CN111613862A

CN111613862A - Substrate design for efficient coupling between package and dielectric waveguide

Info

Publication number: CN111613862A
Application number: CN202010107096.7A
Authority: CN
Inventors: H·O·艾利; J·A·赫本萨梅尔; B·S·库克; V·古普塔; A·林; S·桑卡兰
Original assignee: Texas Instruments Inc
Current assignee: Texas Instruments Inc
Priority date: 2019-02-22
Filing date: 2020-02-21
Publication date: 2020-09-01
Anticipated expiration: 2040-02-21
Also published as: CN111613862B

Abstract

A device includes a multilayer substrate (104) having a first surface and a second surface opposite the first surface. An integrated circuit (102) is mounted on the second surface of the multilayer substrate, the integrated circuit having a transmission circuit configured to process millimeter wave signals. A substrate waveguide (170) having substantially solid walls is formed within a portion of the multilayer substrate perpendicular to the first surface. The substrate waveguide has a first end with walls having edges (118TX) exposed on a first surface of the multilayer substrate. A reflector (174) is located in one of the layers of the substrate and is coupled to an edge of the wall on an opposite end of the substrate waveguide.

Description

Substrate design for efficient coupling between package and dielectric waveguide

RELATED APPLICATIONS

This application claims priority from U.S. provisional application No. 62/809,051 entitled "Substrate Design to enable efficient Coupling Between packages and electrolytic Waveguide" filed on 2019, 2/22, which is incorporated herein by reference.

Technical Field

The present invention relates to a substrate design that enables coupling between a package for millimeter wave applications and a dielectric waveguide.

Background

In electromagnetic communications engineering, the term waveguide may refer to any linear structure that transmits electromagnetic waves between its end points. The most primitive and common means hollow metal tubes used to carry radio waves. In devices such as microwave ovens, radar, satellite communication and microwave radio links, this type of waveguide acts as a transmission line for connecting microwave transmitters and receivers to their antennas.

Dielectric Waveguides (DWGs) are a high frequency alternative to copper wire and fiber optic cables. Dielectric waveguides employ a solid dielectric core rather than a hollow tube. A dielectric is an electrical insulator that can be polarized by an applied electric field. When dielectrics are placed in an electric field, the charges do not flow through the material as they do in a conductor, but only slightly deviate from their average equilibrium position, thereby causing dielectric polarization. Due to dielectric polarization, positive charges are shifted in the direction of the electric field, while negative charges are shifted in the opposite direction. This creates an internal electric field that reduces the overall electric field within the dielectric itself. If the dielectric is composed of weakly bonded molecules, these molecules not only become polarized, but also reorient so that their axes of symmetry align with the electric field. While the term "insulator" means low conductivity, "dielectric" is often used to describe materials with high polarizability; the polarizability is expressed by a number called the dielectric constant (k). The term "insulator" is commonly used to denote resistance, while the term "dielectric" is used to denote the energy storage capability of a material achieved by polarization.

Disclosure of Invention

In described examples, a device includes a multilayer substrate having a first surface and a second surface opposite the first surface. An integrated circuit is mounted on the second surface of the multilayer substrate, the integrated circuit having a transmission circuit configured to process millimeter wave signals. A substrate waveguide having substantially solid walls is formed within a portion of the multilayer substrate perpendicular to the first surface of the substrate. The substrate waveguide has a first end with walls having edges exposed on the first surface of the multilayer substrate. The reflector is located in one of the layers of the substrate and is coupled to an edge of the wall on the opposite end of the substrate waveguide.

Drawings

Fig. 1 is an exploded partial view of a Dielectric Waveguide (DWG) communication system.

Fig. 2 is a plan view of a multilayer substrate used in the system of fig. 1.

Fig. 3 is a cross-sectional view of portions of the system of fig. 1.

Fig. 4 is a schematic diagram of a portion of the DWG communication system of fig. 1.

Figure 5 is a perspective view of a portion of another example substrate of a DWG communication system.

Fig. 6A, 6B are side and perspective views, respectively, of a portion of another example substrate for a DWG communication system.

Figure 7 is a top view of a portion of another example substrate of a DWG communication system.

Fig. 8 is a cross-sectional view of a portion of an example system that includes an interposer positioned between a radiating element of a microelectronic device and a DWG.

Fig. 9 is a flowchart illustrating the operation of the dielectric waveguide system.

Detailed Description

In the drawings, like elements are represented by like reference numerals for consistency.

Waves in open space propagate in all directions in the form of spherical waves. Thus, their power loss is proportional to the square of the distance; that is, at a distance R from the wave source, the power is the wave source power divided by R². Dielectric Waveguide (DWG) usableHigh frequency signals are transmitted over relatively long distances. The waveguide confines the wave to propagating in one dimension so that, under ideal conditions, the wave does not lose power while propagating. The propagation of an electromagnetic wave along the axis of a waveguide is described by wave equations, derived from maxwell's equations, where the wavelength depends on the structure of the waveguide, the material inside it (air, plastic, vacuum, etc.), and the frequency of the wave. One common type of waveguide is rectangular in cross-section, typically not square. The long side of such a cross-section is typically twice as long as the short side. They are useful for carrying horizontally or vertically polarized electromagnetic waves. Another common type of waveguide is circular. Circular waveguides are useful for transmitting circularly polarized electromagnetic waves. Circular dielectric waveguides are readily fabricated using known or later developed techniques.

Common problems that may arise when coupling a DWG to a radiating element include: poor isolation between the transmitting antenna and the receiving antenna in the same microelectronic device; misregistration between the radiating elements and the interconnects; and sub-optimal impedance matching between the antenna and the dielectric waveguide. The underlying reason is the lack of well-defined electrical and mechanical interfaces between the radiating elements on the microelectronic device and the DWG interconnects.

The examples described below improve the interface between the electromagnetic radiation elements on the microelectronic device and the DWG interconnects. An interpolator acting as a buffer can be used to establish two well-defined reference planes that can be optimized independently. The first plane is located between the radiating elements and the interposer, and the second plane is the surface between the interposer and the DWG interconnect. The interposer allows the introduction of features that can improve isolation between the transmitter and receiver antennas in the device, relax alignment tolerances, and enhance impedance matching between the antennas and the dielectric waveguide. An interposer is a piece of material used to join together an antenna in a substrate and a DWG connector. The interposer has defined regions aligned with the antennas and acting as waveguides to conduct signals from the radiating elements on the microelectronic device substrate to the DWG connectors. Several exemplary interpolators are described in more detail in U.S. patent publication 2019-0109362 (entitled "Interporer Between Microelectronic Package Substrate and ElectrocWaveguide Connector for mm-Wave Application"), which is incorporated herein by reference. As will be described in greater detail below, a waveguide integrated within a substrate of a microelectronic device directs backscattered radiation from a radiating element to an interposer mounted on the substrate.

