Tag Archives: Top Story Left

Worldwide silicon wafer area shipments increased during the second quarter 2013 when compared to first quarter 2013 area shipments, according to the SEMI Silicon Manufacturers Group (SMG) in its quarterly analysis of the silicon wafer industry.

Read more: Quarterly semiconductor sales increase 6%, outperforming industry forecast

Total silicon wafer area shipments were 2,390 million square inches during the most recent quarter, a 12.3 percent increase from the 2,128 million square inches shipped during the previous quarter. New quarterly total area shipments are 2.3 percent lower than second quarter 2012 shipments.

"Total quarterly silicon shipment volumes accelerated in the most recent quarter in contrast to the first quarter” said Byungseop (Brad) Hong, chairman of SEMI SMG and director of Global Marketing at LG Siltron. “As such, silicon shipment volumes for the first half of this year are trending at a slightly higher level than the first half of 2012.”

Quarterly Silicon Area Shipment Trends
Semiconductor Silicon Shipments* – Millions of Square Inches

 

Million of Square Inches

 

Q2 2012

Q1 2013

Q2 2013

TOTAL

2,447

2,128

2,390

*Shipments are for semiconductor applications only and do not include solar applications

 Silicon wafers are the fundamental building material for semiconductors, which in turn, are vital components of virtually all electronics goods, including computers, telecommunications products, and consumer electronics. The highly engineered thin round disks are produced in various diameters (from one inch to 12 inches) and serve as the substrate material on which most semiconductor devices or "chips" are fabricated.

All data cited in this release is inclusive of polished silicon wafers, including virgin test wafers, epitaxial silicon wafers, and non-polished silicon wafers shipped by the wafer manufacturers to the end-users.

The Silicon Manufacturers Group acts as an independent special interest group within the SEMI structure and is open to SEMI members involved in manufacturing polycrystalline silicon, monocrystalline silicon or silicon wafers (e.g., as cut, polished, epi, etc.). The purpose of the group is to facilitate collective efforts on issues related to the silicon industry including the development of market information and statistics about the silicon industry and the semiconductor market.

SEMI is the global industry association serving the nano- and micro-electronic manufacturing supply chains.

Crossbar, Inc., a start-up company, unveiled a new Resistive RAM (RRAM) technology that will be capable of storing up to one terabyte (TB) of data on a single 200mm2 chip. A working memory was produced array at a commercial fab, and Crossbar is entering the first phase of productization. “We have achieved all the major technical milestones that prove our RRAM technology is easy to manufacture and ready for commercialization,” said George Minassian, chief executive officer, Crossbar, Inc. The company is backed by Artiman Ventures, Kleiner Perkins Caufield & Byers and Northern Light Venture Capital.

The technology, which was conceived by Professor Wei Lu of the University of Michigan, is based on a simple three-layer structure of silver, amorphous silicon and silicon (Fig. 1). The resistance switching mechanism is based on the formation of a filament in the switching material when a voltage is applied between the two electrodes. Minassian said the RRAM is very stable, capable of withstanding temperature swings up to 125°C, with up to 10,000 cycles, and a retention of 10 years. “The filaments are rock solid,” he said.

 

Crossbar has filed 100 unique patents, with 30 already issued, relating to the development, commercialization and manufacturing of RRAM technology.

After completing the technology transfer to Crossbar’s R&D fab and technology analysis and optimization, Crossbar has now successfully developed its demonstration product in a commercial fab.  This working silicon is a fully integrated monolithic CMOS controller and memory array chip. The company is currently completing the characterization and optimization of this device and plans to bring its first product to market in the embedded SOC market.

Sherry Garber, Founding Partner, Convergent Semiconductors, said: “RRAM is widely considered the obvious leader in the battle for a next generation memory and Crossbar is the company most advanced to show working demo that proves the manufacturability of RRAM.  This is a significant development in the industry, as it provides a clear path to commercialization of a new storage technology, capable of changing the future landscape of electronics innovation.”

Crossbar technology can be stacked in 3D, delivering multiple terabytes of storage on a single chip. Its simplicity, stackability and CMOS compatibility enables logic and memory to be integrated onto a single chip at the latest technology node.

