FIELD OF THE DISCLOSURE

The present disclosure relates to electronic circuits which interact with electromagnetic (EM) waves, particularly at sub-THz and THz frequencies. The present disclosure further concerns the integration of such circuits with waveguides.

BACKGROUND OF THE DISCLOSURE

Sub-THz and THz electromagnetic fields are useful in technical applications such as electronic communication, high-resolution imaging, sensing and atomic clocks. Recent developments in microtechnology have facilitated the development of sub-millimeter-size waveguides which can effectively confine sub-THz and THz electromagnetic fields as a guided wave. When coupled for example to a monolithic microwave integrated (MMIC) transmitter and receiver circuit, such waveguides can be used as highly accurate frequency references or as signal generators and receivers in imaging devices.

It is challenging to generate, manipulate and receive sub-THz and THz electromagnetic waves in waveguides without significant losses. Document WO2016161215 discloses a device where an electromagnetic wave is generated in a waveguide cavity formed in a silicon substrate. The coupling structures which generate the wave in the waveguide are placed on a glass substrate which overlies the waveguide. Document US2019089036 discloses a device where chip on a base die are coupled to an underlying waveguide with interconnecting pillars. A problem with both approaches is that the durability of the device may not be optimal due to the complexity of the stacked structure.

BRIEF DESCRIPTION OF THE DISCLOSURE

An object of the present disclosure is to provide an apparatus which alleviates the above disadvantages. The object of the disclosure is achieved by an arrangement which is characterized by what is stated in the independent claim. The preferred embodiments of the disclosure are disclosed in the dependent claims.

The disclosure is based on the idea of building a waveguide in a first silicon substrate, joining the first silicon substrate to a second silicon substrate, building an electronic circuit on the second silicon substrate and coupling the circuit to the underlying waveguide. An advantage of this arrangement is that a simple and reliable electronic module utilizing high- frequency EM waves can be built at low cost.

BRIEF DESCRIPTION OF THE DRAWINGS

In the following the disclosure will be described in greater detail by means of preferred embodiments with reference to the accompanying drawings, in which

Figures 1a and 1 b illustrate a first wafer with a waveguide and a second wafer placed on top of the first wafer.

Figures 2a and 2b also illustrate a first wafer with a waveguide and a second wafer placed on top of the first wafer.

Figure 3 illustrates a high-frequency electronic module with electric vias and coupling probes.

Figures 4a - 4b illustrate a high-frequency electronic module with grounding vias and coupling probes.

Figure 4c illustrates a high-frequency electronic module with a protected enclosure for the electric circuit.

Figure 5 illustrates a control circuit for a molecular clock.

Figure 6 illustrates a communication module.

Figure 7 illustrates a method for manufacturing a high-frequency electronic module. DETAILED DESCRIPTION OF THE DISCLOSURE

A high-frequency electronic module comprising a first silicon wafer with a top surface. The first silicon wafer comprises a waveguide with a first end and a second end. The walls of the waveguide are coated with an electrically conductive material. The high-frequency electronic module comprises a second silicon wafer with a bottom surface and a top surface. The bottom surface of the second silicon wafer is attached to the top surface of the first silicon wafer. At least the part of the bottom surface of the second silicon wafer which overlies the waveguide is coated with an electrically conductive layer. The high- frequency electronic module comprises an electric circuit on the top surface of the second silicon wafer. The high-frequency electronic module comprises a first electric via which extends through the second silicon wafer and a first radiation-coupling probe which extends from the first electric via into the waveguide. An output of the electric circuit is connected to the first electric via. The high-frequency electronic module comprises a second electric via which extends through the second silicon wafer and a second radiation coupling probe which extends from the second electric via into the waveguide. An input of the electric circuit is connected to the second electric via.

Figure 1a illustrates a first silicon wafer 111 and a second silicon wafer 112. The top surface 1111 of the first silicon wafer defines a component plane which will be illustrated as the xy-plane in this disclosure. The second silicon wafer has a top surface 1121 and a bottom surface 1122.

The expression “top surface” refers in this disclosure to surfaces which are indicated as the upper surface in figures where an xz- or yz-cross-section in shown. Correspondingly, the expression “bottom surface” refers to surfaces which are indicated as the lower surface in figures where an xz- or yz-cross-section in shown. However, expressions such as “top” / “bottom” and up / down merely define two opposing surfaces and directions - they do not imply anything about how the component should be oriented during use or manufacture.

