Search

Integrated passive devices (IPDs) have been widely used in advanced packaging of semiconductor chips, to improve their power integrity and impedance matching. There is a growing demand in guaranteeing signal and power integrity for the chips used in safety-critical products, such as those used in automotive, aviation, industrial, and defense systems, where IPDs help improve quality and reliability of the chips. Therefore, IPD testing and screening itself is essential. Note that the cost of replacing failed IPDs is much higher than the cost of manufacturing them, so screening bad IPDs before mounting is also crucial. In this work, we propose a machine learning (ML) based screening methodology to identifying the IPDs that have potential reliability issues. Based on the parametric data of 360,000 IPDs collected from the wafer probing test, the proposed Semiconductor Quality Net (SQnet) is trained to predict the IPDs which have low breakdown voltage, i.e., low reliability. Keeping the overkill rate below 10%, our method can screen out 6 to 15X more bad dies than the existing industrial methods, i.e., DPAT and GDBC.

This paper introduces a novel test methodology for ultra-tiny silicon-based inductor, which has the capability to regulate voltage and to filter out high frequency noise for power management integrated circuits (PMICs). As 5th Generation (5G) communication network and High Performance Computing (HPC) applications are expected to surge in the coming decade, computing power is projected to grow exponentially and power management is getting more critical than ever for mobile devices and HPC applications to save energy. As compared to traditional Surface Mount Device (SMD) inductors on package substrate, silicon-based inductor integrated in System on Chip (SoC) for better power management performance. And its form factor is also much smaller than discrete inductors. However, its inductance (less than 10 nanohenries) can only be measured by laboratory instruments which cannot meet mass production test throughput requirements for yield learning. In this paper, we propose an ATE-based test methodology that leads to highly accurate and reliable measurements for production test of silicon-based inductors.

This paper introduces an anomaly detection methodology with machine learning for Circuit Probing (CP) using Integrated Passive Device (IPD) as example devices. The IPD can improve the power integrity, performance, and package dimensions of the Integrated Fan-Out Package on Package (InFO-PoP), which is more cost-effective than 3D Integrated Circuits (3DIC) to achieve “More than Moore’s law” for mobile devices. Because a defective IPD can invalidate the entire package, the previous test methods are dedicated to very high-end screening for the underkill/failure-escape of high quality and reliable devices. On the other hand, the overkill issues are not concerned yet, which periodically impact the yield and cost. In this paper, we propose a new flow with machine learning methodologies to detect previously ignored anomalies on site-aware wafer-maps for predictive maintenance. The proposed flow covers the overkill and re-test issues to complete the high-quality and cost-effective test methodology with test defense.

We demonstrate two heterogeneous solutions to improve power delivery to High-Performance Computing (HPC) processors. The scalable HPC vehicle integrates two 7nm CMOS processor chiplets, each with four ARM® Cortex®-A72 cores, that are mounted on a Chip-on-Wafer-on-Substrate (CoWoS®) silicon interposer [1]. In the first solution, Integrated Passive Device (IPD) capacitors are placed directly beneath the interposer to provide more accessible and effective supply noise decoupling. The result is 3.9% higher maximum clock frequency at a core supply of 1.135V. In the second solution, the processor is powered by a laterally mounted in-Package Voltage Regulator (PVR) built in 28nm CMOS augmented with high-per-meability on-die inductors. The processor performance provided by the buck converter-based PVR matches that by an off-package External Voltage Regulator (EVR). As processor power increases with higher core counts, PVRs with on-die inductors will be increasingly compelling for efficient system power delivery.

Redundancy repair is a commonly used technique for memory yield improvement. In order to ensure high repair rate and final product yield, it is necessary to develop a repair scheme for the coming three-dimensional (3D) architecture of stacked DRAM. According to the JEDEC mobile memory technology roadmap, the interface of 3D DRAM, including the Wide I/O and High-Bandwidth Memory (HBM), is mainly classified as channel-based memories. In this paper, we propose a built-off self-test (BOSR) scheme at the controller level for channel-based 3D memory to enhance final product yield after the bonding of a memory cube to its corresponding logic die. The logic die contains the Channel controller, in which the BOSR circuit resides. Experimental results show that the repair rate is high with higher cluster failure ratio due to the flexible algorithm we choose. The area overhead is low and it decreases significantly when the memory size or channel count increases. The performance penalty is also low due to the parallel execution of address comparison and repair. Moreover, the manufacture cost is lower than conventional DRAM architecture due to allocator-based redundancies. Finally, the proposed scheme can easily be applied to other channel-based 3D memories.

Three-dimensional stacked memory stacking logic and memory dies are one of the most promising 3-D integration applications. This paper proposes two memory redundancy schemes to improve the yield of channelbased 3-D stacked DRAM by sharing spare memory across dies and satisfying channel constraints at the same time. The proposed schemes achieve much higher yield with very small area overhead than other memory redundancy schemes.

