EP0165119A1 - Electron multiplier device with electric field localisation - Google Patents

Abstract

1. An electron multiplier device comprising, in a vacuum tube : - an entrance window (FE), - a succession of plane, parallel electrodes comprising small interconnected parallel bars, capable of secondary electrical emission, each dynode stage (D1 ...) comprising two successive planes (D11 , D12 ...) adapted to intercept the electrical trajectories in the manner of a baffle, the width of the bars, in cross-section, being at most equal to 0.5 mm, - an anode capable of localizing the impact of the secondary electrons at its level, and - means (E1 , Ei , R0 -R3 ) connected to these dynode stages (D1 -D10 ) in order to establish between them an electron accelerating electric field, the general direction of which is perpendicular to the electrodes, characterised in that the distance (Z1 ) between two consecutive dynode stages (D1 -D2 ), which is several times greater than the width of the bars, is selected, depending on the electrical field, in such a manner that the secondary electrons originating from the upstream stage (D1 ), in a concentrated distribution, a restricted number of bars of the downstream stage (D2 ), and in that the distance (Z0 ) between the two successive planes of each dynode stage is substantially equal to a quarter of the distance (Z1 ) between dynode stages and is selected, depending on the electrical field prevailing between these two planes, to avoid the recapture of a secondary electron by this dynode stage.

Description

The invention relates to electron multiplier devices, more particularly photomultiplier tubes.

French Patent 78 36148, published under No. 2,445,018, describes an electron multiplier tube capable of "localization". In such a tube, the center of the distribution of secondary electrons on the exit anode corresponds, to a certain extent, to the position of the point of impact of the radiation to be amplified on the entry window of the tube. The word "radiation" is taken here in the broad sense, since it can be photons as well as electrons or other charged particles, capable of causing the extraction of secondary electrons. The electron multiplier previously described gives complete satisfaction, in particular in terms of the spatial resolution that it makes it possible to obtain; but for this it uses the superposition of a magnetic field on the accelerating electric field which the device naturally comprises. The means necessary to obtain this magnetic field tend to complicate the structure of the electron multiplier device, at the same time as increasing its cost. In addition, by their own size, they also tend to reduce the space available for the multiplication of electrons, as well as the width of the input window of the device and / or the access thereof.

As will be seen below, the present invention comes to solve the problem consisting in producing an electron multiplier device, capable of localization, which functions without an added magnetic field, while making it possible to achieve localization properties which are comparable or almost to those of the previously known electric and magnetic field device.

The electron multiplier device comprises, in a vacuum tube, a succession of parallel planar electrodes, which define several stages of dynodes, capable of secondary electronic emission, between an input window and an output anode, and means connected to these electrodes in order to establish between them an electron accelerating field whose general direction is perpendicular to the electrodes.

Furthermore, the proposed electron multiplier device has, in certain respects, a structural relationship with the previous device making use of a magnetic field: in both cases, each dynode stage is defined in two successive planes, each consisting of parallel strips interconnected, and these strips are offset from one plane to another so that these two planes together define a baffle obstacle for the electronic paths which are perpendicular to them. It is important to note immediately that, despite this structural relationship, the functioning is not at all the same in the two cases, because the electronic trajectories obtained by jointly using an electric field and a magnetic field are totally different from those obtained with an electric field alone. In the latter case, the location is essentially defined by the lateral path of the secondary electrons due to the transverse component of the initial velocity. The present invention has made it possible to solve, using an appropriate geometric structure of dynodes, the problem of finding a compromise between gain and spatial resolution, which involve this parameter in opaque directions. This therefore constitutes a first element of the invention.

Besides this, the invention also provides that each dynode stage is arranged so that the majority of the secondary electrons actually leaving a lamella of its foreground does not strike a lamella of its second plane, while the distance between two stages of consecutive dynodes, large compared to the distance between the two planes of the same stage, is chosen, according to the electric field, so that the secondary electrons coming from the stage upstream strike according to a concentrated distribution a number restricted of slats of the downstream floor.

