Microelectronic structure with viscous damping controlled by controlling a thermo-piezoresistive effect

Abstract

Microelectronic structure comprising at least one movable mass that is mechanically connected to a first mechanical element by a first mechanically linking connector and to a second mechanical element (24) by electrically conductive second mechanically linking connector, and a device for electrically biasing the second mechanically linking connector, the second mechanically linking connector being such that they are the seat of a thermo-piezoresistive effect, the second linking connector and the movable mass being placed with respect to each other so that a movement of the movable mass applies a mechanical stress to the second linking connector, wherein the electrically biasing device are DC voltage biasing device and form, with at least the second mechanically linking connector, a thermo-piezoresistive feedback electric circuit.

Claims

1. A resonating microelectronic structure comprising at least one movable mass mechanically connected to at least one first mechanical element by at least a first mechanical connector and to at least one second mechanical element by at least a second electrically conductive mechanical connector, said movable mass being configured to be vibrated by an external stimulus, an electrical bias device for electrically biasing the second mechanical connector, said second mechanical connector being the object of a thermo-piezoresistive effect, the second mechanical connector and the movable mass being disposed with respect to each other such that a movement of the movable mass applies a mechanical stress to the second mechanical connector, wherein the electrical bias device is a DC voltage bias device and forms with at least the second mechanical connector a thermo-piezoresistive feedback electric circuit wherein the first mechanical element is an anchoring pad secured to a support and the movable mass is rotatably hinged with respect to the anchoring pad and wherein the second mechanical element is a second anchoring pad secured to the support.

2. The resonating microelectronic structure according to claim 1, wherein the thermo-piezoresistive feedback electric circuit comprises a current regulator electrically connected in series with the second mechanical connector.

3. The resonating microelectronic structure according to claim 2, wherein the current regulator comprises at least one electrical component configured to assume a determined electrical resistance value.

4. The resonating microelectronic structure according to claim 3, wherein the electrical component is configured to select a determined electrical resistance value from several electrical resistance values.

5. The resonating microelectronic structure according to claim 4, wherein the electrical component is a potentiometer.

6. The resonating microelectronic structure according to claim 3, wherein the second mechanical connector has a given electrical resistance and wherein the electrical resistance of the current regulator has a value lower than or equal to 5 times the electrical resistance value of the second mechanical connector when the movable mass is in the idle state.

7. The resonating microelectronic structure according to claim 3, wherein the second mechanical connector has a given electrical resistance and wherein the electrical resistance of the current regulator is equal to the electrical resistance of the second mechanical connector so as to compensate for the thermo-piezoresistive effect.

8. The resonating microelectronic structure according to claim 2, wherein the current regulator comprises at least one dipole connected in series with the second mechanical connector, said dipole having a negative resistance characteristic in the vicinity of the operating point of the resonating microelectronic structure.

9. The resonating microelectronic structure according to claim 8, wherein the dipole is chosen from a tunnel diode, a Gunn diode, a thyristor and a feedback circuit comprising at least one operational amplifier.

10. The resonating microelectronic structure according to claim 2, wherein the current regulator comprises at least one inductor connected in series with the second mechanical connector.

11. The resonating microelectronic structure according to claim 2, comprising an active controller for actively controlling the bias voltage enabling a constant voltage to be applied to the second mechanical connector.

12. The resonating microelectronic structure according to claim 11, wherein the active controller for actively controlling the bias voltage comprise a voltage feedback loop comprising measurement device for measuring the voltage at the terminals of the second mechanical connector, a comparator for comparing said measured voltage and a reference voltage, and a corrector sending a corrective signal to the electrical bias device.

13. The resonating microelectronic structure according to claim 12, wherein the measurement device is configured to measure the voltage at the terminals of the second mechanical connector by a 4-wire method.

14. The resonating microelectronic structure according to claim 1, wherein the second mechanical connector comprises a beam.

