US20040207456A1

US20040207456A1 - Circuit configuration for DC-biased capacitors

Info

Publication number: US20040207456A1
Application number: US10/752,370
Authority: US
Inventors: Robert York
Original assignee: Individual
Current assignee: Individual
Priority date: 2001-10-31
Filing date: 2004-01-05
Publication date: 2004-10-21
Also published as: WO2003038996A2; EP1451927A4; US6674321B1; JP2005508096A; WO2003038996A3; KR100762672B1; KR20040075861A; EP1451927A2; CN1596506A

Abstract

A capacitive element includes two or more voltage-variable capacitors (varactors). The varactors are configured so that they are coupled in series with respect to an applied AC signal and are coupled in parallel with respect to an applied DC bias voltage. The effective capacitance of the overall capacitive element can be tuned by varying the DC bias voltage.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a continuation of U.S. patent application Ser. No. 10/211,765, “Circuit Configuration for DC-biased Capacitors,” filed Aug. 1, 2002; which claims priority under 35 U.S.C.§ 119(e) to U.S. Provisional Patent Application Serial No. 60/335,191, “Thin-Film Ferroelectric Tuning Circuit,” filed Oct. 31, 2001. This application relates to U.S. patent application Ser. No. 10/144,185, “Voltage-Variable Capacitor with Increased Current Conducting Perimeter,” by Robert A. York, filed May 10, 2002. The subject matter of all of the foregoing is incorporated herein by reference in their entirety.[0001]

BACKGROUND OF THE INVENTION

1. Field of the Invention

This invention relates generally to capacitive elements in which the capacitance can be varied by an applied bias voltage.

2. Description of the Related Art

Circuit elements which have a variable capacitance are a staple component in the design of electronic circuits. In one approach, the capacitance is adjusted by a bias voltage applied to the capacitive element. These voltage-variable capacitors (varactors) can be created using a number of different technologies, including technologies based on thin-film ferroelectric materials.

In the thin-film ferroelectric approach, a thin film of ferroelectric material is sandwiched between conducting electrodes. Examples of suitable ferroelectric materials include barium titanate, strontium titanate, and composites of the two, for example barium strontium titanate (BST). The capacitance value of this structure varies with the applied electric field due to the nonlinear electrical polarization characteristics of the ferroelectric film. The applied electric field is approximately given by ·E=V/d, where E is the electric field, V is the voltage applied across the varactor, and d is the thickness of the ferroelectric film. In practice, electric field strengths of up to 1 MV/cm are required to achieve useful capacitance variations, depending on the specific material composition. However, increasing the field strength further leads to device failure or reliability concerns.

In many applications, the varactor is primarily used to process AC signals and a DC bias voltage is applied across the varactor to set the capacitance of the varactor. The capacitance is tuned by varying the DC bias voltage. However, if the application is limited to low DC voltages (e.g., in battery powered applications) and it is also desirable to tune the varactor over a large range of capacitances (i.e., high “tunability”), then the dielectric film typically must be quite thin in order to achieve the required electric fields. But thin dielectric films result in low breakdown voltages and poor AC power handling. In many circuits such as power amplifiers for wireless applications, the peak AC voltage applied across the varactor can significantly exceed the DC bias voltage. For example, in current cell phones, the battery voltage is typically around 3.5V and the battery produces the DC bias voltage. Therefore, the DC bias voltage typically is limited to 3.5V or less. However, the total voltage (AC+DC) can reach over 7V.

One common approach to increasing the breakdown field in ferroelectric films has been to lightly dope the films with one or more materials. For example, Ti, Mg, Mn, and Zr have been used in BST films to increase the breakdown field. The disadvantage of this approach is that the composite material often has a greatly reduced capacitive tuning for a given applied voltage. This forces the designer to use even thinner films, thus exacerbating the breakdown issue and counteracting gains resulting from the dopants.

Another problem with varactors is that varactors are used as one component in a larger circuit. However, the DC bias voltage typically is applied at the same two terminals which are connected to the external circuit. As a result, the DC bias voltage and corresponding biasing circuitry may not be isolated from the external circuit and interference between the two may result.

