Category Archives: Metrology

April 30, 2002 – Phoenix, AZ – The SEZ Group is opening its first 300mm applications lab in the US. Located in Phoenix, AZ, the 3,000 square foot lab will be a resource for SEZ and its customers and partners worldwide to test 300mm wafer cleaning processes.

“In this extremely fierce and competitive market, it is critical for our customers to reach high yield and volume manufacturing to meet market windows for the next-generation of chips,” commented Kurt Lackenbucker, chief marketing officer for the SEZ Group. “This facility will enable SEZ to be a value-add partner, working closely together with customers to combat cleaning challenges and find effective solutions for their processes.”

Currently running three to four demos per week, SEZ is primarily testing customer wafers for stress relief, polymer removal, and wafer thinning. Additional capabilities include backside film removal for pre-litho yield enhancement, backside and bevel clean for Cu decontamination and frontside polymer clean for metal line, via holes and Cu dual damascene technologies.

The lab has complete 300mm capability, featuring 1,253 square feet of class 10 cleanroom space and currently houses 20 pieces of process equipment. The facility incorporates SEZ’s Spin-Processors 101, 203, 303, 1050, and the 1300 will be installed in May.

SEZ’s investment in the new facility also includes the Axcelis 200ES Asher tool and metrology equipment from Tencor, FEI, Philips, ADE, Nikon, Luxtron, Rigaku, Veeco, and CDE RESMAP.

Optical collimation improves capabilities for Au/Pd/Ni films

BY FRANCIS C. REILLY

A new metrology approach to metal film thickness measurement can be applied to the plated layers on leadframes. The technique, micro-beam X-ray fluorescence (MXRF), combines the fundamental metrology method of X-ray fluorescence with a new X-ray beam generation solution.

MXRF: What Is It?

MXRF is a new approach to metal film thickness determination. It employs a device called an optical collimator that bends X-rays to form an intense spot. These are directed at a sample to induce fluorescence. X-ray fluorescence yields information about what elements are present and how much of each is there. This information is used to calculate film stack thickness and composition.

The mechanism of X-ray fluorescence is shown in Figure 1. An incoming X-ray photon strikes an electron orbiting the atomic nucleus. That electron is ejected from the atom, and an electron from a higher energy orbital replaces that electron. This is known as electron cascading. As that electron drops into the lower orbital, it releases energy in the form of an X-ray photon. This process is called fluorescence. The energy of that X-ray photon is equal to the potential energy difference between the higher and lower energy orbitals. That value is a characteristic of the specific atom, and it allows the element to be identified. The relative intensity of the X-ray fluorescence at different energies allows a determination of the thickness and composition of the materials present. The thickness and composition are determined by comparing the relative intensities of each element to “known” values (thickness standards) or to mathematical values (fundamental parameters). Using thickness standards is called empirical calibration.

Current Plating Metrology Challenges

The use of a tri-layer metal system, in which a thin layer of gold (Au) deposited on a palladium (Pd) structural layer is placed on a nickel (Ni) barrier layer over a copper (Cu) substrate, is rapidly gaining acceptance for plating in high-density packaging applications (Figure 2). Reduction of gold content reduces per-unit costs. In addition, palladium is harder, and in this application, offers better mechanical reliability.


Figure 1. X-ray fluorescence.
Click here to enlarge image

While manufacturers are adopting this material system, it presents two significant challenges for film metrology measurement. The first is the thickness of the films. In general, the metal system employs a protective gold deposition of 2 to 10 nm on a palladium layer of 10 to 50 nm over a 0.3 to 1.0 µm thick nickel layer.

The second challenge is the measurable area of the sample. Usually, working areas are pads that are 100 µm or smaller. Although XRF metrology is the method of choice for leadframe metal system thickness measurement, conventional XRF tools have not been capable of delivering sufficient X-ray flux in sub-100 µm areas to measure nanometer range films with adequate precision. Thus, the two trends of thinner layers and smaller areas typically prevent conventional metrology tools from meeting the requirements in this application.

New Collimation Approach

One solution to the challenges of thin-film, small-area metrology is an MXRF tool using an optical collimator to significantly increase the resulting X-ray flux. A consistent, controlled, and well-defined X-ray beam geometry is critical to measurement accuracy, and the process of shaping the primary beam geometry is called collimation. There are two principal methods of collimation: mechanical and optical.


