Switch with phase change material

Abstract

Switch comprising at least one PCM portion that can be in a conducting or blocked state depending on the amorphous or crystalline state of the PCM that can change state when it is heated, in which the PCM portion is continuous and has an elongated shape such that an input and an output of the switch are connected to two ends of the PCM portion respectively that are separated from each other by a distance corresponding to the largest dimension of the PCM portion, and comprising a control device of the state of the switch capable of passing heating currents through the PCM portion, approximately perpendicular to the largest dimension of the PCM portion, from at least two input points separated from each other and separated from the ends of the PCM portion, to at least two output points separated from each other and separated from the ends of the PCM portion.

Claims

1. A switch comprising at least one PCM portion that can be in a conducting or blocked state depending on the amorphous or crystalline state of the PCM that can change state when it is heated, in which the PCM portion is continuous and has an elongated shape such that an input and an output of the switch are connected to two ends of the PCM portion respectively that are spaced at a distance from each other corresponding to the largest dimension of the PCM portion, and comprising a device to control the state of the switch capable of passing heating currents through the PCM portion, approximately perpendicular to the largest dimension of the PCM portion, from at least two input points at a distance from each other and at a distance from the ends of the PCM portion, to at least two output points at a distance from each other and at a distance from the ends of the PCM portion.

2. The switch according to claim 1, in which the control device comprises a plurality of electronic components that form first current sources each capable of injecting one of the heating currents into the PCM portion from one of the input points, and second current sources each capable of extracting one of the heating currents from one of the output points.

3. The switch according to claim 2, in which the first and second current sources formed by the electronic components of the control device can be controlled simultaneously by a same control signal CTR.

4. The switch according to claim 3, in which each of the first current sources comprises at least one MOS transistor comprising one of its source and drain electrodes connected to one of the input points and the other of its source and drain electrodes to an electrical power supply voltage, and in which the control signal CTR can be applied to the gate.

5. The switch according to claim 3, in which each of the second current sources comprises at least several MOS transistors mounted in a current mirror, at least one of said MOS transistors having a gate to which the control signal CTR can be applied.

6. The switch according to claim 5, in which each of the second current sources also comprises at least one resistive element coupled to the current mirror and capable of forming a resistance with a first or a second value, different from each other and such that the value of the heating current extracted by the second current source is adapted as a function of the state of the switch that will be controlled by the control device.

7. The switch according to claim 3, in which each of the second current sources are formed from at least one digital - analogue converter.

8. The switch according to claim 1, in which the PCM portion is in the form of a rectangular parallelepiped with a length greater than its width and its thickness.

9. The switch according to claim 8, in which the input points and output points are located on side faces of the PCM portion, separated from each other by a distance equal to its width.

10. The switch according to claim 1, comprising several PCM portions located side by side and approximately parallel to each other, the control device being capable of passing heating currents through each of the PCM portions, approximately perpendicular to the largest dimension of each of the PCM portions, from at least two input points separated from each other and separated from the ends of each of the PCM portions, to at least two output points separated from each other and separated from the ends of each of the PCM portions.

11. The switch according to claim 2, in which the switch also comprises a device coupled to the ends of the PCM portions and capable of compensating for parasite capacitances of MOS transistors that will activate or deactivate the first and second current sources.

12. The switch according to claim 2, in which said electronic components comprise active zones formed in at least one semiconducting layer, and in which the PCM portion is located within a metallic interconnections level located above the semiconducting layer and separated from the semiconducting layer by at least one interlayer dielectric.

13. A transmission line matching device, comprising at least: a first portion of electrically conducting material comprising a first end that will be connected to the transmission line and a second end connected to at least one second portion of the electrically conducting material that will be connected to a reference potential, a plurality of switches according to claim 1, the ends of the PCM portions of each of the switches being coupled to the first and second portions of electrically conducting material at different distances from the first end of the first portion of electrically conducting material such that the matching length of the device can be defined by the distance between the first end of the first portion of electrically conducting material and the switch in the conducting state closest to the first end of the first portion of electrically conducting material.