The expression starting "effectively" from a lamella of a given plane of lamellae is used here to take into account the fact that a secondary electron can be recaptured either by the lamella which gave birth to it, or by another lamella of the same plan.

According to another characteristic of the invention, the slats, which are prismatic or cylindrical, have a cross section which projects from the side of the entry window, with two flanks capable of secondary emission and which present themselves in a substantially symmetrical manner relative to the general direction of the electric field; the distance between dynode stages is chosen so that the secondary electrons coming from the upstream stage strike in a substantially balanced manner the flanks of lamellae of the downstream stage which have symmetrical inclinations, which makes it possible to avoid a systematic drift of the location.

In a particular embodiment, which is currently preferred, the cross section of the strips is substantially in the form of an isosceles triangle, where the two equal angles are between 40 ° and 70 ° approximately. It can of course be a curvilinear triangle, or whose sides are deformed in another way, taking into account the machining tolerances applicable to the scale of the lamellae.

According to another particular characteristic of the invention, the secondary electrons coming from a side of a lamella of an upstream stage strike mainly only two lamellae neighboring the first plane of the next downstream stage, and a lamella of the second plan of the same downstream floor.

Advantageously, the distance between stages of consecutive dynodes is chosen to slightly unbalance the impact symmetry, on the downstream stage, of the secondary electrons thus coming from the upstream stage, in order to avoid a shift in the spatial location due to the inclination of the sides.

• - la distance entre étaaea de dynodes consécutifs doit être de l'ordre de huit à dix fois la largeur apparente des lamelles ;
• - la distance entre les deux plans d'un même étage de dynode doit être de l'ordre du quart de la distance entre deux étages de dynodes consécutifs:
• - la largeur apparente (sensiblement la largeur hors tout) des lamelles doit être au plus égale à environ 0,5mm:
• - le champ électrique moyen à l'intérieur du tube électronique doit être au moins égal à environ 500 volts/centimètres;
• - l'énergie initiale des électrons secondaires effectivement émis est, de préférence, au moins égale à 5 électrons-volt environ, et peut aller jusqu'à quelques dizaines d'électrons-volt.

Although these parameters may depend on the particular embodiment concerned, it is currently considered that:

• - the distance between etaaea of consecutive dynodes must be of the order of eight to ten times the apparent width of the strips;
• - the distance between the two planes of the same dynode stage must be of the order of a quarter of the distance between two consecutive dynode stages:
• - the apparent width (appreciably the overall width) of the slats must be at most equal to approximately 0.5 mm:
• - the average electric field inside the electron tube must be at least equal to around 500 volts / centimeters;
• - The initial energy of the secondary electrons actually emitted is preferably at least equal to about 5 electron volts, and can range up to a few tens of electron volts.

All the lamellae of the tube can be parallel, but the localization properties can also be improved by orienting them in different directions along the different stages of dynodes, in a regular manner. The easiest way is to make the slats of a dynode stage perpendicular to those of the previous stage.

The invention also allows good detection for an isolated photo-electron (or an isolated incident charged particle). To this end, it is expected that the electrical voltage prevailing between the two planes of the same stage of dynodes is at most equal to about 50 volts, at least for the first stages of dynodes.

According to another caractéristngue still of the invention, means are provided for adjusting the feed of the electrodes in order to optimize the spatial resolution of the multiplier El ^p ctrons.

Depending on the applications, the latter may include a cathode or a photocathode near the first dynode.

Although a conventional anode is sufficient in certain cases, it in principle comprises, as anode, a divided anode with multiple connections, an electroluminescent surface, a resistive anode or any equivalent means allowing the use of the location property.

• - la figure 1 est une vue en coupe verticale d'un photomultiplicateur selon l'invention;
• - la figure 2 est une vue en coupe horizontale du photomultiplicateur de la figure 1;
• - la figure 3 est un schéma électrique illustrant l'interconnexion des électrodes du même photomultiplicateur;
• - la figure 4 est un schéma partiel de deux étages de dynodes consécutifs du photomultiplicateur des figures 1 et 2; et
• - la figure 5 est un diagramme tendant à permettre une interprétation de la résolution spatiale dans la direction X perpendiculaire à grande dimension des lamelles.

