32b. The housing for the Coil arrangements are stiffened by an implosion pin, not shown here, in order to prevent them from collapsing. As already indicated, the arrangement of Figure 1 is shown rotated by ninety degrees from its operating position; the z-axis is usually horizontal. This representation was chosen to facilitate a comparison with the applications mentioned at the beginning, in which the central plane is horizontal runs.

The coil assemblies 20 and 22 generate in the evacuated reactor chamber 12 a magnetic field, the strength of which decreases with increasing radial distance from the central axis and with increasing radial distance from the Median plane increases along the central axis. Such a field can be generated by a broken series expansion with respect to powers of r and ζ being represented:

B _z (r, z) / B _o = 1 -k (f) ² + 2k (f) ² (1)

in which B the field strength in the middle of the room at the intersection of the Central axis ζ with the central plane (i.e. x = y = z = 0), k a constant less than one, R the maximum radial extent of the field, r a variable representing the radial distance from the central axis and ζ one is the axial distance from the median plane along the median axis is representational variable. To further explain this magnetic field serve Figs. 3 (a) and 3 (b).

609883/0347

FIG. 5 (a) is a graph of the magnetic field radially in FIG Middle plane, based on its 7 / ert at the center of the middle plane. This ratio has its maximum in the central axis, i.e. for r = 0. Fig. 5 (b) is a similar illustration of the magnetic field along the z-axis related to its value in the center of the median plane. It has a minimum there.

The magnetic field of Figs. 3 (a) and 3 (b) had on the migma cell of the initially Registrations sensed the influence that a metastable bondage of the injected ions occurred over so many precession periods that self-collisions of the ion beams occurred, which caused the fusion reactions maintained so that the reaction products enabled the immediate generation of electrical current. Such a metastable The restraint of the ions is due to the energy level of the feed, which foreseeable movements of the fed ions on Orbits caused, as well as on the focusing properties of the magnetic field in the horizontal and vertical direction. In this way, by feeding the ions at an angle to the central plane, the mentioned precession of the ion beams in the form of a figure eight.

The present invention relies on this focusing property of the Magnetic field to guide the fed-in molecular beam once through the center of the room. The invention also uses the Property of the magnetic field from the fact that with a radially inward movement from the circumference of the field to the center of the A constantly increasing field strength can be determined in the central plane of the room. The ion feed takes place here so far coplanar with the Central plane, as is practically desired, which is indicated in Fig. 2 by the feed beam I, which is essentially in the central plane lies.

In FIG. 4, the path of the injected beam I, which has a single pass, is shown in the central plane. It shows a continuous inward displacement along the path §, -0 under the influence of the magnetic field up to the point χ = y = ζ = * 0 of maximum field strength, from where then along the path BO a movement directed radially outward from the center of the room to the exit. A precession ier orbits and a bondage over

GG3B83 / 0347

590 204 - ψ - 104

♦ fr ·

many periods, as in the mentioned registrations, do not occur there the feed takes place in the center plane and no electrostatic field acts on the beam during its movement from 0 to B. forcing him back to the center of the room.

To generate a magnetic field according to equation (1), the coil arrangements have 20 and 22 the shape of a solenoid, but other coil shapes can also be determined, those in an air core magnet system satisfy the conditions of equation (1). In an article in the Journal of Applied Physics, ToI 22, No. 9 "September 1951" Pp. 1091-1107 procedures are described with which from the Coefficients of mathematical equations that produce a desired field profile indicate which magnet systems can be designed. In the book "Particle Accelerators" by Livingstone and Blewett (1962) theoretical methods are discussed with computer support, which are used to determine the physical coil parameters in acceleration devices can serve for charged particles. With magnet systems With an iron core, the field of equation (1) can be achieved by parabolic shaping of the pole piece tips, which are comprised of individual coils (Journal "Nuclear Instruments and Methods" 120, 1974 »

p. 309-319).

So far, the invention of the magnetic field has made use of the migma cell only to focus a single pass after feeding the ions into the midplane of the migma cell (or as coplanar as practically). However, the magnetic field has another use in the so-called Lorentz dissociation of injected molecular ions, a collisionless dissociation caused by magnetic influence.

