Thin film PZT has been used for years in the manufacturing of FeRam and also integrated passive devices: according to Yole Développement “Thin Film PZT for Semiconductor Application Trends & Technology Update” published in January 2014, the market for thin film PZT FeRam and IPD was in 2013 $286M, with a CAGR of more than 5% for the next 5 years.
In addition to these applications, MEMS devices are now more and more using thin film PZT for multiple applications: in addition to the ink jet head business (with Epson mainly) and gyro application (manufactured by Panasonic and SSS mainly), new applications are emerging like auto-focus, infra-red detector, hard disc heads, pumps and valves for µ-fluidics that require effective actuators and some micro-mirrors that could take full benefit of the PZT properties. MEMS thin film PZT applications represented in 2013, according to Yole Développement, a combined market of $350M. But all these new applications are waiting for simpler way to process thin film PZT…
And this is what Mitsubishi Materials Corp. (MMC) and Dainippon Screen Mfg. Co., Ltd. are now proposing, followed by Silex point of view, as a MEMS foundry.
We have interviewed at MMC Mr. Nishiyama, Manager for SG business at the Advanced Materials Division, Electronic Materials & Components Company, and at SCREEN, Mr. Nagashima, Manager for the Product Strategy Department, Business Management Division, Semiconductor Equipment Company.
In order to understand the impact of the new materials and equipment on the manufacturing of devices, we have interviewed at Silex Mr. Ebefors, Chief Technologist, co-founder and VP R&D, Mr. Bauer, SVP Sales and Business Development and Mr. Rimskog, Sales Director, North America and Japan. They give their point of view, as a major MEMS foundry, involved in piezo MEMS technology.
Yole Développement: What is the history of your Sol-Gel PZT development?
Nishiyama: In 1989, MMC has started Sol-Gel R&D for ferroelectric film in CRI and started mass production at SANDA plant (http://www.mmc.co.jp/adv/ele/en/) in Japan in 1999. The first application was FRAM. After this first success, 2005 has been the year for the ramp-up for next application which was IPD for smart phone. The new application has really started in R&D in 2012 for MEMS applications. As a feedback on all the R&D and production investment of MMC and its work with its customer, MMC is now the world leader for the production of Sol-Gel PZT materials.
YD: For what applications is it used for?
N: PZT Sol-Gel is mainly used in volume production for FRAM, IPD and MEMS manufacturing. As FRAM and IPD are well established, MEMS is for MMC the key new growth area with multiple new devices that are now using Sol-Gel PZT like Ink Jet Head, gyro sensor, IR sensor and imager, Micro Miller, energy harvester and multiple new MEMS devices under development.
YD: Why have you created this new PZT-N solution?
N: Today, the MEMS customers are eager to reduce the multi-step deposition of Sol-Gel PZT. By doing so, they can certainly increase yield and decrease manufacturing costs. So we have tried to develop a new solution enabling thicker layer per coat and reducing steps, using our long experience of Sol-Gel. The new PZT-N Sol-Gel material is the product we are proposing, enabling easier processing capabilities, increase in yield and production cost reduction.
YD: What is the advantages of this new PZT-N turnkey solution?
N: So far, PZT Thin film deposition is complex to manufacture for PiezoMEMS, and few company have been successful to stabilize their process. The main reason is that PZT Sol-Gel deposition process is very sensitive to the processing conditions such as spin, temperature uniformity, exhaust condition, atmosphere… and also to equipment configurations.
Hence, MMC has decided to offer total solution of chemical and equipment in order to define all conditions of deposition. The equipment has been developed by SCREEN.
The main advantage of PZT-N TKS for customers is that any PiezoMEMS customers who want mass production can do it by using this combined turnkey solution from MMC and SCREEN.
YD: How does this solution compare with respect to competition (Sputtering…)?
N: Today main competitor for Sol-Gel PZT is sputtering. But we think at MMC that PZT-N TKS realize better uniformity, better through put than sputtering. This is a real breakthrough in the use of PZT materials for MEMS manufacturing.
YD: What is your timing?
N: So 2014 is the year of introduction of MMC and DNS solutions, with multiple presentations and active collaborations with customers in development.
YD: What are your perspectives?
N: We think many MEMS companies can start mass production of PiezoMEMS from now to 2020. We wish our innovative PZT-N TKS will be de facto standard process of PZT thin film deposition. And of course we are expecting a strong growth of MMC sales with these PZT materials!
Yole Développement: Can you tell us more about your cooperation with Mitsubishi Materials?
Nagashima: SCREEN has developed and will provide the best suited equipment (Sol-Gel Coater & Infrared lamp Annealer) for mass production of PZT Sol-Gel, with best process conditions for PZT-N solution of MMC.
The TKS process evaluation using each customer device wafers is available in the Process Technology Center (PTC) of SCREEN (Hikone, Japan). So such support helps the user of the Sol-Gel PZT materials to go into volume production much faster.
YD: What were/are the main motivations to partner with Mitsubishi Materials?
N: MMC is the best and the largest chemical manufacturer that has continued to develop Sol-Gel PZT for long time. Over the years, MMC and SCREEN have been able to build the trust relationship through various cooperations for the evaluations of Sol-Gel PZT in real conditions. And we have decided with MMC to create innovation of Sol-Gel PZT deposition together with MMC.
YD: What are your perspectives in PZT equipment?
N: We think the following points are very important for PZT equipment: good uniformity, high productivity and optimization for best process condition. We can provide the total equipment solution (Sol-Gel Coater and Lamp Annealer) for the Sol-Gel PZT deposition, with all related support to help our customers to go faster from development to volume production.
