Method of applying heat sink to semiconductor package, method of maintaining processor module in server, and data center system

Abstract

The disclosure relates to a method of applying a heat sink to a semiconductor package, a method of maintaining a processor module in a server, and a data center system. One or more sensors are integrated into a processor module that includes a semiconductor package and a heat sink. There are multiple processor modules in the server, and data from each processor module is sent to a digital integration monitoring system, which may be part of a data center hardware monitoring system. The server may be one of a number of servers located in the data center. The data may also be sent to a remote data center maintenance center that simultaneously monitors multiple data centers. When compared to different criteria, the data may be used to determine in real time whether a given processor module needs maintenance.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to a method of applying a heat sink to a semiconductor package, a method of maintaining a processor module in a server, and a data center system. Background Art

[0002] Various computer applications require the use of large numbers of computer processors. Examples of these applications include high-performance computing (HPC), cloud computing, data centers, and artificial intelligence. Maintaining the performance of computer processors is desirable. Summary of the Invention

[0003] Some embodiments of the present disclosure provide a method for applying a heat sink to a semiconductor package. The method includes applying a thermal interface material to the semiconductor package. The method includes aligning at least one pressure sensor between the semiconductor package and the heat sink. The method includes securing the heat sink to the semiconductor package using a fastener system. The method includes reading a signal from the at least one pressure sensor. The method includes comparing the signal from the at least one pressure sensor to a reference to determine whether setup quality is met. Furthermore, the method includes adjusting the fastener system until setup quality is met.

[0004] Some embodiments of the present disclosure provide a method for maintaining a processor module in a server. The method includes reading a signal from at least one sensor in the processor module. The method includes comparing the signal from the at least one sensor with a reference to determine whether maintenance is required. The method also includes alerting an operator when maintenance is required, or automatically performing maintenance.

[0005] Some embodiments of the present disclosure provide a data center system. The data center system includes at least one server, including multiple processor modules, each of which includes: a semiconductor package mounted on a motherboard; a heat sink located above the semiconductor package; a thermal interface material located between the semiconductor package and the heat sink; and at least one sensor located between the semiconductor package and the heat sink. Furthermore, the data center system includes a monitoring system configured to process data received from each sensor in the processor module and alert an operator when maintenance is required on one of the processor modules. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] The following detailed description, taken in conjunction with the accompanying drawings, will provide a better understanding of the concepts of the disclosed embodiments. It should be noted that, in accordance with standard industry practice, the various features in the drawings are not necessarily drawn to scale. In fact, the dimensions of various features may be arbitrarily enlarged or reduced to provide clarity of illustration.

[0007] Figure 1 is a cross-sectional view of an embodiment of a processor module mounted on a motherboard according to some embodiments of the present disclosure.

[0008] Figure 2 It is along Figure 1 A bottom view along line AA in FIG. 1 shows a heat sink with fastener holes and a sensor embedded in the heat sink.

[0009] Figure 3 A schematic diagram of a group of servers in a data center and connections to a real-time monitoring system for monitoring the health of processor modules in the data center.

[0010] Figure 4 is a flow chart illustrating a method of applying a heat spreader to a semiconductor package according to some embodiments.

[0011] Figure 5 This is a side cross-sectional view of a semiconductor package after it is mounted on a main board.

[0012] Figure 6 is a side cross-sectional view showing the alignment of the sensor and heat sink with the semiconductor package.

[0013] Figure 7 is a flow chart illustrating a method of maintaining a processor module in a server according to some embodiments.

[0014] The description of the accompanying drawings is as follows:

[0015] 100: Processor module

[0016] 110: Semiconductor packaging

[0017] 112: Surroundings

[0018] 120: System on Chip / Chip

[0019] 122: Wafer

[0020] 124: Substrate

[0021] 126: Memory chip

[0022] 128: Sealant

[0023] 130: First interconnection layer

[0024] 132: Second interconnection layer

[0025] 134: Bottom filling material

[0026] 136: Third interconnection layer

[0027] 138: Internal thermal interface material layer / first thermal interface material layer