It is known that every charged particle that moves at a speed (ha (speed of light c = 3.10 cm / s) in a magnetic field of strength B is an induced electric field E (in v / cm) according to the equation

E = 3ΟθβΒ (2)

subject. If the field strength is sufficiently large and the particle enters When the molecule is charged, it is dissociated, i.e. broken down into the atoms that make it up. The probability R of the occurrence of a

609 8 83/0347

590 204 - ^ - 104

such magnetic dissociation increases with the induced electrical Field as follows:

R = E ^1/4 .exp (o6E / E _n) (3)

_n )

where o6 = 3 «10 and E is the binding energy in electron volts of one Molecule is in its nth vibrational state.

Deuterium (Dp) molecules can be excited in 26 different vibrational states. For the first state the binding energy E ₁ = 2.69 eV.

For η «25 and for η = 26 the binding energies are Ep ₁ - = 6.2.10 eV

-4
and Ep / -> = 8.2.10 eV. With the Migma cell parameters of, for example, 0.025 and B = 20 kilogauss, equation (3) becomes:

R = 19.68 exp (0.045 / E _n ) (4)

which for η = 25 and η «= 26 results in R = 2.8.1O ⁴ l / s or R = 1.34.1ο ² · ⁵ l / s. Hence, the rate of the Lorentz dissociation is D _? ⁺ (n = 26) about 4.8.10 times that for Dp (n = 25). Under these circumstances, the Lorentz dissociation in the Migma cell is evidently limited to deuterium molecules in the 26th state of excitation.

With a normal distribution of the molecular vibration levels of a deuterium plasma, the probability that a D _? ⁺ Molecule is in its 26th state of excitation, not exactly known. However, the literature shows 0.004 f ° i, which has also been confirmed experimentally. With the above typical migma cell parameters it can therefore be concluded that a fed-in beam of deuterium molecules with normal distribution in the migma cell certainly does not lead to a significant atomic migma through Lorentz dissociation, which reduces the need for dissociation through collisions as a result of high feed-in currents could. For example, a fed-in current of 0.3 mA supplies only 1.2.10 mA atomic migmas, ie 7> 5 · 1Ο dissociations per second of D ₂ molecules in their 26th oscillation state.

While the invention thus provides conditions that are partially suitable for collision-free dissociation - the feeding takes place in the middle of the migma cell space via a path of continuously increasing magnetic field strength, whereby the collision-free dissociation is limited to the area around the middle part of the migma cell - are

603883/0347

590 204 - V> -. 104

the terms, taken together, are not yet for the intended purpose suitable.

To achieve optimal conditions necessary for collision-free Dissociation in the Migma cell are suitable and worth mentioning lead to a reduction in the current fed in, according to the invention the molecules to be introduced are prepared in such a way that the quantity distribution of the molecules in the 26th oscillation state is increased by at least one order of magnitude compared to the normal distribution. To this Purpose, so the experiment shows, a "Duoplasmatron" ion source (e.g. Type 350 from General Ionex Corporation) can be operated in such a way that the molecules are excited in such a way that their molecular bond is weakened without, however, already splitting the molecules into their atoms. These conditions are unusual in the sense that they differ from the normal operating conditions for such ion sources and require a substantial increase in gas pressure for the inert gas in the ion source. Under normal operating conditions where a normal distribution is achieved, the molecules collide ionized with electrons. At the abnormal operation of the ion source According to the invention, the quantity distribution is changed by intermolecular Working together to favor the highest vibrational states. It is believed that the abnormal operation of the ion source follows the interaction of molecules versus molecules favored that of molecules with electrons:

Under normal circumstances the ion source will operate at a pressure of approximately 75 microns. As can be seen from FIG. 5, the highest yield of D _? ⁺ . The invention now makes use of a particularly important property of diatomic deuterium molecules in order to increase the collision-free dissociation in the Migma cell. As is well known, D ₂ ^{+ changes} into a vibrational state of appreciable order when it comes from a splitting of the triatomic D, ⁺ :

D ₃ ⁺ - ^ D ₂ ⁺ * + D ₁ (5)

where D _? represents an ion in its highest excited state. These excited diatomic molecules can be generated by accelerating three atomic molecules to energies in the range of 1 to 2 MeV and then causing them to collide with gas molecules.