Crossbar’s technology will deliver 20x faster write performance; 20x lower power consumption; and 10x the endurance at half the die size, compared to today’s best-in-class NAND Flash memory. Minassian said the biggest advantage of the technology is its simplicity. “That allowed us in three years time to get from technology understanding, characterization, cell array and put a device together,” he said.

Minassian said RRAM compares favorably with NAND, which is getting more complex and expensive. “In 3D NAND, you put all of these thing layers of top of each other – 32 layers, or 64 or 128 in some cases – then you have to etch them, you have to slice them all at once and the equipment required for that accuracy and that geometry is very expensive. This is one of the reasons that 3D has been very difficult for NAND to be introduced.” With the Crossbar approach, “you’re always dealing with three layers. It’s much easier to stack these and it gives you a huge density advantage,” Minassian said.

“The switching media is highly resistive,” explains Minassian. “If you try to read the resistance between top and bottom electrode without doing anything, it’s a high resistance. That’s the off state. To turn on the device, we apply a positive voltage to the top electrode. That ionizes the metal on the top layer and puts the metal ions into the switching media. The metal ions form a filament that connect the top and bottom electrode. The moment they hit the bottom electrode, you have a short, which means that the top and bottom electrode are connected which means they have a low resistance.” The low resistance state is the on state. He said that although silver is not commonly used in front-end CMOS processing, the RRAM memory formation process is a back-end process. “You produce all your CMOS and then right before the device exits the fab, you put the silver on top,” he said. The silver is deposited, encapsulated, etched and then packaged. “That equipment is available, you just have to isolate it at the end,” Minassian said.

The approach is also CMOS compatible, with processes used to fabricate the memory layers all running at less than 400°C.  “This allows you to not only be CMOS compatible, but it allows you to stack more and more of these memory layers on top of each other,” Minassian said. “You can put the logic, the controllers and microprocessors, next to the memory in the same die. That allows you to simplify packaging and increase performance.”

Another advantage compared to NAND is that the controllers used to address the cells can be less complicated. Minassian said that in conventional cells, 30 electrons are required to produce  1 Volt. “If you shrink that to a smaller node, the number of electrons is less. Fewer electrons are much harder to detect. You need a massive controller that does error recovery and complex coding so if the bits are changed, it can still provide you the right program to execute.” Also, because the Crossbar RRAM is capable of 10,000 write cycles, less complicated controllers are needed. Today’s NAND is capable of only 1000 write cycles. “If you write information 1000 times, that cell is destroyed. It will not contain or maintain the information. You have this complex controller that keeps track of how many cells have been written, how many times, to make sure all of them are aged equally,” Minassian said.  

Non-volatile memory, expected to grow to become a $60 billion market in 2013, is the most common storage technology used for both code storage (NOR) and data storage (NAND) in a wide range of electronics applications. Crossbar plans to bring to market standalone chip solutions, optimized for both code and data storage, used in place of traditional NOR and NAND Flash memory. Crossbar also plans to license its technology to SOC developers for integration into next-generation systems-on-chips (SOC).

Michael Yang, Senior Principal Analyst, Memory and Storage, HIS, said: “Ninety percent of the data we store today was created in the past two years.  The creation and instant access of data has become an integral part of the modern experience, continuing to drive dramatic growth for storage for the foreseeable future.  However, the current storage medium, planar NAND, is seeing challenges as it reaches the lower lithographies, pushing against physical and engineering limits.  The next generation non-volatile memory, such as Crossbar’s RRAM, would bypass those limits, and provide the performance and capacity necessary to become the replacement memory solution.”

Later this month, IC Insights’ August Update to the 2013 McClean Report will show a ranking of the top 25 semiconductor suppliers in 1H13.  A preview of the top 20 companies is listed in Figure 1.  The top 20 worldwide semiconductor (IC and O S D—optoelectronic, discrete, and sensor) sales leaders for 1H13 include eight suppliers headquartered in the U.S., four in Japan, three in Europe, three in Taiwan, and two in South Korea.