15. A resonating microelectronic structure comprising at least one movable mass mechanically connected to at least one first mechanical element by at least a first mechanical connector and to at least one second mechanical element by at least a second electrically conductive mechanical connector, said movable mass being configured to be vibrated by an external stimulus, an electrical bias device for electrically biasing the second mechanical connector, said second mechanical connector being the object of a thermo-piezoresistive effect, the second mechanical connector and the movable mass being disposed with respect to each other such that a movement of the movable mass applies a mechanical stress to the second mechanical connector, wherein the electrical bias device is a DC voltage bias device and forms with at least the second mechanical connector a thermo-piezoresistive feedback electric circuit, wherein the first mechanical element is an anchoring pad secured to a support and the second mechanical element is a second anchoring pad secured to the support, and the movable mass is rotatably hinged with respect to the anchoring pad from a pivot hinge comprising two beams extending between the anchoring pad and the movable mass.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The present invention will be better understood based on the description that follows and the appended drawings in which:

(2) FIG. 1 is a schematic representation of an exemplary embodiment of a microelectronic structure according to the invention,

(3) FIG. 2 is a representation of a circuit diagram of the feedback electric circuit of the structure of FIG. 1,

(4) FIG. 3 is a representation of a circuit diagram of a feedback electric circuit comprising stray resistances,

(5) FIG. 4 is a representation of a circuit diagram of another exemplary feedback electric circuit that can be implemented in the structure according to the invention,

(6) FIG. 5 is a representation of a circuit diagram of another feedback electric circuit that can be implemented in the structure according to the invention, comprising an active control for the bias voltage,

(7) FIG. 6 is a representation of another exemplary feedback electric circuit that can be implemented in the structure according to the invention, implementing a dipole having a negative electric resistance characteristic,

(8) FIG. 7 represents the characteristic of a dipole (voltage U as a function of the current I) having a negative resistance behaviour in the vicinity of an operating point I.sub.0, U.sub.0,

(9) FIGS. 8A to 8F are schematic representations of the different steps of an exemplary method for making a microelectronic structure according to the present invention,

(10) FIG. 9A is a representation of the absolute variation of the quality factor QF of a structure close to that of FIG. 1 as a function of the DC bias voltage Vc in mV,

(11) FIG. 9B is a representation of the variation in the delivered current I in μA as a function of time t in s,

(12) FIG. 9C is an enlarged view of FIG. 9B beyond 60 s,

(13) FIG. 10 is a schematic representation of another exemplary embodiment of a microelectronic structure according to the invention comprising two second mechanical connection means,

(14) FIG. 11 is an exemplary circuit diagram applicable to the structure of FIG. 10.

DETAILED DISCLOSURE OF PARTICULAR EMBODIMENTS

(15) In the present application, by “microelectronic structure”, it is meant a resonating structure comprising microelectromechanical elements and/or nanoelectromechanical elements.

(16) In the description that follows, the terms “viscous damping coefficient”, “damping coefficient”, “dissipation coefficient”, “damping factor” and “dissipation factor” are considered as synonymous.

(17) In FIG. 1, an example of a microelectronic structure S according to the invention can be seen. The structure extends along a neutral plane P defined by axes X and Y. The structure or resonator comprises a movable mass 2 able to be moved with respect to a support formed for example by a substrate. The movable mass 2 is connected to the support by the first connection means 6 such that they allow a movement of the movable mass 2 with respect to the support.

(18) In the example represented, the first connection means 6 comprise a pivot hinge having a pivot axis Z orthogonal to the plane P. The movable mass 2 is thus able to be rotatably moved about the axis Z in the plane P. In the example represented, the pivot hinge comprises two beams 10 extending between an anchoring pad 12 and the movable mass 2, the beams 10 having axes intersecting at a point O which is the intersection of the axis Z with the plane P. The beams 10 are flexurally deformed and allow rotation of the mass about the axis Z. It will be understood that the pivot connection could be made differently.

(19) The structure comprises second connection means 14 between the movable mass 2 and a second anchoring pad 24. The second connection means are electrically conductive.

(20) As will be described in the following of the description, the second connection means 14 are such that they exert a force on the movable mass by thermo-piezoresistive back action. The second connection means 14 are thereby oriented with respect to the first connection means 6, such that they are mechanically biased by the moving movable mass 2, and such that the back action force Far is applied at least in the direction along which the mass can be moved. In this example, the movable mass 2 is mainly moved along direction X. The second connection means 14 are thereby such that they exert a back action force Far on the mass along the direction X.

(21) The second connection means 14 are biased by the movable mass 2. Preferably, they are biased in tension and compression, while allowing flexure motions superimposing with the compression or tension.

