Integrated sound transmitter and receiver, and corresponding method for making same

Information

  • Patent Application
  • 20030052355
  • Publication Number
    20030052355
  • Date Filed
    September 16, 2002
    21 years ago
  • Date Published
    March 20, 2003
    21 years ago
Abstract
The transmitter or receiver comprises several transducers made opposite an aperture in a package.
Description


[0001] The invention relates to integrated circuits, and more particularly to integrated semiconductor acoustic transducers, particularly those operating in the ultrasonic range, and to their fabrication.


[0002] At the present time, acoustic transmitters/receivers operating in the ultrasonic range are not “integrable” on a semiconductor chip. This is because these transmitter-receivers generally rely on the piezoelectric effect which deforms a thin layer of a suitable material due to the effect of an electric field. Furthermore, this suitable material, generally a quartz plate, is in practice not integrable in silicon technology.


[0003] The invention aims to remedy this drawback and provides a semiconductor acoustic transducer integrable in silicon technology.


[0004] This integrated semiconductor acoustic transducer according to the invention comprises, in general, a deformable semiconductor membrane capable of being coursed by an electric current and separated from one zone of a semiconductor substrate by a cavity allowing the membrane to deform.


[0005] In a silicon technology, the substrate is made of silicon. Moreover, the membrane is preferably formed from doped silicon so as to allow easier flow of the current therethrough.


[0006] Moreover, according to one embodiment, the height of the cavity is of the order of ten nanometres, while the length of the membrane is of the order of a hundred micrometres. These dimensions are particularly suitable for an ultrasound application of the transducer, thus allowing a resonant frequency of about 30 MHz in the case of the membrane and about 1 MHz in the case of the cavity, which therefore gives the transducer an operating range of between 1 MHz and 30 MHz.


[0007] According to the invention, the transducer can be used both as an element of an acoustic receiver and as an element of an acoustic transmitter.


[0008] When it is used as an acoustic sensor (an element of an acoustic receiver), the transducer according to the invention advantageously comprises a capacitor having a first plate formed by the semiconductor membrane intended to be coursed by an electric current and able to deform due to the effect of an acoustic pressure (the pressure variations resulting from sound propagating through the open air and striking the membrane). The capacitor also includes a second plate formed by a doped zone of the semiconductor substrate and placed opposite the membrane. Moreover, the cavity separating the two plates then contains, for example, a layer of dielectric gas, for example air.


[0009] The invention also provides an acoustic receiver comprising a semiconductor substrate containing at least one transducer as defined above, together with current-generating means, capable of generating the current in the membrane of the transducer, and detection means connected to the capacitor and capable of detecting the variations in the capacitance of the capacitor which are caused by the deformations of the membrane.


[0010] When the transducer according to the invention is used as an element of an acoustic transmitter, the semiconductor membrane intended to be coursed by a modulated electric current is advantageously deformable due to the effect of the Lorenz force generated by a magnetic field lying in the plane of the membrane and perpendicular to the lines of current.


[0011] The invention also provides an acoustic transmitter comprising a semiconductor substrate containing at least one transducer as defined above, together with current-generating means capable of generating a modulated current in the membrane, and means for generating a magnetic field which are capable of generating the said magnetic field lying in the plane of the membrane and perpendicular to the lines of current.


[0012] These means for generating the magnetic field may be, for example, a U-shaped magnet.