Thus, there is a need for capacitive elements which are tunable using low DC bias voltages but which are also capable of handling high AC voltages. It would also be beneficial for the DC bias voltage and circuitry to be isolated from any external circuit in which the capacitive element was used. Capacitive elements based on ferroelectric thin-films typically would have the added advantages of small size, low cost and suitability for mass production.

SUMMARY OF THE INVENTION

The present invention overcomes the limitations of the prior art by providing a capacitive element based on two or more varactors. The varactors are configured so that they are coupled in series with respect to an applied AC signal, thus increasing the AC power handling capability since the total AC voltage swing is divided among all of the varactors. The varactors are coupled in parallel with respect to an applied DC bias voltage, thus maintaining high capacitive tunability with low DC voltages since each varactor experiences the full DC bias voltage.

In one embodiment, the voltage-variable capacitive element includes N (where N>1) varactors that are coupled in series to form a chain. The N+1 nodes in the chain shall be referred to as junction nodes. The capacitive element also includes a first AC node and a second AC node for receiving an AC signal. The first AC node is coupled to the first junction node and the N+1th junction node is coupled to the second AC node. The capacitive element further includes a first DC bias node and a second DC bias node for receiving the DC bias voltage. The first DC bias node is DC coupled to the odd numbered junction nodes and the second DC bias node is DC coupled to the even numbered junction nodes. In some implementations, the DC bias node(s) are coupled to the junction node(s) by AC blocking circuit elements, such as high impedance (i.e., AC blocking) resistors or inductors. In this way, the DC biasing circuitry is isolated from the AC signal. In another aspect of the invention, the AC node(s) are coupled to their respective junction node(s) by DC blocking capacitors, thus isolating the DC bias voltage from any external circuit.

The capacitive elements may be implemented using a wide variety of technologies. For example, discrete components may be used to implement some or all of the capacitive elements. However, the varactors preferably are thin-film ferroelectric varactors and the capacitive element preferably is integrated on a single substrate with the varactor.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention has other advantages and features which will be more readily apparent from the following detailed description of the invention and the appended claims, when taken in conjunction with the accompanying drawings, in which: [0014]
FIG. 1 is a functional diagram of a capacitive element according to the present invention. [0015]
FIGS. 2A and 2B are circuit diagrams illustrating the AC and DC behavior, respectively, of the capacitive element in FIG. 1. [0016]
FIG. 3 is a circuit diagram of one general approach to implementing the capacitive element described in FIGS. 1-2. [0017]
FIG. 4 is a circuit diagram of a capacitive element with four varactors and using resistors as AC blocking circuit elements. [0018]
FIG. 5 is a circuit diagram of another capacitive element with four varactors and using resistors as AC blocking circuit elements. [0019]
FIG. 6 is a circuit diagram of a capacitive element used in a shunt configuration. [0020]
FIG. 7 is a circuit diagram of a capacitive element illustrating the use of any number of varactors in a capacitive element. [0021]
FIG. 8 is a circuit diagram of a capacitive element illustrating the use of inductors as AC blocking circuit elements. [0022]
FIG. 9A is a circuit diagram of another capacitive element. [0023]
FIG. 9B is a top view of a thin-film integrated circuit implementation of the capacitive element of FIG. 9A.[0024]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

