CN115441847A

CN115441847A - Low-loss transverse-excitation film bulk acoustic resonator and filter

Info

Publication number: CN115441847A
Application number: CN202210582259.6A
Authority: CN
Inventors: 布莱恩特·加西亚; 格雷格·戴尔
Original assignee: Resonant Inc
Current assignee: Murata Manufacturing Co Ltd
Priority date: 2021-06-03
Filing date: 2022-05-26
Publication date: 2022-12-06
Also published as: JP2022186631A; JP7416126B2; DE102022113663A1

Abstract

The acoustic resonator device includes a diaphragm comprising a portion of a piezoelectric plate spanning a cavity in a substrate. The conductor pattern on the surface of the piezoelectric plate includes an interdigital transducer (IDT) having a first bus bar, a second bus bar, and a plurality of interleaved fingers extending alternately from the first and second bus bars such that the plurality of interleaved fingers partially overlap the diaphragm. The conductor pattern further includes first and second reflector elements on the membrane proximate and parallel to a first finger of the plurality of interleaved fingers, and third and fourth reflector elements on the membrane proximate and parallel to a last finger of the plurality of interleaved fingers.

Description

Low-loss transverse-excitation film bulk acoustic resonator and filter

Technical Field

The present disclosure relates to radio frequency filters using acoustic wave resonators, and more particularly to filters for use in communication devices.

Background

A Radio Frequency (RF) filter is a two-terminal device that is configured to pass some frequencies and block others, where "pass" means transmitting with relatively low signal loss and "block" means blocking or substantially attenuating. The range of frequencies passed by the filter is referred to as the "passband" of the filter. The range of frequencies blocked by such a filter is called the "stop band" of the filter. A typical RF filter has at least one pass band and at least one stop band. The specific requirements of the pass band or stop band depend on the specific application. For example, "passband" may be defined as a range of frequencies where the insertion loss of the filter is better than a defined value such as 1dB, 2dB, or 3dB. A "stop band" may be defined as a frequency range in which the rejection of the filter is greater than a defined value, for example a value of 20dB, 30dB, 40dB or more, depending on the particular application.

RF filters are used in communication systems that transmit information over wireless links. For example, RF filters can be found in the RF front-ends of cellular base stations, mobile phones and computing devices, satellite transceivers and ground stations, internet of things (IoT) devices, laptops and tablets, fixed-point radio links, and other communication systems. RF filters are also used in radar and electronic and information warfare systems.

RF filters typically require many design tradeoffs to achieve the best compromise between performance parameters such as insertion loss, rejection, isolation, power handling, linearity, size and cost for each particular application. Specific designs and manufacturing methods and enhancements may benefit from one or more of these needs simultaneously.

The performance enhancement of RF filters in wireless systems can have a wide impact on system performance. System performance can be improved by improving the RF filter, e.g., larger cell size, longer battery life, higher data rate, larger network capacity, lower cost, greater security, higher reliability, etc. These improvement points may be implemented individually or in combination at various levels of the wireless system, for example at the RF module, RF transceiver, mobile or fixed subsystem or network level.

High performance RF filters for current communication systems typically incorporate acoustic wave resonators including Surface Acoustic Wave (SAW) resonators, bulk acoustic wave BAW) resonators, thin Film Bulk Acoustic Resonators (FBARs), and other types of acoustic wave resonators. However, these prior art techniques are not suitable for use at higher frequencies and bandwidths that will be required for future communication networks.

To obtain a wider communication channel bandwidth, it is necessary to use a higher frequency communication band. The 3GPP (third generation partnership project) has standardized radio access technologies for mobile telephone networks. The radio access technology for fifth generation mobile networks is defined in the 5G NR (new radio) standard. The 5G NR standard defines several new communication bands. Two of these new communications bands are n77 and n79, where n77 uses a frequency range of 3300MHz to 4200MHz, and n79 uses a frequency range of 4400MHz to 5000 MHz. Both frequency band n77 and frequency band n79 use Time Division Duplexing (TDD), so communication devices operating in frequency band n77 and/or n79 use the same frequency for uplink and downlink transmissions. The bandpass filters for the n77 and n79 frequency bands must be able to handle the transmission power of the communication device. High frequencies and wide bandwidths are also required in the wireless bands of 5GHz and 6 GHz. The 5G NR standard also defines the millimeter wave communication band at frequencies between 24.25GHz and 40 GHz.

Laterally excited thin film bulk acoustic resonators (XBARs) are acoustic resonator structures used in microwave filters. Such an XBAR is described in U.S. Pat. No. 10,491,291 entitled "TRANSVERSELY EXCITED FILM BULK ACOUSTIC RESONATOR". XBAR resonators include interdigital transducers (IDTs) formed on a thin floating layer or membrane having a single crystal of piezoelectric material. The IDT includes a first set of parallel fingers extending from a first bus bar and a second set of parallel fingers extending from a second bus bar. The first set of parallel fingers and the second set of parallel fingers are interleaved. The microwave signal applied to the IDT excites a shear primary acoustic wave in the piezoelectric diaphragm. XBAR resonators provide high electromechanical coupling and high frequency capability. XBAR resonators can be used in a variety of RF filters including band-stop filters, band-pass filters, duplexers, and multiplexers. XBAR is well suited for use in filters for communication bands with frequencies above 3 GHz.

Disclosure of Invention

The invention discloses an acoustic resonator device, comprising: a diaphragm comprising a portion of the piezoelectric plate spanning a cavity in the substrate; and a conductor pattern formed on the surface of the piezoelectric plate, the conductor pattern including: an interdigital transducer (IDT) comprising a first bus bar, a second bus bar, and a plurality of interleaved fingers, wherein the plurality of interleaved fingers extend alternately from the first and second bus bars, portions of the plurality of interleaved fingers overlapping the diaphragm; first and second reflector elements on the membrane proximate and parallel to a first finger of the plurality of interleaved fingers; and third and fourth reflector elements located on the membrane proximate and parallel to a last finger of the plurality of interleaved fingers.

Wherein the first finger extends from the first busbar, the first and second reflector elements extend from the second busbar, the last finger extends from either of the first and second busbars, and the third and fourth reflector elements extend from the other of the first and second busbars.

Wherein pr1 is the center-to-center distance of the first and second reflector elements and the center-to-center distance of the third and fourth reflector elements, p is the pitch of the plurality of interleaved fingers, and 1.2p ≦ pr1< ≦ 1.5p.

Wherein pr2 is the center-to-center distance of the first reflector element and the first finger and the center-to-center distance of the third reflector element and the last finger, and p ≦ pr2 ≦ pr1.

Wherein pr2= (pr 1+ p)/2.

Wherein the markers mr of the first, second, third and fourth reflector elements are configured to increase the Q factor of the device at a predetermined frequency.