In this way, a structure is formed in the substrate on which the millimeter wave transceiver is mounted, which functions as a substrate-integrated waveguide, as an interface to a waveguide-based transmission medium, as a placement guide for such a transmission medium, and as a means for improving isolation between channels.

Fig. 1 is an exploded partial view of a Dielectric Waveguide (DWG) communication system 100. The system 100 includes an Integrated Circuit (IC) die 102, a multilayer substrate 104, and Transmit (TX) and Receive (RX) DWG cable assemblies 106-1 and 106-2 with physical and electrical couplings between the devices. The IC die 102 is physically attached to a first side (e.g., bottom) of the multi-layer substrate 104, such as by a die mounting technique. The TX cable assembly 106-1 and the RX cable assembly 106-2 are physically attached to the side (e.g., top) of the multilayer substrate 104 opposite the IC die 102 by mechanisms such as screws, clips, brackets, and the like. The TX cable assembly 106-1 is axially aligned with respect to a transmit zone 108TX on the substrate 104 and the RX cable assembly 106-2 is axially aligned with respect to a receive zone 108RX on the substrate 104. With such alignment and further attachment of substrate 104 to IC die 102, millimeter electromagnetic waves may be communicated between cable assemblies 106-1 and 106-2 to and from IC die 102, and such waves are communicated through antennas (and respective feed structures) and waveguides configured in multilayer substrate 104.

The IC die 102 has the attributes and dimensions of integrated circuit technology. In this example, the IC die 102 is a 3mm by 3mm square, while the substrate 104 is approximately 6mm by 8 mm. IC die 102 includes a transceiver, shown generally as transceiver 102TXRX (connections not separately shown), that is configured to transmit and receive signals. The operating frequency and bandwidth of the transceiver signals may be selected according to the application, for example, in the millimeter wave range (e.g., 110GHz to 140GHz) for communication along the DWG medium. Although not separately shown, transceiver 102TXRX may include one or more processors (e.g., digital signal processors) and support programming of multiple transmit and receive channels, radio configurations, control, calibration, and mode changes to implement various embodiments. IC die 102 includes a plurality of conductive members, such as die pads, which are shown generally in groups 110-G1 and 110-G2. The conductive regions 110-G1 and 110-G2 (e.g., die pads) are physically positioned in alignment with the conductive regions (e.g., underlying pads) substrate transmit and receive regions 108TX and 108RX, respectively, so that conductors (e.g., copper pillars or other raised structures) may be electrically coupled between opposing ones of these regions. Thus, when the IC die 102 is physically fixed relative to the multilayer substrate 104, it is also electrically coupled to the electrical paths of the substrate 104, including the traces 104TR, which are used for electrical connection between the IC die 102 and the regions 108TX and 108 RX. Through this electrical coupling, millimeter-wave signals may be communicated between IC die 102 and an antenna built into multilayer substrate 104, as described below.

Dielectric waveguide cable assemblies 106-1 and 106-2 each include a respective dielectric cable 112-1 and 112-2, such as a cylindrical outer cladding concentrically surrounding a respective cylindrical inner core 113-1 and 113-2, although other cable configurations including dielectrics are possible. The outer cladding has a dielectric constant (OC) and the inner core has a dielectric constant (IC). In this example, the outer cladding may have a diameter in the range of 3mm to 6mm, and the inner core may have a diameter in the range of 1.5mm to 2 mm. Preferably, the inner core dielectric constant (IC) is sufficiently greater than the outer cladding constant (OC) so that millimeter waves can be coupled to the cable and their energy will be concentrated in the inner core. In this way, even if the cable material is dielectric in nature (which is generally insulating in nature), the dielectric will allow energy to propagate along it without the insulator transferring charge. Thus, the exemplary embodiment allows for the use of a relatively inexpensive dielectric material, such as polyethylene, for the cable core, and may effectively transmit millimeter-wave signals from transceiver 102TXRX to cable 112-1 and thus to devices (not shown) at the distal end of the cable when connected to system 100.

Each of the cable assemblies 106-1 and 106-2 may be paired with a respective interposer 114-1 and 114-2. In the exemplary embodiment, each interposer waveguide 114-x has a cylindrical shape with a height of, for example, 1.5mm and a central axis aligned with the central axis of core 113-x, where both axes are aligned with one of regions 108Tx and 108Rx, respectively. In this example, each interposer waveguide 114-1, 142-2 has a metallic outer wall. The interior of each interposer waveguide 114-1, 114-2 is hollow and therefore typically filled with air, although some examples may be filled with a solid dielectric material. The outer diameter of cable 113-x may vary and, as shown in the illustrative example, may exceed the outer diameter of the corresponding interposer 114-x, while the outer diameter of core 113-x is less than the inner diameter of the corresponding interposer 114-x. With such dimensions and axial alignment, the circular inner diameter of each interposer 114-x is physically aligned with a respective transmit zone 108Tx or receive zone 108Rx, and the cylindrical shape provides a circular cross-section for physical and wave coupling with transmit zone 108Tx or receive zone 108 Rx. Moreover, in the exemplary embodiment, wave communication in system 100 is via circular polarization, and thus the circular cross-section of each interposer 114-x facilitates guiding circularly polarized waves between dielectric cable core 113-x and multilayer substrate 104.

Fig. 2 is a plan view of a multilayer substrate 104 with an IC die 102, represented in this view by a dashed rectangle, positioned below the substrate 104. In this plan view, each of the transmit region 108TX and receive region 108RX of the substrate 104 is further illustrated, and example features in the top metal layer of the substrate 104 are illustrated. For example, in the transmit zone 108TX, just within the perimeter of the region 108TX, a plurality of via waveguide tops 118TX are positioned. Similarly, in the receive region 108RX, just within the diameter of the region 108RX, a plurality of via waveguide tops 118RX are positioned.

Each via waveguide top 118 is a metal pad having the same shape (e.g., circular) in the top metal layer of the substrate 104. Further, the via waveguide tops 118 are equally spaced circumferentially, wherein the number of tops 118 may be selected according to a particular implementation. The via waveguide top and lower waveguide vias are positioned to form a substantially solid circular wall within the substrate 104 that is perpendicular to the surface of the substrate 104. The resulting circular wall structure acts as a waveguide and is referred to herein as a "substrate integrated waveguide" (SIW) 170. Typically, each top will contact an adjacent top. The entirety of the via waveguide vias in regions 108TX or 108RX, together, provide a tapered waveguide to couple to interposers 114-1, 114-2, respectively, as described below. As used herein, the term "substantially solid" means that the distance between adjacent wave vias is less than the diameter of these vias. Thus, the distance between adjacent vias is less than the wavelength of the millimeter signals transmitted and received by the IC 102. In some examples, the diameter of the vias is substantially equal to or less than the pitch of the vias (pitch) such that the walls of adjacent vias contact each other. In the depicted example, the central axis of the SIW is perpendicular to the surface of the substrate 104. "perpendicular" includes a range of plus or minus 20 degrees from 90 degrees from the surface of the substrate 104.