FIG. 1 is a functional diagram of a [0025] capacitive element 100 according to the present invention. The capacitive element 100 includes four nodes, two of which shall be referred to as AC nodes 110A-110B and two of which shall be referred to as DC bias nodes 120A-120B. The element also includes at least two varactors 150A-150N. Varactors are voltage-variable capacitors, meaning that the capacitance of the varactor varies according to a DC bias voltage applied across the varactor.
The [0026] capacitive element 100 also functions as a voltage-variable capacitor. The AC nodes 110 serve as the terminals of the capacitor. An AC signal is applied to the AC nodes 110 and the element 100 primarily has a capacitive effect on the applied AC signal. The actual capacitance of element 100 is determined by a DC bias voltage applied to the DC bias nodes 120. Thus, by changing the DC bias voltage applied to DC bias nodes 120, the effective capacitance between AC nodes 110 can be tuned.
FIGS. 2A and 2B are functional diagrams further illustrating the operation of [0027] capacitive element 100. FIG. 2A illustrates the AC behavior. The circuitry within capacitive element 100 is constructed in a manner so that the N varactors 150 are coupled in series with respect to the AC signal applied to AC nodes 110. Assuming for the moment that there are no other capacitors, as is shown in FIG. 2A, the effective capacitance of element 100 is derived from the series combination of the N varactors 150. If each varactor has a capacitance of C, the effective capacitance is C/N. Furthermore, each varactor 150 experiences an AC voltage drop which is 1/N of the total AC voltage swing applied to the AC nodes 110, thus improving the power handling capability of the element 100.
FIG. 2B illustrates the DC behavior of [0028] capacitive element 100. The circuitry of element 100 is constructed so that the N varactors 150 are coupled in parallel with respect to the DC bias voltage applied to DC bias nodes 120. In this way, the full DC bias voltage is applied across each varactor 150. As a result, lower bias voltages can be used to tune the capacitance of the varactors 150. For example, if the varactors 150 were instead coupled in series with respect to the DC bias voltage and each varactor 150 experienced a DC voltage drop of 1/N of the total DC bias voltage, then the voltage applied across DC bias nodes 120 would have to be N times greater than that required by the parallel configuration shown in FIG. 2B.
It should be understood that FIGS. 1 and 2 are simplified functional descriptions. Various implementations may deviate from this model. For example, the circuits of FIGS. 2A and 2B show perfect series and parallel coupling of the varactors [0029] 150 between the AC and DC bias nodes. Implementations may not achieve this ideal result, for example due to circuit imperfections and/or additional circuit elements which affect the couplings. In most cases, circuit elements in addition to the varactors 150 are used to achieve the functionality shown in FIG. 2 but the resulting capacitive element 100 will not achieve the exact ideal behavior shown in FIG. 2. In some cases, circuit elements may be added to capacitive element 100 in order to achieve other purposes.
As another example, FIG. 1 shows the four nodes [0030] 110 and 120 as separate. This is done for clarity. In various implementations, a physical node may play the role of one or more of the nodes 110 and 120. Conversely, each node 110, 120 may be implemented by two or more physical nodes. Similarly, each varactor 150 is represented by a single symbol. In fact, each varactor 150 itself may include multiple elements, including for example series or parallel combinations of individual varactors. As a final example, it is not required that the varactors 150 be identical or have the same capacitance.
The approach described above can be implemented using many different technologies. Much of this disclosure shall focus on [0031] capacitive elements 100 based on voltage-tunable dielectric materials, including ferroelectric materials or materials derived from ferroelectric materials. The tunable dielectric layer preferably exhibits a field-dependent permittivity in a (non-hysteretic) paraelectric state over a useful temperature range (e.g., −30C to +9C). In the disclosure, particular emphasis will be given to integrated implementations using thin-film ferroelectric varactors 150. However, the invention is not limited to this specific technology. The principles described here are generally applicable to many types of voltage-variable capacitors 150, including those which are not thin-film, not ferroelectric and/or not integrated. The principles are not specific to a particular choice of tunable dielectric material, film thickness or fabrication sequence. Examples of other types of varactor technology include tunable or switchable capacitors using MEMS (micro-electromechanical systems) technology.