The top-20 ranking includes three pure-play foundries (TSMC, GlobalFoundries, and UMC) and four fabless companies.  IC foundries are included in the top-20 semiconductor supplier ranking because IC Insights has always viewed the ranking as a top supplier list, not as a marketshare ranking, and realizes that in some cases semiconductor sales are double counted.  With many of our clients being vendors to the semiconductor industry (supplying equipment, chemicals, gases, etc.), excluding large IC manufacturers like the foundries would leave significant “holes” in the list of top semiconductor suppliers.  Overall, the list shown in Figure 1 provides a guideline to identify which companies are the leading semiconductor suppliers, whether they are IDMs, fabless companies, or foundries.

There were numerous changes within the top-20 semiconductor ranking in 1H13 as compared to the top 20 ranking of 2012.  Some of the companies rising in the ranking included SK Hynix, which moved up three places and into the top 5; Broadcom, which edged into the top 10; Elpida, which was officially purchased by Micron on July 31, 2013, shot up seven places to 17th place; and MediaTek, which jumped up four positions to make it into the top 20 (now ranked 18th).  In contrast, Fujitsu dropped five places and fell out of the top 20 ranking, going from being ranked 17th in 2012 to 22nd in 1H13.  The other company to fall out of the top 20 ranking was fabless supplier Nvidia, which went from being ranked 18th in 2012 to 21st in 1H13, even though the company posted a two percent increase in year-over-year sales.  Another “casualty” in the top 20 ranking was Sony, which fell to 16th place in 1H13 from the 12th position in 2012.

Micron’s acquisition of Elpida was completed on July 31, 2013.  It is interesting to note that if Micron and Elpida’s 1H13 sales were combined, the “new” company would have had $6,699 million in total sales in 1H13 and would have been ranked as the fifth-largest semiconductor supplier worldwide.  Now that the two companies are officially combined, look for Micron to move up in the ranking of top suppliers over the remainder of 2013 and in 2014.

Figure 1

In total, the top 20 semiconductor companies’ sales increased by 4 percent in 1H13 as compared to 1H12, one point better than the total 1H13/1H12 worldwide semiconductor market increase of 3 percent.  It took semiconductor sales of just over $1.9 billion in 1H13 to make the top-20 ranking.

As shown in Figure 2, there was a 64-percentage-point range of growth rates among the worldwide top 20 semiconductor suppliers in 1H13 (from +38 percent for SK Hynix to -26 percent for Sony).  The continued success of the fabless/foundry business model is evident when examining the top 20 semiconductor suppliers ranked by growth rate.  As shown, the top 10 performers included three fabless companies (Qualcomm, MediaTek, and Broadcom) and three pure-play foundries (TSMC, GlobalFoundries, and UMC).

Figure 2

Figure 2 illustrates that two of the three top-20 ranked companies that registered a double-digit sales decline in 1H13 were headquartered in Japan (Renesas and Sony).  Japan-based Fujitsu also registered a double-digit decline (-19 percent) in 1H13 to drop out of the top 20 ranking.  However, it should be noted that the conversion of Japanese company semiconductor sales from yen to U.S. dollars, at 95.47 yen per dollar in 1H13 versus 79.70 yen per dollar in 1H12, had a significant impact on the sales figures for the Japanese companies.  Still, Sony would have logged a double-digit (12 percent) semiconductor sales decline even if its sales results were not converted to U.S. dollars while Renesas would have posted a two percent increase in semiconductor sales if the numbers were expressed in yen.

Unfortunately for AMD, it cannot attribute its extremely poor 1H13 sales performance (-25 percent) to currency conversion issues.  However, the company’s 3Q13/2Q13 guidance is for a 22 percent surge in sales, a significant rebound but one that still may not prevent the company from posting another full-year decline in sales in 2013 (AMD registered a steep 17 percent sales decline in 2012).

More details on the 1H13 top 25 semiconductor suppliers, including a look at the companies’ 3Q13 expectations and guidance, will be provided in the August Update to The McClean Report.

 

Light and proximity sensors in mobile handsets and tablets are set for expansive double-digit growth within a five-year period, thanks to increasing usage by electronic giants Samsung and Apple. Light and proximity sensors can detect a user’s presence as well as help optimize display brightness and color rendering.

