In the past two years, there has been a lot of talk that Moore's Law is dead. But don’t rush to nail the coffin boards, because judging from recent news, it may come back to life again. . .What resurrected Moore's Law is not some exotic gadget, but glass that we can use every day.Recently, Intel released news on its official website that they plan to use glass instead of organic materials for the next generation of advanced packaging substrates.
The reason is not that glass is cheaper or better-looking, but that they found that chips using glass as substrates perform much better than organic materials.
To be more intuitive, using glass as a chip substrate has two advantages:
One is to improve the efficiency of signal transmission in the chip, and the other is to significantly increase the density of the chip, thereby driving better performance.
This is great news now that large models are growing wildly and computing power is in short supply.
Intel officials also made bold claims that they would be able to expand to 1 trillion transistors in one package by 2030.
Shichao pulled out the Moore's Law curve. The current limit of transistors in a package is only 134 billion. Compared with Apple's M2 Ultra chip, the data of 1 trillion is nearly 10 times.
If we check it on the graph, it is quite consistent with Moore's Law. . .
Seeing this, I guess all of you may be thinking something like this: glass is not a rare material, but does it really have such great capabilities?
Before answering this question, we must first understand the basic knowledge of chip substrates.
The chip substrate is the protagonist of the final step of packaging. It is used to fix the wafer (Die) cut from the wafer in the previous step. The more wafers fixed on the substrate, the more transistors the entire chip will naturally have.
For example,The entire packaged chip is equivalent to a city. If the chips on the substrate are like skyscrapers, then the substrate is equivalent to the public transportation connecting these buildings in series, and the transistors are the people living in the buildings.
There are only two ways to get more people in the transistor, that is, in the entire city:
One is to do a good job in urban planning under the existing public transportation resources, which corresponds to improving the process of chip packaging.
The other is to build more and taller buildings. The premise is that the city's public transportation system must be fully upgraded. The corresponding solution is to change the material of the substrate.
Of course, in the process of chip packaging development, these two methods alternate.
Since its inception in the 1970s, chip substrate materials have gone through two iterations., the original chip substrate relied on a lead frame to fix the chip.
Intel 4004 chip
Intel 4004 chip substrate
By the 1990s, ceramic substrates gradually replaced the previous metal lead frames because of their better sealing and good thermal conductivity. Then in the 2000s, our most common organic material substrates appeared.
Compared with ceramic substrates, organic material substrates do not require sintering, are less difficult to process, and are also conducive to high-speed signal transmission.
So so far, organic material substrates are regarded as the leader in the chip field.
But organic materials also have disadvantages, that is, the thermal expansion coefficients between them and the wafer are too different.
It's okay if the temperature is low, but as long as the temperature is slightly too high, one deformation is large and the other is small, the connection between the wafer and the substrate will be broken.
The chip is burned out. . .
Therefore, in order to avoid this situation, the size of the organic substrate is generally not too large.
The size is small, but if you want to increase the number of transistors on it, you can only work hard on the process. To this end, manufacturers in the industry have also used all kinds of skills.
From the original focus on flat packaging, we started to engage in Jenga, which is stacked packaging.
In the field of stacked packaging, it is now rolling out. After many iterations, it has reached the most advanced through silicon via technology (TSV), which is to stack silicon chips and then connect them through holes.
But now, no matter how advanced or awesome the packaging technology is, they have begun to be stretched in the face of the development trend of Moore's Law.
Take TSV technology as an example. Although it can exponentially increase the number of transistors to a certain extent, it also has higher technical requirements, not to mention the cost.
Moreover, the requirement of the next generation packaging technology is that the packaging size must exceed 120mm*120mm.
As mentioned above, since the organic substrate is made of a material similar to synthetic resin, it is easy to bend when heated.
Nowadays, the packaging design of chips requires the chips to be brought together one by one. Heat generation is definitely unavoidable. If you want to make a larger package size, using organic materials is definitely out of the question.
Now the knife has been placed on the head of the organic substrate. Anyway, this life will be changed sooner or later.
How to do it and who can do it?
We already gave the answer at the beginning - glass.The glass here does not mean using pure glass as the substrate, but replacing the synthetic resin-like material in the previous substrate with glass, and the metal edge sealing is still there, similar to the picture below.
Of course, the glass is not the kind of glass we use every day, but will be adjusted to create a glass with properties close to that of silicon.
Compared with the previous organic materials, the glass replaced this time mainly focuses on its three properties: mechanical properties, thermal stability and electrical properties.
The first is mechanical properties. Glass substrates outperform organic substrates in terms of mechanical strength.
When glass serves as a substrate material, holes will be made on it to ensure signal transmission.
Because the glass material is super flat and easier to photolithography or encapsulation, the number of holes opened on it is much more than that on organic materials for the same area.
It is equivalent to saying that public transportation built on glass materials will be denser and have more lines than those built on organic materials.
According to Intel, the spacing between glass core vias can be less than 100 microns, which can directly increase the interconnection density between chips by 10 times.
As the interconnection density increases, more transistors can be accommodated in the same area.
Next is thermal stability. Glass substrates are not prone to warping or deformation due to high temperatures.
In case of a special situation, glass also contains silica, which has properties similar to silicon, and their thermal expansion coefficients are similar. Even if the temperature is too high, the chip and the substrate on the substrate will deform together at the same expansion rate.
Finally, there is the unique electrical performance of the glass core. To be more precise, it is the electrical performance of the glass after opening. Its dielectric loss will be lower, allowing clearer signal and power transmission.
In this way, the power loss during signal transmission will be reduced, and the overall efficiency of the chip will naturally be improved.
When these properties are combined, the final manifestation on the chip is that if it is packaged with a glass core substrate, the number of chips that can be placed is 50% more than other chips.
But there is still a question. Since the performance of glass substrates is so good compared to organic substrates, why not use glass substrates earlier?
In fact, it's not that I don't want to use it, but that it's not that simple to replace a material.Early exploration, mid-term research and development, and late-stage implementation all cost money and time.
Take Intel as an example. It started developing glass-core substrates ten years ago and spent at least one billion dollars on it.
The current result is that a set of testing tools has been assembled. Actual mass production of glass core substrates will have to wait until 2026 and beyond.
Of course, not only Intel, but also many companies in the entire industry are working on the research and development of glass substrates. After all, it is a consensus in the industry that glass replaces organic materials.
For example, more than half a year ago, Japan's DNP also revealed that it was developing glass substrates to replace traditional resin substrates, and they also set a small goal: to achieve sales of 5 billion yen by 2027 with glass substrates.
The first company to enter the glass substrate industry is Absolics, a subsidiary of SKC. Even last year, it had invested US$600 million and planned to build a factory in Covington, Georgia.
According to their plan, not surprisingly, small batches of glass substrates will be produced by the end of this year.
Of course, in a short period of time, the mainstream of the chip substrate market will still be organic materials. After all, it will take a transition period for technology iteration to complete commercialization. Technology cost, yield, etc. are all issues that manufacturers need to solve.
But what is certain is that the importance of organic materials on the stage of chip substrates will gradually be replaced by glass.
Visit the purchase page:
Intel Flagship Store