integrated circuit

1.High cost and low yield
   Substrate capacity limits market supply

    Before manufacturing silicon carbide chips, there are first two steps: substrate manufacturing and epitaxial wafer production, which are important components of silicon carbide devices. From the perspective of the manufacturing cost structure of silicon carbide devices, the substrate cost is the largest, accounting for 47%; The second is the extension cost, accounting for 23%.
The substrate is the embryonic form of a silicon carbide wafer. It generates silicon carbide powder raw materials by mixing high-purity silicon powder and carbon powder, and then undergoes a crystal growth method under specific circumstances to generate cylindrical silicon carbide ingots. After processing, cutting, and obtaining a silicon carbide wafer with a thickness of not more than 1mm, the wafer undergoes grinding, polishing, and cleaning to finally obtain a silicon carbide substrate.
In the process of manufacturing silicon carbide substrates, there are extremely high requirements for the purity of raw materials, environmental control of crystal growth, and later processing. Therefore, the manufacturing of silicon carbide substrates has problems such as slow growth rate, high requirements for crystal shape, and high cutting wear. This directly leads to the problem of low yield and low productivity of the substrate.
The quality of the substrate directly affects the quality of subsequent epitaxial wafer generation, and subsequently affects the performance of finished silicon carbide devices. Therefore, the industry generally believes that the entire silicon carbide industry will still be driven by the production capacity of substrate materials in the next few years.
According to Wolfspeed's prediction, the market size of SiC materials in 2022 will be $700 million, and the device market size will be $4.3 billion. In 2026, the SiC material market will reach 1.7 billion US dollars, and the device market will reach 8.9 billion US dollars. From 2022 to 2026, the composite annual growth rate of the material market size was 24.84%, exceeding the composite annual growth rate of the device market size.
From the perspective of global silicon carbide substrate market share, Wolfspeed, Roma, and II-VI collectively account for 80%, which means that the pace of expansion of these three companies will limit the supply of substrates.
On the other hand, these three enterprises are gradually increasing the proportion of their own materials. For example, Wolfspeed's proportion of their own materials will increase from 40% in 2021 to 56% in 2024, further compressing the capacity that can flow out to the market. In the coming years, there will be greater pressure on global substrate production capacity.
As we can see, Mercedes Benz, Land Rover, Lucid Motors, General Motors, Volkswagen, and others have all chosen to cooperate with Wolfspeed, which indicates that the silicon carbide industry is not only at the device end, but even downstream system suppliers or vehicle companies have made capacity reservations to the upstream supply side.
Gong Xi pointed out that at present, the yield and quality of the substrate are not satisfactory for both the head enterprises and the enterprises located in the middle. This will lead to substrates that do not meet MOSFET requirements impacting the Schottky diode (SBD) market. If the yield and quality of this part of the substrate can be improved, it can help alleviate the constraints on MOSFET production capacity, which depends on the extent to which the yield and quality of domestic substrate enterprises will improve over time.

2.Substrate and Device Process Iteration
   The silicon carbide market is far from mature

       Like MOSFETs, silicon based IGBTs are also used in automotive main drive systems. As a very mature power device, its process iteration is improved around the structure. Gong Xi believes that at this point, the development of silicon carbide devices will also be the same. That is, the evolution from planar to trench architectures will bring growth space in performance and cost.
In addition, moving from a 6-inch substrate to an 8-inch substrate will also bring about significant changes and become a watershed. Wolfspeed estimates that the cost of a single 8-inch bare chip will be 37% of the current 6-inch by 2024, which means a 63% decrease in cost. This decrease in cost includes an increase in yield and an increase in the number of bare films.
Yole pointed out in the report that the 8-inch silicon carbide wafer is considered a key step in expanding production. The goal is obviously to increase production and gain an advantage in the next round of competition. Major IDMs are developing their own 8-inch silicon carbide wafer manufacturing capabilities; As of 2022, some wafer suppliers have started shipping samples. In Yole's power silicon carbide forecast, 6-inch will remain the leading platform for the next five years. However, starting in 2022, the first 8 inches will be considered a strategic resource by market participants.
"Overlapping the impact of device configuration, Wolfspeed predicts that if the 8-inch substrate+trench type approach is adopted, the cost of silicon carbide devices will decrease to 28% of the current 6-inch+planar structure in the next few years.".
With the changes in process and substrate size, the impact on device costs is enormous. Whether in terms of configuration or production of 8-inch substrates, it will have an impact on the existing market pattern in the future. This iterative process is an opportunity for domestic enterprises.

3. Master the upstream material end
    Key points for domestic substitution

          The automotive voltage platform is evolving from 400V-600V-800V, but in fact, the evolution speed is faster than the upstream silicon carbide material end, which means that the window periods of the upstream and downstream are mismatched, especially for the domestic silicon carbide industry.
This will lead to a greater gap between supply and demand. Who can fill it and how? This is a question that we need to think about.
Gong Xi said that many investment institutions simply use the shipment volume of silicon carbide devices or research and development investment in the next few years as a static observation perspective to judge the industry, but the silicon carbide industry still needs to take a dynamic perspective to see the size of the silicon carbide device market in the next few years.
The entire silicon carbide industry chain is divided into different segments. The upstream industry chain includes raw materials, substrate materials, and epitaxial materials. The midstream industry includes chip structural design, chip manufacturing, devices, and modules. The downstream application fields include solar photovoltaic, semiconductor, automotive, rail transportation, 5G base stations, building materials, and steel industries.
The layout of each enterprise in each link varies, from IDM to substrate or epitaxial materials, to epitaxial wafers and devices. Core think tank experts estimate that if the epitaxial+device approach is used, the gross profit will be around 60%, and if the epitaxial chip business is removed, the gross profit will be around 37%. The gross profit of the latter is basically the same as that of silicon based device manufacturers.
This indicates that if you do not master the upstream material end and only make silicon carbide devices, from a gross profit perspective, there is no profit compared to silicon based devices.
Adding to the changes in substrate size and device configuration discussed above, the domestic market will face significant cost pressure in the coming years. Therefore, the key to grasping the opportunities of the future industry is to develop the upstream material industry.
It will take 2-3 years to convert the size of silicon carbide substrates to 8 inches. In the short term, the cost performance of silicon carbide devices based on 6 inch substrates is still high. However, in the medium to long term, the current large-scale silicon carbide MOSFET manufacturers will face challenges and pressures in the future.
Yole said that two main trends are affecting the silicon carbide supply chain: vertical integration of wafer manufacturing and module packaging to generate more revenue in the coming years. In this context, terminal systems companies (such as automotive OEMs) are adopting silicon carbide faster and more flexibly to manage the supply of multiple wafer suppliers on the market.