[0028] 140: Adhesive

[0029] 142: Lid

[0030] 144: Motherboard

[0031] 145: Motherboard hole

[0032] 146: Backplate

[0033] 148: Back panel hole

[0034] 150: Fasteners

[0035] 151: Spring

[0036] 152: Second thermal interface material

[0037] 154: Radiator

[0038] 156: Fastener hole

[0039] 160, 162, 164, 166, 168: Sensors

[0040] 170: Data Center

[0041] 172: Server

[0042] 174: Server-level sensor / first environmental sensor

[0043] 176: (Real-time) Monitoring System

[0044] 178: Data Center Maintenance Center

[0045] 200: Method

[0046] 205,210,215,220,225,230,232,235,240,245,250,252,255,260,265,270: Step 300: Method

[0047] 305,310,315,320,325,330,335: Steps

[0048] AA: Line DETAILED DESCRIPTION

[0049] The following disclosure provides many different embodiments or examples for implementing different features of the embodiments of the present disclosure. Specific examples of components and configurations are described below to simplify the description of the embodiments of the present disclosure. Of course, these specific examples are for illustrative purposes only and are not intended to limit the embodiments of the present disclosure. For example, in the following description, it is mentioned that a first feature is formed on or above a second feature, which means that it may include an embodiment in which the first feature and the second feature are in direct contact, and may also include an embodiment in which an additional feature is formed between the first feature and the second feature, so that the first feature and the second feature may not be in direct contact. In addition, the present disclosure may repeat reference symbols and / or letters in various examples. This repetition is for the purpose of simplicity and clarity, and does not itself limit the relationship between the various embodiments and / or configurations described.

[0050] Furthermore, spatially relative terms, such as "below," "beneath," "lower," "above," "upper," and the like, may be used herein to describe the relationship of one element or feature to another element or feature depicted in the drawings. These spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the drawings. The device may be oriented differently (rotated 90 degrees or at other orientations), and the spatially relative terms used herein should be interpreted accordingly.

[0051] The numerical values ​​in the specification and claims of this application should be understood to include the same numerical values ​​when reduced to the same number of significant figures and the numerical values ​​that differ from the stated value by less than the experimental error of routine measurement techniques of the type described in this application to determine the value. All ranges disclosed herein include the stated endpoints.

[0052] The term "about" can be used to include any value that can vary without changing the basic function of the value. When used in conjunction with a range, "about" also discloses the range defined by the absolute values ​​of the two endpoints, for example, "about 2 to about 4" also discloses a range of "2 to 4". The term "about" can mean plus or minus 10% of the specified number.

[0053] The present disclosure relates to structures composed of different layers. When the terms "on..." or "above..." are used to refer to two different layers (including a substrate), they simply mean that one layer is on or above another layer. These terms do not require that the two layers are in direct contact with each other, and allow other layers to be located between the two layers. For example, all layers of a structure can be considered to be "on" a substrate, even if they are not all in direct contact with the substrate. The term "directly" can be used to indicate that two layers are in direct contact with each other without any layers in between. In addition, when reference is made to performing processing steps on or on a substrate, this should be understood as performing these steps on any layer that may be present on the substrate, depending on the context.

[0054] Embodiments of the present disclosure relate to processor modules with sensors and methods for processing data received from these sensors to improve device performance and reduce performance variations between these modules. In this regard, semiconductor dies can be packaged in a variety of different ways, such as package-on-package (PoP), in which two semiconductor packages are stacked on each other, or chip-on-wafer-on-substrate packaging, in which a semiconductor die is attached to a wafer (e.g., an interposer), which is then attached to a substrate (e.g., a printed circuit board). These may also be referred to as three-dimensional integrated circuit (3DIC) devices. A thermal interface material (TIM) is applied to the semiconductor die to improve thermal coupling with a heat sink. A fastener system is typically used to secure the heat sink in place relative to the die, such as by using screws at the corners of the heat sink. In some TIM processes, the die is exposed in the package and the TIM is applied between the die and the heat sink. In other TIM processes, the package includes a lid (which can act as a heat spreader), and the TIM is applied between the lid and the heat sink. During temperature cycling, TIM pump-out can occur due to package warping or uneven stress on the TIM due to the fastener system applying different forces on the heat sink. Differential stress can also cause die cracking. This can cause the die to stop functioning or performance to vary. In a data center containing multiple servers, each server contains multiple processor modules, which include semiconductor wafers or dies, which can result in high maintenance costs.

[0055] In an embodiment of the present disclosure, a processor module comprising one or more semiconductor dies further includes sensors, particularly pressure sensors and / or temperature sensors, that can provide information about one or more critical parameters. A method for monitoring the processor modules present in each server or data center to improve performance and / or reduce maintenance is also provided.

