Category Archives: Displays

According to the recently published TechSci Research report, “United States Photomask Market Forecast & Opportunities, 2019“, the photomask market in the United States is forecast to reach $474.58 billion by 2019. The Western region dominates the photomask market in the US, in terms of revenue share, due to the presence of a large number of semiconductor and FPD manufacturing facilities in the region.

Besides widespread usage in the semiconductor industry, the FPD application segment accounted for the largest revenue share in the US photomask market in 2013, where Hoya Corporation emerged as leading player in the country, in terms of photomask shipments for FPD application. While photomasks are widely used in the manufacture of TFT-LCD display panels, growth in LED and Laser panel display markets is expected to benefit the overall demand for photomasks in the FPD industry in the US, over the next five years.

Photomasks are key tools used in the manufacture of semiconductors, Flat Panel Displays (FPD), optical devices and other products. The photomask market is largely dependent on growth in the overall semiconductor industry. During the last three years, the United States photomask market witnessed negative growth in terms of unit shipment, due to decline in the country’s semiconductor industry. The industry exhibited a negative growth of 3.6 percent during 2011-12, predominantly due to steep decline in the desktop PC and components markets. However, with anticipated recovery in the desktop PC market over the next five years, the US semiconductor industry is forecast to register moderate growth, thus positively influencing the photomask market.

“Photomasks with soda lime and quartz base glass material are two widely available types, where the former variety accounts for majority revenue share, in both value and volume terms. Soda lime base glass material photomasks are easily affected by temperature disparity, resulting in higher overlay error when loading multiple layers. However, soda lime base glass material photomask segment is expected to witness steady growth over the next five years due to relatively low pricing of this material. On the contrary, quartz base glass material photomasks are popular due to low impurity levels, which offer high optical and thermal characteristics. As a result, quartz base glass material photomask segment is expected to gain traction in the United States photomask market, despite being costlier than soda lime base glass material photomasks,” said Mr. Karan Chechi, Research Director with TechSci Research, a research based global management consulting firm.

“United States Photomask Market Forecast & Opportunities, 2019” has evaluated the future growth potential of the photomask market in the US and provides statistics and information on market structure, market size & share, market trends, etc. The report includes the United States photomask market projections and demand forecasting. The report is intended to provide cutting-edge market intelligence and help decision makers to take sound investment evaluation. Besides, the report also identifies and analyzes the emerging trends along with essential drivers, challenges and opportunities available in the United States photomask market.

Printed, flexible and organic electronic (PFOE) sensors can offer flexible form factors, larger area, lower cost, lower power, and better disposability compared to conventional sensors, key attributes for wearable applications. These attributes will allow them to grow into a $244 million market in wearables, according to Lux Research.

“With players from Apple to Intel to Kickstarter-funded start-ups launching devices, wearables are getting hot, but they still need to add functionality while trimming cost and size to really go mainstream,” said Jonathan Melnick, Lux Research Senior Analyst and lead author of the report titled, “Dial-Up Sensors: Printed, Flexible and Organic Sensors for the Things in the Internet of Things.”

“Printed, flexible, and organic electronic sensors can play a key enabling role for wearables — though many technology developers still need to improve performance, reliability and lifetime,” he added.

Lux Researchers analyzed the market for PFOE sensors across a wide variety of connected applications on the “Internet of Things” (IoT), include wearables, retail, transportation, and buildings. Among their findings:

  • Wearable, retail sensors drive growth. Wearable sensors, particularly for health and fitness, will be the biggest segment for PFOE sensors, but retail sensors — with a $117 million market in 2024 — will clock the fastest growth, a compound annual growth rate (CAGR) of 50% through the next decade.
  • Transportation, buildings remain small. Automotive and buildings, which have accounted for a lot of IoT hype, will be a bust for most PFOE sensors due to performance and reliability disadvantages and a limited addressable market.
  • PFOE sensors face opportunities and challenges. Six types of IoT sensors may be suited for PFOE technologies: motion, pressure, gas, temperature, electromagnetic and optical. For each, the value proposition comes down to manufacturing, form factor or size in each target application.

The report, titled “Dial-Up Sensors: Printed, Flexible and Organic Sensors for the Things in the Internet of Things,” is part of the Lux Research Printed, Flexible, and Organic Electronics Intelligence service.

Flexible LEDs


September 24, 2014

Flexible light-emitting diode (LED) displays and solar cells crafted with inorganic compound semiconductor micro-rods are moving one step closer to reality, thanks to graphene and the work of a team of researchers in Korea.

