As semiconductor manufacturing processes evolve more gradually, 3D packaging emerges as an effective means of prolonging Moore’s Law and enhancing the computational prowess of ICs. Within the realm of 3D stacking technology, the Interuniversity Microelectronics Centre (imec) based in Belgium categorizes 3D integration technologies into four distinct types, each determined by different partitioning locations within a chip: 3D-SIP, 3D-SIC, 3D-SOC, and 3D-IC. Based on our previous discussion of 3D-SIP and 3D-SIC stacking, this article places a spotlight on the other two technologies: 3D-SOC and 3D-IC.
A System on Chip (SOC) involves the redesign of several different chips, all fabricated using the same manufacturing process, and integrates them onto a single chip. 3D-SOC takes this concept to new heights by stacking multiple SOC chips vertically. The image below illustrates the transformation of a 2D System on Chip (2D-SOC), where circuits are redivided into blocks, and then stacked to form a 3D System on Chip (3D-SOC).
imec’s research team previously published a paper on IEEE, outlining the advantages of 3D-SOC and backside interconnects. This technology aims to achieve the integration of diverse chips in a heterogeneous system. By intelligently partitioning circuits, it significantly reduces power consumption and boosts computational performance. In comparison to the trending chiplet technology, 3D-SOC holds a competitive edge.
Eric Beyne, IMEC’s Vice President of Research and Project Director for 3D System Integration, pointed out, “Chiplets involve separately designed and processed chiplet dies. A well-known example are high-bandwidth memories (HBMs) – stacks of dynamic random access memory (DRAM) chips. This memory stack connects to a processor chip through interface buses, which limit their use to latency-tolerant applications. As such, the chiplet concept will never allow for fast access between logic and first and intermediate level cache memories.”
However, it’s essential to acknowledge that 3D-SOC technology comes with apparent drawbacks, primarily higher research and development costs and a longer development timeline compared to 3D-SIP technology. Nevertheless, as applications like AIGC, AR/VR, 8K, and others continue to drive the need for high-speed computing, chips are relentlessly progressing towards higher efficiency, lower power consumption, and smaller size. In this context, 3D-SOC technology will maintain its place in advanced packaging.
Backside Power Delivery Network (BSPDN)
The technology of Backside Power Delivery Network (BSPDN) represents a pivotal development in semiconductor manufacturing, offering several advantages, including more flexible circuit design, shorter metal wire lengths, and higher chip utilization. After transforming a 2D System on Chip (2D-SOC) into a 3D-SOC through layered stacking, the original back sides of the chips become the outer sides of the 3D-SOC. At this stage, the “freed-up” backside of the chips can be utilized for signal routing or as power lines for transistors, in contrast to traditional processes where wiring and power lines are designed on the front side of the wafer.
In the past, backside chips were merely used as carriers, but BSPDN technology allows for more space to be used for logic wafer design. According to simulation results, the transmission efficiency of backside PDN is seven times higher than traditional front-side PDN. Intel has also announced the introduction of this technology in the 20Å and 18Å processes.
To achieve BSPDN, a dedicated wafer thinning process (reducing it to a few hundred nanometers) is required, along with nanoscale through-silicon vias (nTSV) to connect backside power to the front-side logic chip.
Another key technology for BSPDN is the Buried Power Rail (BPR), a miniaturization technique that embeds wires beneath the transistors, with some inside the silicon substrate and others in shallow trench isolation oxide layers. BPR replaces power lines and ground lines under standard cells in traditional processes and further reduces the width of standard cells, mitigating IR voltage drop issues.
The diagram below illustrates BSPDN, where backside PDN’s metal wiring is connected to Buried Power Rails (BPR), and the backside of the chip (BS) is connected to the front side of the logic chip (FS).
The final category, 3D-IC, employs new 3D sequential technology (S3D) or Monolithic technology to vertically stack n-type and p-type transistors, forming a Complementary Field-Effect Transistor (CFET). This technology enables two transistors to be stacked and integrated into the size of a single transistor. This not only significantly increases transistor density but also simplifies the layout of CMOS logic circuits, enhancing design efficiency. As seen in the diagram below, n-type and p-type transistors are integrated vertically to form a CFET.