6G 9883/0347

590 204 -Vl- 104

From Fig.!? it can be seen that the generation of D, ^{+ increases} with the pressure in the ion source, but this increase decreases above two hundred microns.

According to equation (5), the number of D's must be proportional to the number of D's before the split. The number of D "from the splitting (5) is greatest where the increase in the number of D, in Figure 5 ^a m is lowest. This suggests that the ion source should be operated at the highest possible pressure. The number of Dp but also the total number of D ₂ ⁺ proportional that in the ion source of FIG. 5 are present. Generally speaking, the ion source is operated in a pressure range in which the product of the numbers of D ₇ . and D _? a maximum is the greatest number of expected.

It can now be seen from Fig. 6 that at the normal working pressure of fifty to seventy-five microns the yield of D _{2 is} extremely low compared to that which can be obtained in an operating range of between one hundred and two hundred microns. An operation of the ion source with normally set control parameters for the magnet current and the extraction voltage, but with pressures that are increased in the manner described compared to the normal working pressures, thus provides the desired quantity distribution in the present case

On the other hand, the number of D ₂ available in the beam can also be increased by selecting only the D 1 of the ion source by conventional mass spectroscopy or pulse analysis, as can be carried out, for example, with mass analyzers from General Ionex Corporation . Charged particles of a certain energy and mass are deflected by a given magnetic field at a known angle. The D i can therefore be separated from the generated ion beam and accelerated into a chamber between the accelerator and the Migma cell which contains deuterium gas of sufficient pressure to achieve collisional dissociation of the triatomic molecules in D ₂ ⁺ and D ₁ ⁺ . The D thus produced ₂ ⁺ are mostly in the higher Schwi

60S883 / 03A7

590 204 -Vt- 104

The operation of the Migma cell according to this latter aspect of the Invention can only take place with a reduction in the feed current strength if the molecules of the higher excited state have the average Can reach part of the migma cell without prior collision-free dissociation. With the magnetic field formed as described above of the migma cell are the only paths which extend from the circumference of the space to its center and on this 7 / eg a continuously ascending path Have field strength within the midplane. Therefore, the molecules of higher vibrational state would already be dissociated before they reach the middle part of the migma cell when they are fed in on a path that is angularly offset from the median plane. For example, the ions are fed in at an angle of thirty to sixty degrees with respect to the central plane, as in the case of that mentioned Registration incompatible with a central collisionless dissociation.

The output of the ion source 34 in Pig. 2 is with an accelerator 35 connected (e.g. Type KN 3000 from High Voltage Engineering Corporation). The output beam of the accelerator 35 is via a The beam transport device 36 is fed to the Migma cell 10, its vacuum in the reactor chamber 12 partly also by this device 36 upright is obtained. A device suitable for this purpose (Type 50 from Aero-Vac Corporation) is described in the above-mentioned article in the journal "Nuclear Instruments and Methods ", which also outlines the other functions of the system, such as beam focusing and control rotatable mounting of the Migma cell 10 in its central axis and by means of a bellows connection 38, the angle of the ion beam feed into the Reactor chamber 12 can be changed somewhat.

Another aspect of the present invention is that they the reaction rate in a Migma cell of the described Art can be increased by increasing the number of injected particles that are present in the migma cell can. During the single pass movement through the reactor chamber 12, the number of ions (Nsp. Eh.) Present in can take the distance A-O-B in Fig. 4 at any point in time, by the electrostatic repulsive forces between adjacent Ions limited. The ordered movement according to FIG. 4 can therefore only be maintained when the on neighboring ions through

603883/0347

590 204 - 1 * 9 - 104

-4i *

The focusing forces exerted by the magnetic field are those between them exceed electrostatic repulsive forces. This space charge limit is determined as follows:

Nsp.ch. - 1,635.10 ¹⁵ .kT ^5/2 .m ^1/2 / B _o (6)

This number is smaller than the number Ndia. of the ions that can be admitted on the AOB path before the point is reached at which the magnetic field of the ion current exceeds the field of the magnetic system so far
weakens the fact that an orderly particle movement is disturbed, for example by reducing the magnetic field by a third. KcLia.

is given by the following expression:

Ndia. - 9.10 ⁱ⁶ .T ^1/2 .m ^5/2 / B ₀ (7)

The ratio Ndia / ITspch - 0.1817.10 kT / m in equations (6) and (7) indicates that an increase in the number of ions in the reactor chamber by approximately 3 »5 · 1Ο is available if a collapse is exceeded when Nsp.ch. can be avoided. By controlled introduction and maintenance of electrons in the reactor chamber 12
by means of the arrangement shown in FIG. 7, particles can be up to
approximately this limit (Ndia.) or up to a smaller number, which is above the space charge limit.