[0013] The invention also provides a process for fabricating an integrated semiconductor acoustic transducer comprising a deformable semiconductor membrane separated from one zone of a semiconductor substrate by a cavity allowing the membrane to deform, the process comprising the following steps:


[0014] making, in the substrate, an isolating region defining a semiconductor region called the active region;


[0015] depositing, by selective epitaxy on the surface of the said active region, a first layer of a first material, for example germanium or a silicon-germanium alloy;


[0016] depositing, by non-selective epitaxy on the first layer and on the isolating region, a second layer of a second semiconductor material (for example silicon), the first material being selectively removable with respect to the second material;


[0017] locally etching the second layer, the first layer and part of the active region so as to form two lateral trenches which leave a central stack comprising a part of the second layer, a part of the first layer and a part of the active region and which reveal the said part of the first layer along two opposed lateral sides of the stack;


[0018] selectively removing, from the sides, the part of the said first layer so as to form the said cavity which is bounded by the subjacent part of the substrate (forming the said zone of the substrate) and the remaining part of the said second layer (which forms the said membrane); and


[0019] forming lateral spacers in order to close off the cavity beneath the membrane.






[0020] Further advantages and characteristics of the invention will appear upon examination of the detailed description of embodiments and methods of implementation, which are entirely non-limiting, and of the appended drawings in which:


[0021]
FIG. 1 illustrates highly schematically a semiconductor chip comprising a matrix of acoustic transducers according to the invention;


[0022]
FIG. 2 illustrates schematically a package containing a chip as illustrated in FIG. 1;


[0023]
FIG. 3 illustrates an arrangement according to the invention, comprising a U-shaped magnet cooperating with a chip furnished with a matrix of transducers according to the invention;


[0024]
FIG. 4 illustrates schematically the arrangement in FIG. 3 incorporated into a package;


[0025]
FIGS. 5 and 6 illustrate highly schematically one embodiment of an acoustic transducer according to the invention, more particularly one intended to be incorporated into an acoustic transmitter according to the invention;


[0026]
FIGS. 7 and 8 illustrate highly schematically an acoustic transducer according to the invention, more particularly intended to be incorporated into an acoustic receiver according to the invention;


[0027]
FIGS. 9

a
to 9d illustrate schematically the main steps of a fabrication process according to the invention, used for obtaining an acoustic transducer;


[0028]
FIG. 10 illustrates highly schematically an electrical circuit diagram for an acoustic transmitter according to the invention; and


[0029]
FIG. 11 illustrates highly schematically an electrical circuit diagram for an acoustic receiver according to the invention.






[0030] In FIG. 1, the reference P denotes a semiconductor chip or wafer comprising a substrate, for example made of silicon. This chip has several regions called “active regions” ZA, which are mutually isolated by isolating regions STI, for example isolating trenches conventionally produced in a manner known per se by a so-called “shallow trench isolation” process. As will be explained in greater detail below, it is within each of the active regions ZA that an acoustic transducer will be produced, which acoustic transducer, according to the invention, will be intended either to pick up ultrasound or to transmit ultrasound.


[0031] The chip P comprises, moreover, associated with a matrix of transducers MC, a control logic unit LC comprising data processing circuits suitable for the application. This control logic unit may comprise, as will be seen in greater detail below, an oscillator, a modulator and a processor for processing the signal.


[0032] Moreover, the chip P includes, conventionally, input/output pads PES along its edges. These input/output pads are connected via metal tracks and interconnection holes (known by the name “vias”) to the various constituent elements of the chip.


[0033] This chip P is intended to be incorporated into a package BO, like the one illustrated in FIG. 2, comprising, conventionally, a resin encapsulation intended to encapsulate the integrated circuit produced on the chip P, and pins BR conventionally connected in a known manner to the input/output pads PES on the chip.


[0034] Moreover, the package BO also includes an aperture OUV which is made opposite the matrix MC of transducers. Thus, when the package BO contains an acoustic transmitter, the sound produced by the matrix of transducers MC can propagate through the aperture OUV to the open air. Likewise, when the package BO contains an acoustic receiver, the aperture OUV allows the sound-propagating open air to strike the membranes of the transducers of the matrix MC.


[0035] As will be seen in greater detail below, a transducer intended to be incorporated into an acoustic transmitter uses the principle of the Lorenz force. More specifically, a transducer according to the invention comprises a deformable membrane. A current I is intended to flow in the membrane. Moreover, the latter is subjected to the effect of a magnetic induction lying in the plane of the membrane, perpendicular to the lines of current. Consequently, a force normal to the surface of the membrane (the Lorenz force) is generated and this deforms the membrane.