Currently, most flexible electronics and optoelectronics devices are fabricated using organic materials. But inorganic compound semiconductors such as gallium nitride (GaN) can provide plenty of advantages over organic materials for use in these devices — including superior optical, electrical and mechanical properties.

One major obstacle that has so far prevented the use of inorganic compound semiconductors in these types of applications was the difficulty of growing them on flexible substrates.

In the journal APL Materials, from AIP Publishing, a team of Seoul National University (SNU) researchers led by Professor Gyu-Chul Yi describes their work growing GaN micro-rods on graphene to create transferrable LEDs and enable the fabrication of bendable and stretchable devices.

“GaN microstructures and nanostructures are garnering attention within the research community as light-emitting devices because of their variable-color light emission and high-density integration properties,” explained Yi. “When combined with graphene substrates, these microstructures also show excellent tolerance for mechanical deformation.”

Why choose graphene for substrates? Ultrathin graphene films consist of weakly bonded layers of hexagonally arranged carbon atoms held together by strong covalent bonds. This makes graphene an ideal substrate “because it provides the desired flexibility with excellent mechanical strength — and it’s also chemically and physically stable at temperatures in excess of 1,000°C,” said Yi.

It’s important to note that for the GaN micro-rod growth, the very stable and inactive surface of graphene offers a small number of nucleation sites for GaN growth, which would enhance three-dimensional island growth of GaN micro-rods on graphene.

To create the actual GaN microstructure LEDs on the graphene substrates, the team uses a catalyst-free metal-organic chemical vapor deposition (MOCVD) process they developed back in 2002.

“Among the technique’s key criteria, it’s necessary to maintain high crystallinity, control over doping, formation of heterostructures and quantum structures, and vertically aligned growth onto underlying substrates,” Yi says.

When the team put the bendability and reliability of GaN micro-rod LEDs fabricated on graphene to the test, they found that “the resulting flexible LEDs showed intense electroluminescence (EL) and were reliable — there was no significant degradation in optical performance after 1,000 bending cycles,” noted Kunook Chung, the article’s lead author and a graduate student in SNU’s Physics Department.

This represents a tremendous breakthrough for next-generation electronics and optoelectronics devices — enabling the use of large-scale and low-cost manufacturing processes.

“By taking advantage of larger-sized graphene films, hybrid heterostructures can be used to fabricate various electronics and optoelectronics devices such as flexible and wearable LED displays for commercial use,” said Yi.

Researchers from the University of Texas at Austin and Northwestern University have demonstrated a new method to improve the reliability and performance of transistors and circuits based on carbon nanotubes (CNT), a semiconductor material that has long been considered by scientists as one of the most promising successors to silicon for smaller, faster and cheaper electronic devices. The result appears in a new paper published in the journal Applied Physics Letters, from AIP Publishing.

These are optical images of individual SWCNT field-effect transistors. Credit: S. Jang and A. Dodabalapur/University of Texas at Austin

These are optical images of individual SWCNT field-effect transistors.
Credit: S. Jang and A. Dodabalapur/University of Texas at Austin

In the paper, researchers examined the effect of a fluoropolymer coating called PVDF-TrFE on single-walled carbon nanotube (SWCNT) transistors and ring oscillator circuits, and demonstrated that these coatings can substantially improve the performance of single-walled carbon nanotube devices. PVDF-TrFE is also known by its long chemical name polyvinyledenedifluoride-tetrafluoroethylene.

“We attribute the improvements to the polar nature of PVDF-TrFE that mitigates the negative effect of impurities and defects on the performance of semiconductor single-walled carbon nanotubes,” said Ananth Dodabalapur, a professor in the Cockrell School of Engineering at UT Austin who led the research. “The use of [PVDF-TrFE] capping layers will be greatly beneficial to the adoption of single-walled carbon nanotube circuits in printed electronics and flexible display applications.”

The work was done in collaboration between Dodabalapur’s group at UT Austin and Mark Hersam’s group at Northwestern University as part of a Multi-University Research Initiative (MURI) supported by the Office of Naval Research.

A potential successor to silicon chips

Single-walled carbon nanotubes (SWCNT) are just about the thinnest tubes that can be wrought from nature. They are cylinders formed by rolling up a material known as graphene, which is a flat, single-atom-thick layer of carbon graphite. Most single-walled carbon nanotubes typically have a diameter close to 1 nanometer and can be twisted, flattened and bent into small circles or around sharp bends without breaking. These ultra-thin carbon filaments have high mobility, high transparency and electric conductivity, making them ideal for performing electronic tasks and making flexible electronic devices like thin film transistors, the on-off switches at the heart of digital electronic systems.