In FIG. 7, a migma cell 10 * has the structure discussed in connection with FIGS. 1 and 2. It also contains an electron gun 40, potential thresholds 42 and 44, an anode 46 and measuring probes 50, 52, 54 and 56. The gun 40 is of conventional design and has a
Cathode 40a and a grid 40b. The thresholds 42 and 44 are open,
slightly expanded conical sleeves made of electrically conductive material. The anode 46 is cup-shaped and is also made of metal. the
Elements 14 »16 and 18 in the arrangements according to FIGS. 7 and 1 can consist of stainless steel.

When the reactor chamber 12 'is evacuated, the thresholds 42 and 44
and appropriate potentials applied to anode 46, for example
-14 kV at 42, -13 kV at 44 and -12 kV at 46. When the cathode 40a is on
-20 kV and the grid 40b are set to -I5 kV, the electron gun 40 is energized. It supplies a stream of electrons to the reactor chamber 12 ·, which flows there along the z-axis through the potential thresholds
'42 and 44 »cLie anode 46 and the grid 40b is delimited and by the

609883/0347

590 204 - K -. 104

Field of the magnet system from a radially outward movement towards the lower end piece: the chamber wall is prevented. At the beginning, an electrostatic field is formed in axial alignment with the magnetic field in the z-direction of the magnet system. When the area between the thresholds 4 ² and 44 is filled with electrons, the stored electron space charge also forms a radial electrostatic field. Under the influence of these fields, the electron body fed in rotates as a whole in the border around its axis. A cylinder with a radius twice as large as the electron gun can then be filled with electrons up to a limit which is determined by the electron space charge density and the repulsive potentials of the thresholds 42 and 44. The number of electrons in such a cylinder can be calculated as follows, for example

N = Vd / 1.66.10 " ⁷ = 2.1O ¹¹ (8)

where V = 100 kV is the voltage difference due to the electron space charge over the distance d = 30 cm between the central plane and the Electron gun is.

A feedback with the measuring probes 50, 52, 54 and 56, which one Show electron leakage current from the electron cylinder, and a circuit for averaging the voltages on the measuring probes as a result of the Occurrence of such leakage currents and the corresponding change in the potential applied to the grid 40b is used to reduce the number of electrons To keep N essentially constant. If the averaged voltage on the measuring probes indicates that a predetermined number of electrons has exceeded the number N "and escapes from imprisonment this circuit ensures that the negative potential at the grid 40b is correspondingly reduced, so that the supply of electrons to the reactor chamber is reduced or completely discontinued.

After such introduction of electrons into the reactor chamber the molecular ion beam is fed in and dissociated, so that a migma begins to form. Since the migma is positively charged, Its presence in the chamber reduces the electrostatic repulsion between the electrons and causes less electron leakage current from the chamber to the measuring probes and thus again an increase in the supply of electrons from the cannon 40 into the chamber. on

609883/0347

590 204 - 1 / - 104

this way, the velocity of introduction of electrons becomes substantial kept equal to that of the ion injection.

As expected from the classical oscillator theory, the spatial distribution of the introduced electrons a periodic movement: The electrons pass the migma along the z-axis in Fig. 7 · Then they return, attracted by the migma and by the threshold 44 as well the anode 46 is repelled by the migma back towards the Threshold 42 and the grille 40b. Then the electrons come back attracted by the migma and repelled by the sill 42 and grille 40b.

