Integrated Electrical Test Vehicle Co-designed with Microfluidics for Evaluating the Performance of Embedded Cooling

abstract

从信息时代开始，热管理技术在电子系统的持续小型化、性能改进和更高可靠性方面发挥了关键作用。与传统冷却相比，嵌入式冷却的主要优点是减少了热源和散热器之间的热接口数量。大量的研究工作致力于通过嵌入式冷却提高传热效率。然而，缺乏通用的测试平台来表征嵌入式结构的热性能。在这项工作中，提出了一种与嵌入式冷却共同设计的集成电气测试车，用于评估嵌入式微流控结构的冷却性能，该测试器的外形尺寸小于30 x15x2 mm3。它为分析嵌入式冷却的性能提供了一种工具，并为进一步研究嵌入式冷却机制提供了一种方法。

From the dawn of the Information Age thermal management technology has played a key role in the continuing miniaturization, performance improvements, and higher reliability of electronic systems. The key advantage of embedded, versus conventional cooling is a reduction in the number of thermal interfaces between the heat source and the heat sink. Considerable research efforts have been devoted to improving the heat transfer efficiency through embedded cooling. However, there is a lack of a universal test platform to characterize the thermal performance of the embedded structure. In this work, an integrated electrical test vehicle co-designed with embedded cooling is proposed for evaluating the cooling performance of the embedded microfluidic structure and the overall dimension of the vehicle is under 30 x15x2 mm3 mm3. It provides a tool for analyzing the performance of embedded cooling and proposes a method for further study of embedded cooling mechanism.

introduction

随着特征尺寸的减小和集成度的增加，集成电路面临着严重的热挑战。

由于增加了栅极密度以及在单个芯片上制造多核，芯片性能得到了改善。然而，无法遵循电源电压的Dennard定标导致芯片热通量耗散密度高，尤其是在过去十年[1-3]。我们将关注由纳米电子驱动的当前热管理要求，这使封装工程师同时面临高功率、“热点”和3D集成的“三重威胁”[4]。

为了有效传热，可以通过将冷却液循环通过蚀刻在芯片背面的微通道（也称为嵌入式或直接芯片背面微通道冷却）来减少净热阻和热扩散[5]。它满足了应用工程领域对低热阻、低压降和优良传热系数的要求。

与传统的热管理方法相比，嵌入式冷却可以将系统级热阻降低8倍[6]。在微流控结构[7,8]、流形结构[9,10]、相变机制[11]方面进行了大量研究，以优化嵌入式冷却的性能。然而，即使在最先进的方法中，嵌入式冷却中的热电协同设计知识仍然非常有限，并且缺乏有效的分析工具来评估嵌入式结构的冷却性能。本文提出了一种与微流体共同设计的集成电气测试车，用于评估嵌入式冷却的热行为。测试车由带加热单元阵列的热测试芯片和PCB模块组成。在热测试芯片基板上制作嵌入式微通道结构，并通过测试芯片和PCB模块的组装形成闭合的电气和流量测试回路。实验结果表明，测试车可以评估嵌入式结构的冷却性能，并分析最佳工作条件。此外，它为进一步研究嵌入式冷却的传热机理提供了有用的分析工具。

As the feature size decreases and integration degree increases, integrated circuits face a serious thermal challenge.

Chip performance has improved due to increasing gate density as well as fabrication of multiple cores on a single chip. However, the inability to follow the Dennard scaling for supply voltage led to high chip heat flux dissipation densities, especially during the last decade [1-3]. Attention will be devoted to current thermal management requirements, driven by nano-electronics, which confront packaging engineers with the simultaneous “triple threat” of high-power, “hotspots” and 3D integration [4].

For efficient heat transfer, the net thermal resistance and heat spreading can be reduced by circulating the coolant through microchannels etched into the backside of the chip (also termed as embedded or direct chip backside microchannel cooling) [5]. It fulfills the requirements for low thermal resistance, low pressure drop, and excellent heat transfer coefficient in the field of applied engineering.

Compared to the conventional thermal management approaches, embedded cooling can reduce the system level thermal resistance by up to 8x [6]. Considerable research efforts have been worked on microfluidic structures [7, 8], manifold structure [9, 10],phase transformation mechanism [11] to optimize the performance of embedded cooling. However, even in state-of-the-art approaches, the knowledge of the thermal and electrical co-design in embedded cooling is still very limited, and there is a lack of effective analysis tools to evaluate the cooling performance of the embedded structure. In this paper, an integrated electrical test vehicle co-designed with microfluidics is proposed to evaluate the thermal behavior of embedded cooling. The test vehicle is composed of a thermal test chip with heating unit array and PCB module. Embedded microchannel structure is fabricated on the thermal test chip substrate, and a closed electrical and flow test loop is formed through the assembly of the test chip and PCB module. The experiment results indicate that the test vehicle can evaluate the cooling performance of embedded structure and analyze the optimal working conditions. Moreover, it provides a useful analysis tool for further study of the heat transfer mechanism of embedded cooling.