Wherein the device is a series resonator in a ladder bandpass filter circuit having a passband, and mr is selected to increase the Q factor of the device at the upper edge of the passband.

The invention also discloses a band-pass filter, comprising: a plurality of acoustic resonators including one or more series resonators and one or more parallel resonators connected in a ladder filter circuit, wherein each of the plurality of acoustic resonators includes: a respective diaphragm comprising a portion of a piezoelectric plate spanning a respective cavity in a substrate, and an interdigital transducer (IDT) formed on a surface of the piezoelectric plate, the IDT comprising a first bus bar, a second bus bar, and a plurality of interleaved fingers extending alternately from the first and second bus bars, a portion of the plurality of interleaved fingers overlapping the respective diaphragm, wherein at least one of the one or more series resonators further comprises: first and second reflector elements located on the respective membranes proximate and parallel to a first finger of the plurality of interleaved fingers; and third and fourth reflector elements located on the respective membranes proximate and parallel to a last finger of the plurality of interleaved fingers.

Wherein, for each series resonator comprising the first to fourth reflector elements: the first finger extends from the first busbar, the first and second reflector elements extend from the second busbar, the last finger extends from either of the first and second busbars, and the third and fourth reflector elements extend from the other of the first and second busbars.

Wherein, for each series resonator comprising the first to fourth reflector elements: pr1 is the center-to-center distance of the first and second reflector elements and the center-to-center distance of the third and fourth reflector elements, p is the pitch of the interleaved fingers, and 1.2p ≦ pr1 ≦ 1.5p.

Wherein pr2 is the center-to-center distance of the first reflector element and the first finger and the center-to-center distance of the third reflector element and the last finger, and p < ≦ pr2< ≦ pr1.

Wherein pr2= (pr 1+ p)/2.

Wherein, for each series resonator comprising the first to fourth reflector elements: the markers mr of the first, second, third and fourth reflector elements are configured to improve the input/output transfer function of the filter at a predetermined frequency.

Wherein mr is selected to improve the input/output transfer function of the filter at the upper edge of the passband.

Wherein all of the one or more series resonators include respective first through fourth reflector elements.

Drawings

Fig. 1 includes a schematic plan view, two schematic cross-sectional views and a detailed view of a laterally excited thin film bulk acoustic resonator (XBAR).

Fig. 2 is a schematic block diagram of a band pass filter using acoustic resonators.

FIG. 3 is a graph of the Q factor of an XBAR versus the number of fingers in an interdigital transducer (IDT) of the XBAR.

Fig. 4 is a schematic plan view of an IDT having reflector elements.

Fig. 5 is a schematic plan view of another IDT having reflector elements.

Fig. 6 is a graph comparing normalized Q-factors at resonant frequency for XBARs with and without reflector elements.

Fig. 7 is a graph comparing the normalized Q factor at the anti-resonant frequency for XBARs with and without reflector elements.

FIG. 8 is a graph showing the relative Q-factor of a representative XBAR at a 5150MHz frequency as a function of reflector element spacing and index.

Fig. 9 is a graph showing the relative Q factor of an XBAR with two reflector elements at each end as a function of reflector element index and frequency.

Fig. 10 is a graph showing the relative Q factor of an XBAR with one reflector element at each end as a function of reflector element index and frequency.

Fig. 11 is a graph showing the relative Q factor of an XBAR with five reflector elements at each end as a function of reflector element index and frequency.

Figure 12 is a graph comparing the performance of two bandpass filters using XBARs with and without reflector elements.

Fig. 13 is a schematic plan view of another IDT having reflector elements.

Fig. 14A and 14B are graphs comparing the performance of bandpass filters using XBARs with and without reflector elements.

Fig. 15 is a flow chart of a method for manufacturing an XBAR or a filter using an XBAR.

Throughout the specification, elements appearing in the drawings are assigned a three-digit or four-digit reference numeral, where the two least significant bits are specific to the element and one or two most significant bits are the figure number showing the element first. Elements not described in connection with the figures may be assumed to have the same characteristics and functions as previously described elements with the same reference numerals.

Detailed Description

Description of the apparatus

Fig. 1 shows a simplified schematic top view and orthogonal cross-sectional view of an XBAR 100. XBAR resonators such as resonator 100 may be used in a variety of RF filters including band-stop filters, band-pass filters, duplexers, and multiplexers.

XBAR 100 is comprised of a thin film conductor pattern formed on the surface of a piezoelectric plate 110 having parallel front and

back surfaces

112 and 114, respectively. The piezoelectric plate is a thin single crystal layer made of a piezoelectric material such as lithium niobate, lithium tantalate, langasite, gallium nitride, or aluminum nitride. The piezoelectric plate is cut such that the orientation of the X, Y, and Z crystallographic axes relative to the front and back surfaces are known and consistent. The piezoelectric plate may be Z-cut, that is, the Z-axis is perpendicular to the front and back faces 112, 114. The piezoelectric plate may be rotary Z-cut or rotary YX-cut. XBARs can be fabricated on piezoelectric plates having other crystal orientations.

The back side 114 of the piezoelectric plate 110 is attached to the surface of the substrate 120, except that a portion of the piezoelectric plate 110 is not attached to the surface of the substrate 120, wherein the portion of the piezoelectric plate 110 forms a diaphragm 115, the diaphragm 115 spanning a cavity 140 formed in the substrate. The portion of the piezoelectric plate spanning the cavity is referred to herein as the "diaphragm" 115 because this portion is physically similar to the diaphragm of the microphone. As shown in fig. 1, the diaphragm 115 abuts the remainder of the piezoelectric plate 110 around the entire perimeter 145 of the cavity 140. In this case, "adjacent" means "continuously connected without any other article in between". In other configurations, the diaphragm 115 may abut the piezoelectric plate around at least 50% of the perimeter 145 of the cavity 140.

The substrate 120 provides mechanical support for the piezoelectric plate 110. The substrate 120 may be, for example, silicon, sapphire, quartz, or some other material, or a combination of these materials. The back side 114 of the piezoelectric plate 110 may be attached to the substrate 120 using a wafer bonding process. Alternatively, the piezoelectric plate 110 may be grown on the substrate 120 or attached to the substrate in some other manner. The piezoelectric plate 110 may be attached directly to the substrate, or may be attached to the substrate 120 via one or more intermediate layers of material (not shown in fig. 1).

The conventional meaning of "cavity" is "empty space within a solid". The cavity 140 may bebase:Sub>A hole completely through the substrate 120 (as shown in cross-sectionsbase:Sub>A-base:Sub>A and B-B) or may bebase:Sub>A groove in the substrate 120 below the diaphragm 115. For example, the cavity 140 may be formed by selectively etching the substrate 120 before or after attaching the piezoelectric plate 110 to the substrate 120.

