Category Archives: Process Materials

Applied Materials, Inc. today announced a breakthrough in materials engineering that accelerates chip performance in the big data and AI era.

In the past, classic Moore’s Law scaling of a small number of easy-to-integrate materials simultaneously improved chip performance, power and area/cost (PPAC). Today, materials such as tungsten and copper are no longer scalable beyond the 10nm foundry node because their electrical performance has reached physical limits for transistor contacts and local interconnects. This has created a major bottleneck in achieving the full performance potential of FinFET transistors. Cobalt removes this bottleneck but also requires a change in process system strategy. As the industry scales structures to extreme dimensions, the materials behave differently and must be systematically engineered at the atomic scale, often under vacuum.

To enable the use of cobalt as a new conducting material in the transistor contact and interconnect, Applied has combined several materials engineering steps – pre-clean, PVD, ALD and CVD – on the Endura® platform. Moreover, Applied has defined an integrated cobalt suite that includes anneal on the Producer® platform, planarization on the Reflexion® LK Prime CMP platform and e-beam inspection on the PROVision™ platform. Customers can use this proven, Integrated Materials Solution to speed time-to-market and increase chip performance at the 7nm foundry node and beyond.

“Five years ago, Applied anticipated an inflection in the transistor contact and interconnect, and we began developing an alternative materials solution that could take us beyond the 10nm node,” said Dr. Prabu Raja, senior vice president of Applied’s Semiconductor Products Group. “Applied brought together its experts in chemistry, physics, engineering and data science to explore the broad portfolio of Applied’s technologies and create a breakthrough Integrated Materials Solution for the industry. As we enter the big data and AI era, there will be more of these inflections, and we are excited to be having earlier and deeper collaborations with our customers to accelerate their roadmaps and enable devices we never dreamed possible.”

While challenging to integrate, cobalt brings significant benefits to chips and chip making: lower resistance and variability at small dimensions; improved gapfill at very fine dimensions; and improved reliability. Applied’s integrated cobalt suite is now shipping to foundry/logic customers worldwide.

Applied Materials, Inc. (Nasdaq:AMAT) is a leader in materials engineering solutions used to produce virtually every new chip and advanced display in the world.

Physicists developed a way to determine the electronic properties of thin gold films after they interact with light. Nature Communications published the new method, which adds to the understanding of the fundamental laws that govern the interaction of electrons and light.

“Surprisingly, up to now there have been very limited ways of determining what exactly happens with materials after we shine light on them,” says Hayk Harutyunyan, an assistant professor of physics at Emory University and lead author of the research. “Our finding may pave the way for improvements in devices such as optical sensors and photovoltaic cells.”

From solar panels to cameras and cell phones — to seeing with our eyes — the interaction of photons of light with atoms and electrons is ubiquitous. “Optical phenomenon is such a fundamental process that we take it for granted, and yet it’s not fully understood how light interacts with materials,” Harutyunyan says.

One obstacle to understanding the details of these interactions is their complexity. When the energy of a light photon is transferred to an electron in a light-absorbing material, the photon is destroyed and the electron is excited from one level to another. But so many photons, atoms and electrons are involved — and the process happens so quickly — that laboratory modeling of the process is computationally challenging.

For the Nature Communications paper, the physicists started with a relatively simple material system — ultra-thin gold layers — and conducted experiments on it.

“We did not use brute computational power,” Harutyunyan says. “We started with experimental data and developed an analytical and theoretical model that allowed us to use pen and paper to decode the data.”

Harutyunyan and Manoj Manjare, a post-doctoral fellow in his lab, designed and conducted the experiments. Stephen Gray, Gary Wiederrecht and Tal Heipern — from the Argonne National Laboratory — came up with the mathematical tools needed. The Argonne physicists also worked on the theoretical model, along with Alexander Govorov from Ohio University.

For the experiments, the nanolayers of gold were positioned at particular angles. Light was then shined on the gold in two, sequential pulses. “These laser light pulses were very short in time — thousands of billions of times shorter than a second,” Harutyunyan says. “The first pulse was absorbed by the gold. The second pulse of light measured the results of that absorption, showing how the electrons changed from a ground to excited state.”

Typically, gold absorbs light at green frequencies, reflecting all the other colors of the spectrum, which makes the metal appear yellow. In the form of nanolayers, however, gold can absorb light at longer wave lengths, in the infrared part of the spectrum.

“At a certain excitation angle, we were able to induce electronic transitions that were not just a different frequency but a different physical process,” Harutyunyan says. “We were able to track the evolution of that process over time and demonstrate why and how those transitions happen.”

Using the method to better understand the interactions underlying light absorption by a material may lead to ways to tune and manage these interactions.