Inside the circle presented by the via top 118TX or 118RX, four end fed antennas are placed. For example, the transmit block 108TX has four transmit antennas 120TX1A, 120TX1B, 120TX2A, and 120TX2B, each having comparable and generally rectangular (rounded end) members that are physically positioned and preferably aligned 90 degrees apart from each other. Similarly, receive zone 108RX has four receive antennas 120RX1A, 120RX1B, 120RX2A, and 120RX2B, each also physically located preferably 90 degrees apart from each other. As described below, for a set of antennas (transmit antennas or receive antennas), a coupler (e.g., a branch line or other orthogonal coupler within the substrate 104) couples differential signals between the IC die 102 and the feed structure to these antennas so that circularly polarized signals are transmitted by or received from the antennas. The signal is further guided by a waveguide including a waveguide via top 118 so that the signal is transmitted or received in a direction (out of the page) generally perpendicular to the plane of fig. 2. Thus, signal energy is efficiently coupled between the substrate 104 and the corresponding cable assembly 106-x.

The diameter (D)171 of the emission SIW170 and similarly the diameter of the receiver SIW depend to a large extent on the wavelength (in the substrate) at the operating center frequency, but the actual values thereof are determined based on a number of considerations. The first consideration is to ensure single mode operation over the frequency band of interest. The higher order modes increase signal loss and therefore need to be limited. Another consideration is to ensure a proper low frequency cutoff for the mode of interest so that the waveguide can support the propagation of waves within the frequency band of interest. Another consideration is to ensure a linear group delay variation in the frequency band of interest such that the SIV operates in the linear region of the dispersion curve of the propagation constant versus frequency. Another consideration is to ensure that performance is optimal under all substrate manufacturing and system assembly tolerances. For example, a wider diameter may be needed to cover the XY placement uncertainty of the interposer. Although this example illustrates a circular SIW, the same considerations apply to the design of other shapes SIVs, such as a rectangular SIW for a linearly polarized signal.

Fig. 3 is a cross-sectional view of portions of system 100 taken along line 1C-1C across region 108TX of fig. 2. If the 108RX region is truncated, a similar view will also be presented. In general, fig. 3 shows the substrate 104 physically and electrically connected to a circuit substrate 122 (e.g., a Printed Circuit Board (PCB)). For example, the electrical connection may be made using Ball Grid Array (BGA) balls 124 connected to pads 126 on PCB 122. As described above, the IC die 102 is electrically and physically connected beneath the substrate 104. Such electrical connections may be made electrically by die bumps, such as copper pillars 128, along the bottom of the substrate 104 between the IC conductive regions 110 and the contacts (see fig. 1), and may be made physically by various die bonding techniques, such as thermal compression. In addition, an underfill 104UF is placed between the die 102 and the substrate 104. The dashed vertical lines in the upper left and upper right corners of FIG. 3 indicate the desired position of interposer 114-1, which interposer 114-1 may be fixed relative to the upper surface of substrate 104, surrounding transmission region 108 TX. The conductive outer surface of interposer 114-1 is in direct contact with via top 118TX, or there is a slight air gap (e.g., 0.1mm) between the two structures.

The cross-sectional profile of the multi-layer substrate 104 is generally consistent with the evolving technology of substrate packaging. In the example shown, the substrate 104 includes six metal layers, labeled in ascending order from the bottom of the substrate 104 up as layers L1 through L6. Further, for example, each of the metal layers L1-L6 may have the same thickness (e.g., 15 μm), although the metal layer thicknesses may vary, it is desirable (for thermal expansion matching) that each layer be the same distance from the central core 104C (layers L3 and L4 have the same thickness; layers L2 and L5 have the same thickness; layers L1 and L6 have the same thickness). Between successive metal layers is a non-conductive material, not referred to herein as a layer, but is also layered (structurally) between the metal layers. For example, there is a central core 104C between metal layers L3 and L4 that is thicker than the non-conductive material between the other metal layers. Also, for example, the central core 104C may be 200 μm thick, while the non-conductive material between the other metal layers (commonly referred to as stack 104BU) may be 30 μm thick. Stack 104BU is desirably a material with low loss and each layer of the material has the same or similar dielectric constant. Solder masks 104SM1 and 104SM2 are located below metal layer L1 and above metal layer L6, respectively.

Fig. 3 also shows additional structures (in substrate 104) that form a portion of SIW waveguide 170 that terminates at via top 118TX in fig. 2. Specifically, the via 130 is formed through the central core 104C, for example by forming a cylindrical void through the core 104C, which is then filled and/or plated with a conductive material (e.g., a metal), wherein the void has a cross-sectional diameter of 90 μm in this example. Vias 130 also provide electrical contact to the metal in layer L3. Over via 130, metal layer L4 is patterned to form pad 132, e.g., having a circular perimeter and a diameter of 130 μm. Build-up 104BU is formed over pads 132 (and over other portions of metal layer L4), and vias 134 are formed in the build-up (e.g., cylindrical voids are formed through the build-up over layer L4, and filled or plated with a conductive material). The via 134 may have a cross-sectional diameter of 60 μm and it provides electrical contact to the pad 132. Similarly, over via 134, metal layer L5 is patterned to form pad 136 that is similarly shaped but smaller in diameter (e.g., equal to 100 μm) than pad 132. Build-up 104BU is formed on pads 136 (and over other portions of metal layer L5), and vias 138 are formed in build-up 104BU, of the same technology and diameter as vias 134. In addition, metal layer L6 is patterned to form waveguide top 118TX making physical and electrical contact with vias 138. Thus, waveguide top 118TX is part of the physical SIW170 structure and electrical path, which includes various items and paths that pass through at least metal layers L6 through L3. Furthermore, the structure and path are tapered so as to: near the bottom of the substrate 104, the shape/path begins closer to the center 108CTR of the region 108 TX; and as it moves upward, it flexes radially outward from the center. Referring again to fig. 2, the same structure is repeated for each waveguide top 118 TX. The exemplary illustration has a plurality of waveguide tops 118TX inside the transmission region 108TX adjacent to the perimeter, collectively forming a substrate integrated communication area (transmit or receive) waveguide 170 having substantially solid walls. Specifically, each waveguide top 118TX corresponds to the structure shown in fig. 3, which tapers from the top down and toward the center of the launch region. The combination of all of these structures across all of the circumferentially-located through-holes provides a generally funnel-shaped physical profile within the base plate 104, being widest at the top of the base plate 104 (e.g., at level L6) and tapering inward toward the bottom of the base plate 104 (e.g., at level L3). Further, an outer boundary that is substantially circular (or piecewise linear near circular) may be defined along an outermost tangent of the circle of each waveguide top 118 TX. Which is generally aligned with or coincides with the perimeter of the interposer 114-1.