Figure 2. Leadframe metal system with three metal layers above a copper substrate.
Click here to enlarge image

Mechanical Collimation: A mechanical collimator is essentially a pin-hole aperture. A typical mechanical collimator assembly consists of a metal block featuring collimators of various dimensions. A single collimator is aligned directly beneath the path of the primary beam X-rays. X-rays pass through the aperture and emerge in a resultant beam with an initial diameter equal to the diameter of the collimator and then the beam fans out to a larger diameter. This resultant beam is targeted at the sample.

Conventional XRF metrology systems employ a mechanical aperture to govern the resultant beam cross-sectional area (Figure 3a). Because primary beam X-ray output (X-rays emanating from the tube) is constant as a function of the cross-sectional area, decreasing the size of the mechanical element decreases the resultant beam flux and, consequently, the tool's precision. There are many applications for which this precision is still sufficient, so the mechanical collimation approach is appealing for these because it is currently less expensive than optical collimation.


Table 1. MXRF tool capability for Au, Pd, and Ni thickness measurements. Samples were measured three times each by three different operators, and the variation was compared to the allowed tolerance to calculate the %R&R (repeatability and reproducibility). A value of 20 percent or less is considered good and 10 percent or less is excellent.
Click here to enlarge image

Optical Collimation: Optically collimated XRF tools employ an optical element to shape and direct primary beam X-rays. Primary beam X-rays incident upon the surface section of the optical element propagate through the system and emerge in a convergent beam of very high intensity. Optically collimated systems capture and transmit a substantially higher quantity of the primary beam X-rays produced (100 to 1,000 times that of mechanically collimated systems). A higher intensity beam produces significantly higher X-ray count rates from the sample.


Figure 3. a) Mechanical collimation of an X-ray beam diminishes the total flux delivered to the sample. b) Optical collimation of an X-ray beam focuses the energy on a small spot, rather than blocking part of the beam to get a small spot size.
Click here to enlarge image

MXRF tools use the optical element to capture primary beam X-ray output, redirecting its path to sub-100 µm areas at the sample interface (Figure 3b). The higher X-ray flux results in measurement precision gains of one to two orders of magnitude in this kind of system compared to conventional mechanically collimated XRF tools.

Tool Performance

The data shown in Table 1 summarizes the capability of the optically collimated MXRF tool for leadframe applications. The suitability of any tool for a particular application is commonly analyzed by the use of a “Gauge R&R” (repeatability and reproducibility) analysis. In the examples, three operators performed three sets of ten measurements each on samples using the same MXRF tool. Overall sample measurement deviation must fall within specified tolerance (%R&R) for tool performance to be considered capable of controlling a process.

Conclusion

The MXRF is a suitable option for measuring the thickness of plated metal layers on leadframes. The new optical collimation scheme allows enough X-ray flux in the small measurement areas to measure the nanometer scale films found in current plating technology. This metrology capability enables continuing advances in advanced leadframe metallization schemes.

AP

Francis C. Reilly, director of sales and marketing, NeXray, 105 Comac Street, Ron kon koma, NY 11779; 631-738-9300; Fax: 631-738-9329; E-mail: [email protected].

null

null

By M. David Levenson
WaferNews Technical Editor

KLA-Tencor has unveiled a new set of automated tools for process window construction and analysis. The Process Window Monitor series automates the construction of process windows based on focus exposure matrices analyzed using KLA-Tencor CD or optical CD metrology tools.

The increased convenience allows production fabs to locate and optimize shifting process windows with minimal delay or disruption. The PWM system facilitates offloading metrology tasks from exposure tools to CD measurement tools, thus increasing production tool availability.

According to Mike Slessor, product marketing manager for lithography module solutions, the PWM system can detect incipient problems like focus tilt or exposure inhomogenities before yields drop noticeably and can suggest corrective action. Litho cells can be matched and the most productive tool for a given task on a given day can be identified. Data can also be downloaded into a Klarity ProDATA file for further analysis. All of that is reputed to result in dramatic savings in manpower and improved production.