The conductor pattern of XBAR 100 includes an interdigital transducer (IDT) 130. The IDT 130 includes a first plurality of parallel fingers, such as finger 136, extending from the first bus bar 132 and a second plurality of fingers extending from the second bus bar 134. The term "bus bar" refers to a conductor from which the fingers of the IDT extend. The first and second pluralities of parallel fingers are interleaved. The interleaved fingers overlap a distance AP, which is commonly referred to as the "aperture" of the IDT. The center-to-center distance L between the outermost fingers of the IDT 130 is the "length" of the IDT.

First and second bus bars 132, 134 serve as terminals of XBAR 100. A radio frequency or microwave signal applied between the two

bus bars

132, 134 of the IDT 130 excites a primary acoustic mode within the piezoelectric plate 110. The primary acoustic mode is a bulk shear mode in which acoustic energy propagates in a direction substantially perpendicular to the surface of the piezoelectric plate 110, which is also perpendicular or transverse to the direction of the electric field generated by the IDT fingers. XBAR is therefore considered to be a laterally excited thin film bulk wave resonator.

The IDT 130 is placed on the piezoelectric plate 110 such that at least the fingers of the IDT 130 are disposed on a diaphragm 115, the diaphragm 115 spanning or suspended from the cavity 140. As shown in FIG. 1, the cavity 140 has a rectangular shape with dimensions greater than the aperture AP and length L of the IDT 130. The cavities of the XBAR may have different shapes, such as regular or irregular polygons. The cavities of the XBAR may have more or less than four sides, which may be straight or curved.

For ease of illustration in fig. 1, the geometrical spacing and width of the IDT fingers are greatly exaggerated relative to the length (dimension L) and aperture (dimension AP) of the XBAR. A typical XBAR has more than ten parallel fingers in IDT 130. An XBAR may have hundreds, possibly thousands, of parallel fingers in IDT 130. Similarly, the thickness of the IDT fingers and piezoelectric plate is greatly enlarged in the cross-sectional view.

Referring to the detailed cross-sectional view, a front dielectric layer 150 is optionally formed on the front surface of the piezoelectric plate 110. By definition, the "front side" of an XBAR is the surface facing away from the substrate. The front dielectric layer 150 may be formed only between IDT fingers (e.g., IDT finger 138 b) or may be deposited as a cover layer such that a dielectric layer is formed between and over the IDT fingers (e.g., IDT finger 138 a). The front side dielectric layer 150 may be a non-piezoelectric dielectric material such as silicon dioxide, aluminum oxide, or silicon nitride. The thickness of the front dielectric layer 150 is typically less than about one-third of the thickness tp of the piezoelectric plate 110. The front dielectric layer 150 may be formed of multiple layers of two or more materials. In some applications, a backside dielectric layer (not shown) may be formed on the backside of the piezoelectric plate 110.

IDT fingers

138a and 138b can be one or more layers of aluminum, aluminum alloy, copper alloy, beryllium, gold, tungsten, molybdenum, chromium, titanium, or some other electrically conductive material. An IDT finger can be considered "substantially aluminum" if it is made of aluminum or an alloy containing at least 50% aluminum. An IDT finger is considered to be "substantially copper" if it is made of copper or an alloy containing at least 50% copper. A thin (relative to the total thickness of the conductor) layer of another metal (e.g., chromium or titanium) or other thin metal layer may be formed under and/or over the fingers as a layer within the fingers to improve adhesion between the fingers and the piezoelectric plate 110 and/or to passivate or encapsulate the fingers and/or to improve power handling. The bus bars (132, 134 in fig. 1) of the IDT can be made of the same or different material than the fingers.

Dimension p is the center-to-center distance or "pitch" of the IDT fingers, which may be referred to as the pitch of the IDT and/or the pitch of the XBAR. Dimension m is the width or "signature" of the IDT finger. In some embodiments, the spacing and/or marking of the IDT fingers may vary slightly along the length of the IDT. In this case, the dimensions p and m are the average values of the pitch and the mark, respectively. The IDT geometry of XBAR is very different from that used in Surface Acoustic Wave (SAW) resonators. In the SAW resonator, the pitch of the IDT is one-half of the wavelength of the acoustic wave at the resonance frequency. Further, the tag pitch ratio of the SAW resonator IDT is typically close to 0.5 (i.e., the tag or finger width is approximately one-quarter of the acoustic wavelength at resonance). In XBAR, the pitch p of the IDT may be 2 to 20 times the width w of the fingers. The pitch p is typically 3.3 to 5 times the finger width w. Further, the pitch p of the IDTs may be 2 to 20 times the thickness of the piezoelectric plate 210. The pitch p of the IDTs is typically 5 to 12.5 times the thickness of the piezoelectric plate 210. The width m of the IDT fingers is in the XBAR and is not limited to about one quarter of the resonant acoustic wave wavelength. For example, the width of the XBAR IDT fingers can be 500nm or more, so that IDTs can be easily manufactured using optical lithography. The thickness of the IDT fingers can be from 100nm to about equal to the width m. The thickness of the bus bars (132, 134) of the IDT can be equal to or greater than the thickness tm of the IDT fingers.

Fig. 2 is a schematic circuit diagram and layout of a high-frequency band-pass filter 200 using XBAR. The filter 200 has a conventional ladder filter architecture comprising three

series resonators

210A, 210B, 210C and two

parallel resonators

220A, 220B. Three

series resonators

210A, 210B, and 210C are connected in series between the first port and the second port (hence referred to as "series resonators"). In fig. 2, the first and second ports are labeled "In" and "Out", respectively. However, the filter 200 is bi-directional and either port may be used as an input or an output of the filter. The two

parallel resonators

220A, 220B are connected to ground from the node between the series resonators. The filter may contain additional reactive elements, such as capacitors and/or inductors, which are not shown in fig. 2. All parallel resonators and series resonators are XBARs. The inclusion of three series and two parallel resonators is exemplary. A filter may have more or less than five total resonators, more or less than three series resonators, and more or less than two shunt resonators. Typically, all series resonators are connected in series between the input and output of the filter. All parallel resonators are typically connected between ground and the input, output or node between two series resonators.

In the exemplary filter 200, the three series resonators 210A, B, C and the two

parallel resonators

220A, 220B of the filter 200 are formed on a single plate 230 of piezoelectric material bonded to a silicon substrate (not visible). In some filters, the series resonators and the parallel resonators may be formed on different plates of piezoelectric material. Each resonator includes a respective IDT (not shown), at least the fingers of the IDT being disposed above the cavity in the substrate. In this and similar contexts, the term "each" means "relating things to each other", that is, having a one-to-one correspondence. In fig. 2, the cavity is schematically illustrated as a dashed rectangle (e.g., rectangle 235). In this example, each IDT is disposed on a respective cavity. In other filters, IDTs of two or more resonators may be disposed on a single cavity.