Other characteristics and advantages of the invention will appear on examining the detailed description below, and the appended drawings, in which:

• - Figure 1 is a vertical sectional view of a photomultiplier according to the invention;
• - Figure 2 is a horizontal sectional view of the photomultiplier of Figure 1;
• - Figure 3 is an electrical diagram illustrating the interconnection of the electrodes of the same photomultiplier;
• - Figure 4 is a partial diagram of two stages of consecutive dynodes of the photomultiplier of Figures 1 and 2; and
• - Figure 5 is a diagram tending to allow an interpretation of the spatial resolution in the direction X perpendicular to large dimension of the lamellae.

In the present invention, the qeometry of the main constituents of the electron multiplier tube is important. Consequently, the drawings are to be considered as incorporated into the present description, to contribute, where appropriate, to supplement it, as well as to define the invention.

The detailed description below is concerned with a photomultiplier tube. In such a tube, the incident signal is delivered by photons, which we know can excite the dynodes of an electron multiplier, either directly or through a photocathode. However, the present invention may have applications other than photonics, because, more generally, it may be the electrons themselves, or other types of charged particles, which define the input signal of an electron multiplier tube. .

In FIGS. 1 and 2, the photomultiplier tube comprises a vacuum chamber TPM, which houses its main constituents. Figure 1 shows that this enclosure has in the upper part a flat FE entry window. Just behind this window is placed a proximity photocathode denoted PPC. Below this (figure 1), ten stages of d ^y nodes D _l to D _{10 are provided} . These are followed, further down, by an anode divided into "mosaic" This anode comprises a large number of elements such as A and A _i , respectively connected to individual electrical output connections EA ₁ and EA _i . As a whole, the anode elements are denoted A. Finally, other electrical connections such as E ₁ and E make it possible to bring the internal electrodes of the photomultiplier to the potentials suitable for its operation.

FIG. 2 also shows the generally circular shape of the support structure SP which supports the dynodes; this structure is provided with insulating columns such as CP.

FIG. 3 illustrates the electrical diagram associated with the photomultiplier, the TPM enclosure of which is recalled in dashed lines. It is easier to see that each dynode stage such as D ₁ comprises, according to the invention, two levels or planes of electrodes such as D ₁₁ and D ₁₂ , placed one after the other along the axis F electric field tube, and perpendicular to this axis.

The proximity photocathode PPC is connected to a voltage - HT by the electrical connection E ₁ . At the other end, the electrical connection E ₂ is connected to ground. A voltage divider network with resistors is mounted between line E ₂ and line E _{1 in} order to provide each of the dynode planes with an appropriate electrical voltage. The high supply voltage defines the potential difference, therefore the electric field, between the different dynode planes. The resistors are adjusted so that this electric field is made as uniform as possible.

In practice, apart from the end resistances R _O and R _31, a resistance R _{1 is provided} between the first plane of each dynode (for example the plane D ₂₁ of the dynode D ₂ ), and the last plane of the previous dynode (in this case the plane D ₁₂ of the dynode D ₁ ). A lower resistance R ₂ is provided between the two planes of each stage of dynodes (for example between the planes D ₂₁ and D ₂₂ of the dynode D ₂ ). The addition of capacities may possibly be required at certain points of this series resistive network, in particular on the top floors. The anodes A _n are connected to ground by individual resistors.

FIG. 4 illustrates on a larger scale two stages of consecutive dynodes, which are supposed to be stages D ₁ and D ₂ . As previously indicated, the stage D ₁ comprises two planes D ₁₁ and D ₁₂ of dynode elements. The stage D ₂ also includes two planes D ₂₁ and ^D ₂₂ of dynode elements.

Individually, the dynode elements are prismatic or cylindrical lamellae, parallel to each other, and of course coplanar within the same plane of dynodes. These slats are properly treated to have the property of secondary electronic emission, on their faces oriented towards the side of the FE input window, that is to say for any arrival in the direction P of a photon or of a charged particle such as 'an electron. This direction P is parallel or slightly inclined to the general direction of the axis F, along which the electric field inside the tube is established approximately.