“Single-walled carbon nanotube field-effect transistors (FETs) have characteristics similar to polycrystalline silicon FETs, a thin film silicon transistor currently used to drive the pixels in organic light-emitting (OLED) displays,” said Mark Hersam, Dodabalapur’s coworker and a professor in the McCormick School of Engineering and Applied Science at Northwestern University. “But single-walled carbon nanotubes are more advantageous than polycrystalline silicon in that they are solution-processable or printable, which potentially could lower manufacturing costs.”

The mechanical flexibility of single-walled carbon nanotubes also should allow them to be incorporated into emerging applications such as flexible electronics and wearable electronics, he said.

For years, scientists have been experimenting with carbon nanotube devices as a successor to silicon devices, as silicon could soon meet its physical limit in delivering increasingly smaller, faster and cheaper electronic devices. Although circuits made with single-walled carbon nanotube are expected to be more energy-efficient than silicon ones in future, their drawbacks in field-effect transistors, such as high power dissipation and less stability, currently limit their applications in printed electronics, according to Dodabalapur.

A new technique to improve the performance of SWCNTs devices

To overcome the drawbacks of single-walled carbon nanotube field-effect transistors and improve their performance, the researchers deposited PVDF-TrFE on the top of self-fabricated single-walled carbon nanotube transistors by inkjet printing, a low-cost, solution based deposition process with good spatial resolution. The fluoropolymer coated film was then annealed or heated in air at 140 degrees Celsius for three minutes. Later, researchers observed the differences of device characteristics.

“We found substantial performance improvements with the fluoropolymer coated single-walled carbon nanotube both in device level and circuit level,” Dodabalapur noted.

On the device level, significant decreases occur in key parameters such as off-current magnitude, degree of hysteresis, variation in threshold voltage and bias stress degradation, which, Dodabalapur said, means a type of more energy-efficient, stable and uniform transistors with longer life time.

On the circuit level, since a transistor is the most basic component in digital circuits, the improved uniformity in device characteristics, plus the beneficial effects from individual transistors eventually result in improved performance of a five-stage complementary ring oscillator circuit, one of the simplest digital circuits.

“The oscillation frequency and amplitude [of the single-walled carbon nanotube ring oscillator circuit] has increased by 42 percent and 250 percent respectively,” said Dodabalapur. The parameters indicate a faster and better performing circuit with possibly reduced power consumption.

Dodabalapur and his coworkers attributed the improvements to the polar nature of PVDF-TrFE.

“Before single-walled carbon nanotube field-effect transistors were fabricated by inkjet printing, they were dispersed in an organic solvent to make a printable ink. After the fabrication process, there could be residual chemicals left [on the device], causing background impurity concentration,” Dodabalapur explained. “These impurities can act as charged defects that trap charge carriers in semiconductors and reduce carriers’ mobility, which eventually could deteriorate the performance of transistors.”

PVDF-TrFE is a polar molecule whose negative and positive charges are separated on different ends of the molecule, Dodabalapur said. The two charged ends form an electric bond, or dipole, in between. After the annealing process, the dipoles in PVDF-TrFE molecules uniformly adopt a stable orientation that tends to cancel the effects of the charged impurities in single-walled carbon nanotube field-effect transistors, which facilitated carrier flow in the semiconductor and improved device performance.

To confirm their hypothesis, Dodabalapur and his coworkers performed experiments comparing the effects of polar and non-polar vapors on single-walled carbon nanotube field-effect transistors. The results support their assumption.

The next step, Dodabalapur said, is to implement more complex circuits with single-walled carbon nanotube field-effect transistors.

In a major leap forward for gaming, NVIDIA  today introduced the first high-end products based on its Maxwell chip architecture — the new GeForce GTX 980 and 970 GPUs — delivering unmatched performance, major new graphics capabilities and twice the energy efficiency of the previous generation.

Maxwell is the company’s 10th-generation GPU architecture, following Kepler. The engine of next-generation gaming, it solves some of the most complex lighting and graphics challenges in visual computing.

Its new Voxel Global Illumination (VXGI) technology enables gaming GPUs for the first time to deliver real-time dynamic global illumination. Scenes are significantly more lifelike as light interacts realistically in the game environment — resulting in deeper levels of immersion for gamers.