Photovoltaic solar energy cells, for instance, are currently only capable of absorbing a small percentage of the light that hits them. Optical sensors used in biomedicine and photo catalysts used in chemistry are other examples of devices that could potentially be improved by the new method.

While the Nature Communications paper offers proof of concept, the researchers plan to continue to refine the method’s use with gold while also experimenting with a range of other materials.

“Ultimately, we want to demonstrate that this is a broad method that could be applied to many useful materials,” Harutyunyan says.

A new way of enhancing the interactions between light and matter, developed by researchers at MIT and Israel’s Technion, could someday lead to more efficient solar cells that collect a wider range of light wavelengths, and new kinds of lasers and light-emitting diodes (LEDs) that could have fully tunable color emissions.

The fundamental principle behind the new approach is a way to get the momentum of light particles, called photons, to more closely match that of electrons, which is normally many orders of magnitude greater. Because of the huge disparity in momentum, these particles usually interact very weakly; bringing their momenta closer together enables much greater control over their interactions, which could enable new kinds of basic research on these processes as well as a host of new applications, the researchers say.

The new findings, based on a theoretical study, are being published today in the journal Nature Photonics in a paper by Yaniv Kurman of Technion (the Israel Institute of Technology, in Haifa); MIT graduate student Nicholas Rivera; MIT postdoc Thomas Christensen; John Joannopoulos, the Francis Wright Davis Professor of Physics at MIT; Marin Soljacic, professor of physics at MIT; Ido Kaminer, a professor of physics at Technion and former MIT postdoc; and Shai Tsesses and Meir Orenstein at Technion.

While silicon is a hugely important substance as the basis for most present-day electronics, it is not well-suited for applications that involve light, such as LEDs and solar cells — even though it is currently the principal material used for solar cells despite its low efficiency, Kaminer says. Improving the interactions of light with an important electronics material such as silicon could be an important milestone toward integrating photonics — devices based on manipulation of light waves — with electronic semiconductor chips.

Most people looking into this problem have focused on the silicon itself, Kaminer says, but “this approach is very different — we’re trying to change the light instead of changing the silicon.” Kurman adds that “people design the matter in light-matter interactions, but they don’t think about designing the light side.”

One way to do that is by slowing down, or shrinking, the light enough to drastically lower the momentum of its individual photons, to get them closer to that of the electrons. In their theoretical study, the researchers showed that light could be slowed by a factor of a thousand by passing it through a kind of multilayered thin-film material overlaid with a layer of graphene. The layered material, made of gallium arsenide and indium gallium arsenide layers, alters the behavior of photons passing through it in a highly controllable way. This enables the researchers to control the frequency of emissions from the material by as much as 20 to 30 percent, says Kurman, who is the paper’s lead author.

The interaction of a photon with a pair of oppositely charged particles — such as an electron and its corresponding “hole” — produces a quasiparticle called a plasmon, or a plasmon-polariton, which is a kind of oscillation that takes place in an exotic material such as the two-dimensional layered devices used in this research. Such materials “support elastic oscillations on its surface, really tightly confined” within the material, Rivera says. This process effectively shrinks the wavelengths of light by orders of magnitude, he says, bringing it down “almost to the atomic scale.”

Because of that shrinkage, the light can then be absorbed by the semiconductor, or emitted by it, he says. In the graphene-based material, these properties can actually be controlled directly by simply varying a voltage applied to the graphene layer. In that way, “we can totally control the properties of the light, not just measure it,” Kurman says.

Although the work is still at an early and theoretical stage, the researchers say that in principle this approach could lead to new kinds of solar cells capable of absorbing a wider range of light wavelengths, which would make the devices more efficient at converting sunlight to electricity. It could also lead to light-producing devices, such as lasers and LEDs, that could be tuned electronically to produce a wide range of colors. “This has a measure of tunability that’s beyond what is currently available,” Kaminer says.

“The work is very general,” Kurman says, so the results should apply to many more cases than the specific ones used in this study. “We could use several other semiconductor materials, and some other light-matter polaritons.” While this work was not done with silicon, it should be possible to apply the same principles to silicon-based devices, the team says. “By closing the momentum gap, we could introduce silicon into this world” of plasmon-based devices, Kurman says.

Because the findings are so new, Rivera says, it “should enable a lot of functionality we don’t even know about yet.”

At this week’s 2018 IEEE International Interconnect Technology Conference (IITC 2018), imec will present 11 papers on advanced interconnects, ranging from extending Cu and Co damascene metallization, all the way to evaluating new alternatives such as Ru and graphene. After careful evaluation of the resistance and reliability behavior, imec takes first steps towards extending conventional metallization into to the 3nm technology node.

