With a history spanning more than 40 years of expertise and technological innovation, QRT Co., Ltd. has established itself as a leader in the field of semiconductor analysis. The company excels in a wide array of areas, ranging from mechanical and functional evaluations to reliability assessments and intricate chemical analyses, having earned a global reputation for its steadfast commitment to delivering exceptional customer satisfaction.
Considering your experience in both the US and South Korea, can you provide insight into the evolution and reliability of South Korean projects over the years?
Many companies we have spoken to have suggested that despite the challenges of COVID-19, it presented an opportunity for South Korean businesses to develop their core capabilities and build a more robust supply chain. Could you share your thoughts on the current strength of the South Korean supply chain? Additionally, could you discuss the growth and evolution of Small and Medium Enterprises (SMEs) in South Korea over the past decade?
My view on the evolution of Korea, particularly its industrial and technological growth, is rooted in historical contexts. In the 19th century, the concept of "Techno-Science", or science-based technology, came into prominence. Post-World War II, the U.S. quickly advanced, harnessing science to propel technological growth and becoming a global leader in manufacturing. She managed to achieve this despite starting late in the game, as Germany and England had initially led the industrial revolution.
Comparatively, Korea's journey was vastly different. Following the Korean War, the country was decimated, needing to rebuild almost from scratch. Compounding this challenge was the impact of 36 years of Japanese colonization, which stymied the continuity of education and knowledge transfer to the younger generation. This had to be restarted and restructured following the war.
Despite these challenges, the country underwent a significant transformation in the 1980s, driven in part by visionaries or chaebols who channeled resources into certain industrial sectors. This was fueled by a combination of education, hard work, and resilience. The Korean people embraced various opportunities, such as working overseas to raise funds for the nation's growth and producing goods, like shoes, for international clients, including the U.S.
Fast forward to recent times, Korea's technological prowess is highly visible. We have worked with significant international aerospace entities, and our satellite construction capabilities are widely recognized. Education, particularly among the 625 generation (Koreans born around the time of the Korean War), has been a major driver of this growth. Today, Korea's literacy rate is close to 99.5%, a testament to our emphasis on education.
Returning to your original question about the strength of the South Korean supply chain and the evolution of SMEs, I believe we are highly resilient and well-equipped to compete globally. This was evident during the COVID-19 pandemic, when Korea was one of the first countries to develop effective solutions and mitigation strategies, including advanced contact tracing systems.
In the era of Industry 4.0, the challenges for Korea, like any other country, revolve around cost-effectiveness and quality assurance. While technological and infrastructural investments are critical, the skill and knowledge of the people working with these technologies are equally important. We're in a transitional phase where humans must learn to answer the questions that machines are increasingly capable of asking.
That said, I believe that AI and machines can't replicate the innate creativity and cultural identity of humans. Korea's soft power, such as K-pop and Hallyu culture, illustrates this point. It is not about advanced AI or sophisticated technologies, but our unique cultural DNA that resonates globally.
In the context of AI systems like Chat GPT, it is important to recognize their current limitations, such as the inability to disclose the sources of information. This will need to be addressed to ensure proper respect for intellectual property laws in the future. Despite these challenges, AI tools like Chat GPT remain invaluable for information analysis and knowledge discovery.
You are mentioning the use of AI and AI is definitely one of the technology that is making headlines now. I know that you've developed your own AI semiconductor testing in collaboration with your partners.
But I am quite interested as you play a key role in the development and design and the testing of those semiconductors, where do you see the biggest sector for growth in the future from the 4th industrial revolution?
Firstly, I want to emphasize the importance of reliability and safety in the technology industry, particularly for those technologies that involve the use of AI. For instance, in testing companies, significant time and resources are devoted to ensuring the reliability and safety of the equipment. However, these are often reinforced through international standards, as they come at a cost that many companies are not willing to bear unless necessary. Certain organizations, like NASA and the US military, especially in areas related to military communications and satellites, demand high reliability.