Because of the normal speed distribution of those sent out by the cannon Electrons tend to cluster in oscillator devices. The oscillator frequency of an electron beam along the The z-axis can be calculated by first determining the transit time of the electrons between the effective grid and the virtual cathode will. The effective grid consists of the positively charged migma and is essentially through the middle plane of the reactor chamber educated. The virtual cathode lies between this grid and the cathode 40a. The electron transit time results from:

<E- s / (2.io ^7. _V y ² ) (9)

where S is the distance between grid 40b and the central plane and V is the

S is the potential difference between the grid and the cathode of the migma. the

Electron oscillator frequency is

^ _e = 7Γ / £ Τ (ίο, 11)

if ^ C = 2x ^.
e

In order to maintain the ordered particle movement according to FIG. 4, the spatial position of the electron body only needs to be changed in such a way that no permanent imbalance is exerted on the orbiting ions, ie the common field from the periodic electrostatic field of the oscillating electron cloud and from the positive ion space charge must change in such a way that the ions do not have enough time to defocus themselves from their trajectories, so that they only perform small periodic deviations from the 'fiep- AOB. This condition ·? is fulfilled if the above electron Schwinfcunp ·? period ^{r is} at most equal to the period ^l £ -.

■ 609883/0 * 347 *

BAD ORIGHiAL

590 204 - 1 / - 104

the ion movement along the path A-O-B.

In the system defined in this way, the movements of the two types of fully ionized matter are ordered differently in time and space. One type (the electrons) is introduced at an energy level of, for example, 100 kV in order to remain free from the influence of the magnetic system, which strives for an orderly movement, and to oscillate essentially perpendicular to the central plane, in which again the other type (the ions ) moves in an orderly manner under the influence of the magnet system. The electron energy level during the entire operation should not be greater than half the energy level of the ions fed in. While instabilities, as they are known from a plasma, e.g. the so-called two-stream instability, can be avoided by choosing ^ C less than the instability increase time, instabilities must be due to undesired frequencies, i.e. those with an increase frequency (or a harmonic thereof ) in resonance with the electron oscillation frequency, are attenuated by suitable cancellation methods using interference, which are also referred to as asynchronous cancellation. In this way, unwanted oscillations can be canceled by applying a time-dependent electric field of the same frequency but a phase shifted by 180 between thresholds 42 and 44, or by so-called "trembling", in which an electric field is applied, the frequency of which is considerable is higher than the frequency of the instability, which removes energy from the instability. This can be achieved in that the ion beam fed in is scanned at this higher frequency. As has also been confirmed by theoretical investigations, the introduced ions and the introduced electrons move against each other without friction and the temperature of the electrons never reaches that of the ions.

Further details on this can be found in an article in the journal Muclear Instruments and Methods 111 (1973) 213.

609883/0347

Claims (1)

Claims t characterized by the following process steps: (b) Formation of a magnetic field in this vacuum, its field strength decreases with increasing radial distance from a central axis of the vacuum and in the central axis with increasing axial distance increases from a median plane perpendicular to the central axis, (c) producing particles of a given polarity with a quantity distribution in which the density of the particles is higher The oscillation state exceeds that of such particles in a normal distribution, and (d) Feeding these particles in a plane with the median plane from the circumference of the vacuum .aus radially inward into the median Part of the magnetic field. 2. The method according to claim 1, characterized in that before Step (c) particles of opposite polarity are introduced into the vacuum whose energy level is at most equal to half is the energy level of the particles of a given polarity. 609883/0347 204 - t ~ 104 5 · The method according to claim 1 or 2, characterized in that the Particles of predetermined polarity are deuterium molecules and that in step (c) the density of the particles to be fed in of the deuterium molecules in the twenty-sixth vibrational state is enlarged. 4. The method according to any one of claims 1 to 3 *, characterized in that in step (c) a deuterium ion source is operated in a pressure range in which the product of the numbers of D ₃ generated. and D "ions is a maximum for this source" 5. The method according to any one of claims 1 to 4, characterized in that that the magnetic field in its central part is given by the following expression: B _z (r, z) / B _o = 1 - k (|) ² ₊ 2k (|) ² , in which B is the field strength in the middle part of the magnetic field, k is a constant smaller than one »E is the maximum radial expansion of the magnetic field, r is the radial distance from the central axis representing variable and ζ is another variable representing the axial distance from the longitudinal plane. 609883/0 3 47 Blank page