For almost two decades, Cu-based dual damascene has been the workhorse industrial process flow for building reliable interconnects. But when downscaling logic device technology towards the 5nm and 3nm technology nodes, meeting resistance and reliability requirements for the tightly pitched Cu lines has become increasingly challenging. The industry is however in favor of extending the current damascene technology as long as possible, and therefore, different solutions have emerged.

To set the limits of scaling, imec has benchmarked the resistance of Cu with respect to Co and Ru in a damascene vehicle with scaled dimensions, demonstrating that Cu still outperforms Co for wire cross sections down to 300nm2 (or linewidths of 12nm), which corresponds to the 3nm technology node. To meet reliability requirements, one option is to use Cu in combination with thin diffusion barriers such as tantalum nitride (TaN)) and liners such as Co or Ru. It was found that the TaN diffusion barrier can be scaled to below 2nm while maintaining excellent Cu diffusion barrier properties.

For Cu linewidths down to 15–12nm, imec also modeled the impact of the interconnect line-edge roughness on the system-level performance. Line-edge roughness is caused by the lithographic and patterning steps of interconnect wires, resulting in small variations in wire width and spacing. At small pitches, these can affect the Cu interconnect resistance and variability. Although there is a significant impact of line-edge roughness on the resistance distribution for short Cu wires, the effect largely averages out at the system level.

An alternative solution to extend the traditional damascene flow is replacing Cu by Co. Today Co requires a diffusion barrier – an option that recently gained industrial acceptance. A next possible step is to enable barrierless Co or at least sub-nm barrier thickness with careful interface engineering. Co has the clear advantage of having a lower resistance for smaller wire cross-secions and smaller vias. Based on electromigration and thermal storage experiments, imec presents a detailed study of the mechanisms that impact Co via reliability, showing the abscence of voids in barrierless Co vias, demonstrating a better scalability of Co towards smaller nodes.

The research is performed in cooperation with imec’s key nano interconnect program partners including GlobalFoundries, Huawei, Intel, Micron, Qualcomm, Samsung, SK Hynix, SanDisk/Western Digital, Sony Semiconductor Solutions, TOSHIBA Memory and TSMC.

Researchers working at the Department of Energy’s Lawrence Berkeley National Laboratory (Berkeley Lab) coupled graphene, a monolayer form of carbon, with thin layers of magnetic materials like cobalt and nickel to produce exotic behavior in electrons that could be useful for next-generation computing applications.

Andreas Schmid, left, and Gong Chen are pictured here with the spin-polarized low-energy electron microscopy (SPLEEM) instrument at Berkeley Lab. The instrument was integral to measurements of ultrathin samples that included graphene and magnetic materials. Credit: Roy Kaltschmidt/Berkeley Lab

The work was performed in collaboration with French scientists including Nobel Laureate Albert Fert, an emeritus professor at Paris-Sud University and scientific director for a research laboratory in France. The team performed key measurements at Berkeley Lab’s Molecular Foundry, a DOE Office of Science User Facility focused on nanoscience research.

Fert shared the Nobel Prize in Physics in 2007 for his work in understanding a magnetic effect in multilayer materials that led to new technology for reading data in hard drives, for example, and gave rise to a new field studying how to exploit and control a fundamental property known as “spin” in electrons to drive a new type of low-energy, high-speed computer memory and logic technology known as spintronics.

In this latest work, published online Monday in the journal Nature Materials, the research team showed how that spin property – analogous to a compass needle that can be tuned to face either north or south – is affected by the interaction of graphene with the magnetic layers.

The researchers found that the material’s electronic and magnetic properties create tiny swirling patterns where the layers meet, and this effect gives scientists hope for controlling the direction of these swirls and tapping this effect for a form of spintronics applications known as “spin-orbitronics” in ultrathin materials. The ultimate goal is to quickly and efficiently store and manipulate data at very small scales, and without the heat buildup that is a common hiccup for miniaturizing computing devices.

Typically, researchers working to produce this behavior for electrons in materials have coupled heavy and expensive metals like platinum and tantalum with magnetic materials to achieve such effects, but graphene offers a potentially revolutionary alternative since it is ultrathin, lightweight, has very high electrical conductivity, and can also serve as a protective layer for corrosion-prone magnetic materials.

“You could think about replacing computer hard disks with all solid state devices – no moving parts – using electrical signals alone,” said Andreas Schmid, a staff scientist at the Molecular Foundry who participated in the research. “Part of the goal is to get lower power-consumption and non-volatile data storage.”

The latest research represents an early step toward this goal, Schmid noted, and a next step is to control nanoscale magnetic features, called skyrmions, which can exhibit a property known as chirality that makes them swirl in either a clockwise or counterclockwise direction.

In more conventional layered materials, electrons traveling through the materials can act like an “electron wind” that changes magnetic structures like a pile of leaves blown by a strong wind, Schmid said.