It is currently considered preferable to use elements of dynodes whose cross section is in the form of an isosceles triangle. The base B, adjacent to the two equal angles of the isosceles triangle, is perpendicular to the general direction F. It is turned downstream. The two equal sides L and R of the isosceles triangle are made capable of secondary electronic emission, and it is observed that they face symmetrically with the general direction of incidence P. For its part, the anal α is advantageously understood between 40 and 70 ° approximately. In the example illustrated, the lamellae have a cross section in an isosceles right triangle.

The "apparent width" of the slats can be defined as the overall width that they present, perpendicular to the direction F. This width is here equal to the base B of the isosceles right triangle, which measures approximately 0.5 mm in this example. A spacing of 0.5mm is also provided between the adjacent vertices (of angle α) of two strips of the same plane of dynodes. Finally, the lamellae of the second plane of a dynode stage, for example the plane D ₁₂ of the stamen D ₁ , are interspersed with those of the previous plane, here D ₁₁ . Therefore, all the dynode elements of the two planes of the same dynode stage appear as an obstacle, or a baffle, for the (electronic) trajectories parallel to the direction F.

Furthermore, we denote by Z ₀ the distance between two planes of d ^y no ^d D ₁₁ and D ₁₂ of the same floor, distance taken in the direction F. We denote by Z ₁ the distance taken in the same way between two floors consecutive dynodes, that is to say for example between the first plane D ₁₁ of the stage D ₁ and the first plane D ₂₁ of the stage ^q e D ₂ . Preferably, Z ₁ is approximately equal to four times Z ₀ .

In a particular embodiment, we take Z ₀ = 1 mm, and Z ₁ = 4 mm, so that the distance between two stages of dynodes is of the order of eight to ten times the apparent width of the lamellae forming the elements. individual dynodes.

In FIG. 4, we now consider the trajectories of the secondary electrons which leave from the right flank of the lamella D ₁₁₀ . N denotes the normal to this right flank, at the starting point of its electrons.

It is also necessary to define the lower limits of the initial energy of the secondary emissions, as well as of the angle of emission, counted in the trigonometric direction compared to the normal N. This angle of emission is of course limited to the electrons useful secondaries, i.e. those which are not recaptured by the same plane of lamellae. It has been observed that the initial energy must be greater than about 5 electron volts, and that the initial emission angle must be less than 45 °, that is to say that the useful secondary elections are included in a cone with an angular opening of 45 ° from normal.

It has also been observed that the width of the slats must then be at most equal to 0.5 mm. for an electric field of 500 volts / cm. This field value corresponds to a voltage of 50 volts between the two planes D ₁₁ and D ₁₂ of the dynode D ₁ , since Z _o = 1 mm.

Beyond this upper limit for the value, a significant part of the secondary electrons emitted by the coverslip will be recaptured by the emissive surface of origin, due to the strong electric field which prevails. The above takes into account an angular law of emission of secondary electrons around the normal N which is established in cosine θ.

Furthermore, an energy filtering of the electrons takes place, due to the presence of the adjacent lamella D _{111 '.} It has been observed that the maximum energy of the secondary electrons actually leaving the lamella D _{110 is} established at a few tens of electron volts, in this case about 15 electron volts.