[0036] This deformation will generate a sound if the current I flowing in the membrane is modulated.


[0037] In FIG. 3, the reference AI denotes a U-shaped magnet which is one particularly simple means of generating a magnetic field. The magnetic field extends from one arm of the U, for example the arm B1, of the magnet to the other arm of the U, for example the arm B2, in the plane of the chip P. Moreover, as will be seen in greater detail below, the current will flow in the membranes of the transducers perpendicular to the magnetic field lines.


[0038] As may be seen in FIG. 4, the U-shaped magnet AI is also incorporated into the package, the horizontal arm of the U-shaped magnet being located beneath the chip P. The latter is, moreover, placed so that the matrix MC of transducers lies opposite the aperture OUV. The whole assembly is solidly held together by the resin encapsulation.


[0039]
FIGS. 5 and 6 illustrate highly schematically a top view and a sectional view on the line VI-VI of a transducer incorporated into an acoustic transmitter.


[0040] The active region ZA extends, within the silicon substrate SB, between a peripheral isolating region STI. Moreover, a membrane MB, for example also made of silicon, thin enough to be deformed, lies above the active region ZA and rests at two of its ends EX1 and EX2 on the peripheral isolating region STI.


[0041] The lower surface of the membrane MB, that is to say the surface which is located opposite the upper surface of the active region ZA, is separated from this active region by a cavity CV. This cavity CV is closed off to the sides by isolating regions or spacers ESP which, as will be seen in greater detail below, are made in lateral trenches produced in the active region ZA.


[0042] In order for a current I to flow more easily in the membrane MB, the latter is advantageously doped. Consequently, especially if the doping of the membrane MB is done by implantation, the substrate zone ZSB located opposite the membrane is also doped, although the doping of this zone ZSB is not absolutely essential for the operation of the acoustic transducer.


[0043] The transducer also includes several electrical contact pads or vias V1 which come into contact with one of the two ends of the membrane resting on the isolating peripheral region, for example the end EX1. Likewise, several other electrical contact pads V2 are provided which come into contact with the other end of the membrane EX2 resting on the isolating peripheral region STI. Although it would be sufficient to furnish each of the two ends EX1 and EX2 of the membrane with a single via, it is preferable to provide several vias on each of the two ends. This makes it possible, in particular, to reduce the electrical resistance of the set of vias and to obtain better homogeneity of the lines of current in the membrane.


[0044] Each of the vias V1, V2 also comes into contact with a metal track P1, P2 of the first metallization level of the integrated circuit. These metal tracks and vias allow the transducer to be electrically connected. Finally, provision is conventionally made for there to be a first level of insulating material PHD, which insulates the upper surface of the substrate SB from the first metallization level. Conventionally this insulating material is, for example, a doped silicon oxide obtained from tetraethyl orthosilicate (TEOS). The metal tracks of each metallization level are also conventionally coated with another insulating material IMD, for example TEOS-derived oxide, but undoped.


[0045] In operation, a modulated current I flows in the membrane. The magnetic field B generated by the magnet, perpendicular to the lines of current I, creates a Lorenz force perpendicular to the membrane which deforms the latter to a greater or lesser extent depending on the amplitude of the current, thus transmitting sound (or ultrasound).


[0046]
FIGS. 7 and 8 illustrate highly schematically an acoustic transducer incorporated into an acoustic receiver. FIG. 8 is a cross section on the line VIII-VIII of FIG. 7.


[0047] Compared with the transducer illustrated in FIGS. 5 and 6, the membrane MB of the transducer in FIGS. 7 and 8 forms a first plate of a capacitor. The doped zone ZSB1 of the semiconductor substrate, faced opposite the membrane, then forms a second plate of the capacitor. The cavity CV separating the two plates contains a layer of dielectric gas, for example air or nitrogen.