Revenue for the sensors is forecast to reach $782.2 million this year, up a prominent 41 percent from $555.1 million in 2012, according to insights from the MEMS and Sensors Service at information and analytics provider IHS. The market is also expected to grow in the double digits for the next three years before moderating to a still-robust eight percent in 2017. By then, revenue will reach $1.3 billion, as shown in the figure below.

“The continued growth of the smartphone and tablet markets serve as the foundation of a bright future for light sensors,” said Marwan Boustany, senior analyst for MEMS & sensors at IHS. “Market leaders in these areas are driving the growth, with Apple pioneering their adoption and Samsung later taking the lead in their usage.”

Sensor segments

There are three types of light and proximity sensors: ambient light sensors (ALS) that measure the intensity of the surrounding light enveloping a cellphone or tablet to adjust screen brightness and save battery power; RGB sensors that measure a room’s color temperature via the red, green and blue wavelengths of light to help correct white balance in the device display; and proximity sensors that disable a handset’s touch screen when it is held close to the head, in order to avoid unwanted input, and also to turn off the light in the display to save battery power.

Overall, the compound annual growth rate for the sensors from 2012 to 2017 equates to 19 percent.

Driving this growth is the shift in use from ALS to RGB in mid- to high-end smartphones; the growing deployment of proximity sensors with gesture capabilities compared to just simple proximity sensors; and the price premiums associated with such changes in usage.

Aside from their most conspicuous use in wireless communications typified by handsets and tablets, light sensors are also utilized in various other applications. These include consumer electronics and data processing for devices like televisions, laptops and PC tablets; the industrial market for home automation, medical electronics and general lighting; and the automotive space for vehicle displays and car functionalities like rain sensors.

Samsung and Apple are leaders in sensor use

Both Samsung and Apple have made use of light and proximity sensors in recent years, helping the sensor market grow in no small measure.

In 2010, Apple included an RGB and proximity sensor for its iPhone 4 and an RGB sensor in its iPad, even though the sensors were subsequently dropped in the iPhone 4S, iPhone 5 and later iPads. Apple let go of the sensors, which were made available at that time in a combination—or combo package—in favor of discrete solutions consisting of individual proximity as well as ALS sensors for its products. While combo sensors offer the convenience of a single configured package and sourcing from a single supplier, discrete solutions can offer flexibility in the choice of sensor.

Samsung, meanwhile, has gone on to use light and proximity sensors in even larger quantities than Apple. Last year Samsung included an RGB, proximity and infrared (IR) combo sensor, for both its Galaxy SIII smartphone and flagship Galaxy Note 2 device that the company termed as a “phablet.” This year, Samsung deployed a discrete RGB sensor in its latest smartphone, the Galaxy S4, switching from a combo package due to lack of availability of a combo sensor with gesture capability. Samsung’s move toward using RGB sensors in its high-end handsets currently sets the tone for the RGB sensor market given Samsung’s high unit sales. Such a move by the South Korean maker is expected to open the door for other brands to also include RGB sensors in their handsets and tablets, IHS believes.

The new gesture functionality, such as that found in the Galaxy S4, will see especially vigorous growth in the years to come, with revenue enjoying an astonishing 44 percent compound annual growth rate from 2013 to 2017. Maxim Integrated Solutions of California provides the discrete gesture solution for the Galaxy S4, but Japan’s Sharp will be producing a combo sensor product with gesture capabilities by September this year.

Sensor suppliers and buyers tussle

Samsung and Apple are the top buyers of light sensors, accounting for more than 50 percent of light sensor revenue last year. Samsung pulled away from Apple after impressive 90 percent growth in sensor purchases between 2011 and 2012, compared to Apple’s 54 percent growth rate of spend during the period.

This is due to Samsung’s shift toward RGB sensors in its Note 2 and SIII devices, which command higher average selling prices. In third place after Samsung and Apple is a collective group of original equipment manufacturers from China. Included here are global players with significant name recognition like Huawei Technologies, ZTE and Lenovo, as well as a multitude of lesser-known companies such as Coolpad and Xiaomi.

Meanwhile, the top sensor suppliers are Austrian-based ams via its Taos unit in Texas, which supplies to Apple; and Capella Microsystems from Taiwan, the top light sensor supplier to Samsung. Together the two manufacturers furnish more than half of the light sensor market. Other important sensor makers are Avago Technologies from California and Sharp from Japan.