[0056] Figure 11 is a cross-sectional view of an embodiment of a processor module 100 used in servers and data centers of the present disclosure, mounted on a motherboard of a server. First, a semiconductor package 110 is shown. Here, the package includes a system-on-a-chip (SoC) 120, a wafer 122, and a substrate 124. On an SoC, many electronic components are combined on a common substrate. The SoC contains semiconductor dies, shown here as being located between two memory dies 126. Compared to the memory dies 126, the system-on-chip 120 generates a relatively large amount of heat. The SoC can include, for example, a central processing unit (CPU) or a graphics processing unit (GPU). The memory die can be, for example, a high bandwidth memory (HBM). Other electronic components may also be present. In general, any number of dies / dies may be present in the semiconductor package 110 and the processor module 100. The sides of the system-on-chip 120 may be surrounded by an encapsulant 128.

[0057] System-on-a-chip (SoC) 120 is bonded to the top surface of wafer 122. This can be accomplished, for example, via a first interconnect layer 130 comprising electrical contacts such as pads, balls, pins, bumps, pillars, or other similar structures. Wafer 122 can be an interposer substrate formed from a semiconductor substrate such as silicon. Active devices (e.g., transistors) and passive devices (e.g., resistors or capacitors) are sometimes formed on the surface of the wafer. Wafer 122 can also include through-vias.

[0058] The wafer 122 is then bonded to the top surface of the substrate 124 to obtain the semiconductor package 110, shown here as a chip-on-wafer-on-substrate package. Again, this can be accomplished by a second interconnect layer 132 comprising electrical contacts such as pads, balls, pins, bumps, columns or other similar structures. The substrate 124 can be, for example, a printed circuit board (PCB). The substrate 124 can also include other active or passive devices. The bottom fill material 134 is shown here between the system single chip 120 and the wafer 122, and also between the wafer 122 and the substrate 124. The back side of the substrate 124 also includes a third interconnect layer 136, which will be used to connect the semiconductor package to the host board.

[0059] Continuing, a thermal interface material (TIM) is deposited over the semiconductor package 110 to improve thermal coupling. This is referred to herein as an inner thermal interface material layer 138. Suitable TIMs may include polymers that may contain thermally conductive fillers. Some non-limiting examples of thermally conductive fillers may include alumina, boron nitride, aluminum nitride, aluminum, copper, silver, and indium. The TIM may be a film or sheet, including, for example, carbon nanotubes (CNTs) or graphite. The TIM may be in the form of a solid pad, paste, gel, grease, or phase change material, among others. The TIM may be applied continuously over the package. Sometimes, there may be voids or air gaps within the TIM layer, such as between the system-on-chip 120 and the memory chip 126, to reduce lateral thermal interactions. In some embodiments, the thickness of the TIM layer may be in the range of 100 microns (μm) to about 3 millimeters (mm), but other ranges are also within the scope of the present disclosure.

[0060] An adhesive 140 is also disposed on the substrate 124 and surrounds the system-on-chip 120. The adhesive 140 may be, for example, epoxy, silicone, glue, or other adhesives suitable for use with semiconductor devices.

[0061] Lid 142 is attached to substrate 124 and positioned over semiconductor package 110. Lid 142 both physically protects semiconductor package 110 and acts as a heat sink, dissipating heat generated by the SoC over the larger surface area of ​​the lid. The lid is typically made of a material with high thermal conductivity, such as aluminum, steel, stainless steel, copper, or other similar materials. Lid 142 is secured to the substrate via adhesive 140. The adhesive and TIM may need to be heated at an appropriate temperature for an appropriate amount of time to cure.

[0062] Continuing, semiconductor package 110 is then mounted on motherboard 144. A backplate 146 may be present beneath motherboard 144. Backplate 146 provides holes that engage fasteners 150 to secure heat sink 154 in place. Backplate 146 may also provide additional support for motherboard 144. Backplate 146 may be made of or coated with a non-conductive material to prevent short circuits from occurring on the back of motherboard 144.

[0063] A second thermal interface material (TIM) layer 152 is placed on the lid 142, which thermally couples the lid 142 to the heat sink 154. This layer can have the same composition and form as described for the inner thermal interface material layer 138.