But with the new graphene-layered material, its strong electron spin effects can drive magnetic textures of opposite chirality in different directions as a result of the “spin Hall effect,” which explains how electrical currents can affect spin and vice versa. If that chirality can be universally aligned across a material and flipped in a controlled way, researchers could use it to process data.

“Calculations by other team members show that if you take different magnetic materials and graphene and build a multilayer stack of many repeating structures, then this phenomenon and effect could possibly be very powerfully amplified,” Schmid said.

To measure the layered material, scientists applied spin-polarized low-energy electron microscopy (SPLEEM) using an instrument at the Molecular Foundry’s National Center for Electron Microscopy. It is one of just a handful of specialized devices around the world that allow scientists to combine different images to essentially map the orientations of a sample’s 3-D magnetization profile (or vector), revealing a its “spin textures.”

The research team also created the samples using the same SPLEEM instrument through a precise process known as molecular beam epitaxy, and separately studied the samples using other forms of electron-beam probing techniques.

Gong Chen, a co-lead author who participated in the study as a postdoctoral researcher at the Molecular Foundry and is now an assistant project scientist in the UC Davis Physics Department, said the collaboration sprang out of a discussion with French scientists at a conference in 2016 – both groups had independently been working on similar research and realized the synergy of working together.

While the effects that researchers have now observed in the latest experiments had been discussed decades ago in previous journal articles, Chen noted that the concept of using an atomically thin material like graphene in place of heavy elements to generate those effects was a new concept.

“It has only recently become a hot topic,” Chen said. “This effect in thin films had been ignored for a long time. This type of multilayer stacking is really stable and robust.”

Using skyrmions could be revolutionary for data processing, he said, because information can potentially be stored at much higher densities than is possible with conventional technologies, and with much lower power usage.

Molecular Foundry researchers are now working to form the graphene-magnetic multilayer material on an insulator or semiconductor to bring it closer to potential applications, Schmid said.

Nanoscientists at Northwestern University have developed a blueprint to fabricate new heterostructures from different types of 2-D materials. 2-D materials are single atom layers that can be stacked together like “nano-interlocking building blocks.” Materials scientists and physicists are excited about the properties of 2-D materials and their potential applications. The researchers describe their blueprint in the Journal of Applied Physics, from AIP Publishing.

Nanoscientists at Northwestern University have developed a blueprint to fabricate new heterostructures from different types of 2-D materials. The researchers describe their blueprint in the Journal of Applied Physics. In this image: Top: Vertical MoSe2-WSe2 heterostructure, radial MoS2-WS2 heterostructure, hybrid MoS2-WS2 heterostructure and Mose2-WSe2 alloy building block representations and crystal structure models Bottom: Vertical MoSe2-WSe2 heterostructure crystal structure model Credit: Cain, Hanson and Dravid

“We’ve outlined an easy, deterministic and readily deployable way to stack and stitch these individual layers into orders not seen in nature,” said Jeffrey Cain, an author on the paper who was formerly at Northwestern University but is now at Lawrence Berkeley National Laboratory and the University of California.

Cain explained that for nanoscientists, “the dream” is to combine 2-D materials in any order and collate a library of these heterostructures with their documented properties. Scientists can then select appropriate heterostructures from the library for their desired applications. For instance, the computer industry is trying to make transistors smaller and faster to increase computing power. A nanoscale semiconductor with favorable electronic properties could be used to make transistors in next-generation computers.

So far, nanoscientists have lacked clear methods for fabricating heterostructures, and have not yet been able to develop this library. In this work, the scientists looked to solve these fabrication issues. After identifying trends in the literature, they tested different conditions to map out the different parameters required to grow specific heterostructures from four types of 2-D materials: molybdenum disulfide and diselenide, and tungsten disulfide and diselenide. To fully characterize the atomically thin final products, the scientists used microscopy and spectrometry techniques.

The group was inspired by the science of time-temperature-transformation diagrams in classical materials, which maps out heating and cooling profiles to generate precise metallic microstructures. Based on this method, the researchers packaged their findings into one diagrammatic technique — the Time-Temperature-Architecture Diagram.

“People had previously written papers for specific morphologies, but we have unified it all and enabled the generation of these morphologies with one technique,” Cain said.

The unified Time-Temperature-Architecture Diagrams provide directions for the exact conditions required to generate numerous heterostructure morphologies and compositions. Using these diagrams, the researchers developed a unique library of nanostructures with physical properties of interest to physicists and materials scientists. The Northwestern University scientists are now examining the behaviors displayed by some materials in their library, like the electron flow across the stitched junctions between materials.