14. A phase shift device for a differential transmission line, comprising at least: several first delay lines that will be coupled to a first electrically conducting portion of the differential transmission line; several second delay lines that will be coupled to a second electrically conducting portion of the differential transmission line; several switches according to claim 1, the ends of the PCM portions of each of the switches being coupled to one of the first delay lines and to one of the second delay lines such that said first and second delay lines can be electrically connected to each other when the switch is in the conducting state and can be connected to a reference electrical potential when the switch is in the blocked state.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) This invention will be better understood after reading the description of example embodiments given purely for information and that are in no way limitative with reference to the appended drawings on which:

(2) FIG. 1 diagrammatically shows a switch according to one particular embodiment;

(3) FIG. 2 diagrammatically shows a PCM portion of a switch according to one particular embodiment;

(4) FIG. 3 diagrammatically shows some of the electronic elements of a control circuit of a switch according to a first particular embodiment;

(5) FIG. 4 diagrammatically shows an equivalent electrical circuit for a PCM portion in the amorphous state coupled to a control circuit of the switch;

(6) FIG. 5 diagrammatically shows an equivalent electrical circuit for a PCM portion in the amorphous state coupled to a control circuit of the switch, and a substrate on which this PCM portion is located;

(7) FIG. 6 diagrammatically shows the different parasite capacitances connected to the PCM portion and that are connected to the transistors of the control circuit of the switch;

(8) FIG. 7 diagrammatically shows an equivalent electrical circuit of the PCM portion of the switch, coupled to the current sources formed by the control circuit when the PCM is in the crystalline state;

(9) FIG. 8 diagrammatically shows a switch comprising a device for compensation of parasite capacitances of transistors of the switch control device;

(10) FIG. 9 diagrammatically shows a matching device for a transmission line comprising a plurality of switches;

(11) FIGS. 10 and 11 diagrammatically show a phase shift device for a differential transmission line comprising a plurality of switches;

(12) FIGS. 12 and 13 diagrammatically show a part of the electronic elements of a control circuit of a switch according to different variant embodiments;

(13) Identical, similar or equivalent parts of the different figures described below have the same numeric references to facilitate comparison between the different figures.

(14) The different parts shown on the figures are not necessarily all at the same scale, to make the figures more easily understandable.

(15) It must be understood that the different possibilities (variants and embodiments) are not mutually exclusive and that they can be combined with each other.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

(16) Refer firstly to FIG. 1 that diagrammatically shows one particular embodiment of a switch 100.

(17) In this case, the switch 100 is inserted between two metallic lines 102, 104. Line 102 is the RF signal input line into the switch 100 and forms the input to the switch 100. Line 104 is the output line of these RF signals and forms the output from the switch 100.

(18) In this particular embodiment, the switch 100 forms a switch that will or will not allow RF signals to pass from line 102 to line 104, depending on its state (conducting or non-conducting). As a variant, the switch 100 could form several switches capable of routing or directing signals to different conducting paths.

(19) The switch 100 comprises several PCM portions 106 (four portions on FIG. 1, references 106a-106d) each of which has an elongated shape, in other words one dimension (the length) of each portion is larger than its other dimensions (for example its width and thickness). The largest dimension of each PCM portion 106 is approximately parallel to the direction of displacement of RF signals that will pass through the switch 100 from line 102 to line 104. Furthermore, each PCM portion 106 is continuous, unlike the case of several PCMs separated from each other.

(20) The PCM of portions 106 may for example be GeTe, or any other appropriate phase change material, for example comprising at least one of the following materials: Ge.sub.2Sb.sub.2Te.sub.5, GeTeN, GeTeC, GeTeB.

(21) One of the PCM portions 106 is shown in more detail on FIG. 2. In this case the portion 106 is elongated along the direction of the X axis, in other words its largest dimension L named the length is parallel to this X axis. The RF signals will pass through the portion 106 from a first end 108 connected to line 102 to a second end 110 of the portion 106 connected to the line 104, these two ends 108, 110 being separated from each other by a distance equal to the length L.

(22) On the example in FIG. 2, the portion 106 comprises a section in a plane perpendicular to its length L, approximately rectangular in shape and with dimensions ab, where a is the thickness of the portion 106 and b is its width. Considering the different dimensions of portion 106, the following relations are satisfied: L>a and L>b. The L/b ratio is also advantageously greater than or equal to 30.

(23) In general, each portion 106 may comprise a section in a plane perpendicular to its length L, with a shape other than rectangular, the dimensions of this section (sides, diameter, etc.) being less than the length L of the portion 106.