The area of layer L3 within the perimeter of SIW170 remains substantially solid except for the feed-through of the antenna signal, forming reflector surface 174. The via 130 contacts the reflector surface 174 to form a substantially closed end to the SIW 170. Core 104C and build-up 104BU are selected to have respective thicknesses such that height (h)172 of SIW170 is approximately equal to one-quarter wavelength (λ/4) of a transmission signal generated by IC 102 as it propagates through substrate 104. Since the dielectric constant of the substrate is higher than that of air, the lambda/4 value in the substrate is different from that in air. The presence of the metal layer also affects lambda/4. Therefore, for most cases, it may be difficult to exactly match the distance due to manufacturing limitations or small differences in the specified dielectric constant, so the distance is about λ/4, and the reflection is not perfect. Thus, the height 172 is designed to be about λ/4, where "about λ/4" covers a range of λ/4 +/-10%. A given substrate may also be designed to support a certain frequency band, such as the 110-140GHz frequency band. Thus, in some examples, the height 172 may be selected to be within a range of λ/4 +/-10% for the frequency bands within the substrate. In another example, the height 172 may be selected to be in the range of λ/4 +/-20% for a frequency band within the substrate.

Thus, when a signal is transmitted by antenna 120TX, a portion of the radiated signal is transmitted from the front side of the antenna, enters interposer waveguide 114-1, and travels to DWG 106-1. Another part of the radiated signal is emitted from the backside of the antenna and is therefore called "backscattering". The backscattered radiation enters a funnel-type substrate integrated waveguide 170, which generally directs the backscattered signal wave in a direction perpendicular to the upper surface of the substrate 104 until it reaches a reflector 174. The backscattered signal is then reflected by the SIW170 and directed back up the interior of the interpolator 114-1. Since the reflection produces a 180 degree phase shift (λ/2) and the height 172 of the SIW170 is λ/4, the reflected signal arrives in phase (in-phase) with the forward signal after the downward and return movement. The reflected backscattered signal then enters the interpolator 114-1 and combines with the forward signal to enhance the signal transmitted from the antenna 120TX to the DWG 106-1. The tapered SIW170 thus reduces the loss of signal passing between the mismatched impedance material of the interposer and the substrate.

Although such waveguide structures and functions are described above with respect to transmit region 108TX, receive region 108RX (in an example embodiment) has the same structure. Thus, when a signal wave is received from the interposer 114-2, a portion of the signal that is not absorbed by the receive antenna is directed downward (by virtue of the inward taper of the funneled SIW waveguide therein) through the substrate 104 and reflected, and then directed back to the receive antenna, thereby returning to the IC die 102.

Fig. 3 also shows additional feed structures (similarly applied to the antennas 120TX2A and 120TX2B that are not visible in cross-section) feeding the illustrated transmit antennas 120TX1A and 120TX1B in the substrate 104. Specifically, a via 140 is formed through the central core 104C, which may be formed simultaneously with the same dimensions and materials as the via 130 described above in connection with the via waveguide top 118 TX. Vias 140 also provide electrical contact to the metal in layer L3, where metal pads 142 are formed in layer L3 by forming a ring-shaped opening in metal layer L3, so that metal pads 142 remain at the center of the ring, while the open area (eventually filled by the buildup) in metal layer L3 concentrically surrounds metal pads 142. In this manner, the structure associated with the antenna is isolated from the other connections of layer L3, thereby allowing signal path connections to be made through a signal coupler as described below. Further, in this regard, metal pad 142 is connected to conductor portion 144 of metal layer L2 via 146. Conductor portion 144 is part of a signal coupler that communicates with IC die 102 through additional vias 147V and metal layer L1 pads 147P. Returning to via 140, metal layer L4 over it is patterned to form pad 148, metal layer L5 is patterned to form pad 150, and vias 152 of respective layers L4 through L5 and vias 154 of layers L5 through L6 are formed through stack 104BU, all of which may be formed similarly and simultaneously to the respective metal layer (and stack) formation of the horizontal coplanar structure described above in connection with via waveguide top 118 TX. Thus, vias 154 provide electrical contact to transmit antenna 120TX1A, completing the structure and electrical path of the antenna in substrate 104. Accordingly, below the transmit antenna 120TX1A in the substrate 104, the overall physical antenna feed structure provides an electrical path through various items and includes at least metal layers L5 through L2. Preferably, the structure has an overall vertical height of λ/4 from the top of metal layer L3 to the bottom of metal layer L6, where λ is the wavelength of the signal that the antennas 120TX, 120RX are to transmit/receive. Further, in this regard, to use at higher frequencies, λ is proportionally reduced, allowing the antenna height to be reduced and occupy less space within the substrate 104; in such an example, the additional space within the substrate may be used to embed the IC die 102 inside the substrate 104 rather than physically attaching it to an outer surface (e.g., bottom) of the substrate 104. Further, the antenna structure includes a portion closer to the lower substrate surface (the pad 142 and the via 140) than other portions extending toward the upper substrate surface, the portion being closer to the center 108CTR of the region 108 TX. This structure can reduce the loss of signal communication, and avoid or minimize the influence of impedance mismatch generated in the transmission (or reception) signal path. However, since the structure and signal path are considered to pass vertically upward through the substrate 104, the final signal transmitting portion (or receiving portion in the case of region 108 RX) of the transmit antenna 120TX1A is advantageously kept away from the transmit antenna 120TX1B (and from the other two transmit antennas 120TX2A and 120TX2B) to reduce cross-talk that may occur between the transmitted (or received) signals. Although such antenna structures and functions are described above with respect to transmit region 108TX, receive region 108RX (in an example embodiment) has the same structure.

Fig. 3 also shows additional structure forming a signal block structure in the substrate 104 that preferably connects the normally electrically grounded metal layer L1 to the metal layer L3. Specifically, the structure includes a metal pad 156 in layer L1, a via 158 from metal pad 156 to layer L2 metal pad 160, and a via 162 from metal pad 160 to the general plane of layer L3. The structure is repeated and positioned at a number of selected locations between the four signal antennas.

