Tech
Arun Mampazhy
Aug 14, 2022, 04:22 PM | Updated 04:26 PM IST
Save & read from anywhere!
Bookmark stories for easy access on any device or the Swarajya app.
In Part 1 of this series, we examined the possibility of running the 8-inch (200 mm) fab line with 180 nm (0.18 micron) process node at India’s first fab (semiconductor fabrication plant) — Semi-conductor Laboratory (SCL), Mohali — at a higher capacity (close to 4,000 wafers per month).
In this second part, some examples of integrated circuits (ICs) that are, or can potentially be, manufactured globally using 180 nm technology are discussed along with, briefly, market and revenue aspects.
Global Scenario, An Overview
A recent estimate by market analysis group Gartner showed that for foundry business (that is, contract chip manufacturing), out of a global revenue of nearly $108 billion in 2021, the 130/180 nm process nodes together fetched over $10 billion for the top-five companies alone.
Given that many more foundries have the capability to process 130/180 nm nodes (compared to bleeding-edge process nodes like 7 nm and 5 nm), all the foundries together are likely to have about 10 per cent revenue from these two process nodes.
In 2021, the overall semiconductor market, comprising foundry, fabless (fabrication outsourced), and integrated device manufacturers (IDMs), was nearly $600 billion.
Going by the same trend as with foundry, 130/180 nm technology must be about 10 per cent or, perhaps, even more because many of the IDMs do not have the most advanced nodes. Of this, nearly half of the revenue will be from 180 nm and the rest from 130 nm.
Note that the same tools that the SCL owns, as mentioned in Part 1, are likely to be good enough to fabricate 130 nm chips too, provided production-grade technology is obtained (usually through a 'transfer of technology').
So, what chips are “still” made using 180 nm process nodes?
Going by this classification, chips that fall under the 180 nm process node may predominantly fall into four categories — communication, power management, microcontrollers, and sensors. (Note that some products may have a mix of these four categories.)
Flavours Within 180 nm
To get a feel for what are the “flavours” offered within 180 nm, let us take a look at two charts — one from United Microelectronics Corporation (UMC) and the other from Tower Semiconductors, both of which are among the top 10 foundries in the world.
In Part 1, it was said that Tower Semiconductor had helped with the upgrade of SCL to 180 nm technology. However, it is possible that not all the flavours were transferred as part of the transfer agreement then.
The “base technology” is the ability to process wafers (fabricate chips) using a particular process node and will cover some flavours. The cost and ease of adding capability will depend on which one, and also on what kind of customers (or applications and products) the fab is targeting.
In other words, after the technology transfer that happened around 2010-11, Tower Semiconductor and SCL may have developed additional flavours on their own in their respective fabs.
That is to say that not every example from the list (in the charts) that will follow may be doable right now at SCL, but many are likely to be possible (or even already implemented on a small scale). Some of them may be doable with tweaks or additions to the production line.
Examples Of Semiconductor Products Using 180 nm Technology
Radio Frequency Front End Modules (RF FEM or just RFFE) are key devices in communication applications like mobile phones and other devices that receive and send signals.
The exact configuration of the RFFE modules may differ based on applications (5G, RFID reader, automotive, etc), but components like RF switches, low noise amplifiers (LNA), power amplifiers (PA), antenna tuners, analog to digital converters (ADC), and digital to analog converters (ADC) are seen in many or all of these modules.
Here are some examples from within the past four to five years mentioning 180 nm technology — PA for 4G/5G , multiple examples of LNA for the global navigational satellite system (GNSS), examples of various LNAs for GPS, PA and RF switch for health applications, phase-change material switch driver, PA for wireless combo chip, antenna for bio-medical applications and soil moisture monitoring (check sizes), MIMO on chip antenna for 5G, antenna design for millimetre and THz applications, and so on.
Multiple examples can also be found for ADC and DAC using 180 nm. In fact, the SCL already lists many on its website.
The global positioning system (GPS, originally Navstar GPS) is a GNSS, while India’s regional autonomous navigation system NaviC (Navigation with Indian Constellation) is said to be more suited to India’s needs.
COORDINATES, an exclusive monthly magazine on positioning, navigation, and associated technologies and applications, in its January 2022 edition gives the example of a chip fabricated using 180 nm at SCL for NaviC.
Further, in power management ICs (PMICs), even as recently as March 2020, Tower Semiconductor has been offering new process design kits using 180 nm. Tower has a wide range of offering for manufacturing PMICs, including for high voltage using 180 nm.
There are also examples from other fabs, such as X-Fab. Interestingly, these technologies offered are not “statis” either; both have enhanced their 180 nm offering recently. X-Fab added flash memory capability in April 2021 and Tower enhanced its BCD process in May 2022.
Industry website Semiconductor Engineering has estimated (November 2021) that PMIC shortages (made using process nodes ranging from 180 nm up to 40 nm technologies) would last well into 2022.
Earlier, in an article titled 'Chip shortages grow at mature nodes' (July 2021), the website had mentioned shortages of microcontroller units (MCUs) that are fabricated using multiple process nodes (based on complexity), including 180 nm.