EXPERIMENTAL SET-UP A.

Concept of Integrated Electrical Test Vehicle

传统背面连接的微流控散热器通过热界面材料（TIM）与芯片热连接，该材料与基板和散热器底座的厚度相结合，导致热扩散和高热阻。相比之下，嵌入式冷却可以通过在基板中直接制造微流控结构来有效地消除界面热阻（ITR），并缩短热点和冷却剂之间的传热路径[12]。在这项工作中，提出了一种与微流体共同设计的集成电气测试车。测试车由热测试芯片和PCB模块组成。

嵌入式微流控结构蚀刻在热测试芯片的基板上，冷却液可以直接引入芯片底部，与热测试芯片中的热点进行热交换，如图1所示。深度反应离子刻蚀工艺（DRIE）可以实现各种微流控结构的制造，如微通道、微柱、微槽等。

Conventional backside attached microfluidic heat sinks are thermally joined to the chip through a thermal interface material (TIM) which combined with the thickness of substrate and heat sink base, results in heat spreading and high thermal resistance. In contrast, embedded cooling could effectively remove the interfacial thermal resistance (ITR) by directly fabricating the microfluidic structure in the substrate and shorten the heat transfer path between the hotspot and the coolant [12]. In this work, an integrated electrical test vehicle co-designed with microfluidics was proposed. The test vehicle is composed of thermal test chip and PCB module.

The embedded microfluidic structure is etched on the substrate of the thermal test chip, and the cooling liquid can be directly introduced into the bottom of the chip to exchange heat with hot spots in the thermal test chip, as shown in Fig.1. Deep reactive ion etching process (DRIE) can realize the manufacture of different kinds of microfluidic structures, such as microchannels, micropillars, microgrooves, etc.

The main function of the PCB module consists of two parts: (1) supplying power to the thermal test chip and reading the signal; (2) introducing the coolant into the embedded structure. Fig.2 plots its working principle and assembly method. The thermal test chip is assembled on the top layer of PCB. The bonding region are designed on the bottom of thermal test chip and top surface of PCB for filling solder to ensure the soldering strength. In addition, the power supply and signal readout circuits of the test chip are also designed on the top layer of PCB, and the electrical connection between chip and PCB is established through the wire bonding process.

In the process of chip design, the coolant is usually introduced directly into the embedded structure at the bottom of the chip to optimize the size of the embedded cooling chip. Therefore, the coolant delivery channels are machined on the bottom layer of the PCB model. Here, appropriate substrate material can be selected according to the test requirements. Visible polymethyl methacrylate substrate (PMMA) materials can be used to observe boiling phenomena during embedded cooling, on the other hand conventional PCB materials can be selected for reliable electronical expansibility. A closed flow channel is formed by the assembly of the top PCB and the bottom substrate, which connect the coolant I/O interface and the thermal test chip, as shown in Fig.2(b). The glue bonding method is adapted to effectively resist the leakage of coolant.

Moreover, two negative temperature coefficient (NTC) thermistors are designed in the upstream and downstream channels to accurately measure the changes in coolant temperature during the cooling process, as shown in Fig.2(c).

The integrated thermal and electrical co-design test vehicle with microfluidics structure is an open platform with multi-variability and testability. The overall size of the multi-dimensional test vehicle is less than 30×15×2 mm3, meeting the rapidly increasing optimization requirements for the size, weight, power, and cost (SWaP-C) of electronics

PCB模块的主要功能包括两部分：（1）为热测试芯片供电并读取信号；（2）将冷却剂引入嵌入式结构。图2描绘了其工作原理和装配方法。热测试芯片组装在PCB的顶层。焊接区设计在热测试芯片的底部和PCB的上表面，用于填充焊料，以确保焊接强度。此外，测试芯片的电源和信号读出电路也设计在PCB的顶层，芯片和PCB之间的电气连接是通过引线键合过程建立的。