KLA has also announced the availability of an etch modeling function for PROLITH lithography simulation tools. This “Etch for Lithographers” simulator captures the influence of the resist dimensions and profile on the pattern transfer process in two and three dimensions, according to Chris Mack, VP of the Finle Division of KLA-Tencor. The after etch process window can be determined for the first time using PROLITH. Since resist can be re-worked, but etched patterns cannot, the improved predictive power should help companies improve wafer disposition decisions and process understanding.

The new system works for single and multiple film stacks, but requires four parameters to be determined for each material and etch stage. It includes real-world effects such as undercutting shadowing and tapering but omits the etch process controls typical of simulators designed for etch engineers. By providing a connection between the litho and etch modules, this new tool may help mediate – or help escalate – disputes in the fab.
WaferNews

By M. David Levenson
WaferNews Technical Editor

Nanometrics, a company with a long history in film metrology, has introduced an optical CD metrology tool for the sub-130nm market that has made libraries of templates superfluous.

The Nanometrics 9300ocd uses normal incidence polarized light reflectometry to extract the spectroscopic signatures of line-space pattern dimensions, pretty much as do spectroscopic and angle-dependent scatterometry tools. However, the use of normal incidence significantly simplifies the calculations, making libraries of templates unnecessary.

Nanometrics also uses an R-theta stage to position the wafer, rather than the more familiar x-y stage, resulting in a dramatic footprint reduction. Because the orientation of the line-space pattern will vary with theta, they also employ patented polarization and image rotating technology, according to Senior Marketing Manager Peter Gise. The CDs measured optically on the Nanometrics system correlate well with the ensemble average of 20 SEM measurements made within the 50-micron diameter target, but less well with individual SEM measurements, which suffer from noise. While the system has sub-nanometer precision, it is insensitive to line edge roughness, a current metrology concern. According to Prasana Chitturi, senior director of marketing, Nanometrics is developing integrated metrology applications of their OCD technology first, but will also offer the same instrument as a stand-alone system.
WaferNews

By Matt Wickenheiser
WaferNews Editor

For the most part, 4Q was a painful quarter to cap off what’s largely been a painful year. Several companies, however, stood out for one simple fact: Their revenue actually increased from 4Q00 to 4Q01.

Publicly held firms FEI Co., PDF Solutions Inc., and Numerical Technologies Inc. all saw increases in revenue (if not in actual income). Privately held Brewer Science Inc. doesn’t release quarterly statements, but the firm told WaferNews that numbers released for the last quarter show product sales increased by 40% in 2001, and overall sales revenue increased by 38% in 2001.

While most of the industry is taking a bath, why do these few succeed?

In a word – technology.

“Clearly these are all pushing the edge of technology,” FEI CEO Vahe Sarkissian noted of the few successful companies.

For example, FEI’s, “SEM, TEM, and dual beams have created the ability for customers to do things they couldn’t do before in terms of 3D metrology, in terms of defect analysis, and structural metrology and analysis,” explained Sarkissian.

FEI’s technology is comparable to surgery for wafers – it allows an engineer to deftly analyze a small section of a wafer anywhere by cross-sectioning. The analysis FEI’s tools allow is a critical part of a company’s success, as device sizes become increasingly small and complex.

“You go to a defect, at the precision level, and cut it,” he explained. “It’s the ability to cut, look, measure, and analyze.”

This becomes particularly important with the emergence of 300mm wafers – which have greater value than smaller wafers. Any way to analyze a wafer’s problems without destroying the wafer is an advantage. In the last year, FEI’s 300mm tools began selling in the third and fourth quarters, adding to the company’s buoyancy.

Another reason for success, said Sarkissian, is that Hillsboro, OR-based FEI deals in a number of technologies, in different areas of semiconductors – and in other industries entirely.

FEI also plays a role in mask repair – something that’s critical as the cost of masks continue to soar. The latest machine can do repairs at the 0.13nm node, Sarkissian said, and FEI’s entered an agreement with International SEMATECH to develop repair tools for the 0.10 and 0.17 nodes, as well.

FEI is also in the circuit edit business – tuning chips in the design cycle – and deals with metrology and measuring thin film heads in the wafer form, before they’re sliced.

“That’s given us some traction,” Sarkissian noted.

The company’s tools also go to the life sciences and materials sciences sectors.