And a range of new technologies — including multi-frame sampled antialiasing (MFAA), dynamic super resolution (DSR), VR Direct and extremely energy-efficient design — enable Maxwell-based GTX 980 and 970 GPUs to render frames with the highest fidelity at higher clock speeds and lower power consumption than any other GPU in their class.

“Maxwell has been years in the making, inspired by our gamers, and created by the best minds in 3D graphics,” said Jen-Hsun Huang, NVIDIA’s co-founder and chief executive officer. “Its extraordinary performance, efficiency and technologies will empower developers to do their finest work and delight gamers worldwide.”

Toward optical chips


September 19, 2014

Chips that use light, rather than electricity, to move data would consume much less power — and energy efficiency is a growing concern as chips’ transistor counts rise.

Of the three chief components of optical circuits — light emitters, modulators, and detectors — emitters are the toughest to build. One promising light source for optical chips is molybdenum disulfide (MoS2), which has excellent optical properties when deposited as a single, atom-thick layer. Other experimental on-chip light emitters have more-complex three-dimensional geometries and use rarer materials, which would make them more difficult and costly to manufacture.

In the next issue of the journal Nano Letters, researchers from MIT’s departments of Physics and of Electrical Engineering and Computer Science will describe a new technique for building MoS2 light emitters tuned to different frequencies, an essential requirement for optoelectronic chips. Since thin films of material can also be patterned onto sheets of plastic, the same work could point toward thin, flexible, bright, color displays.

The researchers also provide a theoretical characterization of the physical phenomena that explain the emitters’ tunability, which could aid in the search for even better candidate materials. Molybdenum is one of several elements, clustered together on the periodic table, known as transition metals. “There’s a whole family of transition metals,” says Institute Professor Emeritus Mildred Dresselhaus, the corresponding author on the new paper. “If you find it in one, then it gives you some incentive to look at it in the whole family.”

Joining Dresselhaus on the paper are joint first authors Shengxi Huang, a graduate student in electrical engineering and computer science, and Xi Ling, a postdoc in the Research Laboratory of Electronics; associate professor of electrical engineering and computer science Jing Kong; and Liangbo Liang, Humberto Terrones, and Vincent Meunier of Rensselaer Polytechnic Institute.

Monolayer — with a twist

Most optical communications systems — such as the fiber-optic networks that provide many people with Internet and TV service — maximize bandwidth by encoding different data at different optical frequencies. So tunability is crucial to realizing the full potential of optoelectronic chips.

The MIT researchers tuned their emitters by depositing two layers of MoS2 on a silicon substrate. The top layers were rotated relative to the lower layers, and the degree of rotation determined the wavelength of the emitted light.

Ordinarily, MoS2 is a good light emitter only in monolayers, or atom-thick sheets. As Huang explains, that’s because the two-dimensional structure of the sheet confines the electrons orbiting the MoS2 molecules to a limited number of energy states.

MoS2, like all light-emitting semiconductors, is what’s called a direct-band-gap material. When energy is added to the material, either by a laser “pump” or as an electrical current, it kicks some of the electrons orbiting the molecules into higher energy states. When the electrons fall back into their initial state, they emit their excess energy as light.

In a monolayer of MoS2, the excited electrons can’t escape the plane defined by the material’s crystal lattice: Because of the crystal’s geometry, the only energy states available to them to leap into cross the light-emitting threshold. But in multilayer MoS2, the adjacent layers offer lower-energy states, below the threshold, and an excited electron will always seek the lowest energy it can find.

Mind the gap

So while the researchers knew that rotating the layers of MoS2 should alter the wavelength of the emitted light, they were by no means certain that the light would be intense enough for use in optoelectronics. As it turns out, however, the rotation of the layers relative to each other alters the crystal geometry enough to preserve the band gap. The emitted light is not quite as intense as that produced by a monolayer of MoS2, but it’s certainly intense enough for practical use — and significantly more intense than that produced by most rival technologies.

The researchers were able to precisely characterize the relationship between the geometries of the rotated layers and the wavelength and intensity of the light emitted. “For different twisted angles, the actual separation between the two layers is different, so the coupling between the two layers is different,” Huang explains. “This interferes with the electron densities in the bilayer system, which gives you a different photoluminescence.” That theoretical characterization should make it much easier to predict whether other transition-metal compounds will display similar light emission.

“This thing is something really new,” says Fengnian Xia, an assistant professor of electrical engineering at Yale University. “It gives you a new model for tuning.”

“I expected that this kind of angle adjustment would work, but I didn’t expect that the effect would be so huge,” Xia adds. “They get quite significant tuning. That’s a little bit surprising.”