Now, let's envision a future about five years from now. ChatGPT has evolved into "my ChatGPT," assisted by a new AI-based cell chip. This personalized AI knows my habits, dietary preferences, daily activities, and even monitors my physical condition for any abnormalities. It has access to a wealth of data about me and can predict my needs. It schedules my daily activities based on my preferences and energy levels and alerts me to any health anomalies. If I get a hiccup, it can even make a doctor's appointment for me or suggest new medical treatments available.
My life is fully optimized with my ChatGPT's personalized service, and I am living in a way that truly reflects my preferences and potential. But there are potential pitfalls. If my ChatGPT fails to function, and I don't know where my data is stored, I would be lost. The AI chip, which knows where my data resides, could perhaps be my saving grace. This situation is not unlike people today who depend heavily on prominent social networking platforms or search engines; without them, they'd feel lost. Similarly, with autonomous vehicles, even a single technical glitch can have catastrophic consequences. We need to approach this future mindful of these potential challenges.
Considering the ongoing transformation in the automotive industry, where cars are evolving from mere transportation devices to complex "computers on wheels" powered by AI and other advanced technologies, it is projected that in the next 10 to 20 years, about 40 to 50 percent of a car's value will be derived from electronic components. This shift is likely to present numerous testing and analysis challenges due to the increasing number of components in a car, resulting in more radio frequencies and electrostatic discharges, among other things. Furthermore, in the context of autonomous vehicles, a single malfunctioning component can pose a significant risk to life.
In light of these developments, I am curious to know how these changes in the automotive industry, particularly the shift towards hyperconnected and autonomous cars, are impacting your business at QRT. And what kind of solutions are you developing to address these emerging challenges?
As mentioned earlier, reliability is a key aspect we need to consider as we progress in the world of autonomous cars. In developed nations like the U.S, standards are already being established for AI integration in these vehicles.
While fully autonomous driving presents an exciting prospect, in countries like Japan, Korea, Germany, and the U.K., where there is a growing elderly population, the primary concern is not just the novelty of driverless cars, but their safety and preventative capabilities. For instance, the elderly demographic generally believes that human intervention is still crucial in certain situations, such as driving in heavy snow or during the night to prevent major accidents. Additionally, autonomous self-driving cars can provide significant benefits for those who have mobility impairments or disabilities.
The demand for AI in the automotive industry is expected to increase. Considering that the younger generation comprises 60-70% of the population in many countries, with Japan and the U.K. having younger generations that make up over 40% of their total population, their road safety is paramount. Unfortunately, this demographic experiences higher fatality rates due to car accidents, implying that safer cars that can prevent accidents are needed. Therefore, this will be a critical aspect moving forward.
As microchips become more complex and increasingly crucial in people's lives and safety, how do you rise to the continuous challenges this presents?
Indeed, the challenges we face are manifold. For instance, when collaborating with companies like Integrated Device Manufacturers or other Fabless design houses, the process can be complicated as they often can't be straightforward about their issues due to confidentiality concerns.
Likewise, the situation is more stringent when engaging with Silicon Valley firms. These companies have exceptionally rigorous reliability requirements and do not permit us to use their names publicly. Despite this, we have established a robust working relationship over the years. We address their problems, share our internal reliability data, and collaboratively arrive at solutions. It is important to remember that while they may not manufacture the chips, they do package and test them differently. As chip manufacturers, we play a pivotal role as solution providers.
Take the case of Radio Frequency (RF) testing. Traditionally, Direct Current (DC) tests were simple to perform, where you would check whether the temperature voltage was static or changing. But now, there's a demand to provide RF signals and measure the output from these signals across multiple devices simultaneously. It becomes a real challenge to consistently and accurately measure, in real-time, the performance of 600 devices. And when a failure occurs, it is crucial to know precisely when and how it happened.
Also, another predicament arises from the fact that existing standards may become outdated as technology rapidly evolves according to Moore's Law. The standards set by Asian countries often cannot be applied to US-based companies, so we have to maintain regular communication with the Standard Committee to stay up-to-date.