Nevertheless, the key challenge lies in how to vertically integrate each minuscule transistor and address heat dissipation issues under high-speed computing. Major manufacturers are still in the development phase, but the technology’s biggest advantage lies in achieving the highest component density and the smallest node width, even without nodes. With the continuous increase in demand for high-speed computing, 3D-IC technology is set to become a focal point in the industry’s development.
3D Stacking Leading the Global Semiconductor Advancement
imec has outlined a roadmap for 3D stacking, aiming to reduce pitch spacings and increase point density within unit areas. However, imec also emphasizes that the development of 3D packaging technologies does not follow a linear timeline, as depicted in the figure above, as there is no single packaging technology that can cater to all requirements.
With the rapid development of applications such as AIGC, AR/VR, 8K, 5G, and others, a significant demand for computing power is expected to persist. To overcome the bottlenecks in semiconductor process technology, countries worldwide are fully engaged in advanced packaging research, and 3D stacking undoubtedly takes the center stage as the elixir for Moore’s Law continuation.
As semiconductor fabrication technologies continue to advance, the number of transistors in integrated circuits (ICs) has steadily increased. Initially, ICs contained only tens of transistors, but as technology progressed, ICs integrating hundreds of thousands of transistors enabled the realization of 3D animation. ICs with millions of transistors allowed computers to enter households, and today, ICs with hundreds of billions or even trillions of transistors enable digital technology to connect the entire world, profoundly impacting people’s lives.
Over the past 65 years, semiconductor fabrication processes have rapidly evolved, driven by Moore’s Law, gradually reshaping society. However, in recent years, semiconductor processes have approached physical limits, and the failure of Moore’s Law has been a topic of concern. In response, 3D IC stacking and heterogeneous integration technologies have emerged as promising solutions.
3D Stacking Trends
With the rapid development of applications such as AI, AR/VR, and 8K, a significant demand for computation is expected to continue, particularly driving parallel computing systems capable of handling vast amounts of data in a short time. As semiconductor processes slow down, 3D packaging has become an effective means to extend Moore’s Law and enhance IC computing performance.
3D packaging technology offers numerous advantages over traditional 2D packaging. It enables size reduction, with silicon interposer efficiency exceeding 100%, improved connectivity, reduced parasitic effects, lower power consumption, lower latency, and higher operating frequencies. These advantages, along with various benefits of 3D integration and interconnection technologies, make 3D packaging a development direction pursued by major players in the industry.
imec’s Vision for 3D Technology
In the field of 3D stacking technology, imec (imec, the Belgian Interuniversity Microelectronics Centre) defines four categories of 3D integration solutions: 3D-SIP, 3D-SIC, 3D-SOC, and 3D-IC, each requiring different process solutions and 3D integration techniques. Eric Beyne, VP R&D, Director 3D System Integration Program at imec specifically notes that concerning 3D interconnection technology, the scope of 3D interconnection will extend from stack packaging below 1 millimeter (mm), such as Package-on-Package (POP), to below 100 nanometers (nm) with true 3D ICs using transistor stacking, surpassing an interconnect density of 108/mm².
imec identifies three key elements in 3D integration technology: Through-Silicon Via (TSV), die-to-die and die-to-wafer stacking and interconnection, and wafer-to-wafer bonding technology. Beyne points out that TSV miniaturization technology continues to evolve. However, regarding “interconnect gaps,” as TSVs further shrink, microbump technology may struggle to meet higher interconnection demands, making cu-cu hybrid bonding technology a focus of development.
▲The image shows imec’s 3D interconnect technology roadmap, illustrating that as packaging technology continues to advance, node sizes shrink, and density further increases in 3D packaging. (Source：ISSCC 2021)
System-in-Package (SIP), a form of system-level packaging, connects multiple chips that undergo different fabrication processes and preliminary packaging using heterogeneous integration techniques, integrating them within the same packaging shell. 3D-SIP involves vertically stacking multiple SIP chips, including packaging interconnects, fan-out wafer-level packaging, and solder ball bonding.
▲The image on the left is a schematic diagram of 3D-SIP packaging, where the connection points on both sides of the PCB board link the chips that have undergone initial packaging from top to bottom. The image on the right is an actual product illustration. (Source：TrendTorce (Left)，ISSCC 2021(Right))
Currently, the connection pitch in existing solutions is approximately 400 micrometers (µm). imec’s research aims to increase the interconnectivity of such SIPs by 100 times, reducing connection pitch to 40 µm. Common applications of 3D-SIP packaging include RF FEMs, TWS Barbuds SoCs.