IC Insights’ new 250-page Mid-Year Update to the 2013 McClean Report, which is slated to be released by the end of July, describes why a very clear distinction should be made between the IC market (i.e., consumption) in China and IC production within China.  Although China has been the largest individual market for ICs since 2005, it does not necessarily mean that large increases in IC production within China would immediately follow, or ever follow. IC production in China represented only 11.1 percent of its $81 billion IC market in 2012.  Moreover, IC Insights forecasts that this share will increase only about three percentage points to 14.4 percent in 2017.

China-based IC production is forecast to exhibit a very strong 2012-2017 CAGR of 17.6 percent.  However, considering that China-based IC production was only about $8.9 billion in 2012, this growth will come off a relatively small base.  In 2012, SK Hynix, TSMC, and Intel were the major foreign IC manufacturers that had significant IC production in China.  In fact, SK Hynix’ China fab had the most capacity of any of its fabs last year.  In 2012, Intel continued to ramp-up its 300mm fab in Dalian, China (it started production in late October 2010), which is expected to give a noticeable boost to the China-based IC production figures over the next few years.  This fab currently has an installed capacity of 30,000 300mm wafers per month with a maximum capacity of 52,000 wafers per month.

In early 2012, Samsung gained approval from the South Korean government to construct a 300mm IC fabrication facility to produce NAND flash memory in Xian, China.  Samsung started construction of the fab in September of 2012 with production set to begin in the first half of 2014.  The company expects to invest $2.3 billion in the first phase of the fab with $7.0 billion budgeted in total.  This facility is targeting NAND flash production using a 10-19nm feature size process technology.

If China-based IC production rises to $20.0 billion in 2017 as forecast, it would still represent only 5.6 percent of the total forecasted 2017 worldwide IC market of $359.1 billion.  Even after adding a significant “markup” to many of the Chinese producers’ IC sales figures (since many of the Chinese IC producers are foundries that sell their ICs to companies that re-sell these products to the electronic system producers), China-based IC production would still represent less than 10 percent of the global IC market in 2017.

China's IC producers

Historically, the lack of consistent intellectual property protection has been a major deterrent for foreign firms seeking to establish state-of-the-art IC fabrication facilities in China.  The lack of intellectual property protection is also a reason many large fabless IC suppliers (e.g., Qualcomm, Broadcom, etc.) have not brought leading-edge IC designs into China for the indigenous Chinese IC foundries to manufacture.  It should also be noted that, thus far, Chinese IC foundries have also been unable to offer large amounts of IC production using leading-edge feature sizes.

IC Insights believes that the future size of the IC production base in China is more dependent upon whether foreign companies continue to locate, or re-locate, IC fabrication facilities in China than on the success of indigenous Chinese IC producers (e.g., SMIC, Hua Hong Grace, etc.).  As a result, IC Insights forecasts that at least 70 percent of IC production in China in 2017 will come from foreign companies such as SK Hynix, TSMC, Intel, and Samsung.

As LED lighting becomes an $80 billion industry, the market for the epitaxial wafers (epi-wafers) LEDs are made from will grow to $4 billion in 2020, according to Lux Research.

The vast majority of these epi-wafers are gallium nitride (GaN)-on-sapphire today. GaN-on-silicon is the leading emerging technology with a strong economic allure – silicon is just one-eighth the cost of a sapphire substrate – but technical challenges will limit it to only a 10% market share in 2020. GaN-on-silicon carbide (SiC), championed by Cree, will grow to 18 percent market share.

epi wafer market

“Silicon is already widely used for electronics, and some LED die manufacturers are hoping to take advantage of silicon substrates,” said Pallavi Madakasira, Lux Research Analyst and lead author of the report titled, “Dimming the Hype: GaN-on-Si Fails to Outshine Sapphire by 2020.”

“But GaN-on-Si is more prone to cracking than GaN-on-sapphire, and mitigating this mismatch is expensive,” she added.