Agencies such as NASA and JPL conduct specific types of testing for individual groups and areas. The devices they test need to work flawlessly, as they will be sent into space and are required to function without fail for 10 to 15 years. They cannot be replaced mid-mission, making the testing phase crucial, especially for RF devices. RF devices usually consist of silicon carbide and gallium nitride, and these Wide Bandgap (WBG) devices present their own unique challenges.
I'd like to inquire about the subject of compound semiconductors. Indeed, we're seeing new types of substrates, such as gallium arsenide, silicon carbide, and gallium nitride, being used in various applications, including aerospace and power devices.
I am interested to know how these shifts in substrates are influencing your product development strategies. Can you shed some light on how testing these materials differs from testing standard silicon chips?
Silicon chips, first introduced in the 1950s, revolutionized technology thanks to advancements in quantum physics. Notable early adopters like Fairchild used them to create more reliable transistors for projects such as the LGM 30 Minuteman, an American intercontinental missile. Over time, transistors became more compact, leading to developments in silicon characteristics, including gate structures, degradation patterns, voltage, and temperature variations. These parameters continue to guide testing practices and provide accurate results.
However, there's a shift towards Wide Band Gap (WBG) devices, which possess a band gap two to three times larger than silicon devices. These new substrates, such as silicon carbide, offer advantages like high current density and rapid switching time due to their high energy band gap. But they also present new challenges. For instance, it is harder to control doping in certain chip areas, resulting in varied AC parameters across the same wafer. Moreover, WBG devices can withstand extreme temperatures up to 250 °C, making them ideal for missions like Mars exploration.
Currently, these WBG devices are primarily produced in small wafer sizes due to doping and process control difficulties. Several companies, including some in Korea, are attempting to produce larger 8-inch wafers. A significant portion of these WBG devices, particularly those for military and aerospace use, are manufactured in the United States and subject to strict ITAR regulations.
We have been collaborating with the Korean government and other industries for a few years, funded by government grants, to advance silicon carbide technology. We hope to achieve a breakthrough soon with the successful production of 8-inch wafers and improved yields.
As an example of how these technologies are being applied, remember General Electric's wind turbines that initially faced challenges due to the failure of silicon carbide-based IGBT power devices. Today, the quality and reliability of these devices have improved, and they're being widely used in electric vehicles. Compared to IGBTs, they're lighter and more efficient. In fact, tests conducted by Toyota revealed a 20 to 30 percent improvement in mileage. Thus, there's growing interest in these devices, and we're working on their radiation resistance, an aspect currently considered a weak point.
Given their efficiency, WBG devices are poised to play a crucial role in the development of electric smart cities, power stations, and green energy initiatives in the future.
My final query revolves around the unique role your company fulfills within the manufacturing process. It is clear that clients approach you at different stages, from prototype testing to seeking design advice before the prototype is even developed, leveraging your accumulated knowledge. I am curious to understand how you would characterize your company's contribution within this manufacturing process?
Indeed, our interaction with clients can take on several forms, depending on the type of project. In the case of automotive parts, for instance, we usually deal with assemblies rather than single parts. Clients have a pre-existing part database, and they assemble these parts before sending them to us for final testing or reliability assessments. This might involve various tests like drop tests, shock and vibration tests, chemical tests like salt sprays, and electrical evaluations. Sometimes, we even perform highly accelerated tests that push the product until it fails to identify the weak points.
American companies follow a similar practice. They engage us early in the assembly process, right when the first prototypes are ready. We conduct ongoing reliability testing on these prototypes as the client proceeds with manufacturing in successive batches. This iterative testing continues throughout the production phase. We perform preliminary testing, identifying potential weak points before the client proceeds with full life tests – which usually involve running the product for about 2000 hours with a few hundred samples. By the time they reach this stage, they can't afford any 'oops' moments, so trial and error are integral throughout the process, allowing for improvements at each manufacturing step.