Each of the

resonators

210A, 210B, 210C, 220A, 220B in the filter 200 has resonance when the admittance of the resonator is very high and has anti-resonance when the admittance of the resonator is very low. Resonance and antiresonance occur at the resonant frequency and antiresonant frequency, respectively, which may be the same or different for the various resonators in filter 200. In overly simplified terms, each resonator may be considered a short circuit at its resonant frequency and an open circuit at its anti-resonant frequency. At the resonant frequency of the parallel resonator and the anti-resonant frequency of the series resonator, the input-output transfer function will be close to zero. In a typical filter, the resonance frequency of the parallel resonators is located below the lower edge of the filter passband, and the anti-resonance frequency of the series resonators is located above the upper edge of the passband. In some filters, a front dielectric layer (also referred to as a "frequency setting layer") indicated by a dotted rectangle 270 may be formed on the parallel resonators to set the resonance frequency of the parallel resonators lower than the resonance frequency of the series resonators.

The Q-factor of an acoustic resonator is generally defined as the peak energy stored within one period of an applied RF signal divided by the total energy dissipated or lost within that period. The Q factor of an XBAR is a complex function of a number of parameters, including the length or number of fingers in the IDT of the XBAR.

Possible loss mechanisms in acoustic resonators include resistive losses in IDTs and other conductors; viscous or acoustic losses in piezoelectric plates, IDT fingers and other materials; and acoustic energy leaks out of the resonator structure. The peak energy stored in the resonator is proportional to the capacitance of the resonator. In an XBAR resonator, the capacitance is proportional to the number of IDT fingers. Resistive losses and adhesive losses are also proportional to the number of IDT fingers. The acoustic energy that leaks from the resonator in the lateral direction (i.e., the direction parallel to the IDT fingers) is proportional to the length of the resonator and thus proportional to the number of IDT fingers. In contrast, the energy loss of the IDT end in the longitudinal direction (i.e., the direction perpendicular to the IDT fingers) is approximately constant regardless of the number of IDT fingers. As the number of IDT fingers and the peak energy stored in the XBAR decrease, the longitudinally lost acoustic energy becomes an increasing fraction of the stored peak energy.

Fig. 3 is a graph of normalized Q factor of a representative XBAR as a function of the number of fingers in the IDT of the XBAR. The "normalized Q factor" is the Q factor of an XBAR with a finite number of IDT fingers divided by the Q factor of a hypothetical XBAR with the same structure and an infinite number of IDT fingers. In fig. 3, the normalized Q factor is quantized as a percentage of the Q factor of an XBAR with an infinite number of IDT fingers. Specifically, solid line 310 is a plot of the normalized Q factor at the resonant frequency, while dashed line 320 is a plot of the normalized Q factor at the anti-resonant frequency. The data in fig. 3 were obtained from simulations using the finite element method.

Fig. 3 shows that the normalized Q factor of an XBAR with a finite number of IDT fingers is less than 100%, that is, the Q factor of an XBAR with a finite number of IDT fingers is less than the Q factor of a similar XBAR with an infinite number of IDT fingers. Although not shown in fig. 3, the normalized Q factor of the XBAR may gradually approach 100% for a very large number of IDT fingers. As expected, the normalized Q factor depends on the number of IDT fingers. In particular, due to the increasing importance of acoustic energy loss in the longitudinal direction, the normalized Q factor drops sharply for XBARs with less than about 20 IDT fingers.

Fig. 4 is a plan view of an exemplary conductor pattern 400 that reduces acoustic energy leakage in the longitudinal direction of the XBAR tip. The conductor pattern 400 includes an IDT430 and four

reflector elements

462, 464, 466, 468. The IDT430 includes a first bus bar 432, a second bus bar 434, and a plurality of n interleaved IDT fingers extending alternately from the first and second bus bars. In this example, the number n of IDT fingers is equal to 24. In other XBARs, n may range from 20 to 100 or more IDT fingers. IDT finger 436 is the first finger and IDT finger 438 is the nth finger. IDT fingers numbered from left to right (as shown in FIG. 4) are arbitrary, and the names of the 1 st and nth fingers may be reversed.

As shown in fig. 4, odd IDT fingers extend from a first bus bar 432 and even IDT fingers extend from a second bus bar 434. The IDT430 has an even number of IDT fingers such that the first and nth'

IDT fingers

436, 438 extend from different bus bars. In some cases, an IDT may have an odd number of IDT fingers, so that the 1 st and nth' IDT fingers and all reflector elements extend from the same bus bar.

A total of four reflector elements are disposed outside the periphery of the IDT 430. The first reflector element 462 is near and parallel to the first IDT finger 436 at the left end of the IDT 430. The second reflector element 466 is adjacent to and parallel to the nth IDT finger 438 at the right end of the IDT 430. The optional third reflector element 464 is parallel to the first reflector element 462. An optional fourth reflector element 468 is parallel to the second reflector element 466.

The first and

third reflector elements

462, 464 extend from the first bus 432 and are therefore at the same potential as the first IDT finger 436. Similarly, second and

fourth reflector elements

466 and 468 extend from the second bus bar 430 and are therefore at the same potential as the nth' IDT finger 438.

The

reflector elements

462, 464, 466, 468 are configured to confine acoustic energy to the region of the IDT430, thereby reducing acoustic energy loss in the longitudinal direction. To this end, the spacing pr between adjacent reflector elements and between

reflector elements

462 and 466 and adjacent first and nth' IDT fingers, respectively, is generally greater than the spacing p of the IDT fingers. The width or index mr of the

reflector elements

462, 464, 466, 468 is not necessarily equal to the index m of the IDT finger. As will be described later, the index mr of the reflector element may be selected to optimize the Q factor at a particular frequency or range of frequencies.

Fig. 5 is a plan view of another conductor pattern 500, the conductor pattern 500 reducing leakage of acoustic energy in the longitudinal direction at the ends of the XBAR. The conductor pattern 500 includes an IDT530 and four

reflector elements

562, 564, 566, 568. The IDT530 includes a first bus bar 532, a second bus bar 534, and a plurality of interleaved IDT fingers extending alternately from the first and second bus bars as previously described.

IDT fingers

536 and 538 are the 1 st and nth' IDT fingers at the left and right ends of IDT530 (shown in FIG. 5).