Fig. 4 is a schematic diagram of the signal paths from the IC die 102 to the four transmit antennas 120TX1A, 120TX1B, 120TX2A, and 120TX2B of the region 108TX (fig. 2) via the two branch couplers 200-1 and 200-2. Transceiver 102TXRX (fig. 1) includes various circuits for transmitting and receiving millimeter-wave signals (e.g., from differential amplifier 102A). The differential amplifier 102A provides differential outputs 102A-1 and 102A-2, which are represented in FIG. 4 by the corresponding (+) and (-) labels, depicting the differential nature of the outputs and depicting the output signals having a phase difference of 180 degrees. Output 102A-1 is connected to input 202-1 of coupler 200-1 and output 102A-2 is connected to input 202-2 of coupler 200-2. The isolated ports 204-1 and 204-2 of couplers 200-1 and 200-2, respectively, are connected to ground through a matched termination impedance (e.g., Z (PM)). Output 206-1 of coupler 200-1 is connected to antenna 120TX2A and output 208-1 of coupler 200-1 is connected to antenna 120TX1A, antenna 120TX1A being physically oriented and electrically coupled to provide a signal 90 degrees apart from the concurrent signal from antenna 120TX 2A. Output 206-2 of coupler 200-2 is connected to antenna 120TX2B and output 208-2 of coupler 200-2 is connected to antenna 120TX1B, antenna 120TX1B being physically and electrically coupled to provide a signal 90 degrees apart from the concurrent signal from antenna 120TX 2B.

Within the IC die 102, the transmitter circuit generates a millimeter wave signal WS, which is output to the differential amplifier 102A. In response, the amplifier 102A outputs 180 degree separated versions of the input signal (possibly filtered and/or amplified) at its differential outputs 102A-1 and 102A-2. Each of the tap couplers 202-1 and 202-2 operates to receive an input and produce a corresponding 90 degree phase separated output. With respect to branch coupler 200-1, its two outputs are shown as 0 ° (+) and 90 ° (+), representing (+) signals corresponding to output 102A-1 and spaced 90 degrees apart from each other. Thus, the output 0 ° (+) may be considered as a first unit length vector located at 0 ° in the positive direction (typically to the right of the origin of polar coordinates), while the output 90 ° (+) may be considered as a second unit length vector located at +90 degrees with respect to the first vector. Similarly, with respect to branch coupler 200-2, its two outputs are shown as 0 ° (-) and 90 ° (-), representing the (-) signal corresponding to output 102A-2 and being 90 degrees apart from each other. Thus, the output 0 ° (-) can be considered a third unit length vector located at 0 degrees in the negative direction (typically to the left of the polar origin), while the output 90 ° (-) can be considered a fourth unit length vector located at-90 degrees with respect to the first vector or at +90 degrees with respect to the third vector. Thus, in general, fig. 4 encodes an input signal as four waveform vectors, each waveform vector being equally spaced at a distance of 90 degrees and such that each of the four different 90-degree positions is occupied by a respective waveform vector. Thus, the four antennas of fig. 3 collectively produce a circularly polarized output, so that when the millimeter-wave signal WS input to the amplifier 102A varies, each of the resulting four vectors has a constant magnitude but is rotated with time in a plane perpendicular to a plane along which the tops of the transmitting antennas 120TX1A, 120TX1B, 120TX2A, and 120TX2B are aligned. Referring to fig. 1, a circularly polarized signal is rotated vertically upward from the substrate 104 and enters the interior of the interposer 114-1. The rotation may be left-handed or right-handed polarization. In a similar but opposite direction, the receive antennas 120RX1A, 120RX1B, 120RX2A, and 120RX2B are configured to receive four signal components (one for each antenna) of a circularly polarized signal from the interpolator 114-2, and these signal components are connected in the opposite direction to comparable components as shown in fig. 4, so as to decode from the signal components respective output signals representative of the received millimeter wave signals.

The circular polarization achieved by the described physical and electrical configurations of the exemplary embodiments provides a number of benefits. In contrast, for proper signal communication, linear polarization requires fairly accurate planar linear alignment with the signal and the receiver of the signal. Also, by contrast, the exemplary embodiment that provides circular polarization eliminates the need for such planar alignment. For example, with respect to fig. 1, because the DWG cable assembly 106-x is positioned relative to the substrate 104, the circular cross-section of the cable need not be rotationally aligned to a particular position relative to the substrate 104 for purposes of signal communication. Thus, rotational independence is achieved about the axis of each cable relative to the transmit/receive structures in the substrate 104. As another example, circularly polarized signals produce less signal loss in certain environments (e.g., environments with vibration, rotation, or communication gaps).

Additional details of exemplary antenna structures and phase splitters are described in U.S. patent application No. 16/393,809 entitled "circular-Polarized Dielectric Wave mount for Millimeter-Wave High-Speed data communication," filed 24/4 in 2019, which is incorporated herein by reference.

Fig. 5 is a perspective view of a portion of another exemplary substrate of a DWG communication system similar to DWG communication system 100 of fig. 1. The antenna 508 represents a transmitting antenna or a receiving antenna for circularly polarized signals. In this example, SIW570 is constructed using closely spaced stacked vias (e.g., vias 530, 534, 538) in a manner similar to the SIWs shown in fig. 2-3 to form a circular ring of vias 571 in core 504C and an accumulation layer 504BU of multilayer substrate 504 between metal layers L3-L6. However, in this example, the second row of stacked vias forms a second circular ring of vias 572. The circular rings of

vias

571 and 572 together serve to form a more solid conductive wall of SIW 570.

In this example, stacked vias such as

vias

530, 534, 538 are all inline to form a straight-walled SIW570, rather than offset by

vias

130, 134, 138 to form a tapered SIW as shown in fig. 3.

The interposer waveguide 514 has a conductive outer wall 5141. In this example, the hollow interior 5142 is filled with air, which is a dielectric. In another example, the interposer waveguide 514 may have a solid dielectric core instead of being hollow. In another example, the interposer waveguide 514 may be filled with another type of dielectric gas or liquid. In either case, the conductive outer wall 5141 is coupled to the exposed top edge of the SIW 570. In this example, the top edge of SIW570 is exposed by an opening in solder mask layer 504SM as shown at 574. In this example, each opening 574 is filled with a conductive epoxy to permanently couple the conductive walls 5141 of the interposer 514 to the top edge of the SIW 570.