The researchers hope that their blueprint design will be useful for heterostructure fabrication beyond the first four materials. “Our specific diagrams would need revisions in the context of each new material, but we think that this idea is applicable and extendable to other material systems,” Cain said.

Scientists from the universities of Bristol and Cambridge have found a way to create polymeric semiconductor nanostructures that absorb light and transport its energy further than previously observed.

Image showing light emission from the polymeric nanostructures and schematic of a single nanostructure. Credit: University of Bristol

This could pave the way for more flexible and more efficient solar cells and photodetectors.

The researchers, whose work appears in the journal Science, say their findings could be a “game changer” by allowing the energy from sunlight absorbed in these materials to be captured and used more efficiently.

Lightweight semiconducting plastics are now widely used in mass market electronic displays such those found in phones, tablets and flat screen televisions. However, using these materials to convert sunlight into electricity, to make solar cells, is far more complex.

The photo-excited states – which is when photons of light are absorbed by the semiconducting material – need to move so that they can be “harvested” before they lose their energy in less useful ways. These excitations typically only travel ca. 10 nanometres in polymeric semiconductors, thus requiring the construction of structures patterned on this length-scale to maximise the “harvest”.

In the chemistry labs of the University of Bristol, Dr Xu-Hui Jin and colleagues developed a novel way to make highly ordered crystalline semiconducting structures using polymers.

While in the Cavendish Laboratory in Cambridge, Dr Michael Price measured the distance that the photo-exited states can travel, which reached distances of 200 nanometres – 20 times further than was previously possible.

200 nanometres is especially significant because it is greater than the thickness of material needed to completely absorb ambient light thus making these polymers more suitable as “light harvesters” for solar cells and photodetectors.

Dr George Whittell from Bristol’s School of Chemistry, explains: “The gain in efficiency would actually be for two reasons: first, because the energetic particles travel further, they are easier to “harvest”, and second, we could now incorporate layers ca. 100 nanometres thick, which is the minimum thickness needed to absorb all the energy from light – the so-called optical absorption depth. Previously, in layers this thick, the particles were unable to travel far enough to reach the surfaces.”

Co-researcher Professor Richard Friend, from Cambridge, added: “The distance that energy can be moved in these materials comes as a big surprise and points to the role of unexpected quantum coherent transport processes.”

The research team now plans to prepare structures thicker than those in the current study and greater than the optical absorption depth, with a view to building prototype solar cells based on this technology.

They are also preparing other structures capable of using light to perform chemical reactions, such as the splitting of water into hydrogen and oxygen.

Purdue researchers have discovered a new two-dimensional material, derived from the rare element tellurium, to make transistors that carry a current better throughout a computer chip.

Purdue researchers Wenzhuo Wu and Peide Ye recently discovered tellurene, a two-dimensional material they manufactured in a solution, that has what it takes to make high-speed electronics faster. Credit: Purdue University image/Vincent Walter

The discovery adds to a list of extremely thin, two-dimensional materials that engineers have tried to use for improving the operation speed of a chip’s transistors, which then allows information to be processed faster in electronic devices, such as phones and computers, and defense technologies like infrared sensors.

Other two-dimensional materials, such as graphene, black phosphorus and silicene, have lacked either stability at room temperature or the feasible production approaches required to nanomanufacture effective transistors for higher speed devices.

“All transistors need to send a large current, which translates to high-speed electronics,” said Peide Ye, Purdue’s Richard J. and Mary Jo Schwartz Professor of Electrical and Computer Engineering. “One-dimensional wires that currently make up transistors have very small cross sections. But a two-dimensional material, acting like a sheet, can send a current over a wider surface area.”

Tellurene, a two-dimensional film researchers found in the element tellurium, achieves a stable, sheet-like transistor structure with faster-moving “carriers” – meaning electrons and the holes they leave in their place. Despite tellurium’s rarity, the pros of tellurene would make transistors made from two-dimensional materials easier to produce on a larger scale. The researchers detail their findings in Nature Electronics.

“Even though tellurium is not abundant on the Earth’s crust, we only need a little bit to be synthesized through a solution method. And within the same batch, we have a very high production yield of two-dimensional tellurene materials,” said Wenzhuo Wu, assistant professor in Purdue’s School of Industrial Engineering. “You simply scale up the container that holds the solution, so productivity is high.”

Since electronics are typically in use at room temperature, naturally stable tellurene transistors at this temperature are more practical and cost-effective than other two-dimensional materials that have required a vacuum chamber or low operation temperature to achieve similar stability and performance.

The larger crystal flakes of tellurene also mean less barriers between flakes to electron movement – an issue with the more numerous, smaller flakes of other two-dimensional materials.

“High carrier mobility at room temperature means more practical applications,” Ye said. Faster-moving electrons and holes then lead to higher currents across a chip.