A total of four reflector elements are disposed outside the periphery of the IDT 530. The first and

third reflector elements

562 and 564 are adjacent and parallel to the first IDT finger 536 at the left end of the IDT 530. The first and

third reflector elements

562, 564 are connected to each other but not to any of the bus bars 532, 534. The first and

third reflector elements

562, 564 are capacitively coupled to the first IDT finger 536 and are therefore at substantially the same potential as the first IDT finger 536. If the potential between the reflector element and the first IDT finger is less than the potential between adjacent IDT fingers when an RF signal is applied between the bus bars 532, 534, the reflector element is considered to be at substantially the same potential.

Similarly, the second and

fourth reflector elements

566 and 568 are adjacent and parallel to the nth' IDT finger 538 at the right end of the IDT 530. The second and

fourth reflector elements

566, 568 are connected to each other and not to each other or to either

bus bar

532, 534. The second and

fourth reflector elements

566, 568 are capacitively coupled to each other and to the nth 'IDT finger 538, and thus at approximately the same potential as the nth' IDT finger 538.

The

reflector elements

562, 564, 566, 568 are configured to confine acoustic energy to the region of the IDT530, thereby reducing the loss of acoustic energy in the longitudinal direction. To this end, the spacing pr between adjacent reflector elements and between

reflector elements

562 and 566 and adjacent terminal IDT fingers is generally greater than the spacing p of the IDT fingers.

The width or mark mr of the

reflector elements

562, 564, 566, 568 is not necessarily equal to the mark m of the IDT finger. The index mr of the reflector element may be selected to optimize the Q factor for a particular frequency of the frequency range.

Fig. 6 is a graph of normalized Q factor as a function of the number of IDT fingers for another XBAR with and without reflector elements similar to that shown in fig. 4. In particular, the solid line 610 is a plot of the normalized Q factor of an XBAR without reflector elements at its resonant frequency. Dashed curve 620 is a graph of normalized Q-factor at XBAR-like resonant frequencies with two reflector elements on each side of the IDT. In both cases, the piezoelectric plate was 400nm thick lithium niobate, the IDT fingers were 500nm thick aluminum, the IDT pitch p =4 microns, and the IDT finger mark m =1 micron. For XBAR with reflector elements, pr =4.2 microns, mr =0.735 microns. With the reflector element, only a 10 finger XBAR can have a normalized Q factor of up to 80%.

Fig. 7 is a graph of normalized Q factor as a function of IDT fingers for another XBAR with and without reflector elements similar to those shown in fig. 4. In particular, solid line 710 is a plot of the normalized Q factor of an XBAR without reflector elements at its anti-resonant frequency. Dashed curve 720 is a plot of normalized Q factor at anti-resonance frequency for a similar XBAR with two reflector elements on each side of the IDT. In both cases, the piezoelectric plate was 400nm thick lithium niobate, the IDT fingers were 500nm thick aluminum, the IDT pitch p =4 microns, and the IDT finger mark m =1 micron. For XBAR with reflector elements, pr =8 microns, mr =0.80 microns. With the reflector element, only the XBAR of 14 fingers can have a normalized Q factor of up to 80%.

Fig. 8 shows a graph 800 illustrating the relationship between the pitch pr of the reflector elements and the index mr of an exemplary XBAR device at a fixed frequency of 5150 MHz. An exemplary XBAR device has a lithium niobate piezoelectric plate with a thickness of 400nm and an aluminum IDT and reflector element with a thickness of 500 nm. The pitch and mark of the IDT fingers are 4 microns and 1 micron, respectively. There are two reflector elements at each end of the IDT. The lighter

shaded areas

810A, 810B, 810C, 810D identify a combination of pr and mr, where the normalized Q factor is greater than or equal to 85%. The darker

shaded regions

820A, 820B, 820C, 820D identify combinations of pr and mr where the normalized Q factor is greater than or equal to 90%. For comparison, the normalized Q factor of this XBAR without reflector element is 74% at 5150 MHz. Although not identified in fig. 8, there does exist a combination of pr and mr where the normalized Q factor is less than 75%, indicating that improperly configured reflector elements may reduce the XBAR Q factor.

Various combinations of pr and mr increase the normalized Q factor to 85% or 90%. To achieve a normalized Q factor greater than or equal to 90%, pr must be greater than or equal to 1.2 times the IDT finger pitch p. For pr =6 microns (1.5 p), at least four mr values improved the normalized Q factor to over 90%.

Fig. 9 shows a graph 900 illustrating the relationship between the index mr and frequency for the reflector elements of an exemplary XBAR device with two reflector elements on each side of the IDT and pr =5.2 microns. As with the previous example, the exemplary XBAR device has a lithium niobate piezoelectric plate with a thickness of 400nm and an aluminum IDT and reflector element with a thickness of 500 nm. The pitch and mark of the IDT fingers are 4 microns and 1 micron, respectively. The lighter shaded areas (e.g., area 910) identify a combination of frequency and mr where the normalized Q factor is greater than or equal to 85%. The darker shaded areas (e.g., area 920) identify a combination of frequency and mr where the normalized Q factor is greater than or equal to 90%. For comparison, the normalized Q factor of this XBAR without reflector element is 74% at 5150 MHz.

Graph 900 shows that for a particular reflector element spacing pr, reflector element markers mr must be selected taking into account the frequency at which the Q factor of the XBAR requires improvement. For example, selecting mr =0.95 microns may provide a normalized Q factor greater than 90% over a frequency range from about 4980MHz to greater than 5200 MHz. Selecting mr =1.7 microns may provide a normalized Q factor greater than 90% over a frequency range of less than 4700MHz to about 4950 MHz. However, choosing mr =1.7 microns may actually reduce the Q factor at 5200MHz compared to an XBAR without reflector elements.

Fig. 10 shows a graph 11000 illustrating the relationship between index mr and frequency for reflector elements of an exemplary XBAR device, where the XBAR device has one reflector element on each side of the IDT and pr =5.2 microns. The exemplary XBAR device is the same as the previous example. As shown in fig. 9, lighter shaded regions such as region 1010 identify a combination of frequency and mr where the normalized Q factor is greater than or equal to 85%. Darker shaded areas, such as area 1020, identify a combination of frequency and mr where the normalized Q factor is greater than or equal to 90%.

A comparison of fig. 9 and 10 shows that only one reflector element is generally not as effective as two reflector elements in improving the normalized Q factor. However, in some applications, it may be sufficient to have one reflector element at each end of the IDT. In this example, one mr =0.75 micron reflector element (at each end of the IDT) provides a significant improvement in the normalized Q factor for the frequency range of approximately 4770MHz to 4970 MHz.

Fig. 11 shows a graph 1100 illustrating the relationship between the index mr and frequency for the reflector elements of an exemplary XBAR device having five reflector elements at each end of the IDT, and pr =5.2 microns. The exemplary XBAR device is the same as the previous example. In fig. 9, a lighter shaded region, such as region 1110, identifies a combination of frequency and mr where the normalized Q factor is greater than or equal to 85%. Darker shaded areas, such as area 1120, identify a combination of frequency and mr where the normalized Q factor is greater than or equal to 90%.