As per IC Insights, the automotive MCU market is expected to have surged 23 per cent in 2021 despite the shortages, as prices have been on the up.
Sensors could, in general, mean a variety of devices — for example, the magnetic hall (effect) sensor, air flow sensor, temperature sensor, all using 180 nm and so on.
It was reported in 2018 that Tower Semiconductor’s 0.18 um CMOS Image Sensor (CIS) SPAD (single-photon avalanche diode) process had been chosen by LeddarTech for its next-generation automotive LiDAR solutions, combining CMOS, image sensors, and SPAD on the same chip.
Many more examples related to image sensors can be found. Yet another example is of a 128×128 image sensor having been built alongside an embedded convolutional neural network model with programmable weights on a 180 nm CMOS die.
Apart from the four main categories examined so far, there are many other type of chips. For example, implantable devices made using TSMC’s 180 nm (April 2022), the article in Scientific Reports about neural network chip using UMC’s 180 nm (May 2022), and non-volatile memory offerings such as given here. (A comprehensive list is beyond the scope of this article.)
Current Status And Outlook For 200 mm Fabs Worldwide
Here, it is said that “200mm foundries have been operating at near 100% capacity for the last three to four years”. Within the next five years when the new capacity is still in the making, some foundries with 180 nm capability may get more orders than what they can process and SCL may be able to grab some sub-contracts if appropriately positioned.
As per SEMI, “200mm fab activity is strong, very much alive and flexing its muscles”. In its April 2022 report, Ajit Manocha (former CEO of GlobalFoundries and one of the 19 members of the advisory committee for India Semiconductor Mission) said, “Wafer manufacturers will add 25 new 200mm lines over the five-year period to help meet growing demand for applications such as 5G, automotive and Internet of Things (IoT) devices that rely on devices like analog, power management and display driver integrated circuits (ICs), MOSFETs, microcontroller units (MCUs) and sensors."
There may be temporary downtrends, such as the one recently triggered by worries of inflation and potential recession. But, in the medium term (that is over the next four to five years), SCL’s 200 mm fab line can target to process 4,000 wafers per month, as proposed in Part 1.
This capacity should be good enough to make a large number of some or many types of chips both for domestic use and exports; indeed, SCL will have to capture the orders.
How Many Chips And What Kind Of Revenue?
If the chip (die) size is known, the number of dies per wafer can be calculated.
In this chip gallery, pictured are many chips using 180 nm. If we take the size of the “2.5-3.2GHz CMOS differentially-controlled continuously-tuned varactor-less LC-VCO in 180nm CMOS” (0.7mm X 0.7mm), we get more than 45,000 dies per wafer.
Each die is cut out and packaged as an individual chip — SCL is known to have packaging capability as well.
The revenue generated, if all wafers are done in the foundry mode, will depend on the sale price per wafer. As per a 2018 estimate, each processed 180 nm wafer can be sold at $625. Given the increasing demand, the base price may be higher now and special purpose wafers may cost even more.
Going by the $625 per wafer price, 4,000 WPM (48,000 wafers per year) leads to a revenue of $30 million — approximately Rs 225 crore. There is also the MEMS line that can do an additional 2,000 WPM.
The 2020-21 operational budget estimate for SCL was Rs 256 crore; a large share of this figure is supposed to account for employee salary/pension commitments and the annual maintenance contract, or, in other words, fixed costs. The running expense of the fab seems to be a smaller percentage. Indeed, raising the capacity will increase the share, but this variable cost is the part that will fetch the revenue and is the only way to have hopes of getting returns on investments already made.
Moreover, the SCL can do end-to-end activity — that is, design, fabrication, testing, and packaging (as explained here, it is an IDM). Hence, if the SCL is also run with products of its own (in addition to offering the fab as a foundry), it can fetch more revenue.
The design, testing, and packaging capability, too, can be (or is already) offered on a contract basis. If all these are fully utilised and with proper cost control and manpower management, it may be possible to get to profitability.
So...
It can be said that there are many types of chips that can be made using 180 nm technology that SCL Mohali has and which is or will be very much in use. The examples provided in this article provide a feel for what can be achieved.
Whether under the government, or partially or fully as a private industry, the success of SCL's commercialisation will need passionate, decisive, and technically prudent leadership and a similar work force.
In addition to approving a new 300 mm commercial silicon fab(s), the SCL can be turned around as a useful 200 mm fab.
In the next and final part of this series, some points about the existing "R and D fabs" in India and the lack of logic in asking for a new one will be discussed.
Read: Rise (Part 1): Ramping Up India’s First Fab To Higher Capacity
Arun Mampazhy has a BTech from IITM and MS from University of Maryland in semiconductor fabrication and over a decade of industry experience. His dreams of seeing a commercial fab takeoff in India has changed from black and white to colour over two decades. He can be reached via email nanoarun(at)gmail(dot)com or @nano_arun on twitter. Views expressed are personal.