For a given angle of emission, for example = 0 °, _ob holds the kind of trajectories T and T _{lmin lmax.} corresponding respectively to 5 electron volts and 15 electron volts. These trajectories practically strike only the two lamellae D ₂₁₁ and D ₂₁₂ which form part of the first plane D ₂₁ of the dynode stage following D ₂ . The trajectory with energies close to these extreme values hits the same lamellae. On the other hand, part of the intermediate energy trajectories pass between the lamellas D ₂₁₁ and D ₂₁₂ , to strike, in a substantially symmetrical manner, the two sides of the lamella D ₂₂₂ , which is part of the second plane D ₂₂ of the D2 dynode stage. The intermediate trajectory corresponding to an energy of about 10 electron volts has been shown in T _med . Careful observation shows that one of the trajectories T _ex would pass between the lamellae D ₂₁₂ and D ₂₂₂ . In fact, this is a very small fraction (in terms of probability) of the secondary electrons emitted. A secondary electron which propagates along this trajectory would moreover be picked up by the next dynode stage. In addition, it can be estimated that the edge effects produced on the electric field by the tips of the strips D ₂₁₂ and D ₂₂₂ would in fact allow the effective capture of the secondary electron at the level of the dynode D ₂ , as a result of which it can then emit secondary electrons again, as will have done the other trajectories arriving on the dynode D ₂ .

While the above concerns the first plane of a dynode stage, it has been observed that the second plane also allows localization (Figure 4.).

The operating conditions which have just been described only involve the projection of the electronic trajectories onto the X-Z plane. It has been observed, however, that a correct functioning is thus obtained, in terms of location, not only in the X direction, but also in the Y direction.

• - la distance Z₁ entre deux étages de dynodes consécutifs, qui est grande par rapport à la distance Z₀ entre les deux plans d'un même étage, peut être ajustée en fonction du champ électrique de sorte que les électrons secondaires provenant de l'étage amont D₁ frappent selon une distribution concentrée un nombre restreint de lamelles de l'étage aval D₂;
• - de plus, lorsqu'on utillse, comme c'est le cas ici, des lamelles dont la spction droite est symétrique autour de l'axe F, il a été observé que la distance Z₁ peut être choisie de sorte que les électrons secondaires provenant du premier plan de l'étage amont frappent de manière sensiblement équilibrée des flancs des lamelles de l'étage aval qui ont des inclinaisons symétriques. Et ceci s'étend aux électrons secondaires provenant du second plan de l'étage amont.

The above description shows that:

• - the distance Z ₁ between two stages of consecutive dynodes, which is large compared to the distance Z ₀ between the two planes of the same stage, can be adjusted according to the electric field so that the secondary electrons coming from the upstream stage D ₁ strike in a concentrated distribution a limited number of lamellae of the downstream stage D ₂ ;
• - in addition, when using, as is the case here, lamellae whose straight spction is symmetrical around the axis F, it has been observed that the distance Z ₁ can be chosen so that the secondary electrons from from the foreground of the upstream stage strike in a substantially balanced manner the sides of the lamellae of the downstream stage which have symmetrical inclinations. And this extends to the secondary electrons coming from the second plane of the upstream stage.

Furthermore, it has been observed that the distribution in the direction Y parallel to the large dimension of the lamellae is interpreted by a simple convolution of the lateral paths of the secondary electrons at each of the downstream stages.

Reference is now made to FIG. 5.

This shows the binomial probabilistic distribution characterized by p = q, where p and q are the probabilities that a secondary electron strikes the right flank and the left flank, respectively, of lamellae of the following stage. The figures placed inside a circle are proportional to the probability of production of the secondary electrons starting from only one electron starting from the first stage of dynode (n = l), the other stages being numbered in increasing number on l vertical axis pointing down to the anode. The horizontal axis corresponds to distances expressed in units of the average lateral path of the secondary electrons at the level of the next stage. These distances are denoted X (p).

In the X direction, it therefore appears that a very concentrated distribution of the secondary electrons is obtained at the anode, this distribution being practically centered on the initial axis F ₀ . The offset is mainly due to the inclination of the flank of the coverslip which gives the secondary electron preauer. But we do not observe, by surte, a global drift of the flow of secondary electrons by tappport to the axis F ₀ , drift which would amplify from one stage to another (provided that p = q). This finally results in a slight lateral offset, since if the number circled on the left is indeed on the axis F ₀ in Figure 5, the number 126 circled on the right is slightly offset, which corresponds to an offset in the distribution . It has been observed that this offset can be corrected by varying the values of p and q by about 10%. This can be obtained by acting on the distance Z ₁ , as will be understood by those skilled in the art. But this action plays in the same way, whatever the inclination of the face or flank of the coverslip that produced the initial electron.