[0064] The heat sink 154 is placed above the second TIM layer. The heat sink 154 can be made of a material such as aluminum or copper. The heat sink 154 can include fins to increase the surface area. Figure 2As seen in the bottom plan view of FIG. 1 , the heat sink 154 also includes four fastener holes 156 through which fasteners 150, such as spring-loaded screws (springs 151 are also shown), are passed. The screws also pass through the motherboard 144 and engage the backplane holes 148 (see FIG. 1 ). Figure 5 It should be noted that multiple semiconductor packages 110 can be mounted on one host board. However, each semiconductor package 110 will have its own separate heat sink.

[0065] Now refer to Figure 1 and Figure 2 , at least one sensor is also located between the semiconductor package 110 and the heat sink 154. Multiple pressure sensors are shown here. There are five pressure sensors 160, 162, 164, 166, 168, with one sensor located at each corner of the package along the perimeter and one sensor 168 located in the center of the package. Figure 1 As shown, these sensors are embedded in the heat sink 154 and the second thermal interface material layer 152. Figure 2 , the perimeter 112 of the package is shown in phantom. Fastener holes 156 are located outside the perimeter of the package and are located at the corners of the heat sink 154.

[0066] In some specific embodiments, the sensor is configured to measure pressure, force, displacement, stress, strain, temperature, or a combination thereof. In some embodiments where the sensor measures pressure or force, the sensor can be a pneumatic pressure sensor, a hydraulic sensor, a strain gauge, or a capacitive pressure sensor. As previously described, uneven pressure distribution can lead to uneven heat dissipation during TIM pumping and operation, thereby affecting CPU / GPU performance. Displacement sensors can be used to measure horizontal and / or vertical movement, which is related to assembly quality. Displacements as small as 200 to 300 microns can cause contact problems, resulting in package burnout or other damage. Strain sensors can measure the degree to which the package is bent relative to the substrate (which can be the motherboard) and can be used to identify incorrect assembly or die cracking. Temperature sensors can be used to confirm whether heat transfer occurs as expected, and unexpected temperature curves may indicate a problem.

[0067] In general, the sensors may be located on the package in any desired arrangement. For example, the sensors may be located in the center 114 of each side of the package (see Figure 2Any combination of sensors can be used to measure different properties. In some specific embodiments, the at least one sensor includes at least one pressure sensor and at least one temperature sensor. These sensors can measure key performance indicators (KPIs) in real time to ensure pressure uniformity, TIM bond line thickness, and safe stress levels on the package. This helps maintain stable package performance from installation to end of life.

[0068] In other embodiments, the sensors may be embedded in the cover 142. Alternatively, some sensors may be embedded in the heat sink and some sensors may be embedded in the cover.

[0069] The combination of semiconductor package 110 and heat sink 154 is referred to herein as processor module 100. It should be noted that the present disclosure also extends to bare die packages that do not include a lid and thus have only one TIM layer instead of two, and such combinations are also considered processor modules within the scope of the present disclosure.

[0070] Although not shown, each processor module may also include internal cache, memory, input / output controllers, buses for transferring data, and other similar components. Communication channels may include system buses, network connections, wired and wireless systems. Each processor module may include one or more cores. Each processor module executes instructions according to software / programming as needed.

[0071] Continue, as Figure 2 As shown, each sensor (160, 162, 164, 166, 168) is then configured to transmit data to a monitoring system 176. In other words, the monitoring system 176 reads or receives a signal from each sensor. The monitoring system 176 uses this data to monitor the condition of each processor module.

[0072] Now refer to Figure 3 , a data center 170 (shown by a dashed line) includes a set of servers 172, here labeled Server 1 through Server n. Each server acts as a computing node within the data center. Each server 172 contains multiple processor modules 100, which for illustrative purposes are labeled here as GPU-1 through GPU-n or CPU-1 through CPU-n. This forms a high-performance computing (HPC) environment. Generally speaking, sensors are located in the same location in each processor module so that the data collected from them can be compared with each other in a reliable manner.