[0048] The substrate SB, for example a p-type substrate, includes an n-doped well CS, within a first part of which the doped substrate zone ZSB1 is produced. Moreover, this well CS has another part ZSB2, also doped, which is isolated from the membrane and in contact with which the via V2 is produced.


[0049] The current flowing through the capacitor thus enters by the via V1 for example, flows through the membrane and exits, via the zone ZSB1, the well CS and the zone ZSB2, by the via V2.


[0050] The rest of the structure is identical to that described with reference to FIGS. 5 and 6, the reference LM denoting in FIG. 7, just as in FIG. 5, the internal boundary of the layers of insulating materials PHD and IMD.


[0051] The main steps of a fabrication process used for obtaining a transducer according to the invention will now be described with reference more particularly to FIGS. 9a to 9d.


[0052] As shown in FIG. 9a, the active zone ZA, within which the transducer will be produced, is defined beforehand in the substrate SB by conventionally delimiting it by isolating trenches STI. Next, a layer of a first material, for example a layer of germanium or a silicon-germanium alloy, is deposited by selective epitaxy. By way of indication, it will be possible to use, as alloy, an Si1-xGex alloy with 0x<1 or else an Si1-x-yGexCy alloy with 0<x≦0.95 and 0<y≦0.05.


[0053] An SiGe alloy having a high germanium content, and in particular one containing 10 to 30% by weight of germanium, will preferably be used.


[0054] The epitaxy is carried out conventionally and is selective in the sense that the material 1 grows only on the silicon of the active region ZA.


[0055] Next, a layer of a second material 2, for example silicon, is deposited, in a conventional manner known per se, by non-selective epitaxy.


[0056] The conditions under which the selective epitaxy and the non-selective epitaxy are carried out are well known to those skilled in the art.


[0057] A photoresistive resin mask 3, intended especially to define the geometry and the dimensions of the membrane, is also defined and then the process continues with successive anisotropic etching operations to remove, . in succession, as illustrated in FIG. 9b, a part of the silicon layer 2, then a part of the layer of material 1 and finally a part of the substrate ZA.


[0058] After the resin mask has been removed, what is obtained, as illustrated in FIG. 9b, is a central stack comprising, at the top, the residual part 20 of the silicon layer which will form the future membrane. It should be noted here that this residual portion 20 of the silicon layer rests, along its two longitudinal ends (not shown in FIG. 9b), on the isolating peripheral region STI.


[0059] The central stack has a residual portion 10 of the material 1 beneath the residual portion 20.


[0060] The trenches TR, made on either side of the stack right into the substrate SB, thus reveal, along the two opposed lateral sides of the stack, the side walls of the residual part 10 of the layer of the material 1.


[0061] The residual portion 10 is selectively removed laterally, so as to form the cavity CV (FIG. 9c). This lateral selective etching may be carried out using one of the techniques well known to those skilled in the art, such as isotropic plasma etching or etching by means of an oxidizing chemical solution, for example a solution containing 40 ml of 50% HNO3, 20 ml of H2O2 and 5 ml of 0.5% HF.


[0062] It is also possible, although this is not absolutely necessary, to form thin layers of dielectric material 4, for example SiO2, on the inner walls of the cavity CV so as to serve as a protective barrier.


[0063] The process then continues with the implantation of dopants into the membrane 20 (FIG. 9d), which results in dopants penetrating into the substrate zone located opposite the membrane so as to form the doped substrate zone ZSB1.


[0064] As a variant, it would be possible to dope the membrane in situ during the epitaxy of the layer 2.


[0065] Next, the cavity CV is closed off by producing, in the bottom of the trenches TR, isolating regions (for example made of SiO2) or spacers ESP.


[0066] Of course, if the transducer is intended to be used in a receiver, an implantation into the well, especially so as to form the zone ZSB2, is also carried out.