The approach has many advantages. For example, it improves the AC power handling capability of the [0032] capacitive element 100 without reducing the capacitive tuning range achievable by a given range of DC bias voltages. In contrast, consider the alternatives. In thin-film technology, the capacitance of the varactor is the result of a dielectric thin-film. The AC power handling capability of the varactor is typically limited by the breakdown voltage for the thin-film. This breakdown voltage can be increased by increasing the thickness of the film. However, the capacitive tuning is typically dependent on the electric field in the film. Increasing the film thickness results in a lower electric field for a given DC bias voltage, thus reducing the capacitive tuning range. Therefore, increasing the AC power handling capability reduces the capacitive tuning range. In the approach described here, the AC power handling capability can be increased without reducing the capacitive tuning range.
In addition, in many implementations, the DC bias voltage is electrically isolated from the AC signal in order to simultaneously achieve the AC series coupling and DC parallel coupling shown in FIG. 2. As a result, the DC bias voltage and biasing circuitry typically will also be isolated from the external circuit into which the [0033] capacitive element 100 is integrated. This reduces any adverse effects caused by the DC biasing on the external circuit, or vice versa, and allows the capacitive element 100 to be easily incorporated into a wide variety of external circuits.
The combination of good AC power handling and low DC bias voltage makes the [0034] capacitive element 100 suitable for many applications. Implementation using ferroelectric thin-film technology adds the advantages of small size and a high degree of integration. As a result, the capacitive element 100 is well suited for battery-operated devices, which typically require low operating voltages. It is also appropriate for handheld wireless devices, for example to adjust the impedance level in the power-amplifier circuit to maintain high power and high efficiency, and hence long battery life. With the addition of series- or parallel-connected inductors, a tunable resonator can be formed for use in impedance matching and filter networks. The capacitive element 100 can also be used to compensate for manufacturing tolerances, thus improving yields and reducing costs.
Two example applications concern RF communications. In one example, the [0035] capacitive element 100 is used in a tunable reactive matching network for RF power amplifiers. In the other, the capacitive element 100 is used in a tunable filter. In both cases, the capacitive element increases power handling and reduces intermodulation distortion while maintaining low DC bias voltage. Other applications (including applications at frequencies other than RF) will be apparent.
FIGS. 3-9 are examples of different embodiments of [0036] capacitive element 100. In many of these examples, high-impedance resistors and/or inductors are used as AC blocking circuit elements in order to isolate the AC signal from the DC bias circuitry and DC blocking capacitors are used to isolate the DC bias voltage from the external circuit (i.e., the circuit connected to the AC nodes).
FIG. 3 is a circuit diagram of one [0037] general approach 300 to implementing capacitive element 100. In this approach, the varactors 150A-N are coupled in series to form a chain. The chain itself has N+1 nodes 310: the N−1 nodes 310B-310N located between the varactors and the two nodes 310A and 3100 located on either end of the chain. For convenience, these nodes 310 shall be referred to as junction nodes. One of the AC nodes 110A is coupled to the first junction node 310A and the last junction node 3100 is coupled to the other AC node 110B. Each DC bias node 120 is DC coupled to alternate junction nodes 310. In FIG. 3, DC bias node 120A is DC coupled to the odd-numbered junction nodes (i.e., nodes 310A, 310C, 310E, etc.) and DC bias node 120B is DC coupled to the even-numbered junction nodes (i.e., nodes 310B, 310D, etc.). The couplings between nodes 110, 120 and junction nodes 310 may includes additional elements (as generally represented by the dashed lines), for example to provide isolation between the DC and AC portions of the capacitive element 300.
FIG. 4 is a specific example of [0038] capacitive element 300 using four varactors 150A-D, two capacitors 410A,B as DC blocking circuit elements and five resistors 420A,C,E and 430B,D as AC blocking circuit elements. The four varactors 150 are coupled in series to form a chain, as described in FIG. 3. Each end of the varactor chain (i.e., junction nodes 310A and 310E) is coupled to one of the AC nodes 110A, 110B by one of the DC blocking capacitors 410A,410B. The DC bias nodes 120A,120B are coupled to alternate junction nodes by the bias resistors 420 and 430. More specifically, DC bias node 120A is coupled to junction nodes 310A,C,E by bias resistors 420A,C,E; and DC bias node 120B is coupled to junction nodes 310B,D by bias resistors 420B,D.