[0073] Generally speaking, these servers are interconnected through one or more layers of network switches and routers. Not shown here are power supplies, switches, routers, hubs, gateways, firewalls, intrusion detection / prevention devices, computer terminals, printers, memory / storage devices, modems, access points, fire detection and fire suppression systems, wiring, input / output devices, fans, etc. Each server can have server-level sensors 174 that are present in the server rather than in the processor module. In some specific embodiments, the server-level sensors 174 may include one or more environmental sensors configured to measure temperature, humidity, or moisture. In this regard, comparing the server-level temperature with the processor module-level temperature can provide information about the overall heat transfer efficiency or whether a server device needs maintenance (e.g., a cooling fan). The presence of moisture or humidity may indicate a leak in the liquid cooling system and may also cause electrical problems due to leakage current. Monitoring of these sensors can also improve the maintenance of the processor modules within a given server.

[0074] Each sensor in the processor modules 100 of these servers sends or transmits data to a real-time monitoring system 176 within the data center 170. This can be done wired or wirelessly, depending on the setup. The monitoring system 176 receives the data and uses or processes the data to determine whether any particular processor module or multiple processor modules require maintenance. The monitoring system may also include a user interface for communicating with an operator. The monitoring system can be implemented as one or more general-purpose computers, special-purpose computers, programmed microprocessors or microcontrollers and peripheral integrated circuit components, ASICs or other integrated circuits, digital signal processors, hardwired electronic or logic circuits (e.g., discrete component circuits), programmable logic devices (e.g., PLDs, PLAs, FPGAs, graphics card CPUs (GPUs), or PALs), etc. Such devices typically include at least one memory (e.g., RAM, ROM, EPROM) for storing a control program and a processor for executing the control program.

[0075] Data from the sensors can also be transmitted to a data center maintenance center 178. It is contemplated that the data center maintenance center 178 receives sensor data from multiple data centers. The data center maintenance center 178 can be physically located in the data center or at a separate location. The sensor data can be used for a variety of applications. These applications can include establishing benchmarks and identifying correlations that can improve efficiency.

[0076] Figure 4 is a flow chart illustrating a method 200 of applying a heat sink to a semiconductor package according to some embodiments. Some steps of the method are also shown in FIG. Figure 5 and Figure 6The method steps are discussed below with respect to forming a single processor module on a motherboard, and should be understood to be applicable to multiple processor modules on a motherboard as well.

[0077] First, in Figure 4 In step 205, Figure 5 As shown, a semiconductor package 110 is received. Here, the semiconductor package 110 includes a system-on-chip 120, a wafer 122, and a substrate 124. Again, a lid 142 may be present, but is not required.

[0078] Alternatively, semiconductor package 110 may be formed. In step 210, die 120 is attached to wafer 122. In step 215, wafer 122 is attached to substrate 124. These steps will result in a bare die package with a TIM applied.

[0079] If a lid is required, in optional step 220, a thermal interface material (TIM) is applied to wafer 120 to form a first thermal interface material layer 138. In optional step 225, adhesive 140 is placed around the perimeter of substrate 124. The order of forming the adhesive and forming the first thermal interface material layer can be reversed. Then, in optional step 230, lid 142 is attached to substrate 124 using adhesive 140. Lid 142 also contacts first thermal interface material layer 138. This produces semiconductor package 110 as shown, which is shown in step 232.

[0080] Continue, in Figure 4 In step 235, the semiconductor package 110 is then mounted to the sockets of the host board 144. In this regard, the host board 144 may have one or more sockets. In some embodiments, there may be 1 to 32 sockets on the host board, but other ranges are also within the scope of the present disclosure. Next, in Figure 4 In step 240, a thermal interface material (TIM) is applied to the semiconductor package 110 to form a second thermal interface material layer 152. This layer can be formed on the wafer 120 or on the lid 142. The structure after these steps is as shown in FIG. Figure 5 The motherboard holes 145 and the back panel holes 148 are also visible here.

[0081] Next, in Figure 4 In step 245 and as Figure 6 As shown, at least one sensor 160 is aligned between the semiconductor package 110 and the heat sink 154. Figure 6As shown, the sensor 160 is aligned with the heat sink 154 by placing the sensor 160 in the hole in the heat sink 154. The fastener 150 passes through the hole in the heat sink 154 and is shown here as being partially located in the host board hole 145 in the host board 144. This aligns the sensor 160 with the semiconductor package 110. Thus, the sensor 160 is fixed in place relative to the semiconductor package 110 and the heat sink 154. It should be noted that the sensor 160 does not have to directly contact both the semiconductor package 110 and the heat sink 154. Figure 1 As shown, sensors 160 are in direct contact with heat sink 154 and are embedded in second thermal interface material layer 152. Here, at least one of the sensors is a pressure sensor. Although not shown here, sensors 160 are configured so that their output can be read. In certain embodiments, sensors 160 are connected to monitoring system 176 via physical wiring or wirelessly. This is shown in step 252, where data is transmitted in real time to a continuous monitoring system.