[0067] The subsequent steps in the fabrication of the transducer, such as especially the production of the vias and the production of the dielectric materials PHD and IMD, are carried out conventionally.


[0068] The membrane has, for example, a thickness which varies from 5 to 100 nm, preferably from 10 to 20 nm. The surface area of the membrane is, for example, of the order of 100×100 μm. The resonant frequency for these dimensions is about 30 MHz in the case of the membrane and about 1 MHz in the case of the cavity (the speed of the sound waves is of the order of 1 cm/s for a vacuum of 0.01 torr and for a cavity height of 10 nm). Such an acoustic transducer can thus operate within a frequency range of between 1 MHz and 30 MHz.


[0069] Moreover, it is particularly advantageous to provide a matrix MC comprising a relatively large number of ultrasonic transducers, since the acoustic power is proportional to the total surface area of the sensors.


[0070] Moreover, although only a single transducer per active region ZA has been shown, it would have been possible to produce several transducers in each active region ZA by mutually isolating one transducer from another by trenches produced in the substrate and by doping the bottom of the trenches with opposite doping to the doping of the substrate zone of the transducer located beneath the membrane.


[0071]
FIG. 10 illustrates an electrical circuit diagram of an acoustic transmitter according to the invention. The control logic unit comprises, for example, a modulator MDM, of conventional structure, intended to amplitude-modulate a current I. The modulation is controlled by a control processor PC. The output from the modulator is connected to one of the terminals of the membrane MB, while the other terminal of the membrane MB is connected to earth. Under the combined action of the modulated lines of the current and of the magnetic field B, the membrane MB deforms perpendicularly to its plane due to the effect of the Lorenz force, consequently generating ultrasound which will propagate into the air by leaving via the aperture OUV in the package.


[0072] The electrical circuit diagram illustrated in FIG. 11 is an example intended to be implemented in an acoustic receiver.


[0073] One of the terminals of the capacitor of the transducer, for example the terminal connected to the membrane, is connected to a voltage supply STV. The other electrode of the capacitor, in this case the substrate, is connected to the gate of a JFET transistor. The capacitance of the capacitor will be modulated by the mechanical deformation of the membrane generated by a pressure variation resulting from sound propagating through the open air and striking the membrane. The charge transfer between the capacitor C and the gate of the JFET transistor then induces a variation in the drain current, which is converted into a voltage by a resistor. This voltage can then be analysed, after analogue-digital conversion, by a signal processor.


[0074] Other means of detecting the variations in the capacitance of the capacitor are possible. In particular, an inductive-capacitive oscillating circuit or oscillator, the capacitor of which would be formed by the capacitor of the transducer, may be used. The variations in frequency of the oscillator will be representative of the variations in the capacitance of the capacitor of the transducer, and consequently representative of the ultrasonic information received.