(24) The phase change of the material in each portion 106 is controlled by a control circuit 116 (not shown on FIG. 2) forming several heating current sources that will impose electrical currents at several input points of each of the portions 106, the currents passing through each of the portions 106 from these input points to exit from one of the portions 106 at several output points. The control circuit 116 also forms current sources connected to these output points to draw or extract heating currents from these output points to carry them outside portions 106. On the example in FIG. 2, the figure symbolically shows that heating currents i pass through the portion 106 from three input points 112, these heating currents i being output from portion 106 at three output points 114 each approximately facing one of the input points 112.

(25) The input and output points 112, 114 of the heating currents passing through each of the portions 106 are preferably defined, in other words are positioned on portion 106, such that the heating currents pass through the portion 106 parallel to its width b. However, it is possible as a variant that these input and output points 112, 114 could be arranged such that the heating currents that will change the state of the PCM of portion 106 pass through the portion 106 parallel to its thickness a.

(26) The control circuit 116 forms a set of programmable heating current sources that inject local heating currents through the width of the portions 106, and extract these heating currents.

(27) FIG. 3 diagrammatically shows some of the electronic elements of the control circuit 116 according to one particular example embodiment, the shown electronic elements sending a heating current through the portion 106 from one of the input points 112 and extraction of this heating current from one of the output points 114.

(28) These electronic elements include a first MOS transistor 118 with its gate connected to either the source or drain electrode (to the drain in the case of an N type transistor 118 as shown on FIG. 3) itself connected to an electric power supply potential V.sub.DD. The other source and drain electrode of the first transistor 118 is connected to one of the source and drain electrodes of a second MOS transistor 120 (drain in the case of an N type transistor 120), the other source or drain electrode of which is connected to the input point 112. A control signal CTR will be applied onto the gate of the second transistor 120.

(29) The first and second transistors 118, 120 form a first controlled current source capable of injecting a heating current i into the input point 112 when the CTR signal makes the second transistor 120 conducting. The first transistor 118 forms an element insulating the second transistor 120 from the potential V.sub.DD when the second transistor 120 is blocked.

(30) The other electronic elements of the control circuit 116 that can be seen on FIG. 3 form a second controlled current source that can extract or pull the heating current injected into the input point 112, from the output point 114.

(31) These electronic elements include a third MOS transistor 122 of which one of the source and drain electrodes (the drain in the case of an N type transistor) is connected to the output point 114. The control signal CTR will be applied onto the gate of the third transistor 122. The other source and drain electrode of the third transistor 122 is connected to one of the source and drain electrodes (drain in the case of an N type transistor) of a fourth MOS transistor 124, the other source or drain electrode of which is connected to the ground.

(32) The other electronic elements forming the second current source also include fifth and sixth transistors 126, 128 coupled to the third and fourth transistors 122, 124 such that these four transistors 122-128 form a current mirror. Thus, the control signal CTR will also be applied onto the gate of the fifth transistor 126. Moreover, the gate of the sixth transistor 128 is connected to the gate of the fourth transistor 124 and to the source and drain electrode of the sixth transistor 128 that is connected to the fifth transistor 126.

(33) One of the source and drain electrodes of the fifth transistor 126 (that is not connected to the sixth transistor 128, and that corresponds to the drain in the case of an N type transistor) is connected to a resistive element 130 that may be equal to two different electrical resistance values named R1 and R2, depending on a control signal applied on this element. This variable resistive element 130 is also connected to the electrical power supply potential V.sub.DD.

(34) In the configuration shown on FIG. 3, the current mirror comprises a first branch formed by the resistive element 130 and the fifth and sixth transistors 126, 128 in which a reference current circulates, the value of which is imposed by the value of the resistance of the resistive element 130 and by the value of the saturation voltage V.sub.Dsat of the fifth transistor 126 and the value of the gatesource voltage V.sub.GS of the sixth transistor 128. The current mirror also comprises a second branch comprising the third and fourth transistors 122, 124 into which the current circulating in the first branch is duplicated, thus forcing extraction of the heating current from the portion 106.