[0082] Next, in step 250, the heat sink 154 is secured to the semiconductor package 110 using a fastener system. Figure 6 and Figure 2 As shown, a total of four fasteners 150 (e.g., spring-loaded screws) are shown located at each corner of the heat sink 154. In step 255, as the fasteners 150 are tightened to hold the heat sink 154 against the second thermal interface material layer 152, the signals from the pressure sensor(s) are read.

[0083] exist Figure 4 In step 260, the signal from the sensor(s) 160 is compared to a reference to determine whether setup quality is met. In this regard, each fastener 150 may have some tolerance. Typically, when fasteners are tightened manually, a separate torque or force is applied to each fastener, which may also vary depending on the individual performing the tightening. This may result in deviations in the force applied by each fastener to lock the semiconductor package and heat sink to the host board. The different forces at each location may cause the package to tilt. One application of the monitoring system 176 is to improve the uniformity of the pressure applied to the semiconductor package 110.

[0084] Return to reference Figure 2, five pressure sensors 160, 162, 164, 166, 168 are shown. During installation, the benchmark can be whether the pressure is at a specific target pressure value. Alternatively, the benchmark can be whether the pressure is within a target pressure range, such as from about 10 psi (pounds per square inch) to about 40 psi. In some embodiments, the benchmark is the same for each sensor, regardless of its location. However, this may depend on the package size and the planarity of the cover. For example, if the cover 142 is curved, the benchmark of the center sensor 168 may be higher than the benchmarks of the four corner sensors 160, 162, 164, 166. Each individual pressure sensor can be considered to have its own benchmark, and the overall system benchmark is met only when each individual benchmark is met.

[0085] refer to Figure 4 At step 265, if the benchmark is not met, the individual fasteners 150 are adjusted to change the force they apply. This adjustment continues until the set quality is met, or in other words, the benchmark is met. At step 270, the semiconductor package containing one or more CPUs or GPUs is then operated. The resulting structure is as shown in FIG. Figure 1 shown.

[0086] Figure 7 is a flow chart illustrating a method 300 for maintaining a processor module in a server according to some embodiments. Some steps of the method are also shown in FIG. Figure 5 and Figure 6 The following method steps are discussed for a single sensor in a single processor module during its operational life after installation, and should be interpreted as also applicable to multiple sensors in a given processor module and across multiple processor modules. Generally speaking, the monitoring system issues an alarm when maintenance is required. This could occur, for example, due to loose fasteners caused by aging springs, pumping of thermal interface material, or cracked die. Different conditions can be detected by different sensors.

[0087] exist Figure 7 In step 305, a signal from a sensor in the processor module is read. In step 310, the signal from the sensor is compared with at least one reference to determine whether maintenance is required. The comparison can be performed, for example, by the monitoring system 176.

[0088] The benchmark can vary depending on, for example, the location of the sensor and the type of sensor. For example, for a pressure sensor, the benchmark can be whether the pressure is within a target pressure value or a target pressure range. For a displacement sensor, the benchmark can be whether there is an unacceptably high change in length or an unacceptably high change in position. For a temperature sensor, the benchmark can be whether the measured temperature is within a set temperature range. In embodiments where multiple sensors of the same type are present, the benchmark can be related to the difference between the sensor data at different locations. For example, the benchmark can be whether the temperature difference between a sensor located in a corner of the package and a sensor located in the center of the package is within a set temperature range. As another example, the benchmark can be whether the displacement between a sensor located in a corner of the package and a sensor located in the center of the package is within a set range (which may indicate TIM pumping). The benchmark can also be the change in value measured by the same sensor over a specific time period. For example, a sudden change in the value measured by a particular strain sensor may indicate die cracking.

[0089] The benchmark against which the sensor signal is compared need not be static and may change over time. For example, a temperature benchmark could be whether the measured temperature of a given processor module is within one, two, or three standard deviations of the average measured temperature of all processor modules in a given server. This would identify whether a processor module is abnormal and should be inspected by maintenance personnel.