(22) The second connection means 14 comprise at least one thermo-piezoresistive effect zone. In the example represented, the second mechanical connection means comprise a beam 18 anchored to the movable mass and to the second anchoring pad.

(23) The thermo-piezoresistive effect results from a coupling of electrical, thermal and mechanical phenomena within a mechanical structure, like a beam or a system comprised of beams, having all or part of the conductive, piezoresistive and thermoelastic properties. This effect is demonstrated by the appearance of a feedback force produced by the mechanical structure when it is mechanically biased, for example because of a deformation imposed by an external element, and when the mechanical structure is electrically biased. This force has components proportional to the deformation applied and to the rate of this deformation. The mechanism of this effect is described by a thermo-piezoresistive feedback loop.

(24) The beam 18 is thereby made at least partly of a piezoresistive material and at least partly of a material having thermoelasticity properties. The beam 18 can comprise two distinct materials, the one piezoresistive and the other thermoelastic or a single material being both piezoresistive and thermoelastic. For example, silicon and SiGe are both piezoresistive and thermoelastic. In the case where the beam comprises several materials, these can be aluminium having an expansion coefficient of 23 ppm/K, whereas that of silicon is 2 ppm/K, associated with a piezoresistive material such as silicon. It can be contemplated that the beam comprises alternate portions of these materials disposed in the length direction of the beam.

(25) In the present application, it is considered that an element has thermoelastic properties if it has an expansion coefficient at the working temperature higher than 10.sup.−7K.sup.−1.

(26) The second connection means 14 are such that they are able to exert mechanical strains onto the movable mass 2. For that, they have some stiffness and their connexion to the second anchoring pad is such that the force Far preferentially moves the movable mass.

(27) The structure also comprises DC voltage bias means 15 for these second connection means. For example, the bias means 15 comprise a DC voltage generator applying a DC voltage Vc to the thermo-piezoresistive feedback circuit.

(28) The bias means 15 form with the second mechanical connection means a feedback electric circuit.

(29) In the example represented, the movable mass 2, the first mechanical connection means and the first anchoring pad are also electrically conductive and the first anchoring pad is connected to the ground. Thus, they form with the bias means and the second mechanical connection means a closed electric circuit. But it could be contemplated to only bias the second mechanical connection means, for example by connecting to the ground the end of the beam 18 connected to the movable mass 2 with an electrical connection advantageously having a negligible mechanical stiffness along the space directions with respect to the other mechanical connections. The electrical connection would thereby have a limited impact on the mechanical behaviour of the structure.

(30) In the example represented, the thermo-piezoresistive feedback electric circuit C comprises an electrical resistance Rc connected in series with the beam 18 and disposed between a terminal of the bias means 15 and the second anchoring pad 24. The resistance can be directly made on the substrate of the microelectronic structure or be made outside the structure, for example on an application specific integrated circuit ASIC dissociated from the structure.

(31) In FIG. 2, the circuit diagram of the feedback circuit C can be seen. The beam 18 is represented by its variable electrical resistance Rpo+δRp. Rpo is the electrical resistance value of the beam in the idle state (not subjected to a mechanical stress by the movable mass) which is a constant, and δRp is the electrical resistance variation of the beam because of a mechanical stress applied by the movable mass to the beam at least one part of which is of a piezoresistive material. I is the current in the circuit.

(32) The operation of the feedback circuit is the following one.

(33) The bias voltage V.sub.c biases the circuit. The current I thus assumes the following value:

(34) $I=I_0+\delta I\approx \frac{V_c}{R_c+R_{p0}}\left(1-\frac{\delta R_p}{R_c+R_{p0}}\right).$

(35) By considering δRp<<Rc+Rpo, for example δRp/Rpo<3%.

(36) That is

(37) $\delta I=-\frac{\delta R_p}{{\left(R_c+R_{p0}\right)}^2}{V}_c.$

(38) The power delivered to the beam 18 can then be estimated with:
P=(R.sub.p0+δR.sub.p)(I.sub.0+δI).sup.2.