The capacitive element in FIG. 4 functions as follows. With respect to the AC signal applied to AC nodes [0039] 110, the bias resistors 420,430 have high resistance and effectively impede AC current flow through these portions of the circuit. As a result, to the AC signal, the capacitive element behaves like six capacitors 410, 150 coupled in series, thus approximating the AC model shown in FIG. 2A. With respect to the DC bias voltage applied across DC bias nodes 120, the varactors 150 and DC blocking capacitors 410 effectively impede DC current flow through these portions of the circuit. As a result, there is a negligible voltage drop across the bias resistors 420, 430 and each varactor 150 is biased by approximately the full DC bias voltage since the DC bias nodes 120 are coupled to alternating junction nodes 310. This approximates the DC model shown in FIG. 2B. In addition, the DC blocking capacitors 410 effectively isolate the DC bias circuitry from any external circuit connected to the AC nodes 110, thus preventing interference between these two circuits.
Variations on the DC biasing scheme will be apparent. For example, FIG. 5 is a circuit diagram of another capacitive element in which the DC bias resistors [0040] 520 are arranged differently. Rather than coupling each junction node 310A,C,E directly to DC bias node 120A via a bias resistor, the bias resistors 520 are used to couple between the junction nodes 310A,C,E. More specifically, resistor 520A couples the DC bias node 120A to junction node 310A, resistor 520C couples junction node 310A to 310C and resistor 520E couples junction node 310C to 310E. However, the bias resistors 520 generally perform the same function as in FIG. 4. They impede AC current flow and distribute the DC bias voltage to alternate junction nodes.
In FIG. 6, the capacitive element of FIG. 5 is used in a shunt configuration, meaning that one [0041] AC node 110A is tied to ground. The AC ground and DC ground are assumed to be the same. As a result, the circuitry can be somewhat simplified. In particular, the DC blocking capacitor 410A is eliminated since DC blocking is not required at the grounded node. Similarly, the bias resistor 520A is also eliminated since AC blocking is not required. The resulting circuit is shown in FIG. 6.
FIGS. 4-6 have shown capacitive elements with exactly four varactors. However, as described with respect to FIG. 3, other numbers of varactors can be used. FIG. 7 illustrates how the circuit design of FIG. 4 can be extended to any number of varactors [0042] 150. The other circuit designs can be similarly extended.
Furthermore, resistors are only one type of circuit element which is AC blocking. Other types of AC blocking circuit elements can also be used. For example, some or all of the resistors shown in FIGS. 3-7 can be replaced by large-value inductors. As one example, the circuit shown in FIG. 8 is the same as that shown in FIG. 6, except that the [0043] bias resistors 520C,E are replaced by bias inductors 820C,E.
The capacitive elements described can be implemented using individual discrete components mounted on a separate circuit board or carrier. For example, some thin-film resistor schemes have a limited sheet resistance that makes implementation of large value resistors difficult. In these cases, the resistors can be implemented as discrete components instead. [0044]
Preferably, however, the entire capacitive element is implemented as a single integrated circuit, with ferroelectric thin-film technology used to implement the varactors. Standard IC fabrication methods can be used to fabricate the capacitive elements. To reduce costs, inexpensive insulating substrates are preferred, including but not limited to high-resistivity silicon (HR Si), crystalline sapphire (Al[0045] ₂O₃), Aluminum Nitride (AlN), quartz and glass. These substrates are polished for low surface roughness for compatibility with growth of smooth ferroelectric films with high breakdown fields. This approach results in low-cost, small size, reliable components which are suitable for mass production and for integration with additional circuit elements.
FIGS. 9A and 9B illustrate a capacitive element which is implemented in this manner. FIG. 9A shows the circuit diagram and FIG. 9B is a top view of the thin-film integrated circuit implementation. FIG. 9B is a line drawing based on a photograph of the actual capacitive element. Referring to FIG. 9A, the circuit implemented is a two-varactor version of the design shown in FIG. 6. As such, there are five circuit elements: two [0046] varactors 150A and 150B having capacitance C, two bias resistors 430B and 520C having resistance Rb and Rp respectively, and DC blocking capacitor 410B having capacitance Cb.