(35) Although not shown, the control circuit 116 includes electronic elements similar to the elements 118 to 130 described above to form the other heating current sources that will heat the other parts of the portion 106 and the other portions 106. For example, considering the example of portion 106 shown on FIG. 2 in which three heating currents i are injected from three input points 112, the electronic elements of the control circuit 116 heating this portion 106 include three sets of electronic elements, each similar to that corresponding to the electronic elements 118 to 130 shown on FIG. 3.

(36) The number of heating current sources formed by the control circuit 116 for each portion 106, and the distances between the input points 112, between the output points 114, between the input and output points 112, 114 and the ends 108, 110 of the portion 106, depend particularly on the rate at which heat propagates in the PCM, in other words the thermal resistance associated with this PCM, and its thermal capacity. For example, for a GeTe portion 106 with an approximately rectangular parallelepiped shape as shown in FIG. 2, and having a section (area of the ends 108, 110) between about 100 nm.sup.2 and 10000 nm.sup.2, the number of heating current sources is chosen such that the input points 112 are separated from each other by a distance equal to about 10 m, so that a phase change of the PCM of the portion 106 can be obtained after a heating period compatible with temporal specifications required for a state change of a switch used as an RF antenna switch. This distance between two adjacent input points 112 is similar to the distance between two adjacent output points 114. The distance between the ends 108, 110 of the portion 106 and the input and output points 112, 114 closest to these ends 108, 110 is for example equal to half this separation distance between two adjacent input points 112 and between two adjacent output points 114, for example in this case about 5 m.

(37) In order to change the switch 100 from the blocked state to the conducting state, that corresponds to a change of the PCM of the portions 106 of the switch 100 from the amorphous state to the crystalline state, the control signal CTR changes to the high state (V.sub.DD) to make the transistors 120, 122 and 126 conducting. The heating current injected into the portion 106 by each first current source formed by the circuit 116 passes through the width b of the portion 106, this width b for example being of the order of a few hundred nanometers. At the beginning of this state change, due to the short distance (less than 1 m) between the input point 112 and the facing output point 116 and the large distance from the other adjacent input points 112 (for example about 10 m), the impedance between the input point 112 and the output point 116 faced by the first current source formed by the first and second transistors 118, 120, may for example be between about 20 k and 100 k that is less than the impedance between this input point 112 and an adjacent input point 112 that may for example be more than about 1 M. Consequently, the current injected in the input point 112 circulates from this input point 112 to the output point 114 located facing this input point 112. This current circulating in the PCM thus creates self-heating of the PCM, this heating increasing with increasing resistance of the PCM.

(38) The electronic elements of the control circuit 116 are sized such that when the state of the switch 100 changes from the blocked state to the conducting state, for a power supply voltage V.sub.DD of the order of 2V, the heating current named I1 in this case passing through the portion 106 causes a voltage drop between the input point 112 and the output point 114 less than about 1V for transistors comprising active silicon zones, for example made using the 130 nm PD-SOI (partially deserted silicon on insulator) technology on an HR (High Resistivity, e.g. of the order of 1 k.Math.cm) silicon substrate. Furthermore, the value R1 of the variable resistance 130 during this state change may be chosen such that:

(39) R1I1+V.sub.Dsat(126)+V.sub.GS(128)=V.sub.DD, where V.sub.Dsat(126) is the saturation voltage of the fifth transistor 126, and V.sub.GS(128) is the voltage between the gate and the source of the sixth transistor 128.

(40) This heating of the PCM, above the PCM changeover temperature, causes a change in its phase corresponding to the change from the amorphous state to the crystalline state. As soon as the phase change has been made, the current sources formed by the circuit 116 are cut off (when the control signal CTR blocking the transistors 120, 122 and 126 changes to the low (GND) state) to stop heating currents from circulating in portions 106.

(41) The state change of the PCM in portions 106 from the amorphous state to the crystalline state may be detected in different ways, for example by measuring the voltage at the terminals of portions 106 (voltage close to 0 V when the portions 106 are conducting), or by measuring the temperature or impedance in the portions 106.