Xia believes that compounds made from other transition metals, such as tungsten disulfide or tungsten diselenide, could ultimately prove more practical than MoS2. But he agrees that the MIT and RPI researchers’ theoretical framework could help guide future work. “They use density-functional theory,” he says. “That’s a kind of general theory that can be applied to other materials also.”

North America-based manufacturers of semiconductor equipment posted $1.35 billion in orders worldwide in August 2014 (three-month average basis) and a book-to-bill ratio of 1.04, according to the August EMDS Book-to-Bill Report published today by SEMI.   A book-to-bill of 1.04 means that $104 worth of orders were received for every $100 of product billed for the month.

The three-month average of worldwide bookings in August 2014 was $1.35 billion. The bookings figure is 5.0 percent lower than the final July 2014 level of $1.42 billion, and is 26.5 percent higher than the August 2013 order level of $1.06 billion.

The three-month average of worldwide billings in August 2014 was $1.29 billion. The billings figure is 2.0 percent lower than the final July 2014 level of $1.32 billion, and is 19.5 percent higher than the August 2013 billings level of $1.08 billion.

“The SEMI Book-to-Bill ratio has been at or above parity for 11 consecutive months, and both current month bookings and billings continue to trend well above 2013 levels,” said Denny McGuirk, president and CEO of SEMI. “Strong equipment spending growth for the year is observed across the fab and test and assembly segments.”

The SEMI book-to-bill is a ratio of three-month moving averages of worldwide bookings and billings for North American-based semiconductor equipment manufacturers. Billings and bookings figures are in millions of U.S. dollars.

Billings
(3-mo. avg)

Bookings
(3-mo. avg)

Book-to-Bill
March 2014

$1,225.5

$1,297.7

1.06
April 2014

$1,403.2

$1,443.0

1.03
May 2014

$1,407.8

$1,407.0

1.00
June 2014

$1,327.5

$1,455.0

1.10
July 2014 (final)

$1,319.1

$1,417.1

1.07
August 2014 (prelim)

$1,293.3

$1,346.2

1.04

Source: SEMI, September 2014

The data contained in this release were compiled by David Powell, Inc., an independent financial services firm, without audit, from data submitted directly by the participants. SEMI and David Powell, Inc. assume no responsibility for the accuracy of the underlying data.

The data are contained in a monthly Book-to-Bill Report published by SEMI. The report tracks billings and bookings worldwide of North American-headquartered manufacturers of equipment used to manufacture semiconductor devices, not billings and bookings of the chips themselves. The Book-to-Bill report is one of three reports included with the SEMI Equipment Market Data Subscription (EMDS).

The long wait is finally over. From a sapphire industry standpoint, Apple killed the suspense within the first 10 minutes by announcing that its new 4.7” and 5.5” iPhone 6 and iPhone 6 plus will both feature a traditional ion-exchange strengthened glass display cover. Sapphire remains in the camera lens cover and the touch ID sensor, features that were already present on the iPhone5 S.

Yole Développement believes that technical and manufacturing issues at various levels of the supply chain have prevented Apple from using sapphire as the display cover in this year iteration of its iPhone. An important feature of the design of the new iPhone is that the case doesn’t wrap around the display cover. To protect the edges and, more specifically the corners, the part features a complex shape with smooth curves. This design resulted in fairly low finishing yields with sapphire which drove the cost to above $40 per part.

Apple focused the show on the long anticipated smart watch, announced in three customizable versions named “Apple Watch”, “Apple Watch Sport” and “Apple Watch Edition”. Both the “Watch” and “Edition” versions feature a sapphire display cover on the front. On the back, a zirconia ceramic cover with four sapphire lenses protects a heart rate sensor. The “Sport” model however relies on strengthened ion-exchange glass for the display cover and the lenses. The watches come in 2 case sizes of 38 and 42 mm. The sapphire display cover is a 2.5D design with a surface curved toward the edges that blends in smoothly with the watch case.

“We estimate that for the largest model, those covers are manufactured from long sapphire bars of 40 x 34 mm cross section sliced at a pitch of about 1.8 mm,” said Dr Eric Virey, Senior Analyst at Yole Développement. “The 4 lenses on the back appear to have dimensions fairly similar to the one featured on the iPhone camera lens cover,” he adds. The watches however won’t be available until “early 2015.”

illustrations_sapphiremarket_yolepourusinenouvelle_sept2014_page_1

This raises the question of the status of GTAT mega-sapphire plant in Mesa, Arizona. The company might have to wait another year for an opportunity of a design win in the next iPhone. On a positive side, the company and the downstream supply chain might use this additional time to debottleneck their process and possibly take the “Hyperion” lamination technology from advanced R&D to a manufacturing-ready level that could bring a real disruption in the way to use sapphire in displays.