Yole Développement: Can you tell us more about the first results given by this new PZT-N solution with your test products?
Silex: Silex has during late 2013 and 2014 with the support of MMC engineers successfully evaluated the PZT-N solution. With the new N solution we obtained increased through-put using sub-layers in the 180-220 µm range instead of 50-80 nm as with our previously qualified PZT solution (3-4 X increase in deposition through-put). We also obtained improved dielectric properties (epsR), larger piezo effect (e31f) and similar excellent breakdown and defectivity properties as we achieve with our previous standard Sol-Gel PZT material. We can repeatedly obtain breakdown voltages of Silex SG PZT films above 130 V/µm. This is very critical for many actuator applications such as µ-mirrors, IJ heads and other active µ-fluidic devices.
With proper optimization of the seedlayer process we obtain high performance <100> oriented PZT films with a wide process window for the N-type PZT than with previously used materials. Previously there has been issues to obtain repeatable <100> PZT films with good performances (i.e. BD >80 V/µm) and hence <111> PZT films was often preferred. With the optimized bottom electrode and seed layer process, we can now easily tune the PZT orientation to either <100> or <111> depending on our customers need. We can also obtain doped PZT films for further improved piezo properties. Our latest developments have resulted in I-V curves with drastically decreased leakage currents for our new PZT.
YD: What do you see as main advantages of this PZT-N solution?
S: Improved performance while at the same time offering much improved through-put. The higher density allows fewer layers resulting in improved yield, lower material consumption and in the end lower cost.
YD: How original is this PZT-N solution compared to other companies’ offers?
S: We, at Silex, are carefully following the progress of various PZT technologies and new materials. Even though the strategic choice of Silex for the PZT back in 2011 was on the Sol-Gel solution we also carefully follow the development of other technologies like PVD and PLD. We still believe the Sol-Gel technology has superior performance over these alternatives. And with this new PZT-N solution, the old disadvantage of low through-put for CSD technology has been solved.
We have also seen other Asian suppliers offering thick film PZT CSD solution but we haven’t got the same technical support as from the MMC’s team. To bring the new PZT-N solution to the market, a strong ecosystem of partners in various fields of piezo-MEMS has been established at Silex, not only with processing tool vendors and material suppliers but also with metrology experts such as aixACCT from Germany and various research institutes.
YD: Is this new “PZT-N turnkey solution” something you consider implementing in your fab?
S: YES, Silex is currently fine tuning the process parameters for this PZT-N in our 6” line as a complement to the existing Sol-Gel technology in use since 2012. When high volume demands of PZT (>>50 µm/day) is required we will transfer the N-type PZT solution also to our 8” line. We actually recently got granted an EU project with the goal to demonstrate this capability.
YD: For what kind of applications and devices will you use this solution?
S: First application to reach market using thin film PZT is most likely actuators like Hard disc heads, Ink Jet heads and other pumps and valves for µ-fluidics that require effective actuators. For some SLM and micro-mirror application piezo-drive has advantages over traditional electrostatic comb drive actuators and also other optical MEMS devices would gain from effective and fast long term stable piezo actuators. Silex is also a partner in an R&D project developing PZT process technology for energy harvesting. This program may see commercial implementation some 3-4 years from now.
Also for Sensors, where the component designers use a combination of actuation/transmission and sensing/detection with same mechanical element and at the same time require low cross talk we could offer good solutions. PZT has an advantage and Silex has manufactured designs of multi axis motion sensors (gyro / accelerometer) and is currently optimizing process capability for ultrasound transducer elements (p-MUTs) operating at higher frequency than traditional bulk PZT solutions. These innovative MEMS devices are possible to make with the use of Silex proven PZT on Silicon membrane technology.
YD: What is your timing?
S: We are already in the process qualifying new materials at Silex such as the new PZT-N solution that is on-going right now. As a Pure Play foundry we do not have our own roadmap as far as it comes to introduction in products to the end marker but the material is already under evaluation also by our customers.
YD: What will be your business model: Is it a process platform that you are developing?
S: This depends on what you refer to as being a process platform. The technology will be offered to our customers as a building block to be incorporated into their designs. Silex is a true Pure Play foundry, attracting primarily innovation leaders in the MEMS space that are looking for an “IP safe” manufacturing environment. This also means we explicitly refrain from offering application specific process platforms. Our approach at standardization takes place at a more granular level with our library of SmartBlock® process modules that we combine together, facilitating the process of integrating a full manufacturing flow. PZT on C-SOI is already a fundamental SmartBlock® applicable to a wide range of MEMS components. However this is often only part of the total process integration effort, and integration with other MEMS processes from Silex library is in most cases needed to enable the customers’ actual design. Silex also has a proven track record as a leader when it comes to integration of, for example, TSV’s and novel Wafer Level vacuum Packaging technologies. We are already used to integrating the PZT SmartBlock® with such other SmartBlock® technologies as well as with additional process blocks. For us the PZT is a recent addition and goes along well with our other capabilities and our solid experience of fabricating innovative MEMS.
YD: What is your expected business evolution?
S: We see a strong growing interest for thin film PZT in a multitude of existing MEMS applications. We believe the MEMS market in general is also moving into a second wave of adoption and we see a larger crowd of engineers in various industries applying the potential of this rather new mechanical engineering discipline to create innovative solutions in the micro scale technology space that were not possible before.
Silex has invested in a production solution for thin film PZT and is committed to being one of the leading MEMS foundry manufacturing providers for this technology. Thin film PZT clearly opens up a wide range of new opportunities to design novel micro scale structures and Silex true Pure Play business model is perfectly adapted to serve innovation leaders in this space looking to explore the multitude of new solutions made possible.