Due to more aggressive design rules adopted by memories than logic circuits, memories have been considered as the major technology driver of advanced logic circuits, so far as CMOS process technology is concerned. Memory failure pattern identification therefore is important, and is traditionally considered a key task that can help improve the efficiency of memory diagnosis and failure analysis. Critical failure patterns (that are the yield killers), however, may change in different memory designs and process technologies. It is difficult to identify critical failure patterns from high-volume memory failure bitmaps if they are not predefined. To solve this problem, we propose a local parallel search algorithm for efficient memory failure pattern identification. In addition, the proposed system integrates the defect-spectrum-based and coordinate-distance-based methods to identify critical memory failure patterns from a large amount of memory failure bitmaps automatically, even if they are not defined in advance. In our experiment for 132,488 4-MB memory failure bitmaps, the proposed system can automatically identify six critical yet undefined failure patterns in minutes, in addition to all known patterns. In comparison, the state-of-the-art commercial tools need manual inspection of the memory failure bitmaps to identify the same failure patterns.

This paper introduces test solutions for Integrated Fan Out Wafer Level Chip Scale Packaging (InFO WLCSP) which has the promising of being a very cost effective solution to achieve “More than Moore's law” for mobile devices — more so than 3D integrated circuits (3DIC). InFO WLCSP has its own unique test challenges and it is important to have high-quality screen methodology for severe defects and potential reliablity failures. In this paper we propose novel solutions that leads to fast and reliable tests for InFO WLCSP packages. We will show results on some industrial cases to validate our claims.

This article discusses a design-for-test (DFT) architecture for detecting and repairing faulty interconnects in 3-D IC circuits utilizing through silicon via (TSV) and interposer technology. The yield of such circuits depends highly on the ability to have functioning interconnects which connect the various dies. The authors also propose a built-in-self-test (BIST) framework to enable at-speed testing of such interconnects.

A 1 Tbit/s bandwidth PHY is demonstrated through CoWoS™ platform. Two chips: SOC and embedded DRAM (eDRAM), have been fabricated in TSMC 40 nm CMOS technology and stacked on a silicon interposer chip. 1024 DQ buses operating at 1.1 Gbit/s with VDDQ = 0.3 V are proven between SOC chip and eDRAM chip in experimental results with 1 mm signal trace length on the silicon interposer. A novel timing compensation mechanism is presented to achieve a low-power and small area eDRAM PHY that excludes PLL/DLL but retains good timing margin. Another data sampling alignment training approach is employed to enhance timing robustness. A compact low-swing IO also achieves power efficiency of 0.105 mW/Gbps.

We describe custom 6R, 2/4W general-purpose register files (GRF) in an ASIC-based SOC implemented in a N28 CMOS technology, which has roughly a 2∼3 X smaller area, 2 X faster speed, and 5 X lower power than a logic-synthesized version. Synthesized and custom GRFs also have a different read behavior from static and dynamic circuitry used, respectively. This is addressed by modifying a bypass control block. Hardware showed a DVFS window of 0.5 V @circuit, 130 MHz to 0.96 V, 3.2 GHz.

The three-dimensional integrated circuit (3D IC) is considered a promising approach that can obtain high data band-width and low power consumption for future electronic systems that require high integration level. One of the popular drivers for 3D IC is the integration of a memory stack and a logic die. Because the yield of a 3D IC is the product of respective yields of the mounted dies, the yields of the memory dies and logic die must be high enough, or the 3D IC will be too expensive to be manufactured. To obtain a high yield of 3D ICs, efficient test and repair methodologies for memories are necessary. In this paper, we target the channel-based 3D dynamic random access memory (DRAM) and propose two 3D redundancy architectures, i.e., Cubical Redundancy Architectures 1 and 2 (CRA1 and CRA2). We use Wide-IO DRAM as an example for discussion. In CRA1, spares are associated with each DRAM die as in a conven-tional 2D architecture. In CRA2, we use a static random access memory (SRAM) on the logic die as spares. Experimental results show that the CRA1 can achieve up to 18% higher stack yield than traditional redundancy architecture with the same area overhead. On the other hand, the CRA2 can achieve the same yield as the CRA1 with 40% less spares, but 1.3% higher area overhead.

This paper describes an on-chip intellectual property (IP) testing platform, Universal High Frequency Test structure (UHFTs), which makes logic, memory, and analog / mixed-signal IPs at-speed testable in the same testing structure. Any functional testing pattern can be loaded from an external pattern generator or a tester through standard 5-pin JTAG interfaces operating at 10MHz or below. The on-chip multichannel JTAG interface and elastic buffers convert an externally supplied pattern to an on-chip at-speed high-frequency pattern. The pattern can have address, data, and control fields. Each field is applied as input to a DUT in anyone of 16 available DUT sites, fully synchronized to the on-chip global clock. The output from the DUT is captured at-speed and stored in an output buffer. The content of the output buffer is read out to an external tester through the elastic-buffer and JTAG interfaces under a program control. UHFTs, implemented in TSMC 28-nm High Performance CMOS process, has been successfully used in digital, including ATPG, BIST, and vector-based tests with the capability of mixed-signal and analog tests. UHFTs have been designed with a frequency goal of 4 GHz in TSMC 28-nm CMOS process in the slow corner.

