Interconnection structure between microcircuits

Abstract

The invention relates to an interconnect structure between microcircuits. The invention proposes an assembly of two electronic microcircuits joined together by providing a set of electrical couplings between respective connection points of the two microcircuits, the first microcircuit being a converter microcircuit having at least one charge-accumulating photoelectric conversion element, and the second microcircuit is a readout microcircuit having at least one element for reading the charge generated by the respective photoelectric conversion element, and the connection point comprises face-to-face surface elements separated by a dielectric, the coupling between the two circuits being a capacitive coupling. The present invention is particularly applicable to a mixture between two matrix circuits produced using different techniques, one matrix circuit for conversion and the other matrix circuit for reading.

Description

[0001] The present invention relates generally to hybrid circuits and, in particular, to a new method of electrical coupling between two circuits that are made using different technologies and / or that need to be assembled by stacking after fabrication.

[0002] As a practical example, a matrix sensor consists of pixels which essentially contain a converter element and a local readout circuit. The converter element and the local readout circuit can be produced on the same substrate or on two different substrates.

[0003] Realizing the conversion element and the readout circuit on two different substrates is typically dictated by issues of compatibility and optimizing their production. Today, the readout circuit is produced on a silicon substrate using CMOS technology, whereas the conversion element can be produced using a wider range of technologies.

[0004] For example, the conversion elements can be produced on a silicon substrate using a manufacturing method optimized for conversion height. Alternatively, they can be made of less common semiconductor materials, such as III-V semiconductors with wavelengths in the infrared range.

[0005] These include InGaAs substrates for near-infrared applications, particularly for SWIR (short-wave infrared) sensors, InSb substrates for MWIR (mid-wave infrared), MCT (mercury cadmium telluride) for LWIR (long-wave infrared) sensor applications, and CdTe for X-ray detection.

[0006] Such sensors implemented on two assembled substrates are often referred to as "hybrid sensors", the term "hybrid" referring to the method by which the two substrates are assembled together.

[0007] The invention is particularly applicable to hybrid sensors whose conversion element consists of a PN junction, usually a photodiode which converts incident photons into electric charges.

[0008] There are two main types of circuits for reading the charge generated in a photodiode.

[0009] therefore, Figure 1A A voltage follower readout circuit is shown, in which the photoelectric charge generated by the photodiode PD is accumulated in the photodiode junction capacitance and the associated parasitic capacitance. The potential change is read by a voltage follower SF in the readout circuit, the output of which is the pixel signal SP. This structure is called SFP (source follower pixel) and has the advantage of being very simple and compact. One disadvantage is that the polarization of the photodiode can change during exposure, resulting in nonlinear phenomena.

[0010] Figure 1BA more advanced solution is shown, called CTIA (“capacitive transimpedance amplifier”), which is more stable in terms of polarization. One electrode of the photodiode PD, here its anode, is connected to the input of an operational amplifier AMP, with a capacitor C between the other input and the output of the operational amplifier. It offers several advantages: firstly, the polarization is more stable due to the high gain of the operational amplifier; secondly, since the photoelectric charge generated by the photodiode is transferred to the capacitor C, the sensitivity can be modulated by acting on the amplifier gain.

[0011] Figure 1A and 1B It is shown that a direct electrical connection between the photodiode PD and the readout circuit is required. This interconnection involves a mixture with direct electrical contact points at each sensor pixel.

[0012] Figures 2A to 2D Various known technical solutions for such direct mixing are shown, and the common principle is that for each pixel, electrodes are formed on the photodiode substrate and the substrate of the readout circuit. The two substrates are then placed face to face, aligned and mechanically and electrically connected together.

[0013] Figure 2A A method based on the insertion and pressurization of a polymer adhesive PA is shown. This pressure is maintained by removing the polymer adhesive before and during curing. This method is simple in concept, but requires that the height of the electrodes on both substrates is very uniform and that the tools are highly parallel. On the other hand, there is a risk of failure if the stresses generated by the adhesive removal are released when the sensor is exposed to adverse conditions such as high temperature, humidity, etc. For these reasons, this known method is considered unreliable.

[0014] Figure 2B Metal melting methods are shown. Indium electrodes are generally very popular due to their low melting point and high ductility at very low temperatures.

[0015] Figure 2C A newer method called "copper pillars" is shown, consisting of forming electrodes in the form of copper pillars CC on each substrate, one of the pillars being covered with solder material SM. The solder volume compensates for the uneven height of the copper pillars.