The electronic circuit may be configured so that the first radiation-coupling probe can generate in the waveguide an electromagnetic wave with a frequency which lies in the sub THz range, which may be defined as 100 GHz - 300 GHz. Alternatively, the frequency may lie in the THF range, which may be defined as 300 GHz - 3 THz. Figure 1 b illustrates the top surface 1111 of the first silicon wafer 111 in the component plane. Figures 1 a and 1 b both illustrate a waveguide 12 which extends from a first point to a second point in the component plane. The waveguide has a corresponding first end 121 and second end 122. The waveguide 12 also has a waveguide depth D in the direction of the z-axis. In other words, the bottom 128 of the waveguide 12 has been recessed from the top surface 1111 of the first silicon wafer 111 by the recess depth D.

Figure 1 a also illustrates the first through-hole 131 and the second through-hole 132 in the second silicon wafer 112. These through-holes may be aligned with the first 121 and second 122 ends of the waveguide so that each through-hole gives access to one end of the waveguide. There may be one or more additional through-holes at each end of the waveguide. These additional through-holes may be used as grounding through-holes for grounding the electric circuit. The through-hole and/or additional through-holes can be placed near the center of the waveguide in its width-direction. The sidewalls 127 and the bottom 128 of the waveguide and the bottom surface 1122 of the second silicon wafer (at least the area which overlies the waveguide) are coated with an electrically conductive material (which is not illustrated in figures 1 a - 3). The electrically conductive material may be a metal, for example copper, aluminium or gold. It may alternatively be a conductive non-metallic material.

The waveguide has an elongated shape in the xy-plane. The shape may for example be rectangular, as in figure 1 b where the rectangular shape has a width W and a length L. Figures 2a and 2b, where reference numbers 211 - 212, 22, 221 - 222 and 231 - 232 correspond to reference numbers 111 - 112, 12, 121 - 122 and 131 - 132, respectively, illustrate a high-frequency electronic module where the second silicon wafer 212 has been bonded on top of the first silicon wafer 211. The first silicon wafer 211 comprises a waveguide 22. These figures illustrate a waveguide with a meandering shape in the xy- plane. The width W of the meander is illustrated in figure 2b. The length of the meander is equal to the distance between the two ends 221 and 222, measured along the meander.

Either one of the two waveguide shapes illustrated in figures 1 b and 2b could be used in any embodiment presented in this disclosure. Other waveguide shapes are also possible in the xy-plane and in the z-direction. The cross-section of the waveguide in the xz-plane and yz-planes may have a shape which is not rectangular. The cross-sectional width W of the waveguide may for example be in the range 0.1 mm - 5 mm, or in the range 0.2 mm - 1 .5 mm. The length L of the waveguide may for example be in the range 0.1 - 200 mm. The depth D of the waveguide may for example be in the range 0.05 - 2.5 mm, or in the range 0.1 mm - 0.75 mm.

Figure 3 illustrates a high-frequency electronic module where reference numbers 311 - 312 and 32 correspond to reference numbers 111 - 112 / 211 - 212 and 12 / 22, respectively, in figures 1a and 2a. Electric components 361 and 362, which form at least a part of the electric circuit, lie on the top surface of the second substrate 312. The components 361 and 362 may be attached to the top surface of the second substrate 312 by flip-chip methods or by wire bonding.

A first electric via 341 extends from an output of the electric circuit (here illustrated as a layer extending from the via to component 361) through the first through-hole. A first radiation-coupling probe 351 extends from the first electric via 341 into the waveguide 32. An output signal which is generated by the electric circuit is thereby coupled to the first radiation-coupling probe 351. A second electric via 342 extends from an input of the electric circuit (here illustrated as a layer extending from the via to component 362) through the second through-hole and a second radiation-coupling probe 352 extends from the second electric via 342 into the waveguide 32. A high-frequency signal which is generated in the electric circuit can thereby be passed through the waveguide and back to the circuit.

The first and second electric vias may be electrical conductors which fill the corresponding through-hole and thereby facilitate electrical interaction between the circuit which lies on the top of the second wafer 312 and the waveguide which lies beneath it. As above, the sidewalls and bottom of the waveguide are coated with an electrically conductive material (not illustrated in figure 3). The electric circuit may be connected to the coating in the waveguide, as explained in more detail below.

The first electric via may be aligned with the first end of the waveguide, and the second electric via may be aligned with the second end of the waveguide. The first and second radiation-coupling probes may for example be wires which are attached to the bottom of the corresponding electric vias with wire bonds.

The output of the electric circuit which is coupled to the waveguide may for example be an output from a high-frequency signal generator or an antenna which receives a high- frequency signal. The frequency of the signal may lie in the sub-THz or THz range and the signal may be transmitted through the waveguide.