Referring to FIG. 9B, the AC and DC bias nodes [0047] 110 and 120 are implemented as large metal pads 1, 2 and 3 to allow for interconnection to external elements. The leftmost pad 1 is AC node 10B, the rightmost pad 2 functions as both AC node 10A and DC bias node 120A (i.e., the nodes connected to ground in FIG. 9A), and the bottom pad 3 is DC bias node 120B. These pads can be used for wire-bond attachment or solder-bump attachment, for example.
Beginning with the AC signal path, [0048] AC node 110B is connected to AC ground 110A by a DC blocking capacitor 410B and two varactors 150B and 150A coupled in series. In FIG. 9B, this signal path can be seen laid out from left to right. Immediately to the right of the AC node 110B (metal pad 1) are two senes-connected capacitors labeled “2Cb.” Each of these capacitors has a capacitance of 2Cb. Together, they have an effective capacitance of Cb, thus implementing the DC blocking capacitor 410B. The series-connected capacitor pair is used in order to increase the DC breakdown voltage of this element. In this example, the DC blocking capacitors are constructed from the same materials as the varactors, although this is not required.
Since ferroelectric films have a large capacitance density, it is advantageous to use ferroelectric films for the blocking [0049] capacitor 410B even though the resulting variation of capacitance with voltage is not necessarily utilized or desired. Large-value blocking capacitors ordinarily consume a large amount of substrate area on a chip, so the use of ferroelectric materials results in a significant reduction in size. Other dielectric materials can also be used for the blocking capacitor, including for example silicon dioxide (SiO₂), silicon nitride (Si_xN_y), aluminum oxide (Al₂O₃) and titanium oxide (TiO₂).
The two [0050] varactors 150B,150A are labeled as “C” in FIG. 9B. They are implemented using conventional ferroelectric thin-film technology and typically barium titanate, strontium titanate, or a solid-solution of the two is used as the dielectric.
The capacitance Cb of the blocking [0051] capacitor 410B preferably is at least five times greater than the capacitance C of the varactors 150, in order to make effective use of the potential tuning variation of the varactors. In this particular example, the blocking capacitor has a capacitance of at least 20 times the varactor capacitance and the varactors provide a change in capacitance of a factor of 3 in response to DC bias voltages that range from zero volts to near the breakdown voltage of the tunable dielectric material. For many RF applications, the blocking capacitors may be several hundred picoFarads in size, and the capacitance of the varactors may vary by a factor of 2 in response to a DC bias voltage that ranges from 0-5 volts.
The two [0052] bias resistors 430B and 520C are implemented using conventional thin-film technology. They are labeled as “Rb” and “Rp,” respectively, in FIG. 9B. The resistances preferably are large compared to the reactance of the varactors to insure that negligible AC current flows through these bias resistors.
More specifically, [0053] bias resistor 430B preferably has a resistance which is at least ten times greater than the reactance of varactor 150. In other words, Rb>10/ωC, where co is the angular frequency of operation. Similarly, bias resistor 520C preferably has a resistance which is at least ten times greater than the reactance of varactor 150, or Rp>10/ωC.
If bias inductors are used in place of bias resistors, a wide range of values may be used. In some applications, it may be desirable to select the inductance to resonate with the varactors at a certain frequency coo. In other cases, the bias inductor may function primarily as a choke, in which case the reactance of the inductor should be significantly higher than that of the varactor. For example, the ×10 and ×50 rules of thumb described for resistors may also be applied to inductors. [0054]
Although the invention has been described in considerable detail with reference to certain preferred embodiments thereof, other embodiments will be apparent. For example, some implementations do not include blocking capacitors and/or shunt resistors. Referring to FIG. 9A, if [0055] AC node 110B is connected to an RF circuit which is at a DC ground potential, then shunt resistor 520C and blocking capacitor 410B can be eliminated. Therefore, the scope of the appended claims should not be limited to the description of the preferred embodiments contained herein.

Claims

What is claimed is:

1. A voltage-variable capacitive element comprising:

a first DC bias node and a second DC bias node for receiving a DC bias voltage;

a first AC node and a second AC node for receiving an AC signal;

N varactors, wherein N is at least two and each varactor has a capacitance that varies according to a voltage applied across the varactor; and

wherein the N varactors are coupled in parallel between the DC bias nodes with respect to the DC bias voltage and are coupled in series between the AC nodes with respect to the AC signal.