On occasion, we might have to go back to the very beginning, but that's rare. Tweaks in the encapsulation process or device orientation might be needed, but entire design changes would necessitate revisiting the foundry or the fab, which is not usually the case. The process typically involves initial qualification of the device and then progress through multiple levels, with us assisting in the reliability testing at each stage.
When it comes to software, we perform risk runs, which means testing on prototypes or test runs. We carry out preliminary tests on a wafer or part of it. Depending on the project, we might conduct full-prone testing or use similar beams (protons or alpha particles) to gauge the device's performance. If there's existing data, a comparative analysis is carried out. For cases where latch-outs (unwanted triggering of circuit elements) are a concern, we perform thorough screening. Our goal is to identify potential weak points early in the process, like with ESD (Electrostatic Discharge) testing. The last thing anyone wants is to find out about ESD issues late in the process. If ESD testing reveals problems, the packaging is usually the first aspect checked for potential mitigation before considering major design changes.
Would it be accurate to characterize QRT as a comprehensive, all-in-one solution for testing and analyzing at any stage of production and for any type of electronic technology?
Indeed. We perform comprehensive evaluations of new chips and devices, examining them layer by layer. If failures are discovered, such as in displays or other devices, we're capable of conducting both non-destructive and destructive tests. For example, if a device burns out, we can pinpoint the location of the burnout, and physically analyze it. We're able to section off the affected area and assess how deep the damage extends, or if there are ruptures in other regions.
In terms of components like displays or batteries, we're equipped to analyze their chemical compounds, crystal structures, and other facets, all on a minute scale. There's no need for complete disassembly for devices such as smartphones or camera modules. We can utilize 3D X-ray technology to inspect them from various angles and locate any breaks or failures.
Once we have identified where a short or open circuit is, we can proceed with a detailed, destructive analysis. Our services cover a broad spectrum - from mechanical to functional and reliability aspects, as well as chemical analyses.
I'd like to delve deeper into the aspect of Soft Error (SE) and soft error analysis that forms a significant part of your operations. Initially, soft errors became critically important in the '70s due to defects in memory products like DRAM. I am aware of your close association with SK Hynix, an innovator in 3D architecture and multi-layer stacking. They've not only brought the 64-layer AND memories to market but are also launching products with more than 120 layers. Very recently, they announced their work on High Bandwidth Memory 3 (HBM3).
So my queries are two-fold. Firstly, could you share more insights into the evolution of these new multi-layer architectures, particularly HBM3? Secondly, I am interested to understand the specific challenges these innovative multi-layer architectures pose when it comes to testing memory products.
A multi-layer architecture involves devices stacked with Through-Silicon Vias (TSVs). If there is a failure in one TSV, it doesn't necessarily mean the entire system fails because there is a repair strategy in place that allows for routing to healthy TSVs.
However, an important challenge during the stacking process arises if a TSV goes wrong. Although it is a rare occurrence, the repercussions can escalate due to the multiple device layers. Hence, the testability of TSVs is crucial.
When it comes to memory stacking, each memory component has its repair capability. This entails replacing defective cells' addresses with good ones, which must be done for every die. Yet, a 3D device's repair ability is limited to its own wafer repair capability. Therefore, to optimize the yield of the stacked die, it becomes imperative to strategize the die selection process carefully.
Testing becomes more complex with higher complexity. A consideration to keep in mind is neutron testing, as neutrons can penetrate the entire die. Spallation can occur when neutrons hit a nucleus, changing silicon into a different material by producing secondary ionizing particles. The residual energy generates an electron-hole pair that can potentially change the logic state of a cell, which could be perceived as an error by the system.
When dealing with stacked dies, error rates tend to be lower on the top and bottom of the die, but more concentrated in the middle. However, this can depend on the type of testing - heavy ions, for example, lose energy more quickly and thus have a higher failure rate.
Errors in memory cells are generally not an issue due to Error Correction Code (ECC) capabilities. Single errors are not usually problematic, but if the control circuit is upset, it could make multiple cells look bad, leading to a larger system issue.