As the signal passes through the waveguide 32, it is coupled to the input of the electric circuit through the second coupling probe. The electric circuit may for example comprise measurement devices which measure the signal from the input of the electric circuit. The term “output”, without any further specifiers, refers in this discussion only to the output connection from the electric circuit to the waveguide. Correspondingly, the term “input”, without any further specifiers, refers only to the input connection from the waveguide to the electric circuit. The signal which passes through the waveguide is referred to simply as “the signal” or “the high frequency signal”. The output may also be called the signal output, and the input may be called the signal input. As explained in the examples below, in addition to the output and input toward the waveguide, the electric circuit may comprise other inputs and outputs. These other inputs and outputs will be named with a specifier, for example “antenna input”, to distinguish them from the waveguide input and output.

The second silicon wafer may comprise at least one grounding through-hole which extends from the top surface of the second silicon wafer to the bottom surface of the second silicon wafer. The high-frequency electronic module may comprise at least one grounding via which extends from the electric circuit through the at least one grounding through-hole to the electrically conductive layer.

Figure 4a illustrates a high-frequency electronic module where reference numbers 411 - 412, 42, 441 - 442, 451 - 452 and 461 - 462 correspond to reference numbers 311 - 312, 32, 341 - 342, 351 - 352 and 361 - 362, respectively, in figure 3. The bottom surface of the second silicon wafer 412 is at least partly coated with an electrically conductive layer 426, which may be set to a ground potential. The second silicon wafer 412 also comprises additional through-holes, which may be called grounding through-holes, and first and second grounding vias 471 and 472 which extend from the electric circuit to the metal layer 426.

Figure 4b illustrates the same high-frequency electronic module in the xy-plane. The underlying waveguide 42 is illustrated with dashed lines. The first and second radiation coupling probes 451 - 452 can in this arrangement be connected between the waveguide 42 and the electric circuit 461 - 462 in a ground-signal-ground geometry (481 / 482) where both the first and the second electric via 441 - 442 is flanked by two grounding vias. Alternatively, one grounding via can be placed at each end of the waveguide, and/or one or more grounding vias can be distributed along the length of the waveguide.

As mentioned above, in any embodiment of this disclosure the sidewalls of the cavity may be coated with a layer of electrically conductive material, such as 427 in figure 4a, to increase the reflectance of the sidewalls. The same material may be used to coat the floor of the cavity and optionally also the bottom of the second silicon wafer. The electrically conductive material may be a metallic material, for example copper, aluminium or gold. The component may also comprise a third silicon wafer which is attached to the top surface of the second silicon wafer. The third silicon wafer may comprise a circuit cavity which is placed around the electronic circuit to seal said circuit in a protected enclosure.

Figure 4c illustrates a third silicon wafer 413 which overlies the second silicon wafer 412 and has been bonded on top of the second silicon wafer 412 to seal the electric circuit in a protected enclosure. The circuit is enclosed in a sealed circuit cavity 49 formed in the third silicon wafer 413. The walls of the circuit cavity may optionally be coated with a metallic layer 491 .

Examples

The high-frequency electronic module may for example be a molecular clock which can provide a highly accurate clock reference signal. The waveguide is in this case filled with gas comprising polar gas molecules and hermetically sealed. The gas may be called a measurement gas and it may comprise dipolar gas molecules. The measurement gas may for example be carbonyl sulfide ( ¹⁶ 0 ¹² C ³² S) or any other gas suitable for this purpose.

When a high-frequency signal is passed through the waveguide, the electromagnetic wave can excite rotational transitions in the polar gas molecules. The frequency of a given rotational transition (referred to as the transition center frequency or peak absorption frequency below) is an invariant physical constant. When the molecular clock is targeted to a given rotational transition in the measurement gas the frequency of that rotational transition, which is known from the literature, is selected as a predetermined transition center frequency. The control circuit of the molecular clock can keep the frequency of the high-frequency signal fixed to this predetermined transition center frequency.

More specifically, the bottom surface of the second silicon wafer may be hermetically sealed to the first silicon wafer and the waveguide may be filled with a measurement gas. The electric circuit may comprise an RF signal generator which is coupled to the output of the electric circuit and configured to generate a high-frequency signal.