[0016] at last, Figure 2DA new method for molecular bonding is presented, which is theoretically more reliable and more reproducible. It is based on the fact that at fairly moderate temperatures (<300°C) covalent bonds can be formed between two dielectric surfaces (typically made of silicon dioxide) in close contact. This method requires both surfaces in contact to be atomically flat and extremely clean, since the initial bonding involves Van Der Walls forces. Therefore, the cost of implementation is potentially high.

[0017] Two key points of conventional hybridization are the need for a large number of direct electrical contacts between all pixels on the sensor, and the need for alignment, which becomes more demanding as the pixel pitch decreases.

[0018] These difficulties are far from being solved today, and manufacturing yields are still relatively low due to a high percentage of defective products, especially for high-resolution sensors. We note that the greatest difficulty in hybridization is making reliable electrical contact between the photodiode and the readout circuit, while precise alignment has become relatively easy to ensure thanks to advances in digital microscopy. Summary of the invention

[0019] The object of the present invention is to propose a new mixing principle which solves some or all of the above-mentioned problems associated with direct contact mixing.

[0020] To this end, an assembly of two electronic microcircuits joined together by providing a set of electrical couplings between respective connection points of the two microcircuits is proposed, the first microcircuit being a converter microcircuit having at least one charge accumulating photoelectric conversion element, and the second microcircuit being a readout microcircuit having at least one element for reading the charge generated by the respective photoelectric conversion element, and the connection points comprising face-to-face surface elements separated by a dielectric, the coupling between the two circuits being a capacitive coupling.

[0021] This circuit advantageously but optionally includes the following additional features, alone or in any technically compatible combination:

[0022] * The or each photoelectric capture element is a photodiode, one electrode of which is connected to a corresponding surface element on the first microcircuit and the other electrode of which is connected to a reference potential.

[0023] * The or each surface element on the second microcircuit is connected to the input of a reading element comprising an infinite or quasi-infinite DC input impedance circuit and an associated reset switch.

[0024] *A circuit with infinite or quasi-infinite DC input impedance is a voltage follower circuit.

[0025] * The or each surface element on the second microcircuit is connected to an input of a reading element comprising a capacitive transimpedance circuit associated with a reset switch.

[0026] * The or each surface element on the second microcircuit comprises a first portion connected to the input of the capacitive transimpedance circuit and a second portion belonging to a feedback link, the first and second portions forming a single capacitor with the corresponding surface element of the first microcircuit.

[0027] * The first microcircuit comprises a matrix of charge-accumulating photoelectric conversion elements and the second microcircuit comprises a corresponding matrix of elements for reading the charge generated by the corresponding photoelectric conversion elements.

[0028] * The assembly comprises at least one direct electrical connection between the first and second microcircuits for control signals common to the various matrix elements.

[0029] * The first and second microcircuits each comprise at least one additional surface element, wherein two opposing surface elements provide additional capacitive coupling for control signals common to the various matrix elements.

[0030] * Additional surface elements around the connection point.

[0031] * The additional capacitive coupling is able to carry a reset signal for the optoelectronic conversion element of the first microcircuit.

[0032] *Dielectrics include the bonding adhesive between two microcircuits.

[0033] *The adhesive contains additives that increase the dielectric constant.

[0034] *The dielectric consists of a layer of dielectric material joined to two microcircuits by molecular bonding.

[0035] * The or each photo-capturing element comprises an avalanche photodiode, capacitive coupling providing isolation from a bias voltage of the or each photodiode. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Other aspects, objects and advantages of the present invention will become more apparent upon reading the following detailed description of preferred embodiments of the invention given by way of non-limiting example and with reference to the accompanying drawings, in which:

[0037] Figure 1A and Figure 1B Schematic diagram of a hybrid sensor pixel with two conventional readout techniques.

[0038] Figure 2A , Figure 2B , Figure 2C and Figure 2DFour known hybrid technologies between the conversion element substrate and the reading element substrate for such hybrid sensors are schematically shown,

[0039] Figure 3 is a circuit diagram of a pixel in a hybrid sensor according to a first embodiment of the present invention,

[0040] Figure 4 yes Figure 3 A more detailed diagram of the circuit shown,

[0041] Figure 5 It is shown Figure 4 The timing diagram of the various signals involved in the circuit shown,

[0042] Fig. 6A shows the mixing of two substrates in the hybrid sensor according to the present invention,

[0043] Figure 6B Show Fig. 6A A mixed variant shown,

[0044] Figure 6C Show Fig. 6A Another embodiment of the mixture shown,

[0045] Fig.6D Show Figure 6B Another embodiment of the mixture shown, and

[0046] Figure 7 is a circuit diagram of a pixel in a hybrid sensor according to a second embodiment of the present invention. DETAILED DESCRIPTION

[0047] Hereinafter, two possible implementations of the hybrid image sensor according to the present invention will be described, one being a combination of each pixel of a photodiode and an SFP follower source readout circuit, and the other being a combination of each pixel of a photodiode and a CTIA amplifier readout circuit.