But for now, smart watches won’t provide enough upside to hit initial revenue targets. Apple currently sources sapphire for this product from multiple suppliers, most located in China. “Even if GTAT was supplying a significant fraction of Apple smart watches, we estimate that total revenue derived from this opportunity wouldn’t exceed $45m in 2014, far from the company’s initial revenue guidance for its sapphire business,” said Eric Virey. “If Apple wants to help GTAT, it could also shift more of its other sapphire needs away from its current suppliers. Under those circumstances, it will also be interesting to see if GTAT revises guidance for 2014 and if it receives the last US$139M installment of the $578M of prepayment promised by Apple before the end of the year. If so, this would show that Apple still supports GTAT and the overall project. GTAT was initially expected to start paying back this sum over quarterly installments starting in 2015 and we’ll also see if Apple tries to enforce this,” he commented.

Apple has nevertheless generated a lot of excitement in the sapphire industry since the announcement of its partnership with GTAT. It remains to be seen if, after adopting sapphire for the camera lens cover in 2012 and in the home button in 2013, the Apple Watch announcement is the last stop or just another step of the journey. Various OEMs have recently introduced smartphones featuring sapphire display covers. Kyocera introduced its sapphire Brigadier to the US Market in August and on September 4th, Huawei announced its “P7 Sapphire edition.”

Xiaomi, a fast growing Apple competitor on the Chinese market also plans a limited edition featuring sapphire.

To the exception of Kyocera’s Brigadier which specifically targets the market of rugged smartphone, most of those announcements were intended at testing the market and showing capabilities ahead of a possible Apple announcement. None will come anywhere close to the iPhone in term of volume potential.

This September 9th event might therefore signals the death of sapphire as a display cover in smartphones. But after the tremendous buzz generated by the Apple and GTAT partnership, and following the positive reviews received by the first sapphire smartphones, the lack of a “sapphire iPhone” announcement might instead create a vacuum that Apple competitors will be eager to fill before Apple potentially enters the market.

illustrations_sapphiremarket_yolepourusinenouvelle_sept2014_page_2

“In any case, 2015 will be a pivotal year: the idea of using the material for smartphones display covers will either materialize in multiple devices and transform the industry, or just crash and burn,” Eric Virey concluded.

By Christian Gregor Dieseldorff, Industry Research & Statistics, SEMI (September 8, 2014)

The general consensus for the semiconductor industry is for this year’s positive trend to continue into 2015 as both revenue growth and unit shipment growth are expected to be in the mid- to high- single digit range. SEMI just published the World Fab Forecast report at the end of August, listing major investments for 216 facilities in 2014 and over 200 projects in 2015.  The report predicts growth of 21% for Front End fab equipment spending in 2014 (including new, used, and in-house), for total spending of US$34.9 billion, with current scenarios ranging from 19% to 24%.

Front end fab equipment spending is projected to grow another 20% in 2015 to $42 billion.  According to the SEMI World Fab Forecast data, this means that 2015 spending could mark a historical record high, surpassing the previous peak years of 2007 ($39 billion) and 2011 ($40 billion).

About 90% of all equipment spending is for 300mm fabs, and, interestingly, the report also shows increased fab equipment spending for 200mm facilities, growing by 10% in 2014.  Equipment spending for wafer sizes less than 200mm is also expected to grow by a healthy 12% in 2015 which includes LEDs and MEMS fabs.

According to the World Fab Forecast, the five regions spending the most in 2014 will be Taiwan ($9.7 billion), Americas ($7.8 billion), Korea ($6.8 billion), China ($4.6 billion), and Japan ($1.9 billion). In 2015, the same regions will lead: Taiwan ($12 billion), Korea ($8 billion), Americas ($7.9 billion), China ($5 billion), and Japan ($4.2 billion). Spending in Europe is expected to nearly double to $3.8 billion.

Seven companies are expected to spend $2 billion or more in 2014, representing almost 80% of all fab equipment spending for Front End facilities. A similar pattern will prevail in 2015.

Worldwide installed capacity falls below 3% mark

World_fab_chart

Figure 1 illustrates fab equipment spending since 2003 and the change of installed capacity (excluding Discretes and LEDs).