The researchers anticipate that because tellurene can grow on its own without the help of any other substance, the material could possibly find use in other applications beyond computer chip transistors, such as flexible printed devices that convert mechanical vibrations or heat to electricity.

“Tellurene is a multifunctional material, and Purdue is the birthplace for this new material,” Wu said. “In our opinion, this is much closer to the scalable production of two-dimensional materials with controlled properties for practical technologies.”

Researchers using powerful supercomputers have found a way to generate microwaves with inexpensive silicon, a breakthrough that could dramatically cut costs and improve devices such as sensors in self-driving vehicles.

“Until now, this was considered impossible,” said C.R. Selvakumar, an engineering professor at the University of Waterloo who proposed the concept several years ago.

High-frequency microwaves carry signals in a wide range of devices, including the radar units police use to catch speeders and collision-avoidance systems in cars.

The microwaves are typically generated by devices called Gunn diodes, which take advantage of the unique properties of expensive and toxic semiconductor materials such as gallium arsenide.

When voltage is applied to gallium arsenide and then increased, the electrical current running through it also increases – but only to a certain point. Beyond that point, the current decreases, an oddity known as the Gunn effect that results in the emission of microwaves.

Lead researcher Daryoush Shiri, a former Waterloo doctoral student who now works at Chalmers University of Technology in Sweden, used computational nanotechnology to show that the same effect could be achieved with silicon.

The second-most abundant substance on earth, silicon would be far easier to work with for manufacturing and costs about one-twentieth as much as gallium arsenide.

The new technology involves silicon nanowires so tiny it would take 100,000 of them bundled together to equal the thickness of a human hair.

Complex computer models showed that if silicon nanowires were stretched as voltage was applied to them, the Gunn effect, and therefore the emission of microwaves, could be induced.

“With the advent of new nano-fabrication methods, it is now easy to shape bulk silicon into nanowire forms and use it for this purpose,” said Shiri.

Selvakumar said the theoretical work is the first step in a development process that could lead to much cheaper, more flexible devices for the generation of microwaves.

The stretching mechanism could also act as a switch to turn the effect on and off, or vary the frequency of microwaves for a host of new applications that haven’t even been imagined yet.

“This is only the beginning,” said Selvakumar, a professor of electrical and computer engineering. “Now we will see where it goes, how it will ramify.”

A new class of adsorbent materials offer high capacity storage and safe delivery of dopant gases

BY J. ARNÓ, O.K. FARHA, W. MORRIS, P. SIU, G.M. TOM, M.H. WESTON, and P.E. FULLER, NuMat Technologies, Skokie, USA J. MCCABE, M. S. AMEEN, Axcelis Technologies, Beverly, MA

Metal-Organic Framework (MOF) materials are a new class of crystalline adsorbents with broad applicability in electronics materials storage, delivery, purification, and abatement. The adsorbents have unprecedented surface areas and uniform pore sizes that can be precisely customized to the specific properties of electronic gases. ION-X® is a sub-atmospheric dopant gas delivery system designed for ion implantation, and the first commercial product that uses MOFs (ION-X® is commercially available through an agreement between NuMat Technologies and Versum Materials). The performance of ION-X deliv- ering arsine (AsH3), phosphine (PH3), and boron triflu- oride (BF3) was evaluated in high current implanters at the Axcelis Advanced Technology Center and compared to the incumbent delivery systems. In-process and on-wafer results of the MOF-based dopant gases compared positively to conventional source gases. Flow, pressure, and beam stability were undistinguishable from conven- tional gas sources throughout the lifetime of the cylinder. Beam and wafer contamination levels (both surface and energetic) were below specification limits, matching the performance of the reference qualified products.

Dopant gas safety challenges

The storage and delivery of hazardous gases creates signif- icant environmental, health, and safety challenges. Their usage requires implementation of stringent safety control systems to minimize the risks of exposure to humans and the environment. The dangers associated with handling toxic gases are the result of both the inherent chemical hazard of the molecule and the kinetic energy stored in the vessel in the form of compression. In essence, the lethality of a toxic release is magnified exponentially by the energetic force of the high-pressure storage. Historically, one way to mitigate these risks was to dilute the hazardous material with inert gases in an effort to attenuate the toxicity effects. Depending on the concentration, this solution provides a safety factor improvement of 10 or 100 by virtue of reducing the molecular density of the hazardous gas to 10% or 1% mixtures, respectively. This approach is commonly used in the electronics manufacturing industry for gases that are known to have extreme toxicity. Hydride gases (i.e. arsine, phosphine, germane, or diborane) are examples of such highly toxic gases used as source materials in a number of electronic manufacturing processes. While this dilution method is effective at reducing the toxicity levels, these mixtures are typically produced at cylinder pressures significantly higher than the pressures of the pure toxic gases. In a release event, this solution reduces the lethality of the dose at the expense of a higher release rate.