[0090] If maintenance is required, the monitoring system can perform at least two different actions. First, Figure 7 In step 325, the monitoring system can alert the operator that maintenance is required. This can be done, for example, by activating a visual alarm and / or an audio alarm, generating an error message, or sending a text message. The output of the monitoring system can provide information about which processor module may need maintenance, which sensor has been triggered, which benchmark has been violated, etc. Figure 7 In step 330, the monitoring system can automatically perform maintenance. For example, if the monitoring system indicates that a spring-loaded screw on a heat sink needs maintenance because it has become too loose, the monitoring system can instruct a robotic screwdriver to tighten the loose screw until the sensor signal falls within a reference range. As another example, if the temperature of a processor module is too high, the monitoring system can shut off power to the processor module or increase the coolant flow rate to the processor module.

[0091] Similarly, if Figure 7As shown in step 315 in , a signal from the first environmental sensor 174 within the server can also be read. In step 320, the signal from the first environmental sensor 174 is compared to at least one benchmark to determine whether maintenance is required. In steps 325 and 330, if maintenance is required, the monitoring system can alert the operator or perform automatic maintenance. It is expected that these steps will be performed at the server level rather than at the level of each individual processor module, although this is not required. For example, if the temperature measured by the environmental sensor is much higher than the average temperature of the processor modules in the server, this may indicate a problem with the server's cooling system. As another example, a measured humidity value that is much higher than the humidity of the surrounding environment may indicate a leak in the liquid cooling system used in the server. In step 335, if maintenance is not required and all benchmarks are met, monitoring continues.

[0092] The monitoring system can also use data obtained from sensors on multiple servers to identify potential maintenance issues. For example, if the average temperature of the processor modules in server A is significantly higher than the average temperature of the processor modules in server B, this may indicate a maintenance issue in server A. Similar benchmarking can occur in Figure 3 A data center maintenance center 178 where data is compared between different servers and different data centers to identify potential maintenance issues and identify operational improvements that can be made.

[0093] An overall goal of the disclosed embodiments is to improve data center efficiency. This can be measured, for example, in terms of power required to achieve the same cooling efficiency or wafer temperature at the same power input. Consequently, servers and data centers that utilize processor modules with sensors to measure one or more parameters, particularly pressure and / or temperature, improve thermal performance, increase reliability, and reduce defects.

[0094] Therefore, some embodiments of the present disclosure relate to a method for applying a heat sink to a semiconductor package. The method includes applying a thermal interface material to the semiconductor package to form a thermal interface material layer. The method includes aligning at least one pressure sensor between the semiconductor package and the heat sink. The method includes securing the heat sink to the semiconductor package using a fastener system. The method includes reading a signal from the at least one pressure sensor. The method includes comparing the signal from the at least one pressure sensor to a reference to determine whether a setup quality is met. In addition, the method includes adjusting the fastener system until the setup quality is met.

[0095] In some embodiments, the at least one pressure sensor is configured to transmit data wirelessly. In some embodiments, the at least one pressure sensor comprises a plurality of pressure sensors spaced along a perimeter of the semiconductor package. In some embodiments, the at least one pressure sensor comprises a pressure sensor located above the center of the semiconductor package. In some embodiments, the at least one pressure sensor is embedded in a heat sink or in a lid of the semiconductor package. In some embodiments, the at least one pressure sensor is a pneumatic pressure sensor, a hydraulic pressure sensor, a strain gauge, or a capacitive pressure sensor.

[0096] Some other embodiments disclosed herein relate to a method for maintaining a processor module in a server. The method includes reading a signal from at least one sensor in the processor module. The method includes comparing the signal from the at least one sensor to a reference to determine whether maintenance is required. The method also includes alerting an operator when maintenance is required, or automatically performing the maintenance.

[0097] In some embodiments, the method further comprises: reading a signal from a first environmental sensor located within the server, wherein the first environmental sensor is configured to measure temperature, humidity, or moisture; comparing the signal from the first environmental sensor to a reference to determine whether maintenance is required; and alerting an operator when maintenance is required, or automatically performing maintenance. In some embodiments, the at least one sensor is configured to measure pressure, force, displacement, stress, strain, temperature, or a combination thereof. In some embodiments, the at least one sensor comprises a pressure sensor and a temperature sensor.