在芯片设计过程中，通常将冷却剂直接引入芯片底部的嵌入式结构中，以优化嵌入式冷却芯片的尺寸。因此，在PCB模型的底层加工冷却液输送通道。这里，可以根据测试要求选择合适的基板材料。可见聚甲基丙烯酸甲酯基板（PMMA）材料可用于观察嵌入冷却过程中的沸腾现象，另一方面，可以选择常规PCB材料以实现可靠的电子膨胀性。如图2（b）所示，由顶部PCB和底部基板组装而成的闭合流道连接冷却剂I/O接口和热测试芯片。胶接方法适用于有效防止冷却液泄漏。

此外，在上游和下游通道中设计了两个负温度系数（NTC）热敏电阻，以准确测量冷却过程中冷却液温度的变化，如图2（c）所示。

具有微流体结构的集成热电协同设计测试车是一个具有多变性和可测试性的开放平台。多维测试车辆的总体尺寸小于30×15×2 mm3，满足快速增长的电子设备尺寸、重量、功率和成本（SWaP-C）优化要求

Design and Fabrication of thermal test chip

提出的热测试芯片旨在再现大多数垂直功率器件的热行为[13]。在这项工作中，设计并制造了具有多个加热单元的热测试芯片，如图3（a）所示。热测试芯片的总体尺寸为6㼿7.㼿0.4 mm3，分为4㼿5阵列1㼿该区域中央交叉处的1 mm2平方单位。每个单元包含多个平行的硅电阻条，可以模拟局部热点的加热行为，芯片尺寸如表所示。1、与铂微加热器相比，并联沉积硅电阻片具有更高的温度稳定性和良好的工艺兼容性，可以作为一种低成本的微热源解决方案。

加热阵列单元的柔性连接可以通过引线键合实现，每个单元可以用作热点，根据测试需要进行组合。不同尺寸、形状和相对位置的热点在嵌入式冷却过程中的传热效应

在这项工作中，多个平行微通道被设计为嵌入式微流控结构，位于加热单元阵列下方，如图3（b）所示。表1显示了结构的详细尺寸。与其他微流控结构相比，微通道具有更好的层流特性，确保了冷却液在热交换过程中的热稳定性。

上述结构是在热测试芯片的基板上通过DRIE工艺处理的，图3（b）中的扫描电子显微镜（SEM）图像显示了芯片的实际形貌。结构周围的一部分区域保留为键合环，以便于PCB和芯片之间的键合。

the proposed of thermal test chip is designed to reproduce the thermal behavior of most vertical power devices [13]. In this work, the thermal test chip with multi-heating units is designed and fabricated, as shown in Fig.3(a). The overall size of thermal test chip is 6㼿7㼿0.4 mm3, which divided into a 4㼿5 array of 1㼿1 mm2 square units in the central cross of the region. Each unit contains multiple parallel silicon resistance strips, which can simulate the heating behavior of local hotspots, the dimensions of the chip are shown in Table.1. Compared with the Pt micro-heater, the deposit silicon resistance strips in parallel has higher temperature stability and excellent process compatibility, which can be used as a low-cost micro-heat source solution.

The flexible connection of the heating array units can be realized by wire bonding, and each unit can be used as a hot spot to be combined according to the needs of the test. The heat transfer effect of hot spots with different sizes, shapes, and relative positions in the process of embedded cooling can be accurately simulated. It accords with the thermal behavior of electronic chip with single hotspot and multi-hotspot.

in this work, multiple parallel microchannels is designed as embedded microfluidic structure, which located below the heating cell array, as shown in Fig.3(b). Table1 illustrates the detail dimensions of the structure. Compared with other microfluidic structures, microchannel has better laminar flow characteristics, which ensures the thermal stability of the coolant during the heat exchange process.

The above structure is processed by DRIE process on the substrate of thermal test chip, scanning electron microscopy (SEM) imagine in Fig.3(b) shows the actual morphology of the chip. A part of the region around the structure is reserved as a bonding ring to facilitate the bonding between the PCB and the chip.

构建流量回路是为了测量试验车辆的热工水力特性，如图所示。4、使用去离子水作为冷却剂，温度控制在20℃。输液泵（Y-600，XYHY）可以提供高达1L/min的稳定体积流量，最大系统压力为550 kPa。NTC热敏电阻（NTHS0402，VISHAY）和差压传感器（DPG409，OMEGA）用于监测冷却液的温度和压降。利用低纹波电源和编程系统为热点供电。将红外摄像机放置在测试车辆上方，并重点研究了热点的温度分布。值得注意的是，在实验之前对红外温度测量设备进行了校准