In 1993, ATMI (now an Entegris company) introduced a different approach to reduce the toxic gas storage hazards [1]. The technology involves using nano-porous adsor- bents to condense the gas molecules onto their surfaces. This process effectively reduces the kinetic energy of the gas, thus reducing the pressure in the gas cylinder. The large available surface areas within these materials result in gas storage capacities comparable to the high-pressure cylinders. The intrinsic safety advantages of adsorbed gas cylinders are derived from the reduction in pressure within the cylinder. Typically, these vessels are filled to sub-atmospheric pressures (measured at room temperature) in order to inhibit an outward gas release in the event of a leak.

The first sub-atmospheric dopant gas delivery systems used zeolites (SDS® 1) while the second and third genera- tions (SDS® 2 and SDS® 3) evolved to activated carbon adsorbent materials. These gas cylinders store and deliver dopant precursor gases (primarily arsine, phosphine, and boron trifuoride) predominantly for ion implantation processes. In its third generation, and in order to further improve gas storage capacities, SDS 3 evolved by creating a highly dense monolithic adsorbent that nearly eliminated void volumes in the cylinder.

In this paper, we describe a new sub-atmospheric gas delivery system (ION-X ®) that uses a novel ultra-high surface area class of materials called metal-organic frame- works (MOFs). In addition, the implant process perfor- mance using the new product delivering arsine, phosphine, and boron trifluoride was evaluated in a major ion implant OEM facility will be described.

MOF overview: The next generation in nano- porous adsorbents

MOF are three-dimensional crystalline structures assembled with metal-containing nodes connected by organic links (FIGURE 1). The resulting highly organized molecular structures generate nano-pores with record surface areas [2-4]. In addition, the large number of available metal nodes and organic linkers provide unpar- alleled molecular design flexibility to tailor the chemical and physical properties of the adsorbent material to fit the application. Since their discovery in the early 1990’s, MOFs have evolved from an academic curiosity to a widely recognized new class of materials with practical applications in energy, specialty chemicals, military, medical, pharmaceutical, and electronics industries. MOFs are one of the fastest growing classes of materials, with thousands of experimental structures now being reported.

For gas storage and delivery applications, MOFs’ design flexibility provides advantages over traditional adsorbents (FIGURE 2). Pore size, surface area, and chemical stability can be tailored to the specific properties of the adsorbed gases. Compared to zeolites and activated carbon adsorbents, MOFs have significantly larger surface areas (up to 7,000m2/g has been reported[5]. This property, combined with bulk density, is critical in gas storage applications where capacity is measured in terms of vessel volume rather than adsorbent mass. Pore size tunability is also an important parameter in efforts to match the dimensions of the MOF cavities to the molecular sizes of the target adsorbates. This parameter impacts adsorption capacities (how much gas can be loaded) and desorption characteristics (how much can be delivered as a function of pressure). Unlike the broad pore size distributions found in activated carbon adsorbents, MOFs’ crystallinity results in more “usable” pores. This pore size uniformity also results in higher gas quality, as impurities are selectively size excluded.

Preventing reactions between the adsorbent and the target gas is extremely important in electronics applications. Adsorbent/gas interactions will contribute to gas decomposition, leading to impurities and unwanted dopant gas composition changes that could affect the process. The molecular composition of zeolites and carbon adsor- bents are limited to a few elements (typically carbon, aluminum, and silicon) and their oxides. MOFs, on the other hand, can be synthetized from a large range of organic and inorganic constituents, offering more options for creating stable gas/ adsorbent interactions.

MOF-based gas delivery system for ion implant gases ION-X (FIGURE 3) is a sub-atmospheric dopant gas storage and delivery system designed for ion implantation [6]. ION-X uses individual MOF structures with tailored pore sizes to effectively and reversibly adsorb arsine, phosphine, and boron trifluoride gases. The pressure in filled ION-X cylinders is below one atmosphere, significantly reducing the health and environmental impact of an accidental gas release. Furthermore, MOFs’ ultra-high surface areas and uniform structures provide capacity and deliverable advantages compared to existing carbon adsorbent-based products (FIGURE 4). It is important to note that the first-generation ION-X cylinders utilize granulated MOFs with similar adsorbent bulk density to the first-generation carbon product: for the same mass of adsorbent, MOFs provide 40% to 55% higher gas delivery by virtue of their superior surface area and pose size uniformity. Analogous to the evolution of SDS®2, MOF densification inside the cylinder will further increase the gas capacity in next-generation ION-X products.