Fig. 6A, 6B are respective side and perspective views of a portion of another exemplary multi-layer substrate 604 for a DWG communication system (e.g., the system 100 of fig. 1). In this example, the SIW 670 is a solid ring of material that is placed in a circular groove 676 that is milled through the core 604C and build-up layer 604BU of the multilayer substrate 604. In this example, the grooves 676 are milled using a rotating tool. In another example, circular grooves may be formed in multilayer substrate 604 using known or later developed chemical or plasma etching processes. After milling the trench 604, a conductive material is then inserted into the trench to form the solid wall SIW 670. In this example, the conductive material is applied using an electroplating process. In another example, a solid SIW may be created within the trench 676 of the multi-layer substrate 604 using known or later developed techniques. For example, the trench may be filled with a conductive paste. In another example, the trench may be filled using an additive manufacturing process that utilizes three-dimensional printing techniques.

The interposer waveguide 614 has a conductive outer wall 6141. In this example, the hollow interior 6142 is filled with air, which is a dielectric. In another example, the interposer waveguide 614 may have a solid dielectric core instead of being hollow, or filled with another type of dielectric gas or liquid. In either case, the conductive outer wall 6141 is coupled to the exposed top edge of the SIW 670. In this example, the top edge of SIW 670 is exposed through an opening in solder mask layer 604SM as shown at 674. In this example, each opening 674 is filled with a conductive epoxy to permanently couple a conductive wall 6141 of interposer 614 to a top edge of SIW 670.

Fig. 7 is a top view of a portion of another example multilayer substrate 704 for a DWG communication system. In this example, rectangular DWG cables are used to transmit linearly polarized signals, so SIWs 770, 771 are rectangular rather than circular. Rectangular interposer waveguides (not shown) would then be coupled to the exposed edges of the SIWs 770, 771. In order for an interposer to provide a standardized interface, it is useful to define a set of waveguide dimensions suitable for various frequencies. For example, the Electronics Industry Alliance (EIA) RS-261-B "rectangular waveguides (WR 3-WR 2300)" has standardized waveguides of various sizes to facilitate interchangeability of metal waveguides. WR-6 (rectangular waveguide) is the standard dimension (about 0.83 mm. times.1.7 mm) for the operating band of about 110 and 170 GHz. WR-5 is a standard size of 140 and 220GHz (about 0.65 mm. times.1.3 mm). In this example, the waveguide regions 770, 771 have a rectangular cross-section and are sized to conform to the WR-6 standard for operation in the 110- 170GHz band. Other exemplary interpolators may include waveguide regions of larger or smaller standard sizes for systems operating in different frequency bands.

Rectangular SIWs 770, 771 are fabricated in a manner similar to that described above for circular SIW170, using one or more rows of closely spaced vias or milled grooves filled with a conductive material to form a waveguide with substantially solid walls.

Fig. 8 is a cross-sectional view of a portion of an example system 800 that includes an interposer 810 positioned between

antennas

821, 822 of a microelectronic device 825 and a dielectric waveguide interconnect 830. In this example, antenna 821 is a transmit antenna and antenna 822 is a receive antenna. However, in other examples, there may be two or more transmit antennas, two or more receive antennas, or various combinations.

In this example, the

antennas

821, 822 are orthogonal antennas sized to transmit or receive circularly polarized Radio Frequency (RF) signals having frequencies in the range of approximately 110-140 GHz. However, in other examples, higher or lower frequencies may be used by appropriately sizing the

antennas

821, 822. As used herein, the term "antenna" refers to any type of radiating element or transmitting structure that may be used to transmit or receive high frequency RF signals.

As described in more detail above with respect to fig. 1-7, the substrate integrated

waveguides

870, 871 are fabricated within the substrate 820. In this example,

SIWs

870, 871 are circular and are fabricated using stacked vias to form a substantially solid wall, as described with respect to fig. 4. In another example,

SIWs

870, 871 can be fabricated as solid walls as described with respect to fig. 6A, 6B. In another example,

SIWs

870, 871 can have a rectangular shape, or another shape suitable for the type of RF signal being transmitted.

In this example, BGA substrate 820 provides a substrate on which IC die 823 is mounted in a "dead bug" inverted manner. The

antennas

821 and 822 are fabricated on the top side of the BGA substrate 820 by patterning the copper layer using known or later developed fabrication techniques. In this example, IC die 823 includes a transmitter and receiver coupled to respective transmitter and

receiver antennas

821, 822 through differential signal paths fabricated on BGA substrate 820. Solder balls 824 are used to connect signal pads and power pads on BGA substrate 820 to corresponding pads on substrate 840 using a known or later developed soldering process.

Together, BGA substrate 820 and IC die 823 may be referred to as a "BGA package," an "IC package," an "integrated circuit," an "IC," a "chip," a "microelectronic device," or similar terminology. BGA package 825 may include an encapsulating material to cover and protect IC die 823 from damage.

Although IC die 823 is mounted in dead bug (dead bug) fashion in this example, in other examples interposer 810 may be suitably modified to mount an integrated circuit containing RF transmitters and/or receivers on the top side of BGA substrate 820 to allow for mechanical clearance. In this example, IC die 823 is wire bonded to BGA substrate 820 using known or later developed fabrication techniques. In other examples, various known or later developed package configurations (e.g., QFN (quad flat no lead), DFN (dual planar no lead), MLF (micro lead frame), SON (low profile no lead), flip-chip die, dual in-line package (DIP), etc.) may be attached to the substrate and coupled to one or more antennas on the substrate.

Substrate 840 may have additional circuit devices mounted thereon and interconnected with BGA package 825. The substrate 840 may be single-sided (one copper layer), double-sided (two copper layers), or multi-layered (outer and inner layers). Conductors on different layers may be connected to the vias. In this example, substrate 840 is a Printed Circuit Board (PCB) having a plurality of conductive layers that are patterned using known or later developed PCB fabrication techniques to provide interconnecting signal lines for various components and devices mounted on substrate 840. Glass epoxy is a major insulating substrate; however, various examples may use various types of PCBs, either now known or later developed. In other examples, substrate 840 may be fabricated using various known or later developed techniques, such as from ceramic, silicon wafers, plastic, and the like.

The interposer 810 is a block of material that is shaped to provide well-defined reference planes 813, the reference planes 813 being located near the top surface 826 of the BGA substrate 820. The second well-defined reference plane 814 is located near the DWG interconnect 830. In this example, interposer 810 includes two defined

regions

811, 812 that form an interface waveguide between reference plane 813 and reference plane 814. In this example, the

interposer waveguide areas

811, 812 are open and thus filled with air or other ambient gas or liquid. In this example,

interposer waveguide areas

811, 812 are lined with

conductive layers

815, 816 such that

interface waveguide areas

811, 812 act as metal waveguides. In another example, the

interposer waveguide regions

811, 812 can be filled with a dielectric material to act as dielectric waveguides. In this example, interposer 810 is made of a non-conductive material (e.g., plastic, epoxy, ceramic, etc.).