The second category, 3D-SIC (Stacking IC), involves the stacking of individual chips on top of each other. 3D-SIC is achieved by stacking chips on an interposer or wafer, with the finished chips bonded to the top of the wafer. Chips are interconnected through TSVs and microbumps, with industry solutions achieving pitch sizes as small as 40 µm. The technology is applied to products like 3D-DRAM and logic chips, connected alongside optical I/O units on the interposer. Currently, 3D-SIC technology is widely used in High-Bandwidth Memory (HBM) manufacturing.
▲The image depicts a schematic diagram of 3D-SIC, which utilizes cu-cu hybrid bonding technology to connect the upper and lower layers of ICs. (Source:imec)
3D stacking packaging is leading the global semiconductor industry, and imec has outlined a development blueprint focused on reducing interconnection pitch and increasing contact density per unit area, positioning 3D stacking as a solution to continue Moore’s Law amid slowing semiconductor processes.
As applications like AIGC, 8K, AR/MR, and others continue to develop, 3D IC stacking and heterogeneous integration of chiplet have become the primary solutions to meet future high-performance computing demands and extend Moore’s Law.
Major companies like TSMC and Intel have been expanding their investments in heterogeneous integration manufacturing and related research and development in recent years. Additionally, leading EDA company Cadence has taken the industry lead by introducing the “Integrity 3D-IC” platform, an integrated solution for design planning, realization, and system analysis simulation tools, marking a significant step towards 3D chip stacking.
Differences between 2.5D and 3D Packaging
The main difference between 2.5D and 3D packaging technologies lies in the stacking method. 2.5D packaging involves stacking chips one by one on an interposer or connecting them through silicon bridges, primarily used for assembling logic processing chips and high-bandwidth memory. On the other hand, 3D packaging is a technology that vertically stacks chips, mainly targeting high-performance logic chips and SoC manufacturing.
CPU and HBM Stacking Demands
With the rapid development of applications like AIGC, AR/VR, and 8K, it is expected that a significant amount of computational demand will arise, particularly driving the need for parallel computing systems capable of processing big data in a short time. To overcome the bandwidth limitations of DDR SDRAM and further enhance parallel computing performance, the industry has been increasingly adopting High-Bandwidth Memory (HBM). This trend has led to a shift from the traditional “CPU + memory (such as DDR4)” architecture to the “Chip + HBM stacking” 2.5D architecture. With continuous growth in computational demand, the future may see the integration of CPU, GPU, or SoC through 3D stacking.
3D Stacking with HBM Prevails, but CPU Stacking Lags Behind
HBM was introduced in 2013 as a 3D stacked architecture for high-performance SDRAM. Over time, the stacking of multiple layers of HBM has become widespread in packaging, while the stacking of CPUs/GPUs has not seen significant progress.
The main reasons for this disparity can be attributed to three factors: 1. Thermal conduction, 2. Thermal stress, and 3. IC design. First, 3D stacking has historically performed poorly in terms of thermal conduction, which is why it is primarily used in memory stacking, as memory operations generate much less heat than logic operations. As a result, the thermal conduction issues faced by current memory stacking products can be largely disregarded.
Second, thermal stress issues arise from the mismatch in coefficients of thermal expansion (CTE) between materials and the introduction of stress from thinning the chips and introducing metal layers. The complex stress distribution in stacked structures has a significant negative impact on product reliability.
Finally, IC design challenges from a lack of EDA tools, as traditional CAD tools are inadequate for handling 3D design rules. Developers must create their own tools to address process requirements, and the complex design of 3D packaging further increases the design, manufacturing, and testing costs.
How EDA Companies Offer Solutions
Cadence, during the LIVE Taiwan 2023 user annual conference, highlighted its years of effort in developing solutions. They have introduced tools like the Clarity 3D solver, Celsius thermal solver, and Sigrity Signal and Power Integrity, which can address thermal conduction and thermal stress simulation issues. When combined with Cadence’s comprehensive EDA tools, these offerings contribute to the growth of the “Integrity 3D-IC” platform, aiding in the development of 3D IC design.