(39) That is in the δR.sub.p first order

(40) $P\approx {R_{p0}\left(\frac{V_c}{R_c+R_{p0}}\right)}^2\left(1+\delta R_p\frac{R_c-R_{p0}}{R_{p0}\left(R_c+R_{p0}\right)}\right).$

(41) The power delivered to the beam TPBA takes the form:

(42) $P=P_0+\mathrm{\delta P}$$\mathrm{By}\mathrm{setting}{I}_0=\frac{V_c}{R_c+R_{p0}}.P_0={R_{p0}\left(\frac{V_c}{R_c+R_{p0}}\right)}^2=R_{p0}{I}_0^2$$\mathrm{and}$$\delta P=\eta \frac{\delta R_p}{R_{p0}}$$\mathrm{with}\eta =\frac{R_c-R_{p0}}{R_c+R_{p0}}.$

(43) the force generated by the thermo-piezoresistive effect or the feedback force is written as

(44) 0$F\left(\omega \right)=\mathrm{\eta E}\alpha R_{\mathrm{th}}{I}_0^2{\pi }_g{R}_{p0}\frac{x-j{\tau }_{\mathrm{th}}\stackrel{.}{x}}{1+{\left({\mathrm{\omega \tau }}_{\mathrm{th}}\right)}^2}.$

(45) Thus, by choosing the Rc value, the coefficient η and thus the feedback force can be modified. For example, it is possible to vary the coefficient η from −1 to +1 when Rc>0.

(46) The resistance thus enables the current to be regulated in the piezoresistive feedback electric circuit C and thus the back action effect to be set.

(47) By acting on the value of the resistance R.sub.c with respect to the resistance R.sub.p0, it is possible to vary the parameter η from −1 to +1.

(48) It is thus possible to control the sign and presence of the thermo-piezoresistive action effect, by adjusting the R.sub.c value with respect to R.sub.p on a wide operating frequency range of the microelectronic structure.

(49) In the circuit diagram, the resistance Rc has a fixed value.

(50) Preferably, Rc is equal to at most 5 times the Rp0 value, in order to limit energy consumption.

(51) The resistance Rc is a resistance whose value is determined, either it is a fixed value that cannot be modified, or it is a modifiable value as will be described in the following. The resistance Rc is a resistance deliberately added in the thermo-piezoresistive feedback electric circuit and whose effect on the back action force is determined, the resistance Rc is distinct from the stray resistances which exist for example at the access tracks, but whose value is not determined.

(52) Rc values having a particular interest are given below.

(53) For Rc=Rpo, the coefficient q has the value η=0. In this case, the effect is compensated for. The resonator gets back to its natural dissipation coefficient, there is no longer a thermo-piezoresistive back action effect.

(54) For Rc>>Rpo, the coefficient n has the value η=1. In this case, the feedback has the same sign as the usually obtained effect when the bias is a current bias.

(55) For Rc=0 or Rc<<Rp0, the coefficient η has the value η=−1. In this case, the feedback has a sign opposite to the usually obtained effect when the bias is a current bias. In this particular case, the structure is made such that the stray electrical resistances in series with the beam are reduced. In FIG. 3, these stray resistances are designated by Rpar1 and Rpar2. These stray resistances can result from the access tracks to the beam 18, indeed, their resistivity is not always negligible. They can reach a few tens of percents of the resistance value of the beam, or even exceed this value. The resistances could be structurally reduced for example by using strong-conductivity materials up to the beam anchoring, by dimensioning the access tracks . . . .

(56) By virtue of the invention, the action on the feedback force can occur on a wide operating frequency range, from null frequencies up to high frequencies.

(57) The upper limit of the frequency is set by stray capacitances Cpar. A high frequency

(58) ${\omega }_c=\frac{1}{R_c{C}_{\mathrm{par}}}$
can limit the operating range of the system.

(59) Advantageously, the resistance R.sub.c is integrated to the substrate of the microelectronic structure in the proximity of the second connection means 18 in order to reduce the stray capacitance to a value lower than 100 fF. The upper limit of the frequencies can thus be brought up to a value in the order of about one hundred Megahertz.

(60) In FIG. 4, another exemplary embodiment of a very advantageous feedback circuit can be seen, wherein the resistance Rc comprises a potentiometer enabling the feedback sign to be controlled in situ and very precisely. It is also possible in this way to activate or deactivate the back action effect in the microelectronic structure. When the resistance Rc is increased by handling the potentiometer, the average current delivered tends to decrease, which results in reducing the back action effect. Advantageously, the bias voltage is adjusted such that the average current delivered is constant. For example, the value of the bias voltage can be linked to the Rc value. However, an independent control can be maintained on the voltage source.