[0098] Some other embodiments herein also describe a data center system. The data center system includes at least one server, including multiple processor modules, each processor module including: a semiconductor package mounted on a motherboard; a heat sink located above the semiconductor package; a thermal interface material located between the semiconductor package and the heat sink; and at least one sensor located between the semiconductor package and the heat sink. Furthermore, the data center system includes a monitoring system configured to process data received from each sensor in the processor module and alert an operator when one of the processor modules requires maintenance.

[0099] In some embodiments, the at least one sensor in each processor module is configured to measure pressure, force, displacement, stress, strain, temperature, or a combination thereof. In some embodiments, the at least one sensor in each processor module includes a pressure sensor and a temperature sensor. In some embodiments, the at least one sensor in each processor module is configured to transmit the data wirelessly. In some embodiments, the at least one sensor in each processor module includes a plurality of pressure sensors spaced along a perimeter of the semiconductor package. In some embodiments, the at least one sensor in each processor module is embedded in a heat sink or in a lid of the semiconductor package. In some embodiments, the at least one sensor in each processor module is a pneumatic pressure sensor, a hydraulic pressure sensor, a strain gauge, or a capacitive pressure sensor. In some embodiments, the data center system further includes a first environmental sensor located within the at least one server, the first environmental sensor configured to measure temperature, humidity, or moisture. In some embodiments, the at least one server includes a plurality of servers, each server transmitting data to a monitoring system. In some embodiments, the data center system is further configured to transmit the data to a data center maintenance center.

[0100] The above summarizes the features of many embodiments so that those skilled in the art to which the present disclosure belongs can better understand the various embodiments of the present disclosure. Those skilled in the art to which the present disclosure belongs should understand that other processes and structures can be easily designed or changed based on the embodiments of the present disclosure to achieve the same purposes and / or achieve the same advantages as the embodiments introduced herein. Those skilled in the art to which the present disclosure belongs should also understand that these equivalent structures do not depart from the concept and scope of the present disclosure. Various changes, substitutions and modifications can be made to the embodiments of the present disclosure without departing from the concept and scope of the appended claims.

Claims

1. A method of applying a heat sink to a semiconductor package, the method comprising: applying a thermal interface material to the semiconductor package; aligning at least one pressure sensor between the semiconductor package and the heat sink; securing the heat sink to the semiconductor package using a fastener system; reading a signal from the at least one pressure sensor; comparing the signal from the at least one pressure sensor with a reference to determine whether a setting quality is met; and The fastener system is adjusted until the set quality is met. 2 . The method of claim 1 , wherein the at least one pressure sensor comprises a plurality of pressure sensors spaced apart along a perimeter of the semiconductor package. 3 . The method of claim 1 , wherein the at least one pressure sensor comprises a pressure sensor located over a center of the semiconductor package. 4 . The method of claim 1 , wherein the at least one pressure sensor is embedded in the heat sink or in a lid of the semiconductor package.

5. A method for maintaining a processor module in a server, the method comprising: reading a signal from at least one sensor in the processor module; comparing the signal from the at least one sensor to a reference to determine whether maintenance is required; as well as Alerts operators when maintenance is required, or performs maintenance automatically.

6. The method of claim 5, further comprising: reading a signal from a first environmental sensor located within the server, wherein the first environmental sensor is configured to measure temperature, humidity, or moisture; comparing the signal from the first environmental sensor to another reference to determine whether maintenance is required; as well as Alerts operators when maintenance is required, or performs maintenance automatically.

7. A data center system, comprising: At least one server includes a plurality of processor modules, each of the plurality of processor modules includes: a semiconductor package mounted on a motherboard; a heat sink located on the semiconductor package; a thermal interface material located between the semiconductor package and the heat sink; and at least one sensor located between the semiconductor package and the heat sink; and A monitoring system is configured to process data received from each of the sensors in the plurality of processor modules and alert an operator when one of the plurality of processor modules requires maintenance. 8 . The data center system of claim 7 , further comprising a first environmental sensor located within the at least one server, the first environmental sensor being configured to measure temperature, humidity, or moisture. 9 . The data center system as claimed in claim 7 , wherein the at least one server comprises a plurality of servers, and each of the plurality of servers sends data to the monitoring system.

10. The data center system as claimed in claim 7, wherein the data center system is further configured to send the data to a data center maintenance center.

Images