In another example, a Photonic Band Gap (PBG) structure may be used to define a portion of interposer 810 between

antennas

821, 822 and/or a portion of substrate 840 between

antennas

821, 822. Manufacture of PBG structures is described in more detail in U.S. Pat. No. 10/371,891 entitled Integrated Circuit with Electrical Waveguide Connector Using Photonic band Structure, granted on 6.8.2019, which is incorporated herein by reference. The purpose of the PBG is to create a high impedance path, avoiding or reducing wave propagation between two points (or regions). In this particular application, it is desirable to reduce crosstalk between the transmitter antenna 821 and the receiver antenna 822 and to increase isolation therebetween. A portion of the interposer material may include a matrix of interstitial nodes that may be filled with a material other than the bulk interposer material. The nodes may be arranged in a three-dimensional array of spherical spaces that are in turn separated by a lattice of interposer material. The photonic bandgap structure formed by the periodic nodes can efficiently guide electromagnetic signals through the PBG waveguide.

The

interposer waveguides

811, 812 may have a circular cross-section, for example, for circularly polarized signals. In another example, the

interposer waveguides

811, 812 may have a rectangular cross-section. For example, the long side of the cross-section may be twice as long as the short side. This is useful for transmitting horizontally or vertically polarized electromagnetic waves. Waveguide dimensions of about 1.5mm x 3.0mm can work well for sub-terahertz signals, for example, in the range of 130-150 gigahertz.

Interposer 810 includes cavities 817 designed to allow the interposer to rest securely on substrate 840 while leaving a small gap between top surface 826 of BGA package 825 and surface 813 of interposer 810. In this manner, BGA package 825 is isolated from stresses or movements of interposer 810 that may affect the connection reliability of solder balls 824.

DWG interconnect 830 is shaped to be coupled to interposer 810 to align one or more DWGs (e.g., DWGs 831, 832) with

waveguide regions

811, 812. Each DWG831, 832 includes a core 833 and a cladding 834. In this example, each DWG831, 832 is also covered by an outer shielding material 835 to provide protection against wear.

At the reference plane 813, the dimensions of the

waveguide areas

811, 812 are approximately matched to the characteristic impedance of the

antennas

821, 822 in order to provide good coupling efficiency. At the reference plane 814, the

waveguide regions

811, 812 flare out to provide transitions to the

DWGs

831, 832 in order to provide good coupling efficiency to the

DWGs

831, 832.

The signal may be launched into the interposer waveguide 811 by a transmitter antenna 821 generated by a transmitter circuit in the IC die 823 using known or later developed techniques. Interface waveguide 811 can then conduct the signal to datum 814 on the other side of interposer 810 with minimal radiation loss. In this way, the insertion loss between the transmitter on IC 823 and DWG831 may be maintained at an acceptable level. For example, if the communication link has a total insertion loss budget of 22dB, it is desirable to keep the insertion loss from the transmitter within the IC 823 to the DWG831 less than 3 dB. Similarly, it is desirable to keep the insertion loss from the DWG 832 to the receiver within the IC 823 to be less than 3 db. Even if the system has a loss budget above 22dB, it may be desirable that the insertion loss of the transition should not exceed a modest percentage, e.g., 10%, of the loss budget.

The DWG interface 830 may include an interlock mechanism that may interlock with the interposer 810, thereby holding the DWG interface 830 securely in place. In this example, DWG interface 830 includes a socket configuration that mates with interposer 810. The interlocking mechanism may be a simple friction scheme, a ridge or lip that interlocks with a recess on the interposer 810, or a more complex known or later developed interlocking scheme. In this example, barbs 836 extend from DWG interface 830 to mechanically interact with interposer 810. In other examples, DWG interface 830 may have a different configuration. For example, DWG interface 830 may be threaded onto substrate 840 or interposer 810, may be snapped onto interposer 810, may be soldered to PCB 840, or the like.

In this way, the interpolator acting as a buffer is used to establish two well-defined reference planes that can be optimized independently. The first plane is located between the radiating elements and the interposer, and the second plane is the surface between the interposer and the DWG interconnect. The interposer allows the introduction of features that can improve isolation between transmitter and receiver antennas in the device, relax alignment tolerances, and enhance impedance matching between the antenna and the dielectric waveguide.

Fig. 9 is a flow chart of a method of interfacing a dielectric waveguide with an antenna on an integrated circuit using an interposer. A dielectric waveguide system such as DWG system 100 (fig. 1) may be operated in this manner.

At 902, a frequency band and antenna configuration to be used on a transceiver IC is selected or defined. For example, it may be determined that the transceiver IC will operate in the RF bands of 120-140 GHz. As described in more detail with reference to fig. 2-4, a four antenna configuration may be selected for the transmit and receive antennas for circularly polarized signals. The antenna may be designed to have a characteristic impedance using known or later developed antenna design techniques.

At 904, a dielectric waveguide interface configuration is selected from a set of available options or a new DWG interconnect structure is designed. In general, the core size and shape, the cladding thickness, and the dielectric constants of the core and cladding will determine the characteristic impedance of the DWG.

An interposer is interposed between the transceiver IC and the DWG interconnect structure and provides two reference planes that can be optimized for the respective interfaces. At 906, an impedance of the interposer waveguide received in the first interface region of the interposer is matched to an impedance of the antenna. This may be accomplished by selecting the size, configuration, and materials for the interposer and interface waveguide regions. For example, to match the 110- > 140GHz operating band of a transceiver IC that uses circularly polarized transmission signals, a circular interposer waveguide having a diameter of about 1.85 mm may be fabricated based on known or later developed waveguide design techniques. In another example, to match the 120- > 140GHz operating band selected for the transceiver IC, the EIA standard WR-6 configuration waveguide region may be fabricated.

At 908, the characteristic impedance of the interposer waveguide at the second interface region of the interposer is matched to the characteristic impedance of the dielectric waveguide. This may be achieved, for example, by tapering the ends of the waveguide region, as shown in figure 3.

At 910, the first interface region is coupled to the second interface region through an interposer waveguide within the interposer. The waveguide may be open (air) or filled with a dielectric. The open waveguide region may be coated with a conductive coating to form a metallic waveguide.

At 912, a front-side signal is launched from the front side of the antenna structure into the interposer waveguide using the first interface region.

At 914, backscattered radiation emitted from the backside of the antenna structure is directed from the antenna structure to the reflective surface using a substrate integrated waveguide as described in more detail above. Due to reflection, the backscattered radiation undergoes a phase shift of λ/2.

At 916, the reflected backscattered signal is directed from the reflective surface back to the first interface. The length of the substrate integrated waveguide is selected to be about λ/4 of the signal in the substrate material. When the reflected backscatter signal reaches the first interface, it is phase shifted by λ/4+ λ/2+ λ/4 ═ λ, so it combines in an aggressive manner with the front-side signal to enhance the signal delivered to the DWG.