(61) Alternatively, the resistance could be replaced by a device capable of assuming several resistance discrete values, for example three values, for example null, equal to Rpo and a value much higher than Rpo.

(62) The potentiometer can be integrated on the chip of the microelectronic structure or in an ASIC from a Junction Field Effect Transistor (JFET) technology for example used for the following components: Vishay AN105®, Fairchild®, Philips semiconductor BF245C®.

(63) In FIG. 5, another exemplary feedback circuit comprising means for actively controlling the bias voltage can be seen.

(64) As has already been explained above, the circuit can comprise stray resistances the effect of which is preferably reduced. For that, the circuit comprises means for actively controlling the bias voltage implementing a voltage feedback loop BR.

(65) The minimum resistance in series with the resistance Rp of the beam 18 is reduced to R.sub.par1+R.sub.par2. These stray resistances limit the value of the coefficient η to:

(66) $\eta \left(R_c\right)=\frac{R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}-R_{p0}}{R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_{p0}}\ge \frac{R_{\mathrm{par}1}+R_{\mathrm{par}2}-R_{p0}}{R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_{p0}}>-1.$

(67) The means for actively controlling the bias voltage comprise means 21 for probing the voltage Vs at the terminals of the beam 18, for example by the 4-wire method, well known to those skilled in the art, which enables the stray resistances effect to be dispensed with.

(68) For that, access tracks dissociated from the bias circuit are made up to the ends of the beam 18. The use of a voltage probe having a high impedance with respect to the impedance of these tracks and that of the beam 18 joined together enables the voltage across the beam 18 to be precisely measured.

(69) The control means also comprise a comparator 22 for comparing this voltage measured across the beam to a reference value V.sub.consigne, and a corrector 26 which drives the voltage generator V.sub.c based on the signal emitted by the comparator in order to remove the error signal ∈=V.sub.consigne−V.sub.s.

(70) Thanks to these active control means, the voltage across the beam 18 is maintained constant and equal to V.sub.consigne. The back action effect obtained is substantially that obtained for η=−1 because the stray resistances and Rc have no longer influence.

(71) In FIG. 10, another exemplary embodiment of a microelectronic structure according to the invention can be seen, which differs from the structure of FIG. 1, in that it comprises a second beam 18′ suspended between the mass 2 and a pad 24′. The beam 18′ and the pad 24′ are symmetrical with the pad 24 and the beam 18 with respect to a plane passing through the axis of rotation of the mass. This structure can be used in an inertial device, for example an accelerometer, the beams forming differential connected gauges.

(72) In FIG. 11, the circuit diagram of the structure of FIG. 10 can be seen. Both beams are connected in parallel to the bias source 15 and both beams are electrically connected. This connection advantageously enables the bias voltage to be maintained constant across the beams by putting the access resistances in common.

(73) R.sub.p1 is the resistance of the beam 18 without mechanical load and ΔR.sub.1 is the variation in the resistance Rp.sub.1 when it is subjected to a stress.

(74) R.sub.p2 is the resistance of the beam 18′ without a mechanical load and ΔR.sub.2 is the variation in the resistance Rp.sub.2 when it is subjected to a stress.

(75) The resistances Rp.sub.1 and Rp.sub.2 have opposite resistance variations, indeed when one of beams is compressed, the other is stretched.

(76) R.sub.a1 and R.sub.a2 are the access resistances. In this configuration, the access resistances R.sub.a1 and R.sub.a2 are put in common for both piezoresistive elements 18 and 18′. The equivalent resistance of the dipole formed by both parallel elements 18 and 18′ does not depend on the first order of the resistance variation ΔR which occurs in an opposite way in each of both these elements.

(77) Indeed, by considering R.sub.p1=R.sub.p2

(78) $R_{\mathrm{eq}}=\frac{R_{p1}}{2}-\frac{\Delta R_1^2}{2R_{p1}}.$

(79) Under these conditions, the voltage across both piezoresistive elements remains constant in the ΔR.sub.1 first order.