[0048] In both cases, the photodiode is reset to zero at the beginning of each exposure, for example, by a switching transistor or by a virtual ground created by an operational amplifier.

[0049] According to the invention, a capacitive hybrid is provided between the photodiode substrate and the readout circuit substrate, making it possible to read the potential changes on these photodiodes while ensuring that the photodiodes are reset to zero between two exposures. In this way, each photodiode is connected to its readout circuit without direct contact, but via a capacitor formed between two electrodes which are face to face and at a short distance from each other when the two substrates are bonded together.

[0050] Since the signal passes through the capacitor, there is no risk of signal disappearing, as described in the introduction, when the contact is defective, nor in direct contact mode.

[0051] In reality, there will always be a certain capacitance between a photodiode and its readout circuitry, and only variations in capacitance at a pixel can affect the readout performed on the pixel in question. And such non-uniformities, when present, are usually easy to correct using digital post-processing.

[0052] The capacitor is formed by flat electrodes of appropriate size and shape being placed opposite each other on two substrates which are then bonded together, typically using conventional techniques for making conductive contact plates.

[0053] First, we will describe how the present invention works with a source follower circuit SFP.

[0054] Considering a pixel comprising a photodiode and a readout circuit SFP, we understand that if a capacitor is inserted between the photodiode and the readout circuit, it is actually possible to capture the potential changes caused by charge accumulation during exposure. However, it is still necessary to ensure that the photodiode is reversely repolarized (pixel reset) between two exposures, and a solution for this is described below.

[0055] Figure 3 A schematic diagram of a pixel with SFP readout circuitry is shown, where two substrates are joined together by capacitive hybridization, namely the photodiode substrate SPD (also called conversion substrate) and the readout circuit substrate SCL.

[0056] One electrode of the photodiode PD, in this case its anode, is connected to an electrode ET which, together with the electrode EL on the readout substrate side, forms a capacitor known as a CH hybrid capacitor, which is formed when the two substrates are joined together. The cathode of the diode receives a signal RAZ enabling it to be reset to zero between exposures, as we will see later.

[0057] On the readout side, the corresponding electrode EL of the hybrid capacitor is connected to the input of a so-called source follower amplifier SF, while a restore switch MRST, typically a simple transistor, can selectively connect the input of the amplifier SF to a reference voltage VREF. The amplifier output delivers the pixel signal SP.

[0058] When the restore switch MRST is turned on, the electrode EL of the hybrid capacitor CH is connected to the voltage VREF (ground or a fixed voltage source), and the photodiode PD and the hybrid capacitor CH together form a rectifying circuit.

[0059] In this configuration, if an AC or pulse signal RAZ is applied to the photodiode PD cathode during this time, a large amount of charge accumulates in the hybrid capacitor CH due to the asymmetric conduction of the photodiode junction. This amount of charge serves to reverse bias the photodiode in order to reset it.

[0060] For more details and reference Figure 4 and 5 , if a reset pulse RAZ is applied to the photodiode at the same time as the control pulse from transistor MRST is applied (see Figure 5 ), then the photodiode becomes forward biased, and becomes reverse biased after the pulse is removed.

[0061] If the reset transistor MRST is turned off immediately after the pulse RAZ has been removed from the substrate, the change in the photodiode potential (voltage Vpd) caused by the incident light is then transmitted to the source follower SF (voltage Vin) via the hybrid capacitor CH, and as photoelectric charges are formed in the photodiode PD, the voltage Vin gradually increases (starting from its starting point VREF).

[0062] At the output of the follower SF, the voltage Vs reproduces this evolution, except that the starting point is equal to VREF plus the offset voltage Vos of the follower SF. This voltage Vs is the pixel output signal SP.

[0063] It can be understood that the capacitive transfer gain depends on the ratio between the hybrid capacitance and the input capacitance of the voltage follower. If the input capacitance of the follower is low, then the gain will be very close to 1, and this largely compensates for the variation in capacitance of the hybrid capacitor CH.