Fig. 11 shows that five reflector elements do not provide any significant improvement compared to two reflector elements.

Figure 12 is a graph of performance of exemplary XBAR bandpass filters with and without reflector elements. Specifically, solid line 1210 is S for the band n77 filter ₂₁ A plot of the magnitude of the (input-output transfer function) with two reflector elements at each end of the XBAR in the filter. The reflector elements of the parallel resonators are optimized for a frequency of 3.35GHz and the reflector elements of the series resonators are optimized for a frequency of 4.2 GHz. These frequencies are at or near the edge of the n77 band, where it is generally most difficult to achieve the lowest S ₂₁ The requirements of (2). Dashed line 1220 is S for the same filter without reflector elements on the XBAR ₂₁ An amplitude map. All data were obtained by simulating the filters using the finite element method.

Comprising a reflector element for reflecting S at 3.35GHz ₂₁ The improvement is 0.2dB, and the improvement is 0.4dB under 4.2 GHz. However, note that the inclusion of reflector elements may be used to reflect S at other frequencies ₂₁ The reduction is as much as 0.25dB, indicating that trade-offs need to be made during the design of the XBAR filter. In the exemplary bandpass filter of fig. 12, the reflector elements of all parallel resonators are selected to obtain the maximum Q-factor at the same frequency (3.35 GHz). The reflector elements of all series resonators were chosen to obtain the maximum Q-factor at the same frequency (4.2 GHz). The filter transfer function may be further improved if the reflector elements of each resonator are optimized independently.

Fig. 13 is a plan view of another exemplary conductor pattern 1300 that reduces acoustic energy leakage in the longitudinal direction of the XBAR tip. The conductor pattern 1300 includes an IDT 1330 and first, second, third and

fourth reflector elements

1362, 1364, 1366, 1368. The IDT 1330 includes a first bus bar 1332, a second bus bar 1334, and a plurality of interleaved IDT fingers extending alternately from the first and second bus bars. In this example, the number of IDT fingers is equal to 24. In other XBARs, the number of fingers may range from 20 to 100 or more IDT fingers. The overlapping portions of the interleaved IDT fingers and the first, second, third and

fourth reflector elements

1362, 1364, 1366, 1368 are on the membranes of the XBAR. The IDT finger 1336 is the first finger and the IDT finger 1338 is the last finger. The names of the first and last IDT fingers are arbitrary and the names of the first and last fingers may be reversed.

A total of four reflector elements are provided outside the periphery of IDT 1330. The first and

second reflector elements

1362, 1364 are adjacent to and parallel to the first IDT finger 1336 at the left end of the IDT (as shown). The third and

fourth reflector elements

1366, 1368 are adjacent to and parallel to the last IDT finger 1338 at the right end (as shown) of the IDT 1330.

The first and

second reflector elements

1362, 1364 extend from different bus bars than the first IDT fingers. In this example, the first and

second reflector elements

1362, 1364 extend from the second bus bar 1334 and the first IDT finger 1336 extends from the first bus bar 1332. Thus, the electrical potential on the first and

second reflector elements

1362, 1364 is opposite the electrical potential on the first IDT finger 1336. Similarly, the third and

fourth reflector elements

1366, 1368 extend different bus bars than the last IDT finger 1338. Thus, the potential on the third and

fourth reflector elements

1366, 1368 is opposite the potential on the last IDT finger 1338.

The

reflector elements

1362, 1364, 1366, 1368 are configured to limit acoustic energy to the area of the IDT 1330 and thus reduce acoustic energy loss in the longitudinal direction. To this end, the center-to-center distance pr1 between adjacent reflector elements (i.e., between the first and second reflector elements and between the third and four reflector elements) is typically greater than or equal to 1.2 times the pitch p of the IDT fingers and less than or equal to 1.5 times p.

The center-to-center distance pr2 between a reflector element and an adjacent IDT finger (i.e., between the first reflector element and the first IDT finger, and between the third reflector element and the last IDT finger) is typically greater than p and less than pr1. In some cases, pr2 may be the average of p and pr1.

The width or mark mr of the

reflector elements

1362, 1364, 1366, 1368 is not necessarily equal to the mark m of the IDT finger. As previously mentioned, the index mr of the reflector element may be selected to optimize the Q factor for a particular frequency or range of frequencies.

Fig. 14A is a plot 1400 of S21 (input/output transfer function) amplitude versus frequency for a first pair of XBAR filters with and without reflector fingers. These filters are intended to pass the 5G NR band n79. A frequency range of the frequency band n79 is identified. The solid line 1410 is the magnitude of S21 for a filter with reflective fingers on all series resonators. For each series resonator, pr1=1.3p and pr2=1.15p. mr is optimized independently for each resonator to minimize loss at the upper band edge (5.0 GHz). The dashed line 1420 is the magnitude of S21 for the same filter without a reflective finger on any resonator. The presence of the reflective fingers on the series resonators (solid line 1410) reduces the loss at the edge of the upper band (i.e., increases S21) by about 0.4dB compared to a filter without reflective fingers (dashed curve 1420).

Fig. 14B is a graph 1450 of the S21 (input/output transfer function) amplitude as a function of frequency for a second pair of XBAR filters with and without reflective fingers. The design of these filters is independent of the design of the first pair of filters to pass the 5G NR band n79. A frequency range of the frequency band n79 is identified.

The solid line 1460 is the magnitude of S21 for a filter having reflective fingers on all series resonators. For each series resonator, pr1=1.3p and pr2=1.15p. mr is optimized independently for each resonator to minimize loss at the upper band edge (5.0 GHz). The dashed curve 1470 is the magnitude of S21 for the same filter without a reflective finger on any resonator. The presence of the reflective fingers on the series resonators (solid line 1460) reduces the loss at the edge of the upper band (i.e., increases S21) by about 0.3dB compared to a filter without reflective fingers (dashed curve 1470).

Description of the method

Fig. 15 is a simplified flow diagram summarizing a process 1500 for fabricating a filter device incorporating XBARs. In particular, process 1500 is used to fabricate a filter device including a plurality of XBARs, some of which may include a frequency setting dielectric layer. The process 1500 begins at 1505 where a device substrate and a sheet of piezoelectric material are disposed on a sacrificial substrate. The process 1500 ends at 1595 when the filter device is completed. The flow chart of fig. 15 includes only the main process steps. Various conventional process steps (e.g., surface preparation, cleaning, inspection, baking, annealing, monitoring, testing, etc.) may be performed before, during, after, and during the steps shown in fig. 15.