(80) $V_p=\frac{R_{p1}{V}_0}{R_{p1}+R_{a1}+R_{a2}}.$

(81) Advantageously: resistances R.sub.par1 and R.sub.par2. substantially identical within 20%, opposite variations of the resistances ΔR.sub.1 and ΔR.sub.2 with modules identical within 20%, are chosen. Thereby

(82) $\begin{array}{c}{V}_p=\frac{R_{\mathrm{eq}0}{V}_0}{R_{a1}+R_{a2}+R_{\mathrm{eq}0}}\left(1-ϵ\frac{R_{p1}-R_{p2}}{R_{p1}+R_{p2}}\left(1-\frac{R_{\mathrm{eq}0}}{R_{a1}+R_{a2}+R_{\mathrm{eq}0}}\right)\right)\mathrm{with}\\ \square \\ R_{\mathrm{eq}0}=\frac{R_{p1}{R}_{p2}}{R_{p1}+R_{p2}}\mathrm{and}ϵ=\frac{\Delta R_1}{R_{p1}}=-\frac{\Delta R_2}{R_{p2}}\end{array}$
which are supposed opposite in the present case.

(83) The factor

(84) $\frac{R_{p1}-R_{p2}}{R_{p1}+R_{p2}}$
characterises the unbalanced searched for.

(85) Alternatively, only one of both access resistances can be put in common.

(86) In FIG. 6, another exemplary feedback circuit according to the invention implementing a dipole 28 having a negative resistance characteristic Rn can be seen. The dipole 28 is electrically connected in series with the beam 18.

(87) The use of such a dipole advantageously enables the presence of stray resistances to be dispensed with in the feedback circuit.

(88) The dipole 28 has a negative resistance characteristic, at least in the vicinity of an operating point U.sub.0, I.sub.0 (FIG. 7). This behaviour type can be achieved with dipoles of the tunnel diode, Gunn diode, thyristor type, by a feedback circuit based on an operational amplifier or any component known to those skilled in the art having a negative resistance characteristic.

(89) For a rated current I.sub.0 desired in the beam 18 at the idle state, the generator voltage to be applied can be estimated with V.sub.c=U.sub.0+(R.sub.c+R.sub.par1+R.sub.par2+R.sub.p0)I.sub.0. In the vicinity of this point, the dipole has the following behaviour:
U=U.sub.0+R.sub.n(I−I.sub.0).

(90) Thus, in the presence of a variation in the resistance δR.sub.p for the beam, the new balance is obtained for I=I.sub.0+δI
V.sub.c=U.sub.0+(R.sub.c+R.sub.par1+R.sub.par2+R.sub.p0+R.sub.n+δR.sub.p)(I.sub.0+δI)−R.sub.nI.sub.0.

(91) That is

(92) $0\approx \delta R_p{I}_o+\left(R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_{p0}+R_n\right)\delta I$$\frac{\delta I}{I_o}\approx \frac{\delta R_p}{\left(R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_{p0}+R_n\right)}.$

(93) The power modulation across the beam 18 has the following expression:

(94) $P_o+\delta P=\left(R_{p0}+\delta R_p\right){\left(I_o+\delta I\right)}^2$$\frac{\delta P}{P_o}=\frac{\delta R_p}{R_{p0}}-2\frac{\delta R_p}{\left(R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_{p0}+R_n\right)}.$

(95) Finally:

(96) $\frac{\delta P}{P_o}={\eta }_n\frac{\delta R_p}{R_{p0}}$

(97) With:

(98) 0${\eta }_n=\frac{R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_n-R_{p0}}{R_c+R_{\mathrm{par}1}+R_{\mathrm{par}2}+R_n+R_{p0}}.$

(99) The force exerted on the system then takes the form:

(100) $F\left(\omega \right)={\eta }_{n}E\alpha R_{\mathrm{th}}{I}_0^2{\pi }_g{R}_{p0}\frac{x-j{\tau }_{\mathrm{th}}\stackrel{.}{x}}{1+{\left({\mathrm{\omega \tau }}_{\mathrm{th}}\right)}^2}.$