“We get into the details of the structure,” Sarkissian said. “We’re in the heart of nanotechnology – in semiconductors, materials sciences, and life sciences.”

Numerical Technologies, San Jose, CA, provides lithography equipment that enables the continuation of advanced geometry shrinks – and the technology is what’s at the core of the company’s success, according to CEO Buno Pati.

Numerical was founded in 1995, with the bet that a certain technology gap would open up between tools that existed and tools that were needed for the continuation of Moore’s Law. For several years, said Pati, Numerical was “making cures for diseases no one had.” But that gap opened in 1999, said Pati, and the result is Numerical’s current success.

“During any bad environment, the investments in technology have to be there – so companies that are focused on the technology side and have the good fortune of having a majority of their business tied to the technology side, are going to do relatively well,” Pati said.

Officials from Rolla, MO-based Brewer Science, on the other hand, quickly point to customer service as one of the keys to that company’s success, and an outright avoidance of the fire/hire cycles that affect many work forces in the industry.

Brewer’s two segments of chemicals and equipment are both doing well, according to Keith Strassner and John Kellas, two division managers who work with chemicals and equipment marketing, respectively.

“We listen to the customers, to give them something that they really need right now,” said Kellas.

For example, Strassner explained, Brewer’s employees were talking with customers about four years ago about what chemicals were working, and what weren’t, and at the time recognized the emergence of dual damascene. Brewer set up products to capitalize on those needs for the photolithography area, and today holds a sizable share of the dual damascene market.

On the equipment side, said Kellas, customers have voiced the need for bridge tools – not just 300mm tools – and Brewer’s responded with its line of spin coaters.

“It’s a less risky investment,” explained Kellas.

Looking ahead, Brewer is hoping to meet a future need by working on a contact planarization program – an improved version of CMP.

Brewer maintains customer visits throughout downturns, said Loretta Wallis, corporate marketing coordinator, to maintain those relationships.

Strassner noted that there are 235 people on Brewer’s payroll, and that the work force has increased during this downturn – and there are job openings now. Keeping a steady, trained work force provides momentum through downturns into upturns, Strassner said. It also avoids the necessary retraining that must follow a massive re-hiring to build personnel levels back up following layoffs.

San Jose-based PDF Solution’s success comes from its ability to help companies with process and design integration to, “build designs that are inherently more manufacturable,” as David Joseph, exec. VP for sales, marketing, and business development puts it.

“A big part of the power of our solution is helping customers understand, during integration, where the product yield loss will come from,” Joseph told WaferNews. “We help them focus and tell them where to change.”

Around the 0.18-micron node, the product complexity has become significantly great, and the fundamental material changes have also become fundamentally great, said Joseph. Random and defect yield losses are joined by feature-driven yield loss, which is now the dominant mechanism, he explained.

“Because the number of contacts and vias is going up, it’s no longer sufficient just to keep up,” he noted. “The problem wasn’t this big, traditionally – the fab ramp and product ramp was much slower than it is today.”

The time to bring a 0.13-micron fab online today is half the time that it was for a half-micron fab, said Joseph – because you can’t leave that capital idle. Additionally, moving from the commercial-driven market to a consumer market means that new, hot products have to be built fast – to take advantage of the market.

It took seven years for the millionth PC to be sold, said Joseph. It took 10 months for the millionth Playstation 1. Playstation 2’s millionth model sold in two days, he said.

“If you’ve got a winning product, you have to build it quick,” he said.

WaferNews

March 7, 2002 – Woodbury, NY – Veeco Instruments Inc. has entered a distribution agreement with Epion Corp., a JDS Uniphase company, to distribute Epion’s gas cluster ion beam (GCIB) product to the data storage, MRAM, MEMs, photomask, and wireless telecommunications markets.

GCIB is an emerging technology for atomic scale surface smoothing and etching, and is important for improved surface characteristics, increased uniformity, defect reduction, and enhanced performance, the company said. GCIB is available as an integrated module on a cluster tool or as a stand-alone system.

Edward Braun, Veeco’s chairman, CEO, and president, commented, “GCIB is an important new technology vital for the development of next generation high areal density 80 GB/in(2) thin film magnetic heads. It is complementary to Veeco’s broad line of etch, deposition, and metrology systems and may be integrated into existing Veeco cluster tools.”