• - la largeur de lamelles ^£ est choisie de façon que o (E,Z= ℓ/2) soit plus grand que ℓ/2 (pour un gain élevé), mais que p(E,Z=Z₁) soit le plus petit possible (pour une bonne localisation).
• - la distance Z₁ est également choisie par un compromis entre la résolution, o (E,Z=Z₁), et la largeur de la distribution des électrons,qui est également proportionnelle àc, et qui doit être suffisamment qrande par rapport à ℓpour éviter la dérive systématique en X.

The mean lateral path, p (E, Z) of the secondary electrons plays an essential role in this device. It turns out that the geometry of dynodes can be defined from this parameter; for example :

• - the width of the slats ^£ is chosen so that o (E, Z = ℓ / 2) is greater than ℓ / 2 (for a high gain), but that p (E, Z = Z ₁ ) is the smallest possible (for good localization).
• - the distance Z ₁ is also chosen by a compromise between the resolution, o (E, Z = Z ₁ ), and the width of the distribution of the electrons, which is also proportional to c, and which must be sufficiently large with respect to ℓfor avoid systematic drift in X.

• - hauteur environ 65 mm;
• - diamètre extérieur 139 mm
• - fenêtre d'entrée de diamètre 100 mm, munie d'une photocathode de proximité:
• - étages de dynode comme décrit plus haut, avec une différence de potentiel d'environ 50 volts entre deux plans d'un même étage de dynode, et une différence de potentiel d'environ 200 volts entre étages de dynodes:
• - anode divisée en 164 éléments d'environ 7 x 7mm², séparés d'un intervalle d'environ 0,5mm;
• - on obtient ainsi un gain de 10⁶ à 10⁷ pour dix étages de dynodes.

A photomultiplier device constituted as described above can be loocated in a tube constituted as follows:

• - height about 65 mm;
• - outside diameter 139 mm
• - 100 mm diameter entry window with proximity photocathode:
• - dynode stages as described above, with a potential difference of around 50 volts between two planes of the same dynode stage, and a potential difference of around 200 volts between dynode stages:
• - anode divided into 164 elements of approximately 7 x 7mm ² , separated by an interval of approximately 0.5mm;
• - This gives a gain of 10 ⁶ to 10 ⁷ for ten stages of dynodes.

Basically, the resolution obtained is approximately 12 mm in the X direction transverse to the large dimension of the lamellae, and approximately 10 mm in the Y direction parallel to the large dimension of the lamellae. In fact, the same resolution is obtained in these two directions X and Y, although the structure of a given plane of lamellae is not at all isotropic.

To further equalize the resolution in X and in Y, it is of course possible to alternately cross the direction of the lamellae in the stages of successive dynodes. The optimal spatial resolution can be easily obtained by adjusting the high voltage, which acts globally on the electric field, or even by a finer action on the electric field at the level of each of the stages and the dynode planes.

The photomultiplier thus obtained has a very large sensitive surface, for a sensitivity which can become comporable to that of the prior device. Indeed, an improved spatial resolution can be further obtained by reducing the dimension r of the dynode strips, and by acting in a corresponding manner on the electric field and the vertical dimensions (or longitudinal) of the device.

Such resolution characteristics are sufficient for a large part of the applications. They are particularly suitable for applications such as X-ray or Y-ray imaging.

For example, when performing Y-ray imaging using an Anger-type camera, consisting of a sodium iodide crystal with a thickness of 10 mm, and, as a detector, an array of 2 inch (50 mm) photomultipliers, coupled directly to the crystal, the spatial resolution obtained after calculation of the barycenter is at best of the order of 4mm. Under these conditions, it is observed that the spatial resolution is dominated by the resolution of the detector, approximately 50 mm, which is too small compared to the size of the spot of the scintillation beams which is approximately twice the thickness of the crystal. , or 20 mm.

At this level, even a limited resolution of the detectors can improve the final resolution by an important factor. Indeed, with a resolution of 10 mm from the photodetector, one can reach a final resolution of 1.6 mm.