(101) The factor η.sub.n can actually reach −1 when R.sub.c=0 and R.sub.n compensates for R.sub.par1+R.sub.par2. This exemplary embodiment enables the problem of the presence of stray resistances to be overcome. Further, this system enables the reverse back action effect to be enhanced. Indeed, for a negative resistance R.sub.n with an absolute value which approaches the sum of the resistances making up the circuit (R.sub.c+R.sub.par1+R.sub.par2±R.sub.p0), the coefficient η.sub.n diverge to infinity by a negative value: that occurs for R.sub.c+R.sub.par1+R.sub.par2+R.sub.n+R.sub.p0˜0. The self-oscillation appears as soon as the damping is compensated for. In the presence of a negative damping coefficient, in addition to the self-oscillation, the mechanical response time decreases with an increasing η.sub.n intensity and the system reaches quicker its self-oscillation regime after a given disturbance.

(102) Preferably, R.sub.c is chosen close to 0 in order to reduce the Joule effect dissipation.

(103) By virtue of the invention, it is possible to compensate for the air dissipation. Indeed, the invention enables the dissipation coefficient to be amplified by a factor higher than 1 in absolute value, the values of the air dissipation coefficient and the back effect dissipation coefficient thereby approach each other.

(104) According to one alternative, the resistance R.sub.c can be replaced or completed by an inductor with the value L. The inductor indeed produces a low impedance at a low frequency and an impedance which diverges at a high frequency (cutoff frequency

(105) $\frac{R_p}{L}).$
That enables the current to be stabilised above the cutoff frequency. By associating an inductor to the resistance Rc, the dissipation effect can be differently driven on two vibration modes located below and above the cutoff frequency. Below the cut off frequency, the thermo-piezoresistive effect is driven by the resistance R.sub.c and above the cutoff frequency, the thermo-piezoresistive effect is driven by the inductor.

(106) As has been indicated above, the back action effect can be theoretically compensated for when Rc=Rp0. But in practice, this compensation can occur for an Rc value different from R.sub.p0, in particular if a modulation effect of the thermal conductivity at the mechanical stress is also present. Indeed, the temperature modulation in the TPBA beam depends not only on: the power modulation

(107) $\delta P=\eta 1_0^2{\pi }_{g}R\frac{x}{L}$ but also the resulting thermal resistance modulation

(108) $\delta R_{\mathrm{th}}=\frac{{\pi }_{\sigma }{R}_{\mathrm{th}0}x}{L}$
or π.sub.σ is a coefficient which characterises the sensitivity of the thermal conductivity to the mechanical deformation

(109) $\frac{x}{L}.$ L being the length of the beam 18. The thermoelastic force can be written as:

(110) $F\left(\omega \right)=E{\alpha R}_{\mathrm{th}0}{R}_{p0}{I}_0^2\frac{\eta {\pi }_g+{\pi }_{\sigma }}{L}\frac{x-{\tau }_{\mathrm{th}}\stackrel{.}{x}}{1+{\left(\omega {\tau }_{\mathrm{th}}\right)}^2}.$

(111) The sign and intensity of the feedback are in particular controlled by the factor:

(112) $\eta +\frac{{\pi }_{\sigma }}{{\pi }_g}$

(113) The η value can be adjusted so as to compensate for the term

(114) $\frac{{\pi }_{\sigma }}{{\pi }_g}.$
In practice, the sensitivity of the thermal conduction coefficient π.sub.σ is low with respect to the piezoresistivity coefficient π.sub.g which is involved in the dissipated power modulation: the R.sub.c value which enables the thermo-piezoresistive effect to be deactivated will be close to R.sub.p within a few tens of percents.

(115) In the example of FIG. 1, the mass is rotatably movable in the plane of the structure. Any structure the mass of which is translationally and/or rotationally movable in-plane and/or off-plane is within the scope of the present invention.

(116) The present invention can be applicable to any structure comprising a thermo-piezoresistive effect beam. For example, it can be applied to a structure comprising two mechanically connected movable masses, a resistance Rc could be placed in series with the second mechanical connection means between both masses, two connection means connecting the resistance Rc to the other mass, these other connection means having for example no thermo-piezoresistive effect, and/or having a negligible stiffness. It can be contemplated to dispose

(117) As regards the second mechanical connection means, they are preferentially formed by a rectilinear beam, but it can be contemplated that they comprise several beam portions the axes of which form an angle with each other. The angle is chosen low such that the back action in the second connection means is preferentially used to apply a strain on the anchorings and not to deform the beam in itself. Further alternatively, it can be contemplated that the second mechanical connection means are in the shape of an arc of circle.