“3D IC” represents a critical design trend in semiconductor development. However, it presents greater challenges and complexity than other projects. In addition to the challenges in Logic IC design, there is a need for analog and multi-physics simulations. Therefore, cross-platform design tools are indispensable. The tools provided by EDA leader Cadence are expected to strengthen the 3D IC design tool platform, reducing the technological barriers for stacking CPU, GPU, or SoC to enhance chip computing performance.
This article is from TechNews, a collaborative media partner of TrendForce.
The surge in AIGC and new technologies such IoT, AI, 5G, AR/VR are driving a huge demand for computational power of high-end chips. This has been even outpacing the performance increase offered by the long-standing Moore’s Law, ushering in a “post-Moore” era where revolutions in advanced chip design are crucial.
Over recent years, chiplet design has seemingly become the mainstream approach for upgrading high-end chips. The concept is to allow more transistors on a single chip, effectively increasing the production yield of high-end chips while reducing overall costs.
By the large, major IC players have all jumped on board. Even Apple has joined the game by releasing their M1 Ultra SoC using the chiplet concept, which doubles computational performance by integrating two M1 Max units in a single chip.
The CPU sector is definitely a clear demonstration of this trend:
AMD took the leap with chiplet design in their 2nd-gen EPYC CPUs, doubling the computing cores from 32 to 64 within two years, while slashing costs by up to half. The company has extended this approach to their 4th-gen EPYC CPUs and even pioneered the GPU Navi 31, the first of its kind to use chiplets.
Intel started incorporating chiplets into their Lakefield series SoC in 2020. Looking ahead, their upcoming CPUs like the Meteor Lake set for 2023, and Arrow Lake and Lunar Lake scheduled for 2024, will all use chiplet design.
Transition from Bumping to Hybrid Bonding
Our analysis in “Chiplet Design: A Real Game-Changer for Substrates” laid out the comprehensive impact of the evolution of chiplet technology on substrates. In fact, chiplets have already caused a significant disruption to the most advanced semiconductor packaging technologies, necessitating the transition towards advanced 2.5D and 3D packaging technologies.
The bottleneck of advanced packaging lies in the chiplets’ interconnections, with bump and microbump still being the key technology for linking chips and forming I/O joints. These connection densities are hard to enhance, thus limiting the overall chip’s transmission speed. In addition, the more chiplets being stacked, the bigger the chip volume gets. The challenge is how to limit the chip size within a specific range, considering the current technical constraints.
Therefore, copper-to-copper hybrid bonding, also known as DBI (Direct Bond Interconnect), has been emerging as the key technology route that overcomes major hurdles in chiplet integration from the bottom-up.
Unlike bumping technology, hybrid bonding significantly shrinks the I/O joint space. The future transmission demand requires the I/O joint space between chiplets to be less than 10µm. While bumping is limited to around 20µm, hybrid bonding can take this down to an impressive 1µm or even less. This also means more I/O joints can be fitted in the same chip size – even reaching up to millions on a mere 1cm2 chip.
On top of this, hybrid bonding only adds an extra 1-2µm of thickness, compared to the 10-30µm of microbump, thereby helping reduce the thickness of stacked chips.
To put it simply, hybrid bonding can boost transmission efficiency, minimize energy usage with higher density of copper joints, manage chip volume, and even cut down on material costs.
The Race for Advanced Packaging Is Kicking Off
Moving forward, hybrid bonding is set to become the key technology supporting the continuous development of chiplet design and 3D packaging. This has been exemplified by TSMC’s front-end So IC packaging technology which is based on hybrid bonding. This puts AMD, a key customer of TSMC, in a favorable position to get ahead.
From AMD’s roadmap of 3D V-Cache technology, they have stacked SRAM on top of CCX (CPU Complex), and gradually integrated it into Milan-X series, the EPYC server CPUs, and Ryzen series, the consumer-grade CPUs, over the past two years. This has significantly improved performance and power consumption as a whole.
Not to be outdone, this year Intel also launched their Foveros Direct packaging technology, which is also based on hybrid bonding route. Assuming everything proceeds smoothly, we can anticipate the release of CPUs utilizing Foveros Direct technology by 2024.
As we look at the current products, AMD’s hybrid bonding apparently focuses on stacking SRAM and computing units at the moment. However, as CPU leaders deepen their understanding of this technology, the application field is expected to further expand. In other words, the future of hybrid bonding solutions stacking multiple computation units is just around the corner.