Lux Research analysts studied the market for GaN-on-sapphire, GaN-on-SiC, GaN-on-bulk GaN, and GaN-on-Si epi-wafers, evaluating each technology’s economic prospects as the industry moves to larger wafer sizes. Among their findings:

  • Choice and cost of LEDs will determine adoption. Where GaN-on-sapphire is suited to all applications, GaN-on-bulk GaN will be relegated to niche commercial lighting and GaN-on-Si, with unproven performance, will be better suited to cost-sensitive residential applications.
  • Four-inch wafers will rule, though six-inch wafers start to come into vogue. Four-inch wafers will peak at 62 percent market share with $2.1 billion in 2017 sales. Later, the LED industry will move towards 6” epi wafers, which will take a 35% share, equivalent to $1.4 billion, in 2020.
  • Technology will advance sapphire substrates. Sapphire substrate manufacturing technology has advanced significantly with specialists such as Rubicon and Monocrystal demonstrating substrates up to 12 inch in diameter. New methods like hydride vapor phase epitaxy (HVPE) will further improve throughput and cut costs, keeping sapphire highly competitive for the rest of the decade.

The report, titled “Dimming the Hype: GaN-on-Si Fails to Outshine Sapphire by 2020,” is part of the Lux Research Energy Electronics Intelligence service.

Engineering samples of The Hybrid Memory Cube (HMC) are expected this summer, with high volume manufacturing coming next year. It will be one of the first high volume devices employing 3D integration and through silicon vias (TSVs), employing a bottom logic layer and 4-8 stacked DRAM layers.

The HMC is the result of a consortium formed in late 2011 by Micron, Samsung, Altera, Xilinx and Open-Silicon to define an industry interface specification for developers, manufacturers and architects of high-performance memory technology. The consortium has grown to 110 members, including SK Hynix, IBM and ARM. Analysts are projecting the TSV-enabled 3D market to be a $40billion market by 2017, or roughly about 10% of the global chip business.

We caught up with Micron’s Scott Graham, General Manager, Hybrid Memory Cube, at Semicon West. “Today, we’re very close to delivering our engineering samples this summer to our lead customers that are taking the technology into their system designs,” Graham said.  The lead applications are in high performance computing, such as supercomputers, as well as the higher end networking space. “Those will be the early adopters. As we move forward in time, we’ll see that technology evolve as costs come down for TSVs and manufacturing technology, it will enter into future space where traditional DDR type of memory has resided. Beyond DDR4, we can certainly see this of memory technology being a mainstream memory,” Graham said.

Since the HMC is an open specification in terms of the architecture of the device, it will be up to each memory manufacturer to decide how it might be customized and manufactured. “The way it’s done today is we source the substrate, we source the logic layer and then we bring those in-house and we complete the finishing of those logic wafers as well as all the slicing, dicing, stacking, assembly and test,” Graham said. “What we end up providing for the customer is a known good cube, or known good piece of memory, just like we would if it was a DDR device or wide I/O device,” he said. He added that the HMC is designed so that it has not only the repair capability during manufacturing but also out in the field. “It’s very flexible and very robust, so reliability is very high with this device,” he said.

The consortium delivered its first specification earlier this year. “We’ve since extended the consortium to work on both future generations of the HMC technology in both the short-reach and ultra-short reach configurations,” Graham said.

The HMC was designed to get high density and high bandwidth in a relatively small package. The team adopted an off-the-shelf SERDES I/O and that’s based on IBM’s 32nm process. “With that, we can achieve 10 Gbps, 12.5 Gbps, or 15 Gbps for those SERDES links,” Graham said. “If you look at a 2 GByte or a 4GByte HMC device, those first devices will deliver a total aggregate bandwidth of 160GBytes/sec. I want to emphasize those are bytes not bits. It’s a very high bandwidth and low energy per bit device that is something that can be designed into a multitude of systems.”

The consortium has several generations of the HMC device planned (this summer’s engineering samples are Gen2). “As we move forward, you’ll see us moving into the 28 Gbps SERDES as far as the I/O goes,” Graham said. Bandwidths are going to be 320 Gigabytes/sec and higher, and the density will be in 4Gbyte and 8 Gbyte configurations.