As mentioned above, the aim is to keep the frequency of the high-frequency signal fixed to the transition center frequency. This can be achieved with many different arrangements. One option is that the electric circuit may comprise a lock-in amplifier which is coupled to the input of the electric circuit. The electric circuit may further comprise a control circuit which is coupled to the RF signal generator and to the lock-in amplifier. The control circuit may be configured to lock the frequency of the high-frequency signal to a known transition center frequency in the measurement gas. In addition to sealing the first silicon wafer to the second silicon wafer hermetically, the first and second electric vias - and any optional grounding vias which may also be present - should also be hermetically sealed in this embodiment. The fabrication steps can for example be performed in a chamber filled with the measurement gas.

Figure 5 illustrates the control circuit which generates a frequency-reference output signal 573. The control circuit may comprise a voltage-controlled crystal oscillator 535 (VCXO) which is connected to a tunable RF signal generator 536. The voltage-controlled crystal oscillator 535 sends a clock signal 576 to the RF signal generator 536, and the RF signal generator 536 generates a high-frequency signal 572. The signal 572 is transmitted to the output of the electric circuit.

The electric circuit comprises a lock-in amplifier 533 which is connected to the input of the electric circuit. The lock-in amplifier 533 also receives the high-frequency signal 572 directly from the RF signal generator 536. The lock-in amplifier provides an error signal 574 as an in-phase output. A phase-locked loop can then be formed in the control circuit which also comprises a loop filter 534 which provides a control signal 575 to the voltage- controlled crystal oscillator 535 to adjust the frequency of the clock signal 576 which is transmitted to the tunable RF signal generator 536.

In other words, the high-frequency signal produced by the RF signal generator 536 is connected with the lock-in amplifier 533 both with a direct connection within the electric circuit and through the connection provided by the waveguide.

The frequency of the high-frequency signal 572 can thereby be maintained at a peak absorption frequency of a selected rotational spectral line in the measurement gas which fills the waveguide 52. The frequency of the clock signal 576 will then also be fixed to a given value. The voltage-controlled crystal oscillator 535 may generate a frequency- reference output signal 573 which will form a stable frequency reference. The clock signal 576 and the frequency-reference output signal 573 will typically have a frequency which is below the sub-THz or THz frequency of the high-frequency signal. The RF signal generator 536 may comprise a frequency multiplier, so that the ratio between the frequency of the high-frequency signal 572 and the frequency of the clock signal 576 is fixed by a (known) multiplier constant. The clock frequency provided by the voltage-controlled crystal oscillator 535 in the clock signal 576 may for example be in the MHz range. The electric circuit may be monolithic microwave integrated circuit (MMIC).

In one form of operation the RF signal generator transmits sub-THz or THz signals at full transmission power at various frequencies within a defined band around an expected rotational transition frequency at which the transmission efficiency of the gas cell in the waveguide is minimal (absorption of the signal is maximal). When the lock-in amplifier finds the transition center frequency where absorption is maximal, it calculates the difference between the probing frequency and the transition center frequency and outputs a signal which indicates the magnitude of this difference. When the difference is zero, the frequency-reference signal 573 is locked to the transition center frequency. The lock-in amplifier provides an error signal 574 to the loop filter 534 to regulate the probing wave 572 via the tunable signal generator 536.

The high-frequency electronic module may alternatively be a communication module which receives an external electromagnetic signal and converts the data contained in this signal into digital communication data. Figure 6 illustrates this embodiment schematically.

In figure 6 the electric circuit 681 is coupled to an antenna 68 which is configured to receive an external signal 67 from the environment which surrounds the high-frequency electronic module and generate a high-frequency signal 672. The high-frequency signal 672 generated by the antenna 68 is coupled to the output of the electric circuit and passes through the waveguide 62. The input of the electric circuit is coupled to a communication unit in the electric circuit.

Figure 7 illustrates a method for manufacturing any high-frequency electronic module described above. The method comprises the steps of (71 ) forming the waveguide in the first silicon wafer by deep silicon etching, (72) depositing an electrically conductive material on the sidewalls and bottom of the waveguide, (73) forming the first and second electric vias in the second silicon wafer and depositing an electrically conductive material on at least a part of the bottom surface of the second silicon wafer which will overlie the waveguide,, (74) forming the first and second radiation-coupling probes on the bottom surface of the second silicon wafer, (75) bonding the bottom surface of the second silicon wafer to the top surface of the first silicon wafer, (76) depositing and patterning electrical connections on the top surface of the second silicon wafer, and (77) attaching electric circuit components on the top surface of the second silicon wafer.

Optionally, the first and second substrates may be thinned and polished either before step 71 or before step 76. Optionally, the sidewalls and bottom of the waveguide made by smoothened before step 72. An electrically conductive material may be deposited on the bottom surface of the second silicon wafer before step 75.