Other embodiments

In described examples, a transceiver implemented in a BGA package having a multilayer substrate is described. Other examples may use other known or later developed integrated circuit packaging techniques to provide a transceiver including one or more antennas located on a surface of a multilayer substrate.

In the described example, a transceiver having a multi-layer substrate size of 6mm x 8mm is described with two antennas operating in the 120-. In other examples, transceiver packages of different sizes and shapes may be accommodated by sizing the interposer accordingly. Operation in different frequency bands can be accommodated by selecting different sized waveguide regions for the SIW and different thicknesses of the multilayer substrate such that the length of the SIW is approximately λ/4 of the signal within the substrate.

The thickness and overall shape of the SIV and interposer can be selected to provide the desired mechanical and electrical characteristics of the selected DWG interconnect structure.

In the described example, copper is used as the conductive layer in the multilayer substrate. In other examples, other types of conductive metal or non-metallic conductors may be used to pattern the signal lines and antenna structures, for example.

In described examples, copper is used to fill vias and/or mill grooves to form substrate integrated waveguides. In other examples, other types of conductive metal or non-metal conductors may be used to fill vias and/or trenches to form SIWs.

In the described examples, circular and rectangular substrate integrated waveguides are described. In another example, the oval or other shape of the SIW can be fabricated to support various types of signal propagation.

In described examples, the stacked vias forming the SIW include contact pads on each layer of the multilayer substrate. In another example, the contact pads may be eliminated or reduced to a minimum size to provide less notching of the SIW, thereby reducing parasitics and thereby improving performance.

In the described examples, a transmitter antenna and SIW and a receiver antenna and SIW are described. In other examples, there may be only one antenna and SIW, or there may be multiple transmitter and/or receiver antennas with respective SIWs.

In this specification, the term "couple" and its derivatives refer to indirect, direct, optical, and/or wireless electrical connections. Thus, if a first device couples to a second device, that connection may be through a direct electrical connection, through an indirect electrical connection via other devices and connections, through an optical electrical connection, and/or through a wireless electrical connection.

Modifications may be made in the described embodiments, and other embodiments are possible within the scope of the claims.

Claims

1. A device, comprising:

a multilayer substrate having a first surface and a second surface opposite the first surface;

an integrated circuit mounted on the second surface of the multilayer substrate, the integrated circuit having a transmission circuit configured to process millimeter wave signals; and

a substrate waveguide having a substantially solid wall formed within a portion of the multilayer substrate, the substrate waveguide having an open end surrounded by an edge of the substrate wall, the edge exposed on the first surface of the multilayer substrate.

2. The device of claim 1, further comprising an antenna configured to radiate the millimeter-wave signals from the multilayer substrate, the antenna being formed on the first surface of the multilayer substrate within the substrate waveguide, the antenna being coupled to the transmission circuit.

3. The device of claim 2, further comprising a reflector in one of the layers of the multilayer substrate, the reflector coupled to the wall of the substrate waveguide.

4. The device of claim 3, wherein a thickness of the multilayer substrate between the reflector and the antenna is one quarter of a wavelength of the millimeter-wave signal in the substrate.

5. The device of claim 1, wherein the multilayer substrate includes a trench, wherein the substrate waveguide wall is a conductive material located within the trench.

6. The device of claim 1, wherein the substrate waveguide wall is formed by at least one row of closely spaced vias.

7. The device of claim 1, wherein the substrate waveguide has a circular cross-section.

8. The device of claim 1, wherein the substrate waveguide has a rectangular cross-section.

9. The device of claim 1, further comprising an interposer mounted on the first surface of the multilayer substrate, the interposer having a first surface and a second surface opposite the first surface, the interposer having an interposer waveguide with a conductive perimeter wall between the first surface and the second surface, wherein an edge of the perimeter wall at the first surface of the interposer is coupled to an edge of the substrate waveguide wall on the first surface of the multilayer substrate.

10. The device of claim 9, wherein the second surface of the interposer is configured to mate with a dielectric waveguide.

11. The device of claim 2, further comprising an interposer having an interposer waveguide with a conductive perimeter region coupled to the substantially solid wall of the substrate waveguide.

12. The device of claim 11, wherein the interposer comprises:

a mass of material having:

a first interface region engaged with the antenna;

a second interface region joined to the dielectric waveguide DWG; and

wherein the interposer waveguide is formed by a defined region within the block of material located between the first interface region and the second interface region.

13. The device of claim 12, wherein the defined region is an opening through the block of material.

14. The device of claim 12, wherein the opening is coated with a conductive material.

15. The device of claim 12, wherein the opening is filled with a dielectric material.

16. The device of claim 12, wherein the defined region is formed from a photonic band gap structure.

17. A system, comprising:

an integrated circuit mounted on the second surface of the multilayer substrate, the integrated circuit having a transmission circuit configured to process millimeter wave signals;

a substrate waveguide having substantially solid walls formed within a portion of the multilayer substrate, the substrate waveguide walls having edges exposed on the first surface of the multilayer substrate;

an antenna configured to radiate millimeter wave signals from the multilayer substrate, the antenna formed on the first surface of the multilayer substrate within the substrate waveguide, the antenna coupled to the transmission circuit; and

a reflector in one of the layers of the multilayer substrate, the reflector coupled to the wall of the substrate waveguide.

18. The system of claim 17, further comprising an interposer mounted on the first surface of the multilayer substrate in alignment with the waveguide, the interposer having a first surface and a second surface opposite the first surface, the interposer having an interposer waveguide with a conductive perimeter wall between the first surface and the second surface, wherein an edge of the perimeter wall at the first surface of the interposer is coupled to an edge of the substrate waveguide wall on the first surface of the multilayer substrate.

19. The system of claim 18, further comprising a Dielectric Waveguide (DWG) having a dielectric core, the DWG having an end coupled to the multilayer substrate such that an axis of the core is aligned with the interposer waveguide.

20. A method of operating a device, the method comprising:

transmitting a first portion of a signal from an antenna located on a surface of a multilayer substrate, wherein the first portion of the signal propagates away from the substrate;

transmitting the second portion of the signal from the antenna, wherein the second portion of the signal propagates into the multilayer substrate;

directing the second portion of the signal to a reflective surface within the multilayer substrate using a substrate waveguide integrated within the multilayer substrate;

reflecting a second portion of the signal through the reflective surface to form a reflected signal;

directing the reflected signal toward the surface of the multilayer substrate using the substrate waveguide; and

combining the reflected signal with the first portion of the signal to form an enhanced signal.