Although fig. 15 generally describes a process for fabricating a single filter device, multiple filter devices may be fabricated simultaneously on a common wafer (consisting of piezoelectric plates bonded to a substrate). In this case, each step of process 1500 may be performed simultaneously on all filter devices on the wafer.

The flow chart of fig. 15 captures three variations of the process 1500 for fabricating XBARs that differ in when and how the cavities are formed in the device substrate. A cavity may be formed at step 1510A, 1510B, or 1510C. In each of the three variations of the process 1500, only one of the steps is performed.

The piezoelectric plate may be, for example, lithium niobate or lithium tantalate, any of which may be Z-cut, rotary Z-cut, or rotary YX-cut. The piezoelectric plate may be some other material and/or some other cutout. The device substrate may preferably be silicon. The device substrate may be some other material that allows for the formation of deep cavities by etching or other processes.

In one variation of process 1500, one or more cavities are formed in a device substrate at 1510A, and then a piezoelectric plate is bonded to the substrate at 1515. A separate cavity may be formed for each resonator in the filter arrangement. The one or more cavities may be formed using conventional photolithography and etching techniques. Typically, the cavity formed at 1510A does not penetrate the device substrate.

At 1515, the piezoelectric plate is bonded to a device substrate. The piezoelectric plate and the device substrate may be bonded by a wafer bonding process. Typically, the mating surfaces of the device substrate and the piezoelectric plate are highly polished. One or more layers of an intermediate material, such as an oxide or a metal, may be formed or deposited on the mating surfaces of the piezoelectric plate and/or the device substrate. One and/or both mating surfaces may be activated using, for example, a plasma process. The mating surfaces may then be pressed together with considerable force to establish molecular bonds between the piezoelectric plate and the device substrate or intermediate material layer.

At 1520, the sacrificial substrate can be removed. For example, the piezoelectric plate and the sacrificial substrate may be wafers of piezoelectric material that have been ion implanted to create defects in the crystal structure along a plane that defines a boundary that will become between the piezoelectric plate and the sacrificial substrate. At 1520, the wafer can be split along the defect plane, e.g., by thermal shock, separating the sacrificial substrate and leaving the piezoelectric plate bonded to the device substrate. After the sacrificial substrate is stripped, the exposed surface of the piezoelectric plate may be polished or otherwise treated.

Sheets of single crystal piezoelectric material laminated to a non-piezoelectric substrate are commercially available. During application, both the lithium niobate and lithium tantalate sheets may be bonded to a variety of substrates, including silicon, quartz, and fused silica. Sheets of other piezoelectric materials may be present or may come into existence in the future. The thickness of the piezoelectric plate may be between 300nm and 1000 nm. When the substrate is silicon, a layer of SiO may be provided between the piezoelectric plate and the substrate ₂ .

Steps

1510A, 1515, and 1520 of process 1500 are not performed when using commercially available piezoelectric plate/device substrate laminates.

At 1545, a first conductor pattern is formed by depositing and patterning one or more conductor layers on the front surface of the piezoelectric plate, including the IDT and reflector elements of each XBAR. The conductor layer may be, for example, aluminum, an aluminum alloy, copper, a copper alloy, or some other conductive metal. Optionally, one or more layers of other materials may be disposed below the conductor layer (i.e., between the conductor layer and the piezoelectric plate) and/or above the conductor layer. For example, a thin film of titanium, chromium, or other metal may be used to improve adhesion between the conductor layer and the piezoelectric plate. A second conductor pattern of gold, aluminum, copper, or other higher conductivity metal may be formed over portions of the first conductor pattern (e.g., the interconnects between the IDT bus bars and the IDTs).

Each conductor pattern may be formed 1545 by sequentially depositing a conductive layer and optionally one or more other metal layers on the surface of the piezoelectric plate. Excess metal may then be removed by etching through the patterned photoresist. The conductor layer may be etched, for example, by plasma etching, reactive ion etching, wet chemical etching, or other etching techniques.

Alternatively, each conductor pattern may be formed at 1545 using a lift-off process. A photoresist may be deposited on the piezoelectric plate and patterned to define a conductor pattern. A conductor layer, and optionally one or more further layers, may be deposited in sequence on the surface of the piezoelectric plate. The photoresist may then be removed, which removes excess material, leaving behind the conductor pattern.

At 1550, one or more frequency setting dielectric layers can be formed by depositing one or more layers of dielectric material on the front surface of the piezoelectric plate. For example, a dielectric layer may be formed on the parallel resonators to lower the frequency of the parallel resonators relative to the series resonators. The one or more dielectric layers may be deposited using conventional deposition techniques, such as physical vapor deposition, atomic layer deposition, chemical vapor deposition, or some other method. One or more photolithographic processes (using a photomask) may be used to limit the deposition of the dielectric layer to selected areas of the piezoelectric plate. For example, a mask may also be used to restrict the dielectric layer to cover only the parallel resonators.

At 1555, a passivation/tuning dielectric layer is deposited over the piezoelectric plate and conductor pattern. The passivation/tuning dielectric layer may cover the entire surface of the filter except for pads for electrical connection to circuitry external to the filter. In some examples of process 1500, the passivation/tuning dielectric layer can be formed after the cavity in the device substrate is etched at 1510B or 810C.

In a second variation of the process 1500, one or more cavities are formed in the back side of the device substrate at 1510B. A separate cavity may be formed for each resonator in the filter arrangement. One or more cavities may be formed using anisotropic or orientation-dependent dry or wet etching to open up from the back side of the device substrate to the piezoelectric plate. In this case the resulting resonator device will have a cross-section as shown in fig. 1.

In a third variation of the process 1500, one or more cavities in the form of grooves may be formed in the device substrate by etching the substrate using an etchant introduced through openings in the piezoelectric plate at 1510C. A separate cavity may be formed for each resonator in the filter arrangement. The cavity or cavities formed at 1510C will not penetrate the device substrate.

Ideally, most or all of the filter devices on the wafer will meet a set of performance requirements after the 1510B or 1510C cavities are formed. However, normal process tolerances can result in variations in parameters such as thickness variations of the dielectric layers formed at 1550 and 1555, thickness and line width variations of the conductor and IDT fingers formed at 1545, and piezoelectric plate thickness variations. These variations result in deviations of the filter device performance from the performance requirements.

To improve the yield of filter devices meeting performance requirements, frequency tuning can be performed by selectively adjusting the thickness of the passivation/tuning layer deposited on the resonator at 1555. The frequency of the passband of the filter arrangement may be reduced by adding material to the passivation/tuning layer and increased by removing material from the passivation/tuning layer. In general, the process 1500 is biased to produce a filter device having a passband that is initially below the desired frequency range but that can be tuned to the desired frequency range by removing material from the surface of the passivation/tuning layer.