As Figure 1 illustrates, before the last economic downturn, most equipment spending was for adding new capacity. The World Fab Forecast report shows that in 2010 and 2011, fab equipment spending growth rates increased dramatically, but installed capacity grew by only 7% in both years. Then in 2012 and 2013, growth for installed capacity sagged even further with only 2% and even less growth. Previously, growth rates less than 2% have been observed only during severe economic downturns (2001 and 2009).

Industry segments, such as foundries, see continuous capacity expansion, though other segments show much lower growth — thus pulling down the total global growth rate for installed capacity to below the 3% mark. Although spending on equipment, some leading-edge product segments experience a loss of fab capacity and, looking closer at this phenomenon, two major trends are observed.

First, coming out of the 2009 downturn, SEMI reports that companies are spending much more on upgrading existing fabs.  From 2005-2008, yearly average spending on upgrading technology was about $6 billion compared to the period of 2011-2015 when the yearly average increased to $14 billion for upgrading existing fabs.  Second, leading-edge fabs experience a loss of capacity when transitioning to leading-edge technology. This is largely observed with nodes below 30/28nm with the increasing complexity and process steps resulting in a -8% to -15% reduction in capacity for fabs.

In addition to foundries, the World Fab Forecast report captures capacities across all industry segments as well as System LSI, Analog, Power, MEMS, LED, Memory and Logic/MPUs. The Logic/MPU sector is also expected to see some positive capacity expansion for 2014 and 2015. Flash capacity is expected to increase by 4% in 2014. Although we see more DRAM capacity coming online, DRAM is now slowly coming out of declining territory with -3% in 2014 and reaching close to zero by end of 2015.

More DRAM capacity?

Over the past three to four years, some major players (such as Samsung, Micron, and SK Hynix) have switched fabs from DRAM to System LSI or Flash.  In addition, other companies stopped DRAM production of some fabs completely, contributing to declining DRAM capacity. Equipment spending levels for DRAM fabs in 2012 and 2013 were near the $4 billion mark annually and are described by some industry observers as being at “maintenance level.”  Increased spending is expected for DRAM in 2014 and 2015, yet although more capacity is being added — the rates are still negative until the end of 2015.  See Figure 2.

Figure 2: Fab equipment spending is compared to the change rate of capacity for DRAM.

Figure 2: Fab equipment spending is compared to the change rate of capacity for DRAM.

As discussed above, SEMI reports that leading-edge DRAM fabs undergo a double-digit capacity loss when upgraded due to an increase in processing steps and complexity. Since the end of last year, Samsung is in the process of adding additional DRAM capacity with two new lines — Line 16 (ramping up this year) and its new Line 17 (the first new DRAM fab ramped since the last economic downturn). In addition SK Hynix is ramping up its M14 DRAM line in 2016. We expect the impact to overall DRAM capacity expansion to occur in 2015 when this fab begins to ramp up. Even if this fab ramps to about half of its potential, the change rate for installed DRAM capacity would still not be positive by end of next year.

Over $6 billion for Fab construction projects

The SEMI World Fab Forecast also provides detailed data about fab construction projects underway. Construction spending is expected to total $6.7 billion in 2014 and over $5 billion in 2015.  Leading regions in spending for 2014 will be Taiwan, Americas, and Korea.  In 2015, the highest spending will be seen in Europe/Mideast, followed by Taiwan and Japan.

Only five companies show strong spending numbers for new fabs or refurbishing existing fabs. Their combined fab construction spending accounts for 88% of all worldwide fab construction spending for Front End facilities.

In 2014, the SEMI report shows 16 new fab construction projects (six alone for 300mm) and 10 fab construction project in 2015 (four for 300mm). Most construction spending in 2014 is for Foundries ($3.1 billion) followed by Memory ($2.5 billion) and Logic. In 2015, Memory will have most spending with ($2.3 billion) closely followed by Foundries ($2.2 billion).

The report lists currently 1150 facilities with 68 future facilities with various probabilities which have started or will start volume production in 2014 or later. See Figure 3.

Figure 3: Count of known facilities (Volume fabs to R&D) in the World Fab Forecast report with various probabilities which are expected to start production in 2014 to 2020.

Figure 3: Count of known facilities (Volume fabs to R&D) in the World Fab Forecast report with various probabilities which are expected to start production in 2014 to 2020.

As it looks right now, SEMI reports that the outlook is positive for 2014 for the chip-making industry compared to the previous few years and the outlook for 2015 also remains healthy.  However, given the current investment trends for spending at the advanced technology nodes and the decline in construction related activity, we continue to expect worldwide capacity expansion to remain in the low-single digits in the next three to five years.