Three-dimension ICs (3D-ICs) are the current trend due to their improvement in heterogeneous integration, performance, power consumption, silicon area, and form factors. However, the consequent new challenges are interconnects between dies, i.e., Through-Silicon-Vias (TSVs) and micro-bumps (μ-bumps). Therefore, many interconnect test, diagnosis, and repair schemes were proposed, such as double TSVs & double μ-bumps schemes. In this paper a novel DFT technique is presented based on the double resource schemes. Challenges to two-die, multi-tier, and numerous interconnects are handled by proposed testable, diag-nosable, repairable, and scalable element, structure, and flow.

A fully-pipelined tile-able 1MB SRAM IP with a 0.127um2 cell in a HKMG 28nm bulk technology has an area of 1.39mm2/MB with 79.2% array efficiency. It operates with 2-cycle latency up to 1GHz. The no-repair hardware has a circuit limited yield of 99.92 and 53% at 100 and 850MHz, respectively with 0.75V VDD. A Data Retention Voltage of 0.42V has been measured.

Recent advances in semiconductor process technology especially interconnects using Through Silicon Vias (TSVs) enable the heterogeneous system integration where dies are implemented in dedicated, optimized process technologies and stacked in a 3D form. TSMC has developed the CoWoS™ (Chip on Wafer on Substrate) process as a design paradigm to assemble silicon interposer-based 3D ICs. To reach quality requirements for volume production, several test challenges related to 3D ICs need to be addressed. This paper describes the test and debug strategy used in designing a CoWoS™ based stacked IC. The 3D design presented in the paper contains three heterogeneous dies (a logic, a DRAM, and a JEDEC Wide-I/O compliant DRAM) stacked on the top of a passive interposer. For passive interposer testing, a novel test methodology called Pretty-Good-Die (PGD) test is presented, while for inter-die test, a novel scalable multi-tower 3D DFT architecture is presented. Silicon results show that most of the test challenges can be solved efficiently if planned properly; and 3D ICs are reality and not a fiction anymore.

A novel test module (NTM) for daisy chain test is proposed. In the daisy chain test, the defected chains and components can be easily observed and repaired rapidly by the proposed technology. Besides, the cost of proposed module is much cheaper than the price of tester improvement. Finally, the measured results of the test with NTM are also proposed for verification. The proposed module can save not only debugging time but also the cost in the daisy chain test.

High-density probing is a main trend of the test technology. The warping issues of probe card are caused by the high-density test. The metal backer and patches are applied to solve this problem and the optimized sizes of backer and patches are decided by the proposed flow. Using the probe card with optimized backer and patches, the stability of test can be ensured and the test yields are increased. 26.56% test-yield improvement can be obtained.

Through-Silicon-Via (TSV)-based three-dimensional ICs (3D-ICs) have gained increasing attention due to their potential in reducing manufacturing costs and capability of integrating more functionality into a single chip. One of the most important factors that affect 3D-IC yield is the integrity of interconnects which connect different dies in a 3D-IC. This paper proposes a Design-for-Test (DIT) scheme that can 1) detect faulty interconnects in 3D-ICs, 2) pinpoint open defect locations to help yield learning, and 3) repair faulty interconnects caused by open defects to improve the 3D-IC yield. Experimental results show that the proposed scheme can achieve a diagnosis resolution of 84% for open defects. With the interconnect repair mechanism, the 3D-IC yield is improved by 10%. In addition, cost-benefit analysis reveals that the proposed technique can significantly increase the net profit, especially when the natural interconnect yield is low.

Editor’s note: To reduce the manufacturing cost of heterogeneous 3-D integration, the Integrated Fan-Out Wafer-Level Chip-Scale Packaging (InFO WLCSP) is one of the emerging packaging technologies. In this article, the authors propose a cost model for InFO WLCSP, which can be used for analyzing the total test cost with respect to the test configuration and for optimizing the test configuration and procedure. —Jin-Fu Li, National Central University

Three-dimensional integrated circuit (3DIC) technologies with the vertical stacking schemes offer the promising performances but are sensitive to the post-bond probe in the testing reliability. In order to overcome this test challenge, the direct probe interface is applied and the performances of chip are also demonstrated. By using the direct probe interface, the post-bond chips have gained with 48% bandwidth enhancement and the test cost is also reduced in the whole test flow due to the reusable characteristics.