(42) In order to change the switch 100 from the conducting state of to the blocked state, that corresponds to a change of the PCM of the portions 106 of the switch 100 from the crystalline state to the amorphous state, the control signal CTR changes to the high state to make the transistors 120, 122 and 126 conducting. The heating current injected into the portion 106 by each current source formed by the control circuit 116 passes through the width b of the portion 106. In the same way as when changing from the blocked state to the conducting state described above, due to the short distance (less than 1 m) between the input point 112 and the facing output point 116 and the large distance from the other adjacent input points 112 (for example about 10 m), the impedance between the input point 112 and the output point 116 faced by the first current source formed by the first and second transistors 118, 120, may for example be between about 1 m and 10 m that is less than the impedance between this input point 112 and an adjacent input point that may for example be between about 20 m and 100 m. Consequently, the current injected in the input point 112 circulates from this input point 112 to the output point 116 located facing this input point 112. This current circulating in the PCM thus creates self-heating of the PCM, this heating increasing with increasing the value of this current.

(43) The electronic elements of the control circuit 116 are sized such that when the state of the switch 100 changes from the conducting state to the blocked state, for a power supply voltage V.sub.DD of the order of 2V, the heating current named I2 in this case passing through the portion 106 causes a voltage drop between the input point 112 and the output point 114 equal to less than about 1V for transistors like those described above. Furthermore, the value R2 of the variable resistance 130 during this state change may be chosen such that:
R2I2+V.sub.Dsat(126)+V.sub.GS(128)=V.sub.DD.

(44) This heating of the PCM above the PCM changeover temperature causes a change in its phase from the crystalline state to the amorphous state. As soon as the phase change has been made, the current sources formed by the control circuit 116 are cut off (when the control signal CTR blocking the transistors 120, 122 and 126 changes to the low state) to stop heating currents from circulating in portions 106.

(45) When the PCM changes to the amorphous state, the voltage at the terminals of each of the portions 106 will tend towards V.sub.DD, which stops operation of current sources formed by the control circuit 116. The state change of the PCM in portions 106 from the crystalline state to the amorphous state may be detected for example by measuring the voltage at the terminals of portions 106 (voltage close to V.sub.DD when the portions 106 are blocked), or by measuring the current output by the current sources.

(46) FIG. 4 diagrammatically shows an equivalent electrical circuit of one of the portions 106 coupled to the current sources formed by the control circuit 116 when the PCM is in the amorphous state. The example in FIG. 4 corresponds to a portion 106 through which two heating currents will pass from two input points 112 to two output points 114.

(47) The resistances 140, 142 and 144 correspond to different parts of the PCM of portion 106. Thus, the resistance 140 corresponds to the PCM located between the first end 108 and the first input and output points 112, 114 to which the control circuit 116 is connected. The resistance 144 corresponds to the PCM located between the second input and output points 112, 114 to which the control circuit 116 is connected, and the second end 110. The resistance 142 is formed by the PCM located between the first and second input and output points 112, 114. For example, the value of each of these resistances 140, 142 and 144 is of the order of 1 M.

(48) Each of the capacitances 146, 148, named Coff, corresponds to the sum of the parasite capacitances formed by the transistors in the control circuit 116 forming the current sources connected to the input and output points 112, 114, and particularly the drain-source capacitances and the gate-source capacitances of these transistors.

(49) Taking account of the substrate on which the portions 106 are made, the equivalent electronic circuit in FIG. 4 may be modified as illustrated in FIG. 5. Considering the substrate, three additional resistances 150, 152 and 154 and two additional capacitances 156, 158 have been added in this equivalent electrical diagram.

(50) The input impedance Zin of the circuit shown on FIG. 5 (impedance between the first end 108 and the ground) is such that:

(51) $\mathrm{Zin}=\underset{k=0}{\overset{n}{.Math.}}\left(\mathrm{Zk}+\frac{1}{J.Math.\mathrm{Coff}}+\frac{\mathrm{RsubV}}{\left(1+\mathrm{subV}\right)}\right)$

(52) Where Zk: resistances of the different parts of the PCM of portion 106 (corresponding to resistances reference 140, 142 and 144 on FIG. 5);

(53) RsubV: sum of the resistances formed by the different parts of the substrate (corresponding to resistances reference 152, 154 on FIG. 5);

(54) subV=RsubVvalue of the sum of capacitances 156 and 158.

(55) The value of this input impedance Zin is equal to several M and the electrical schematic in FIG. 5 can be considered like an open circuit when the PCM is in the amorphous state.