SEMI World Fab Forecast Report

The SEMI World Fab Forecast uses a bottom-up approach methodology, providing high-level summaries and graphs, and in-depth analyses of capital expenditures, capacities, technology and products by fab. Additionally, the database provides forecasts for the next 18 months by quarter. These tools are invaluable for understanding how the semiconductor manufacturing will look in 2014 and 2015, and learning more about capex for construction projects, fab equipping, technology levels, and products.

The SEMI Worldwide Semiconductor Equipment Market Subscription (WWSEMS) data tracks only new equipment for fabs and test and assembly and packaging houses.  The SEMI World Fab Forecast and its related Fab Database reports track any equipment needed to ramp fabs, upgrade technology nodes, and expand or change wafer size, including new equipment, used equipment, or in-house equipment. Also check out the Opto/LED Fab Forecast. Learn more about the SEMI fab databases at: www.semi.org/MarketInfo/FabDatabase and www.youtube.com/user/SEMImktstats

 

Front End fab equipment spending is projected to increase up to another 20 percent in 2015 to US$ 42 billion, according to most recent edition of the SEMI World Fab Forecast.  In 2015, equipment spending could mark a historical record high, surpassing the previous peak years of 2007 ($39 billion) and 2011 ($40 billion). In 2014, the report predicts growth of approximately 21 percent for Front End fab equipment spending, for total spending of $34.9 billion.

Seven companies are expected to spend $2 billion or more in 2014, representing almost 80 percent of all fab equipment spending for Front End facilities; a similar pattern is expected in 2015. About 90 percent of all equipment spending is for 300mm fabs.

According to the World Fab Forecast, in 2014, the five regions with the highest forecast spending on equipment are: Taiwan ($9.7 billion), Americas ($7.8 billion), Korea ($6.8 billion), China ($4.6 billion), and Japan ($1.9 billion). In 2015, the same regions will lead: Taiwan ($12.0 billion), Korea ($8.0 billion), Americas ($7.9 billion), China ($5.0 billion), and Japan ($4.2 billion). Spending in Europe is expected to nearly double (from 2014 to 2015) to $3.8 billion.

As Figure 1 illustrates, before the last economic downturn, most equipment spending was for new additional capacity. SEMI reports that in 2010 and 2011, fab equipment spending growth rates increased dramatically, but installed capacity grew by only 7 percent in both years. In 2012 and 2013, installed capacity grew 2 percent or less. Some industry segments, such as foundries, see continuous capacity expansion, while other segments show much lower growth — pulling down the total global growth rate for installed capacity to below the 3 percent mark.

World_fab_chart

Figure 1 illustrates fab equipment spending since 2003 and the change of installed capacity (excluding Discretes and LEDs).

 

In addition to foundries, the World Fab Forecast report captures capacities across all industry segments as well as System LSI, Analog, Power, MEMS, LED, Memory and Logic/MPUs.

DRAM is now slowly coming out of a declining trend with -3 percent in 2014 and reaching close to zero by end of 2015. Over the past three to four years, some major players have switched fabs from DRAM to System LSI or Flash while others have discontinued DRAM production completely, contributing to declining DRAM capacity.

The SEMI World Fab Forecast also provides detailed data about fab construction projects, with spending expected to total $6.7 billion in 2014 and over $5.0 billion in 2015. In 2014, the leading regions for construction spending are Taiwan, Americas, and Korea.  In 2015, the highest spending is expected in Europe/Mideast, followed by Taiwan and Japan.

Learn more about the SEMI World Fab Forecast which uses a bottom-up approach methodology, providing high-level summaries and graphs, and in-depth analyses of capital expenditures, capacities, technology and products by fab. Additionally, the database provides forecasts for the next 18 months by quarter. These tools are invaluable for understanding how the semiconductor manufacturing will look in 2014 and 2015, and learning more about capex for construction projects, fab equipping, technology levels, and products.

The SEMI Worldwide Semiconductor Equipment Market Subscription (WWSEMS) data tracks only new equipment for fabs and test and assembly and packaging houses.  The SEMI World Fab Forecast and its related Fab Database reports track any equipment needed to ramp fabs, upgrade technology nodes, and expand or change wafer size, including new equipment, used equipment, or in-house equipment. Learn more about the SEMI fab databases at: www.semi.org/MarketInfo/FabDatabase and www.youtube.com/user/SEMImktstats