[0064] Reference now Fig. 6A In a first embodiment, the application of the reset pulse RAZ to the SPD substrate can be achieved by connecting the pad CD directly to the substrate SPD. As shown in the figure, such a direct connection can be achieved by, for example, a wire bond on the face of the conversion substrate SPD exposed to light. Alternatively, one or more direct electrical contacts can be provided on the opposite faces of the substrate SPD and SCL. In this case, it is advantageous to provide multiple contacts to limit the risk of a faulty contact.

[0065] Reference now Figure 6B , the signal RAZ may alternatively be routed via a coupling capacitor CRAZ between the photodiode substrate SPD and the readout circuit substrate SCL. This capacitor is advantageously implemented with a large coupling area compared to the area of ​​the capacitor CH present at each pixel in order to provide a sufficient amount of charge for all pixels to reliably perform a reset.

[0066] Figure 6BAn advantage of this embodiment is that no direct electrical connection is required between the two substrates. Assembling the hybrid sensor then becomes a simple bonding operation, ensuring that the dielectric thickness (the adhesive itself) between the electrode pairs of each capacitor is reasonably uniform.

[0067] Additionally, there is no direct electrical contact between the photodiodes and their respective readout circuitry, reducing noise associated with direct contact, thereby improving image quality.

[0068] A variation of the above follower design is now described with the goal of increasing the operating dynamic range.

[0069] Recall that the reset pulse RAZ on the photodiode substrate allows the photodiode to be initialized at the beginning of exposure while the recovery pulse MRST acts on the readout circuit. However, if the light is too strong, then some photodiodes can be completely discharged, in which case the photocharge of the saturated photodiode will overflow to the adjacent photodiodes, causing a "blooming" phenomenon.

[0070] A well-known solution to this problem is to implement reverse-polarized PN junctions (AB or "anti-glare" junctions) around the photodiode. These reverse-polarized junctions absorb the photoelectric charge that overflows the saturated photodiode to limit the glare propagation. However, these AB junctions must remain reverse biased even when the pulsed RAZ is in effect.

[0071] One disadvantage of this solution is the reduced photoelectric conversion efficiency, since these AB junctions absorb a portion of the photoelectric charge generated by the photons. But one advantage of this solution is that it provides a logarithmic response after the photodiode is saturated. This logarithmic response gives the converter a dynamic range that can exceed 120dB.

[0072] refer to Figure 7 , we will now describe the implementation and operation of the present invention with a CTIA reader circuit.

[0073] In this release, Figure 1B The direct electrical contact between the photodiode and the CTIA amplifier is shown to be replaced not by a single capacitor but by two capacitors CH1 and CH2.

[0074] The first capacitor CH1 and Figure 3 and 4 The capacitor CH in plays the same role. A second capacitor is added to ensure the continuity of the feedback required for the amplifier operation.

[0075] When the amplifier AMP is restored via the switch PRST, a negative reset pulse is applied to the photodiode substrate, just as in the case of a pixel with a readout circuit SFP. Again, this pulse RAZ creates a reverse bias on the photodiode and then the switch MRST of the amplifier opens.

[0076] The feedback allows to keep the bias voltage on the photodiode constant and to transfer the accumulated charge during exposure to the integrating capacitor consisting of capacitor CH2.

[0077] Advantageously, the capacitors CH1 and CH2 consist of a single electrode ET on the substrate side of the photodiode and two adjacent electrodes EL1 and EL2 opposite the electrode ET on the readout substrate side.

[0078] The respective surface areas of capacitors CH1 and CH2 are determined by the desired amplification gain.

[0079] We will now describe how to achieve bonding between the substrate of the photodiode SPD and the readout circuit substrate SCL, thereby fixing them together. This bonding can be achieved in a variety of ways, for example using polymer adhesives, molecular bonding, solvents, etc.

[0080] If a polymer binder is used, e.g. Fig. 6A and 6B As schematically shown, additives can then be incorporated into the adhesive to increase its dielectric constant. A high dielectric constant increases the capacitance between the photodiode and the readout circuit, thereby improving the quality of signal transmission between the photodiode and the readout circuit. For example, we can use insights from the article "Improved dielectric permittivity of NBCTO / epoxy composite films with low dielectric loss", Yanli Su, Chengdong Ba, Fei Wang, SN Applied Sciences (2019), Springer Nature Switzerland AG.