(56) The output impedance Zout of the circuit shown on FIG. 5 (impedance between the second end 110 and the ground) is such that:

(57) $\mathrm{Zout}=\underset{k=n}{\overset{0}{.Math.}}\left(\mathrm{Zk}+\frac{1}{J.Math.\mathrm{Coff}}+\frac{\mathrm{RsubV}}{\left(1+\mathrm{subV}\right)}\right)$

(58) The value of this output impedance is similar to the value of the input impedance Zin and therefore the circuit formed can be considered like an open circuit.

(59) The value of the impedance formed between the ends 108, 110, denoted Ziso, is such that:

(60) $\mathrm{Ziso}=\underset{k=0}{\overset{n}{.Math.}}\frac{\mathrm{Zk}\left(\frac{1}{J.Math.\mathrm{Coff}}+\mathrm{Rsubk}\right)}{\mathrm{Zk}+\mathrm{Rsubk}+\frac{1}{J.Math.\mathrm{Coff}}}$

(61) Rsubk: resistances formed by the different parts of the substrate between the different sets of input and output points (corresponding to the resistance reference 150 on FIG. 5).

(62) Thus, as the values of Rsubk and 1/j.Coff increase, the value of the impedance formed between the ends 108, 110 becomes closer to the sum of the resistances of the different parts of PCM in portion 106. Considering that the value of Rsubk is determined by the resistivity of the substrate and is much less than about 1 k, it is judicious to have a very low value Coff.

(63) FIG. 6 shows the different parasite capacitances connected to the portion 106 and connected to the transistors of the control circuit 116. The capacitances Cdsub.sub.118, Cdsub.sub.120, Cdsub.sub.122 and Cdsub.sub.124 correspond to the capacitances between the substrate and the drain respectively of transistors 118, 120, 122 and 124. The capacitances Cds.sub.118, Cds.sub.120, Cds.sub.122 and Cds.sub.124 correspond to the drainsource capacitances of each of the transistors 118, 120, 122 and 124.

(64) After reduction, the equation determining the value of 1/Coff as a function of parasite capacitances becomes:

(65) $\frac{1}{\mathrm{Coff}}=\frac{\left(\mathrm{Cds}120+1\right)\left(\mathrm{Cds}122+1\right)}{\begin{array}{c}\left(\mathrm{Cds}120+1\right)x\left(\left(\mathrm{Cds}122+1\right)\mathrm{Cdsub}122+1\right)+\\ \left(\mathrm{Cds}122+1\right)\left(\left(\mathrm{Cds}120+1\right)\mathrm{Cdsub}120+1\right)\end{array}}$

(66) Thus, the parasite capacitances of the transistors 118 and 124 do not have any impact on the value of 1/Coff. Transistors 120 and 122 are preferably made such that their parasite capacitances can increase the impedance between the portion 108 and the ground, and thus reduce losses in the switch 100.

(67) FIG. 7 diagrammatically shows an equivalent electrical circuit of the portion 106 coupled to the current sources formed by the control circuit 116 when the PCM is in the crystalline or conducting state.

(68) Unlike the circuits shown on FIGS. 4 and 5, the equivalent circuit in FIG. 7 does not include an electrical resistance since the PCM is in the crystalline state. Thus, this equivalent electrical circuit only includes capacitances 146, 148 with values Coff, and capacitances 160, 162 also with the value Coff.

(69) In order to make the switch 100 described above, the electronic elements of the control circuit 116 may be made in the front-end part of the integrated circuit of which the switch 100 forms part. PCM portions 106 may be made within one of the metallic interconnection levels of the back end part of the circuit. In this case the electronic components (particularly transistors) of the control circuit 116 are connected to the PCM portions 106 through conducting vias passing through the interlayer dielectric and possibly one or several other levels of metallic interconnections located between the semiconducting layer in which the electronic components are fabricated and the metallic interconnections level in which the PCM portion 106 is formed.

(70) As a variant to the particular embodiment described above, the control circuit 116 may comprise a current or voltage digitalanalogue converter or more generally any electronic means of imposing circulation of a heating current through the PCM portions 106, instead of the transistors 122-128 forming the current mirror.

(71) FIG. 12 shows such a variant in which the control circuit 116 comprises current digitalanalogue converters 164, instead of the current mirror comprising transistors 122, 124, 126 and 128 and the resistive element 130, each comprising an output 166 connected to one of the output points 114 through the MOS transistor 122. In this variant, the value of the heating current circulating through the PCM portion 106 depends on the value of an input signal, for example an 8-bit signal, applied to an input 168 of the digitalanalogue converter 164.