To mitigate these potential problems, the system includes microcontrollers for management. Techniques such as CrossLink (CXL) are used to improve performance and reliability. For high-reliability applications, such as space exploration or databases, both individual die reliability and stack reliability are tested.
The approach to testing these architectures requires a different methodology that can accommodate both hardware failures and software issues. This might involve improving firmware or hardware to mitigate any identified issues.
As a South Korean company with a strong presence in the United States, China, and Japan, you’ve successfully gained global market share. In light of the escalating demands and essential capabilities required to maintain competitiveness, I would like to inquire about the primary factors that contribute to your global competitiveness. What do you believe are the key reasons behind your ability to thrive in the international market? Furthermore, what would you identify as your key competitive advantages over other companies in the same industry?
Our CEO, Kim Young Boo, has made substantial investments in R&D, demonstrating unwavering commitment. However, what he has invested in more than anything else is his patience. It is true that we entered the scene relatively later compared to others. Taiwan, for example, has created an ecosystem around TSMC that greatly favors the semiconductor industry, surpassing any other country in this regard.
Nevertheless, there are crucial elements missing in Taiwan's ecosystem that are essential for future competitiveness. Firstly, they lack an automotive industry, an aerospace industry, as well as the required infrastructure, know-how, and human resources for these sectors. While they possess different resources, TSMC is currently shifting its focus towards building memory-based industries, an area in which we already excel.
Over the past few years, we have been diligently addressing our own challenges, and we will continue to do so. As a country, we cannot solely rely on other nations to solve our problems because we possess an abundance of creative and talented minds, surpassing any other country in this regard.
In addition to conducting business in China and Japan, your company ventured into the US Silicon Valley through a joint venture in 2019. With an eye on the future, which specific regional market area do you perceive as holding the greatest growth potential for your firm?
China is seen as having the highest growth potential for our firm due to its significant economic growth and the ability to export products globally. However, one major hindrance to their exports is related to intellectual property issues. For instance, a cell phone may contain thousands of patents, and if royalties are not paid for those patents, the device cannot be exported. Clearing these IP issues is crucial.
China is particularly focused on becoming a market leader in the EV industry, but they face challenges when it comes to exporting to the US. While they may be able to sell parts, exporting complete systems requires resolving IP infringement issues, which are tied to reliability and technology concerns.
China has also committed to reducing carbon emissions and phasing out fossil fuel-based automotive production. This shift necessitates extensive testing and reliability-related work in key components such as RF devices, solid-state drives (SSDs), and power devices.
Another area of interest for us is the satellite industry, which has witnessed exponential investment growth in recent years. Satellite reliability, RF communications, and innovative functions like direct communications without the need for SIM cards are evolving. Such advancements can have a profound impact on our daily lives, similar to the potential of RF devices in smartphones, like the life-saving 911 call function.
Furthermore, developments in NanoSats or CubeSats, small and RF-enabled devices, hold great potential not only for military applications but also for sectors such as agriculture, fishing, security, and disaster management. These advancements require rigorous testing, reliability assessments, and analysis to ensure that life-changing technologies deliver on their promises.
In summary, China's growth potential, coupled with advancements in satellite technology, RF communications, and emerging nanosets, present promising opportunities for our firm in terms of market expansion and transformative technological developments.
Imagine that we return for another interview with you in five years' time. Is there a specific goal or ambition that you would like to have accomplished by then?
Within five years' time, my ambition is to continue working with QRT and contribute to its success. As I mentioned earlier, our CEO has invested not only substantial financial resources and human capital but also demonstrated immense patience over the past four to five years. This has positioned us as latecomers in the industry, but we have now reached a point where we can actively address reliability issues and aspire to become leaders in the reliability aspect of the Fourth Industrial Revolution.
Achieving this goal is crucial not only for our company but for any country aspiring to excel in the industry. While we may not be fully prepared at present, we possess a talented workforce and abundant resources to overcome challenges, learn, innovate, and drive progress. My aim is to be part of this journey, contributing to the company's growth and making a meaningful impact in the realm of reliability in the Industry 4.0.
0 COMMENTS