[0081] In the case of bonding without the addition of adhesives, and as Figure 6C and 6D As shown, the substrate-side electrode of the photodiode and / or the substrate-side electrode of the readout circuit is coated with a thin layer of material, typically an oxide OX, intended to form the dielectric of the capacitor formed by mixing, the bonding being carried out by molecular bonding.

[0082] We will now describe the special case of a hybrid sensor according to the invention, in which the conversion photodiode is of the avalanche type.

[0083] Avalanche photodiodes are very useful for improving the sensitivity and response speed of optoelectronic systems such as single photon detection sensors or ranging sensors (LIDAR). As is known, in order to operate these avalanche photodiodes, very high bias voltages are required, for example, bias voltages of more than 100 V for silicon-based avalanche photodiodes. Such high bias voltages are potentially dangerous for readout circuits that can only withstand a few volts.

[0084] With a design that takes this into account, compatibility between avalanche photodiodes and readout circuitry is correct. But problems arise when one or more of the photodiodes are faulty during manufacturing or during use. High bias voltages can then reach the readout circuitry via uncontrolled and unpredictable paths, causing component damage. The capacitive hybrid of the present invention provides electrical isolation that effectively protects the readout circuitry from these high voltages by avoiding direct electrical contact between the photodiodes and their readout circuitry.

[0085] This means that both of the above designs with SFP and CTIA readout circuits, respectively, can be operated using avalanche photodiodes without any special adjustments.

[0086] Of course, the invention is in no way limited to the embodiments described and shown. In particular, the capacitive hybrid of the invention can be implemented with different types of readout circuits and with photoelectric conversion elements other than simple photodiodes.

Claims

1. An assembly of two electronic microcircuits joined together by providing a set of electrical couplings between respective connection points of said two microcircuits, the first microcircuit being a converter microcircuit having at least one charge accumulating photoelectric conversion element and the second microcircuit being a readout microcircuit having at least one element for reading the charge generated by the respective photoelectric conversion element, and said connection points comprising face-to-face surface elements separated by a dielectric, said coupling between said two circuits being a capacitive coupling.

2. The assembly according to claim 1, characterized in that The or each photoelectric capturing element is a photodiode having one electrode connected to a respective surface element on the first microcircuit and having another electrode connected to a reference potential.

3. The assembly according to claim 1 or 2, characterized in that The or each surface element on the second microcircuit is connected to the input of a reading element comprising an infinite or quasi-infinite DC input impedance circuit and an associated reset switch.

4. The assembly according to claim 3, characterized in that The circuit having infinite or quasi-infinite DC input impedance is a voltage follower circuit.

5. The assembly according to claim 1 or 2, characterized in that The or each surface element on the second microcircuit is connected to the input of a reading element comprising an infinite or quasi-infinite DC input impedance circuit and an associated reset switch.

6. The assembly according to claim 5, characterized in that The or each surface element on the second microcircuit comprises a first portion connected to the input of a capacitive transimpedance circuit and a second portion belonging to a feedback link, the first and second portions forming a single capacitor with a corresponding surface element of the first microcircuit.

7. An assembly according to any one of claims 1 to 6, characterised in that The first microcircuit comprises an array of charge-accumulating photoelectric conversion elements and the second microcircuit comprises a corresponding matrix of reading circuits of elements for reading the charges generated by the respective photoelectric conversion elements.

8. The assembly according to claim 7, characterized in that The assembly comprises at least one direct electrical connection between the first microcircuit and the second microcircuit for control signals common to the various matrix elements.

9. Assembly according to any one of claims 7 and 8, characterised in that The first microcircuit and the second microcircuit each comprise at least one additional surface element, two opposing surface elements providing additional capacitive coupling for control signals common to the various matrix elements.

10. The assembly according to claim 9, characterized in that The additional surface element surrounds the connection point.

11. Assembly according to any one of claims 9 and 10, characterised in that Said additional capacitive coupling is capable of carrying a signal for resetting said optoelectronic conversion element of said first microcircuit.

12. Assembly according to any one of claims 1 to 11, characterised in that The dielectric comprises a bonding adhesive between the two microcircuits.

13. The assembly according to claim 12, characterized in that The adhesive contains an additive that increases the dielectric constant.

14. Assembly according to any one of claims 1 to 10, characterised in that The dielectric comprises a layer of dielectric material joined to the two microcircuits by molecular bonding.

15. Assembly according to any one of claims 1 to 14, characterised in that The or each photo-capturing element comprises an avalanche photodiode, the capacitive coupling providing isolation from a bias voltage of the or each photodiode.

Images