The flow loop is constructed to measure thethermo-hydraulic characteristics of test vehicle, as shown inFig.4. Deionized water was used as coolant and the Temperature is controlled at 20 ć. Infusion pump (Y-600,XYHY) can provide steady volumetric flow rates up to 1L/min with maximum system pressures of 550 kPa. NTCthermistors (NTHS0402, VISHAY) and Differential pressure transducer (DPG409, OMEGA) are used tomonitor temperature and pressure drops of coolant. Utilizedlow-ripple power supply and programming system to power hotspots. An IR camera is placed above the test vehicle and focused on the heat spot to probe the temperature distribution. Notably, the IR temperature measurement equipment was calibrated before experiments

THEORETICAL ANALYSIS

本节描述了用于预测上述试验车辆热阻组成的简化数值模型，如图1所示。由于实验研究中复杂的流体动力学特性和成本硅处理技术，使用这些模型是获得初步传热数据的有效方法。热点的输入电功率P由励磁电压U和测量的热点电阻Rh获得

This section describes simplified numerical models topredict the thermal resistance composition of above testvehicle, as shown in Fig.1. Employing these models is an effective way to obtain preliminary heat transfer data due to sophisticated fluid dynamic features and cost silicon processing techniques in experimental research.The input electrical power P of the hotspot is obtained from the excitation voltage U and measured hotspot resistance Rh

对具有嵌入式结构的终端测试车辆进行了评估。实验重点研究了测试载体ˈ的热阻组成以及影响嵌入式微流控结构冷却的主要因素。热测试芯片中的四个加热单元通过引线键合连接为热点，以改变热点的局部热流密度，如图5所示。微通道结构作为嵌入式微流控结构蚀刻在热测试芯片的底部，第Ċ（B）节说明了详细参数。图6绘制了不同流速下热通量密度和热点最高温度之间的关系。随着热通量的增加，最大热点温度呈线性上升，这一结果与之前的研究[5]非常一致。在125毫升/分钟的体积流量下，热点的最高温度从0到800瓦/平方厘米上升约70℃。此外，随着冷却液流速的降低，这种现象变得更加明显。可以预测，随着流速继续降低或热流继续增加，芯片进入两相沸腾状态。尽管沸腾传热具有较高的换热能力，但气液混合状态的引入很容易导致冷却过程的不稳定。因此，在当前的热设计要求中，最好在传热过程中保持单相冷却状态。根据热阻的定义，可以计算试验车辆在冷却过程中的热阻，如图7所示。在单相冷却状态下，总热阻随着热流密度的增加而减小，最终趋于稳定常数。同时，冷却液的体积流量对热阻的变化趋势没有显著影响。当热流密度为800w/cm2，体积流量达到125ml/min时，热试验车的总热阻达到最小值0.104k/W。随着流速的增加，总热阻更容易达到平衡状态。试验车辆的热阻组成如图8和图9所示。对于嵌入式微通道结构，对流热阻占总热阻的主导地位。当冷却剂流速为125毫升/分钟，热流密度为800瓦/平方厘米时，对流传热的热阻占总热阻的81%。因此，通过优化嵌入式微流控结构以降低对流热阻，可以有效提高嵌入式结构的冷却性能。

五、结论

本文提出了一种与微流体共同设计的集成电气测试车，用于评估嵌入式冷却的热性能。热测试芯片和PCB模块结构是平台的主要组成部分，在热测试芯片基板上制作嵌入式微流控结构，在PCB模块上加工闭合的电气和流量回路。在实验中，分析了热阻的组成和影响冷却性能的主要因素。实验数据表明，测试车辆可以准确地表征嵌入式冷却的热行为。它为分析嵌入式冷却的性能提供了一种工具，并为进一步研究嵌入式冷却机制提供了一种方法