This is precisely what the photomultiplier device described in detail above can do.

Finally, it should also be noted the excellent properties obtained by the electron multiplier according to the invention, ib terms of response time and linearity of the tin, in addition to the spatial resolution already mentioned.

Claims (13)

1. Electron multiplier device, comprising, in a vacuum tube, a succession of parallel planar electrodes, which define several stages of dynodes (D ₁ -D ₁₀ ) capable of secondary electronic emission, between an input window (FE) and an output anode (An) and means (E ₁ , E ₁ , R ₀ -R ₃ ) connected to these electrodes (D- ₁ -D ₁₀ , An) in order to establish between them an electric field electron accelerator, the general direction of which is perpendicular to the electrodes, characterized in that each dynode stage (D ₁ -D ₁₀ ) is defined in two successive planes (D ₁₁ , D ₁₂ , ...) each made up of lamellae interconnected parallels, and that these lamellae (D ₁₁₀ , D ₁₂₀ ...) are offset from one plane to another so that these two planes together define a baffle obstacle for the electronic trajectories which are perpendicular to them, and
in that, each dynode stage (D ₁ -D ₁₀ ) being arranged so that the majority of the secondary electrons actually leaving a lamella (D ₁₁₀ ...) from its foreground does not strike a lamella (D ₁₂₁ , D ₁₂₂ ) in the background,
the distance (Z ₁ ) between two stages of consecutive dynodes (D ₁ -D ₂ ), large compared to the distance (Z _O ) between the two planes (D ₁₁ -D ₁₂ ) of the same stage, is chosen, as a function of the electric field, so that the secondary electrons coming from the upstream stage (D ₁ ) strike selcn a concentrated distribution a restricted number of lamellae of the downstream stage (D ₂ ). 2. Saddler device of claim 1, characterized in that the lamellae, prismatic or cylindrical, have a cross section which projects from the side of the entrance window, with two sides (L, R) capable of emission secondary and substantially symmetrical with respect to the general direction (F) of the electric field, and in that the distance (Z ₁ ) between dynode stages is chosen so that the secondary electrons coming from the upstream stage (D ₁ ) strike in a substantially balanced manner from the sides of the lamellae of the downstream stage which have symmetrical inclinations (D ₂₁₁ R, D ₂₁₂ L, ^D ₂₂₂ ^R and ^L ). 3. Device according to claim 2, characterized in that the cross section of the lamellae is substantially in the form of an isosceles triangle, where the two equal angles are between 40 and 70 ° approximately. 4. Device according to one of claims 2 and 3, characterized in that the distance (Z ₁ ) between stages of consecutive dynodes is chosen to slightly unbalance the impact symmetry, on the downstream stage, of the secondary electrons coming from the upstream floor, in order to avoid a shift in spatial location due to the inclination of the sides. 5. Device according to one of claims 1 to 4, characterized in that the apparent width of the strips is at most equal to about 0.5 mm. 6. Device according to one of claims 1 to 5, characterized in that the average electric field is at least equal to about 500 v / cm. 7. Device according to one of the preceding claims, characterized in that the initial energy of the secondary electrons actually emitted is at least equal to 5 electron volts, enviton. 8. Device according to claim 7, characterized in that the initial energy of the secondary electrons actually emitted is limited to a few tens electron volts. 9. Device according to one of the preceding claims, characterized in that at least two stages of consecutive dynodes have their lamellae oriented in different directions, preferably perpendicular. 10. Device according to one of the preceding claims, characterized in that the electrical voltage prevailing between the two planes of the same dynode stage is at most equal to approximately 50 volts, at least for the first stages, which allows a good detection of an isolated photo-electron. 11. Device according to one of the preceding claims, characterized by means for adjusting the supply of the electrodes in order to optimize the resolution. 12. Device according to one of the preceding claims, characterized in that it comprises a cathode or photocathode (PPC) near the first dynode. 13. Device according to one of the preceding claims, characterized in that it comprises, as anode (An), a divided anode with multiple connections, an electroluminescent surface, or a resistive anode.

Images