Implant performance characterization

The performances of ION-X dopant delivery systems were recently evaluated using a PurionH 300 mm high current ion implanter at Axcelis’ Advanced Technology Center (Beverly, MA, USA). The test plan included flow, mass spectral, and metal contamination analyses (both at the surface and at implanted depth). The experiments were repeated using commercially available and well-estab- lished sub-atmospheric dopant gas sources in order to provide a basis for comparison.

Cylinder installation and setup was seamless, requiring no modifications to the existing gas box hardware or software. Flow rate stability for all three gases (AsH3, PH3, and BF3) was demonstrated in the 3.5 to 8 sccm ranges down to cylinder pressures of 20 torr (spec limit). For arsine, the flow experiment continued through a full cylinder depletion, showing a stable flow rate down to cylinder pressure below 3 torr.

The beam energy, purity, and stability were evaluated by analyzing the mass spectra generated during the implantation processes. In all cases, the target dose was 5 x 1015 at/cm2 with beam energies of 40 keV, 20 keV and 15 keV for As+, P+, and BF¬2+ ion implants respectively. The stability and purity of the target doping ion beams were within specifications and very similar to the ones produced by the reference gas sources. Based on the mass spectra, ION-X did not generate any impurities derived from either gas or MOF decomposition.

Neutral and energetic metal contamination levels were thoroughly investigated in this study. All metal analyses were performed by sampling wafers produced using the recipes described in the previous paragraph. Vapor Phase Decomposition-inductively coupled Plasma-Mass Spectrometry (VPD-ICP-MS) was used to monitor the contamination from key trace metals at the wafer surface. Particular attention was placed on monitoring zinc and iron, metals used in the hydride and BF3 ION-X MOF adsorbents respectively. Results show that all metal levels were within specification limits and compared well to the levels detected in control wafers. In all cases, zinc and iron surface contamination levels were below their corresponding detection limits of 0.03 and 0.05 x 1010 atoms/cm2.

Energetic metal contamination is of special interest in ion implantation as even low levels of impurities could affect the performance of the electronic devices. The depth profile of the metals used in ION-X’s MOFs composition were measured using Secondary Ion Mass Spectrometry (SIMS). Wafers used for SIMs analyses were doped using both ION-X and incumbent gas sources using the same ion implant tool and previously stated recipes. The zinc and iron metal concentration profiles for the hydride and boron implants were well within specifications and show no discernable differences between the incumbent and the MOF-based gas sources (FIGURE 5). These results, combined with the previous surface contamination tests, conclusively establish the gas and ion purity of the dopant species extracted from ION-X adsorbents. Moreover, the results are consistent with extensive gas analyses performed at NuMat after subjecting the MOF adsorbent materials to accelerated aging, vibration, and cycle testing.

Summary

This article provides process and on-wafer performance of ION-X, a new MOF-based dopant gas delivery system. The adsorbents used in these cylinders have surface areas, stability, purity, and pore sizes ideal for the storage and delivery of ion implant dopant gases. In-process and on-wafer performance of boron trifluoride, arsine, and phosphine dopant sources compared positively to conven- tional source gas cylinders. The issue of contamination was investigated in detail, demonstrating that the new adsorbents do not contribute to surface or energetic metal impurities. The results published in this article provide independent evaluation of the new product, supporting the safe use of this product in mainstream ion implant applications. To that end, ION-X is already qualified and being used at an electronics manufacturing site with confirmed high stability and purity performance.

References

  1. Olander, K. and Avila, A., “Subatmospheric Has Storage and Delivery: Past, Present, and Future”, Solid State Technology, Volume 57 (2014), pp 27-302.
  2. Y. Cui, B. Li, H. He, W. Zhou, B. Chen, and G. Qian, “Metal–Organic Frameworks as Platforms for Functional Materials,” Accounts of Chemical Research, vol. 49, pp. 483-493, 2016/03/15 2016.
  3. H. Furukawa, K. E. Cordova, M. O’Keeffe, and O. M. Yaghi, “The Chemistry and Applications of Metal-Organic Frameworks,” Science, vol. 341, 2013.
  4. P. Silva, S. M. F. Vilela, J. P. C. Tome, and F. A. Almeida Paz, “Multifunc- tional metal-organic frameworks: from academia to industrial applications,” Chemical Society Reviews, vol. 44, pp. 6774-6803, 2015.
  5. Omar K Farha et al., “Metal-Organic Framework Materials with Ultrahigh Surface Areas: Is the Sky the Limit?” J. Am. Chem. Soc. (2012), Vol. 134, pp 15016−15021
  6. G. M. Tom et al., “Utilization of Metal-Organic Frameworks for the Management of Gases Used in Ion Implantation”, 2016 21st International Conference on Ion ImplantationTechnology (IIT),Tainan, 2016, pp. 1-4.