At 1560, a probe card or other device may be used to make electrical connections with the filter to allow Radio Frequency (RF) testing and measurement of filter characteristics such as input-output transfer function. Typically, the RF measurements are made on some or most of the filter devices that are fabricated on the same piezoelectric plate and substrate at the same time.

At 1565, global frequency tuning may be performed by removing material from the surface of the passivation/tuning layer using a selective material removal tool, such as a scanning ion mill as previously described. "global" tuning is performed with a spatial resolution equal to or greater than a single filter means. The purpose of global tuning is to move the pass band of each filter means towards the desired frequency range. The test results obtained in 1560 may be processed to generate a global contour map that indicates the amount of material removed according to the two-dimensional location on the wafer. Material is then removed from the contour map using a selective material removal tool.

At 1570, local frequency tuning may be performed in addition to or in lieu of the global frequency tuning performed at 1565. The "local" frequency tuning is performed with a spatial resolution smaller than the individual filter means. The test results obtained in 1560 may be processed to generate a map indicating the amount of material to be removed at each filter arrangement. Local frequency tuning may require the use of a mask to limit the size of the area of material to be removed. For example, a first mask may be used to limit tuning to only parallel resonators, while a second mask may be subsequently used to limit tuning to only series resonators (or vice versa). This will allow independent tuning of the lower band edge (by tuning the parallel resonators) and the upper band edge (by tuning the series resonators) of the filter arrangement.

After frequency tuning is completed at 1565 and/or 1570, the filter arrangement is completed at 1575. Actions that may occur at 1575 include forming bond pads or solder bumps or other means for establishing a connection between the device and external circuitry (if such pads are not formed at 1545); cutting individual filter devices from a wafer containing a plurality of filter devices; other packaging steps; and additional testing. After each filter device is completed, the process ends at 1595.

Concluding sentence

Throughout the specification, the illustrated embodiments and examples should be considered as examples, and not as limitations on the disclosed or claimed devices and processes. Although many of the examples provided herein involve specific combinations of method acts or system elements, it should be understood that those acts and those elements may be combined in other ways to accomplish the same objectives. With regard to flow diagrams, additional steps and fewer steps may be taken, and the illustrated steps may be combined or further refined to implement the methods described herein. Acts, elements and features discussed only in connection with one embodiment are not intended to be excluded from a similar role in other embodiments.

As used herein, "plurality" refers to two or more. As used herein, a "set" of items may include one or more of such items. As used herein, the terms "comprising," "including," "carrying," "having," "containing," "involving," and the like, whether in the written detailed description or in the claims, are to be construed as open-ended, i.e., to mean including, but not limited to. With respect to the claims, only the transition phrases "consisting of" and "consisting essentially of" 823030 ", are closed or semi-closed transition phrases. Ordinal terms such as "first," "second," "third," etc., used in the claims are used to modify a claim element and do not by itself connote any priority, precedence, or order of one claim element over another or the order in which acts of a method are performed, but are used merely as labels to distinguish one claim element having a same name from another element having a same name (but for use of the ordinal term) to distinguish the claim elements. As used herein, "and/or" means that the listed items are alternatives, but alternatives also include any combination of the listed items.

Claims

1. An acoustic resonator device comprising:

a diaphragm comprising a portion of the piezoelectric plate spanning a cavity in the substrate; and

a conductor pattern formed on a surface of the piezoelectric plate, the conductor pattern comprising:

an interdigital transducer (IDT) comprising a first bus bar, a second bus bar, and a plurality of interleaved fingers, wherein the plurality of interleaved fingers extend alternately from the first and second bus bars, portions of the plurality of interleaved fingers overlapping the diaphragm;

first and second reflector elements on the diaphragm proximate and parallel to a first finger of the plurality of interleaved fingers; and

third and fourth reflector elements located on the membrane proximate and parallel to a last finger of the plurality of interleaved fingers.

2. The apparatus of claim 1,

the first finger extends from the first busbar,

the first and second reflector elements extend from the second busbar,

the last finger extends from either of the first and second bus bars and the third and fourth reflector elements extend from the other of the first and second bus bars.

3. The apparatus of claim 1,

pr1 is the center-to-center distance of the first and second reflector elements and the center-to-center distance of the third and fourth reflector elements,

p is the pitch of the plurality of interleaved fingers, and

1.2p≤pr1<≤1.5p。

4. the apparatus of claim 3,

pr2 is the center-to-center distance of the first reflector element and the first finger and the center-to-center distance of the third reflector element and the last finger, and p ≦ pr2 ≦ pr1.

5. The apparatus of claim 4,

pr2＝(pr1+p)/2。

6. the apparatus of claim 1,

the indicia mr of the first, second, third and fourth reflector elements are configured to increase the Q factor of the device at a predetermined frequency.

7. The apparatus of claim 6,

the device is a series resonator in a ladder bandpass filter circuit having a passband, and mr is selected to increase the Q factor of the device at the upper edge of the passband.

8. A band pass filter comprising:

a plurality of acoustic resonators including one or more series resonators and one or more parallel resonators connected in a ladder filter circuit, wherein

Each of the plurality of acoustic resonators includes:

a corresponding diaphragm comprising a portion of the piezoelectric plate spanning a corresponding cavity in the substrate, and

an interdigital transducer (IDT) formed on a surface of the piezoelectric plate, the IDT including a first bus bar, a second bus bar, and a plurality of interleaved fingers extending alternately from the first and second bus bars, portions of the plurality of interleaved fingers overlapping the corresponding diaphragm, wherein

At least one of the one or more series resonators further comprises:

first and second reflector elements located on the respective membranes proximate and parallel to a first finger of the plurality of interleaved fingers; and

third and fourth reflector elements located on the respective membranes proximate and parallel to a last finger of the plurality of interleaved fingers.

9. The filter of claim 8, wherein for each series resonator comprising first through fourth reflector elements:

the first finger extends from the first busbar,

the first and second reflector elements extend from the second busbar,

10. The filter of claim 8, wherein for each series resonator comprising first through fourth reflector elements:

p is the pitch of the interleaved fingers, an

1.2p≤pr1≤1.5p。

11. The filter of claim 10,

pr2 is the center-to-center distance of the first reflector element and the first finger and the center-to-center distance of the third reflector element and the last finger, and p < ≦ pr2< ≦ pr1.

12. The filter of claim 11,

pr2＝(pr1+p)/2。

13. the filter of claim 8, wherein for each series resonator comprising first through fourth reflector elements:

the markers mr of the first, second, third and fourth reflector elements are configured to improve the input/output transfer function of the filter at a predetermined frequency.

14. The filter of claim 13,

mr is selected to improve the input/output transfer function of the filter at the upper edge of the passband.

15. The filter of claim 8, wherein all of the one or more series resonators include respective first through fourth reflector elements.