(118) The second mechanical connection means are chosen such that their mechanical stiffness in the motion direction of the mass is comparable to or higher than the stiffness of the movable masse coupled with the first mechanical connection elements along the same, direction.

(119) FIGS. 9A to 9C represent experimental results obtained on a structure close to the structure of FIG. 1. A voltage bias of a thermo-piezoresistive effect beam, of 5e19 cm.sup.−3 p-doped silicon, is made in the presence of a resistance Rc much lower than the resistance Rp.

(120) A reduction effect of the friction coefficient is observed, which is reflected in FIG. 9A by an increase in the quality factor QF. By virtue of the invention, a reduction in the dissipation for beams of p-doped silicon is obtained, whereas in the state of the art, this reduction was only contemplatable for n-doped beams at least at a low frequency (<100 kHz). It is noticed that for a bias voltage from 900 mV, the thermo-piezoresistive effect fully compensates for the intrinsic losses of the resonator.

(121) FIG. 9B shows the time sequence of the MEMS structure a 1V bias of which is made at about t=20 s. An exponential growth of the amplitude of the current delivered is observed. The amplitude is stabilised then because of the system non-linearities which re-balance the energy exchanges of the resonator with the outside. FIG. 9C shows an enlarged view of the stabilisation zone.

(122) An example of a method for making the structure according to the invention will now be described, of which different steps can be schematically seen in FIGS. 8A to 8F.

(123) A SOI (Silicon on insulator) structure is for example used, comprising a substrate 402, a buried oxide (BOX) layer 404, and a silicon layer 406. The silicon layer 406 has for example a thickness in the order of 200 nm. Generally, the layer 406 can be polycrystalline or single crystal Si, SiGe or Ge.

(124) A structuring of the silicon layer 406 is then made, for example by photolithography and etching with a barrier on the oxide layer 406, which enables the connection means to be defined.

(125) The element thus obtained is represented in FIG. 8A.

(126) During a following step, a deposition of an oxide layer 410 is made on the silicon layer 406 to fill the previously etched zones 408, then an etching of the oxide layer 410 is made to only leave an oxide portion deposited on the silicon layer 406 and connecting the oxide filling the trenches. The etching can be made by dry etching with a barrier on Si or by wet etching, for example using a sulphuric acid based solution.

(127) The element thus obtained is represented in FIG. 8B.

(128) During a following step, a deposition of a silicon layer 412 is formed. The layer 412 is for example obtained, by epitaxial growth, and has a typical thickness from 1 to 50 μm, for example 10 μm. More generally, the layer 412 can consist of polycrystalline or single crystal Si, SiGe, Ge, or a metal material; the deposition can be made by epitaxy or by physical/chemical vapor deposition (PVP/CVD) type methods.

(129) The element thus obtained is represented in FIG. 8C.

(130) During a following step, electrical contacts 414 are made. For that, a deposition of a metal layer (AlSi or Au for example) is made, the zones to be removed and preserved are distinguished by photolithography. Then, the metal layer is etched by dry etching with a barrier on Si or by wet etching selective to Si to only preserve the contacts 414.

(131) The element thus obtained is represented in FIG. 8D.

(132) During a following step, the silicon layers 406 and 412 are structured to define the movable mass and the first connection means and the second connection means, for example by photolithography and deep etching with a barrier on the oxide layers 404 and 410.

(133) The element thus obtained is represented in FIG. 8E.

(134) During a following step, the movable mass and the first and second connection means are released, for example by wet etching of the oxide 404, for example by means of a liquid and/or vapor hydrofluoric acid (HF). That is an etching over time. The hydrofluoric acid is left in contact with the oxide layer for the time necessary to release the movable mass, the first and second connection means while leaving the oxide layer between the substrate and the fixed parts.

(135) The element thus obtained is represented in FIG. 8F.

(136) The MEMS/NEMS micro and nanostructures according to the invention that offer means for controlling the viscous damping can for example be implemented in MEMS/NEMS micro-sensors and micro-actuators.