(72) FIG. 13 shows another variant in which the digitalanalogue converters 164 operate in voltage and impose values of the heating currents passing through the PCM portion 106 via values of output signals 166 from the converters 164 applied to the gates of the MOS transistors 124.

(73) The switch 100 may be coupled to a compensation device 200 as shown in FIG. 8, to reduce constraints on the size of transistors, for example when it is impossible to reduce transistor sizes to reduce their parasite capacitances.

(74) The device 200 comprises a first capacitive element 202, a first electrode of which is connected to the output 104 from switch 100, and a second capacitive element 204, a first electrode of which is connected to the input 102 of the switch 100. The device 200 also comprises a first inverter 206, the input of which is connected to the second electrode of the capacitor 204 and the output of which is connected to the second electrode of the capacitor 202, and a second inverter 208 in antiparallel to the inverter 206, the input of which is connected to the second electrode of the capacitor 202 and the output of which is connected to the second electrode of the capacitor 204.

(75) The values of the capacitances of the capacitive elements 202, 204 are preferably similar to the values of the parasite capacitance that is to be compensated, for example between 0.8 and 1.2 times the value of this parasite capacitance.

(76) The device 200 also comprises a resistive element 210 that acts as a polarisation point for the inverters 206, 208 when the switch 100 is in the OFF state.

(77) The addition of such a device 200 to compensate for the parasite capacitances of MOS transistors that will activate or deactivate the first and second current sources formed by the control device 116 reduces constraints on the design of the transistors, to prevent the presence of large parasite capacitances in these MOS transistors.

(78) FIG. 9 diagrammatically shows an example embodiment of a device 1000 for matching a transmission line, also named a variable stub, comprising several switches 100 similar to that described above. The device 1000 comprises a first conducting line 1002 or first portion of electrically conducting material, and a second conducting line 1004 or second electrically conducting portion, between which several switches 100 are inserted. The conducting lines 1002, 1004 and the portions 106 of switches 100 are advantageously made in the same level of metallic interconnections of a circuit. The second line 1004 partly surrounds the first line 1002. A first end 1006 of the first conducting line 1002 will be connected to the transmission line (not visible on FIG. 9) the impedance of which is to be matched by the device 1000. A second end 1008 of the first conducting line 1002 is connected to the second conducting line 1004.

(79) The ends 108, 110 of the PCM portion(s) 106 of each of the switches are coupled to the first and second conducting lines 1002, 1004 at different distances from the first end 1006 of the first conducting line 1002 such that the matching length of the device 1000 can be determined from the distance between the first end 1006 of the first conducting line 1002 and the closest switch 100 in the conducting state to the first end 1006.

(80) Unlike a matching device using transistors to determine the matching length of the device, the use of switches 100 comprising PCM portions can reduce the value of the electrical resistance of the matching device in the conducting state (Ron), thus increasing the quality factor that can be obtained. Such a matching device, for example used in a power amplifier, can improve the qualities of this amplifier.

(81) Switches 100 may also be used to make a phase shift system for a differential transmission line 2000, for example used in an array of antennas. An example embodiment of such an RF phase shift system 2000 is shown in FIGS. 10 and 11.

(82) This system 2000 comprises a differential transmission line formed from two electrically conducting portions 2002. The RF signals will pass through its portions 2002 parallel to their length L. First delay lines 2004 are arranged under one of the portions 2002, and second delay lines 2006 are arranged under the other of the portions 2002. The wave transmitted in the portions 2002 is more or less slowed, depending on the number of delay lines 2004 connected to the ground or to each other.

(83) The delay lines 2004 are connected to each other by switches 100 in contact with two delay lines 2004 each located under one of the portions 2002. Other switches 100 are inserted between each of the delay lines 2004 and the conducting portions 2008 connected to the ground 2006. As a variant, instead of being connected to the ground through conducting portions 2008 and switches 100 connected between the delay lines 2004, 2006 and these conducting portions 2008, the delay lines 2004, 2006 may be connected to the ground through a ground connection formed directly on the switches 100 inserted between the delay lines 2004, 2006.