Graham said one of the main challenges they had to overcome was the stacking. “We’re stacking a logic layer on top of a substrate and then four to eight DRAM on top of those logic layers,” he said. “We have over 2000 TSVs in this package and it was a challenge to stack these ultrathin die and make sure that what we end up with is a high performance and very reliable package.” Graham declined to comment on the exact TSV process flow used at Micron, saying only that it was leading edge. “We had to make sure our equipment partners were up to speed and could deliver us the technology that would allow us to manufacture this in high volume,” he said.  

Because customer can customize the HMC design, another challenge it to make sure that the design capabilities are available at the foundry for that logic layer, Graham said.  

Heat dissipation in the device is achieved through a metal lid, and through the TSVs which acts as chimneys (in addition to conducting electricity). The photo shows two Gen2 HMC devices. The larger one, in a 31mm x31 mm package, is a 4 link device that will achieve 160 Gig-bytes per second. The smaller one is a two link device capable of 120 Gigabytes/sec, measuring 16mm x 19.5 mm. “Both are being manufactured now in our plant and we’re doing the whole debug phase,” explained Aron Lunde, program manager, DRAM solutions group at Micron in Boise. He said the metal lid was in contact with not only the top layer, but different internal layers. “We call it an integrated heat spreader. It makes contact at more than one level and that’s what really helps,” he said.

Although manufacturers such as Micron, Samsung and SK Hynix must now handle the manufacturing, assembly and test process, Graham believes that it could eventually evolve to the point where select foundry partners would be able to provide volume manufacturing services for these HMC cubes.

Graham said DDR4 will likely be the last DDR device. “Beyond DDR4, you have to move to managed memory like HMC technology,” he said.  “We’re solving the memory wall problem with HMC-like architecture and what’s really going to be happening in the future is that you’ll be running into a CPU wall. That’s going to be the barrier to system progress as we move forward.”

Graham expects some challenges with scaling of conventional memory at sub-20nm process nodes. “We get into physical challenges of meeting the timing requirements and the 12 pages of JEDEC specifications to be able to yield properly and to be able to provide a cost-effective memory device moving forward,” he said.  

Although the HMC is now designed around DRAMs, Graham said it would be possible to use other types of memories, and even a mixed set of memories. He noted Micron is looking at alternatives to the conventional DRAM cell, such as spin torque and resistive memories. “Micron is investing heavily in research in those technologies and of course the HMC team here at Micron is looking at future technologies that we can take HMC architecture and be able to utilize different DRAM or even flash types of memory,” he said. “As the technology matures and it becomes lower cost, we can see this technology certainly evolving into more global applications and utilizing different memory types in that stack – and perhaps even multiple memory types in that stack.”

HMCs could eventually make their way into mobile devices, but Graham said that is likely to be three or four years away. Mobile applications presently employ low power DDR3 solutions, which will be used for several years. “We’ll see quite a few interesting designs start spinning when the mobile folks see they can differentiate with a managed memory solution. It’s not going to be HMC as we know it today, it will have to be optimized for mobile,” Graham said.

Solid State Technology and SEMI announced the recipient of the 2013 “Best of West” Award — Mentor Graphics — for its Tessent TestKompress with Cell-Aware ATPG.  The award recognizes important product and technology developments in the microelectronics supply chain. Held in conjunction with SEMICON West, the largest and most influential microelectronics exposition in North America, the Best of West finalists were selected based on their financial impact on the industry, engineering or scientific achievement, and/or societal impact.

The Mentor Graphics Tessent TestKompress with Cell-Aware ATPG significantly improves on the standard process for testing digital integrated circuits and reduces failure escape rate by detecting defects at the transistor level that are missed by traditional automatic test pattern generation (ATPG) techniques. The Mentor Graphics booth is in the North Hall, Booth #6243. Their winning entry was entered in the “Silicon Test Solutions, Facilities & Software” category.

The Best of West Award winner was announced during SEMICON West on Wednesday, July 10, 2013 at 1:00pm.

When Ajit Manocha, GlobalFoundries CEO, polled his audience during his keynote address on Tuesday at SEMICON West 2013, nearly 60 percent of the audience believed that the biggest challenge facing the semiconductor industry was the economy. However, during his presentation, Manocha seemed to suggest otherwise.