Epion has established expertise in GCIB technology based on work performed at the Ion Engineering Experimental Laboratory at Kyoto U. in Japan, the company said.

Jan. 17, 2002 – Fremont, CA – Therma-Wave Inc., a worldwide developer and manufacturer of process control metrology systems, has completed its acquisition of Sensys Instruments Corp.

Under the terms of the agreement, Therma-Wave. will exchange 4,470,514 shares of common stock for shares of Sensys Instruments, plus assume Sensys warrants for an additional 329,486 Therma-Wave shares.

There was no cash paid as consideration in the acquisition. As a result of the acquisition Sensys has become a wholly owned subsidiary of Therma-Wave, focusing primarily on the emerging market for integrated metrology.

Dec. 19, 2001 – Fremont, CA – Therma-Wave Inc. has signed a definitive agreement to acquire Sensys Instruments Corp.

Upon consummation of the transaction, which is subject to various conditions including approval by Sensys shareholders, Sensys will become a wholly owned subsidiary of Therma-Wave, focusing primarily on the emerging market for integrated metrology.

The transaction will involve the exchange of shares of common stock of Therma-Wave and, under certain circumstances, cash and promissory notes of Therma-Wave. The transaction is expected to close early in 1Q02.

Sensys Instruments Corp., a privately held company based in Santa Clara, CA, was started in 1996 to develop and commercialize integrated metrology systems for the semiconductor industry. The company has introduced integrated metrology systems for CMP and photo resist coater/developer applications.

Dec. 14, 2001 – Chelmsford, MA – Brooks Automation Inc., a supplier of tool and factory hardware and software automation solutions, has completed the purchase of the automation systems group of Zygo Corp.

Zygo’s automation systems group is a manufacturer of reticle automation systems including reticle sorters, reticle macro inspection systems, and reticle handling solutions.

The purchase method of accounting will be used to account for the cash transaction of approximately $11 million.

“The Zygo automation systems acquisition is an important part of Brooks Automation’s strategy of supplying an integrated suite of lithography automation systems,” said Robert Therrien, president and CEO of Brooks. “The cost of advanced photomask sets can now exceed $500,000, causing semiconductor manufacturers to turn increasingly toward automated reticle handling systems to eliminate problems that may be caused by manual operations. This acquisition will enable Brooks to provide a more complete solution for lithography automation.”

Zygo will retain its core automation business in Delray Beach, FL, for metrology integrated solutions and photonic and electronic module assembly.

In other Brooks Automation news, Brooks and PRI Automation Inc.

have received a request for additional information pertaining to Brooks’ pending acquisition of PRI — a $380 million deal that, if completed, will create the world’s largest automation company, Brooks-PRI Automation.

The request was made by the antitrust division of the Department of Justice (DOJ).

The waiting period applicable to the pending acquisition under the US Hart-Scott-Rodino Antitrust Improvements Act will expire 30 days after substantial compliance with the request for additional information, unless terminated earlier by the DOJ.

The companies said they intend to respond as quickly as possible to the second request. However, the companies cannot estimate how long it will take to achieve substantial compliance with the second request. Subject to completion of the Hart-Scott-Rodino process, and satisfaction of other customary closing conditions contained in the definitive merger agreement, including Brooks and PRI stockholder approval, the companies anticipate the acquisition will close in 1Q02, as announced previously.

New PI Facility Complete


December 10, 2001

December 10, 2001 — PALM-BACH, Germany — NanoAutomation/MicroPositioning company PI (Physik Instrumente),

Germany has built a new 140,000 square foot factory to meet an increased demand for its ultra-precision products in industries such as bio-technology, fiber optics, semiconductors, data storage and aerospace engineering.

The new factory in Palm-Bach provides R&D and

manufacturing with increased cleanroom capacity and a

metrology lab, for testing and calibrating the company’s

precision-positioning products to sub-nanometer accuracy.

PI pioneered piezoelectric (PZT) Nanopositioning Technology in the

early 1970’s, when applications could only be found in university- and

government funded research labs.

Today, many high-tech industries require controlled motion on a

nano-meter scale, to produce, test or inspect their products.