Termal test vehicle with embedded structure was evaluated. The experiment focused on the thermal resistance composition of the test vehicleˈand the main factors affecting the cooling of the embedded microfluidic structure . Four heating units in the thermal test chip are connected by wire bonding as hotspots to change the local heat flux density of the hotspots, as shown in Fig.5. The microchannel structure is etched on the bottom of the thermal test chip as embedded microfluidic structure, SectionĊ(B) has illustrated the detail parameters. Fig.6 plots the relationship between the heat flux density and the maximum temperature of the hotspot at different flow rates. As the heat flux increases, the maximum hotspot temperature rises linearly and this results are in good consistent with the previous research [5]. Under the volume flow rate of 125 mL/min, the maximum temperature of the hotspot from 0 to 800 W/cm2 rise about 70 ć. Moreover, this phenomenon becomes more obvious as the coolant flow rate decreases. It can be predicted that as the flow rate continues to decrease or the heat flux continues to increase, the chip enters a two-phase boiling state. Although boiling heat transfer has higher heat exchange capacity, the introduction of gas-liquid mixing state can easily cause instability in the cooling process. Therefore, in the current thermal design requirements, it is preferred to maintain the single-phase cooling state in the heat transfer process. According to the definition of thermal resistance, the thermal resistance of the test vehicle during the cooling process can be calculated, as shown in Fig.7. In the single-phase cooling state, the total thermal resistance decreases with the increase of heat flux, and finally tends to a stable constant. At the same time, the variation trend of thermal resistance is not significantly affected by the volume flow of coolant. The total thermal resistance of the thermal test vehicle reaches the minimum value of 0.104 K/W when the heat flux is 800 W/cm2 and volume flow reaches 125 mL/min. With the increase of flow rate, the total thermal resistance is easier to reach the equilibrium state. The composition of the thermal resistance of the test vehicle is described in Fig.8 and Fig.9. For the embedded microchannel structure, convective thermal resistance dominates the total thermal resistance. At a coolant flow rate of 125 mL/min and heat flux of 800 W/cm2 , the heat resistance of convective heat transfer accounted for 81% of the total thermal resistance. Therefore, the cooling performance of embedded structure can be effectively improved by optimizing the embedded microfluidic structure to reduce the convective thermal resistance. V. CONCLUSION In this paper, an integrated electrical test vehicle co-designed with microfluidics is proposed for evaluating the thermal performance of embedded cooling. The thermal test chip and PCB module structure are the main components of the platform, where embedded microfluidic structure is fabricated on the thermal test chip substrate, and the closed electrical and flow loop is processed on the PCB module. In the experiment, the composition of the thermal resistance and the main factors affecting the cooling performance is analyzed. Experimental data shows that the test vehicle can accurately characterize the thermal behavior of embedded cooling. And it provides a tool for analyzing the performance of embedded cooling and proposes a method for further study of embedded cooling mechanism

Integrated Electrical Test Vehicle Co-designed with Microfluidics for Evaluating the Performance of相关推荐

Domain adaptation from daytime to nighttime: A situation-sensitive vehicle detection and traffic…（翻）
Title:Domain adaptation from daytime to nighttime: A situation-sensitive vehicle detection and traff ...
Introduction for i-Teams
Introduction for i-Teams Jan 17 文章目录 Introduction for i-Teams Online platform Covid Zoom rooms Amy W ...
车载无线自组织网络的介质访问控制协议研究
摘要车载无线自组织网络技术(Vehicular Ad Hoc Network, VANET),作为未来智能交通系统的基础之一,通过车与车.车与路边节点的相互通信来构成无线通信网络,用于传递辅助驾 ...
ITU标准介绍及下载索引
ITU是International Telecommunication Union (国际电信联盟)的英文缩写.该组织是为了全球各成员国或地区之间进行电信事业发展和通信合理使用的合作而成立的.其工作主 ...
European Journal of Operational Research 2023年第307卷第1期论文目录
Title:Prelim p. 2; First issue - Editorial Board Title:初步p. 2;第一期-编辑部 Author:Date: DOI:https://doi.o ...
NLP与AI会议期刊详细整理「CCF, SCI」
NLPer-Conferences-Journals-Survey 本文旨在调查所有可用的 NLP+AI 会议和期刊(主要针对自然语言处理的学者),包括CCF等级.SCI.影响因子.主页和投稿介绍. ...
有机和混合多结和串联太阳能电池的数值优化
content Introduction Numerical simulation method Implementation and Results A. Optimization of Org ...
今日arXiv精选 | Survey/ICCV/ACM MM/ICML/CIKM/SIGIR/RecSys/IROS
关于 #今日arXiv精选这是「AI 学术前沿」旗下的一档栏目,编辑将每日从 arXiv 中精选高质量论文,推送给读者. Scalable pragmatic communication via ...
pymc3 贝叶斯线性回归_使用PyMC3进行贝叶斯媒体混合建模，带来乐趣和收益
pymc3 贝叶斯线性回归 Michael Johns, Zhenyu Wang, Bruno Dupont, and Luca Fiaschi 迈克尔·约翰斯,王振宇,布鲁诺·杜邦和卢卡·菲亚斯基 ...
pytorch 全局变量_Pytorch如何通过深度学习展现全局
pytorch 全局变量 by Déborah Mesquita 由DéborahMesquita Pytorch如何通过深度学习展现全局 (How Pytorch gives the big pic ...

Integrated Electrical Test Vehicle Co-designed with Microfluidics for Evaluating the Performance of