The technology business is booming, according to Manocha, who shared with SEMICON attendees that the mobile business is forecast to be double the size of the PC market in 2016. The mobile business drives many new requirements, said Manocha, including power, performance and features, higher data rates, high resolution multicore processors and thinner form factors.

This incredible growth is driving new dynamics, said Manocha, and pushing the industry to the new technology node each year, which is presenting the industry with what Manocha deems the Big Five Challenges. Manocha believes these challenges are: cost, device architectures, lithography and EUV, packaging and the 450mm wafer transition.

Cost, said Manocha, continues to be the underlying challenge of the entire industy, because, without focusing on wafer cost, even in good times, a company can enter into what he called “profitless prosperity.” Unfortunately, with the introduction of a new technology node each year, advanced technology costs are rapidly rising.

“Fab cost alone escalates 40 percent year after year,” said Manocha.

To keep wafer costs down, what Manocha believes the industry needs for success is a new foundry model altogether. His model, which he calls Foundry 2.0, hinges on industry collaboration rather than wafer price competition. By encouraging the industry to work together on products and meet the same goals, the industry can see a faster rate of change and tap into global R&D talent.

“The best solutions rarely originate from an insultated team,” he said. “It’s critical that we understand what customers need.”

SEMI recognizes GLOBALFOUNDRIES CEO

Click here for more news from SEMICON West or be sure to follow us on Twitter.

What do you think? Tell us in the comment section below.

Since 2008, the majority of integrated circuit production has taken place on 300mm wafers.  In terms of surface area shipped (i.e., on a normalized 200mm-equivalent wafer basis), 300mm wafers represented 56 percent of worldwide installed capacity in December 2012.  Production using 300mm wafers is forecast to steadily increase and reach 70.4 percent in 2017, according to IC Insights’ Global Wafer Capacity 2013 report (see figure).

300mm wafers

For the most part, 300mm fabs are, and will continue to be, limited to production of high-volume, commodity-type devices like DRAMs and flash memories, and very recently image sensors and power management devices; complex logic and microcomponent ICs with large die sizes; and products manufactured by foundries, which can fill a 300mm fab by combining wafer orders from many sources.

The list of companies with the most 300mm wafer capacity includes DRAM and flash memory suppliers like Samsung, SK Hynix, Toshiba, Micron, Elpida, and Nanya; the industry’s biggest IC manufacturer and dominant MPU supplier Intel; and two of the world’s largest pure-play foundries TSMC and GlobalFoundries.  These companies offer the types of ICs that benefit most from using the largest wafer size available to best amortize the manufacturing cost per die.

It is interesting to point out that when (or if) the pending acquisition of Elpida by Micron goes through as expected, the merged company will have the industry’s second-largest share of 300mm wafer fabrication capacity, trailing only fellow memory chip manufacturer Samsung.

Meanwhile, the share of the industry’s monthly wafer capacity represented by 200mm wafers is expected to drop from 32 percent in December 2012 to 21 percent in December 2017, as seen in the figure. Fabs running 200mm wafers will continue to be profitable for many more years and be used to fabricate numerous types of ICs, such as specialty memories, image sensors, display drivers, microcontrollers, analog products, and MEMS-based devices.  Such devices are certainly practical in fully depreciated 200mm fabs that were formerly used in making devices now produced on 300mm wafers.

A significant trend with regard to the industry’s IC manufacturing base, and a perhaps worrisome one from the perspective of companies that supply equipment and materials to chip makers, is that as the industry moves IC fabrication onto larger wafers in bigger fabs, the group of IC manufacturers continues to shrink in number.  There are about 61 percent fewer companies that own and operate 300mm wafer fabs than 200mm fabs.  The distribution of worldwide 300mm wafer capacity among those manufacturers is very top-heavy.  Essentially, there are only about 15 companies that comprise the entire future total available market (TAM) for leading-edge IC fabrication equipment and materials, according to the Global Wafer Capacity 2013 report.  When 450mm wafer fabrication technology comes into existence, this manufacturer group is predicted to shrink even further to a maximum of just 10 companies, and a few of those are questionable.  Despite growing momentum, IC Insights expects that 450mm wafer capacity will account for only one-tenth of a percent of global IC capacity in December 2017.