Claims
  • 1. Integrated semiconductor acoustic transducer comprising a deformable semiconductor membrane (MB) intended to be coursed by an electric current and separated from one zone of a semiconductor substrate by a cavity (CV) allowing the membrane to deform due to the effect of an acoustic pressure or of a Lorenz force.
  • 2. Transducer according to claim 1, characterized in that the membrane (MB) is formed from doped silicon, in that the substrate is made of silicon and in that the height of the cavity (CV) is of the order of ten nanometres, while the length of the membrane is of the order of a hundred micrometres.
  • 3. Transducer according to either of claims 1 and 2, characterized in that it comprises a capacitor having a first plate formed by the semiconductor membrane (MB), intended to be coursed by an electric current and able to deform due to the effect of an acoustic pressure, and a second plate formed by a doped zone (ZSB1) of the semiconductor substrate, this zone being placed opposite the membrane, and in that the cavity (CV) separating the two plates contains a layer of dielectric gas.
  • 4. Transducer according to claim 3, characterized in that the membrane (MB) rests at two of its ends (EX1, EX2) on an isolating region (STI), in that the substrate includes a well (CS) within a first part of which the said doped substrate zone (ZSB1) is formed and in that it includes several electrical contact pads (V1) coming into contact with one of the said two ends of the membrane and several other electrical contact pads (V2) coming into contact with a second part (ZSB2) of the well.
  • 5. Transducer according to claim 1 or 2, characterized in that the semiconductor membrane (MB) intended to be coursed by a modulated electric current is deformable due to the effect of the Lorenz force generated by a magnetic field (B) lying in the plane of the membrane and perpendicular to the lines of current (I).
  • 6. Transducer according to claim 5, characterized in that the membrane (MB) rests at two of its ends (EX1, EX2) on an isolating region (STI) and in that it includes several electrical contact pads (V1, V2) placed at each of the said two ends, respectively.
  • 7. Acoustic receiver, characterized in that it comprises a semiconductor substrate (SB) containing at least one transducer according to claim 3 or 4, current-generating means (STV) capable of generating the current in the membrane of the transducer, and detection means (TR) connected to the capacitor (C) and capable of detecting the variations in the capacitance of the capacitor which are caused by the deformations of the membrane.
  • 8. Receiver according to claim 7, characterized in that the substrate contains several transducers and in that it includes a package (BO) containing the substrate and having an aperture (OUV) made opposite the transducers.
  • 9. Acoustic transmitter, characterized in that it comprises a semiconductor substrate (SB) containing at least one transducer according to claim 5 or 6, current-generating means (MDM) capable of generating a modulated current in the membrane, and means (AI) for generating a magnetic field which are capable of generating the said magnetic field lying in the plane of the membrane and perpendicular to the lines of current.
  • 10. Transmitter according to claim 9, characterized in that the means for generating the magnetic field comprise a U-shaped magnet (AI).
  • 11. Transmitter according to claim 10, characterized in that the substrate contains several transducers and in that it includes a package (BO) containing the substrate and having an aperture (OUV) made opposite the transducers, the substrate being located between the two arms of the U of the magnet.
  • 12. Process for fabricating an integrated semiconductor acoustic transducer, comprising a deformable semiconductor membrane separated from one zone of a semiconductor substrate by a cavity allowing the membrane to deform, comprising the following steps: making, in the substrate (SB), an isolating region (STI) defining a so-called active semiconductor region; depositing, by selective epitaxy on the surface of the said active region, a first layer (1) of a first material; depositing, by non-selective epitaxy on the first layer (1) and on the isolating region (STI), a second layer (2) of a semiconductor material, the first material being selectively removable with respect to the second material; locally etching the second layer (2), the first layer (1) and part of the active region (ZA) so as to form two lateral trenches (TR) which leave a central stack comprising a part of the second layer (20), a part of the first layer (10) and a part of the active region, and which reveal the said part (10) of the first layer along two opposed lateral sides of the stack; selectively removing, from the sides, the part (10) of the said first layer so as to form the said cavity (CV) bounded by the subjacent part of the substrate forming the said zone of the substrate and the remaining part (20) of the said second layer which forms the said membrane; and forming lateral spacers (ESP) in order to close off the cavity beneath the membrane.
  • 13. Process according to claim 12, characterized in that the said first layer (1) is formed from germanium or from a silicon-germanium alloy and in that the second layer (2) is formed from silicon.
  • 14. Process according to claim 12 or 13, characterized in that the said membrane (MB) is doped.
  • 15. Process according to one of claims 12 to 14, characterized in that the said substrate zone (ZSB1) located opposite the membrane is doped.
  • 16. Process according to one of claims 12 to 15, characterized in that a step of oxidizing the internal walls of the cavity (CV) is carried out.
Priority Claims (2)
Number Date Country Kind
99/15410 Dec 1999 FR
00/03547 Mar 2000 FR
PCT Information
Filing Document Filing Date Country Kind
PCT/FR00/03408 12/6/2000 WO