Search

MBE ready for upcoming mass production opportunities – An interview with Riber

COVID has accelerated technological innovation and the demand for more semiconductor devices. In the race to add more functionalities and take the technological lead, More than Moore (MtM) devices, such as power GaN, Power SiC, and VCSELs, are in demand at unprecedented levels. Epitaxy equipment is the enabler of such devices, and hence the device demand directly translates into equipment demand.

Molecular beam epitaxy (MBE), one of the three types of epitaxy equipment, has historically remained a niche and irreplaceable solution for manufacturing high-performance devices. However, less known is the fact that some leading foundries are MBE-exclusive in the production environment.

With the rise of applications such as microLEDs and high-power VCSELs, where higher device specifications are not only a differentiator but also a key criterion, MBE would be a key enabler for volume manufacturing.

This analysis is well detailed in the new semiconductor manufacturing report released by Yole Développement (Yole) last month: Epitaxy equipment for More than Moore 2021. For this article, Yole’s Analyst Vishnu Kumaresan interviewed, Philippe Ley, Riber’s CEO. Riber is the MBE market leader, to understand the MBE ecosystem and the different macroeconomic factors in play.

Vishnu Kumaresan (VK): First, could you please introduce yourself, Riber, and its activities to our readers?

Philippe Ley (PL): My name is Philippe Ley, and I have been Riber’s CEO since 2018. Riber is the global market leader for Molecular Beam Epitaxy (MBE) equipment, used to manufacture compound semiconductors and complex materials under ultrahigh vacuum. Riber’s equipment enables the production of very high-performance electronic and photonic components. As an MBE pure player, our technological expertise is recognized worldwide, with more than 55 years of experience, around 750 MBE machines in operation, and over 90% of revenues generated abroad. Riber’s workforce comprises 120 employees, 80% of whom are engineers. The company has all the necessary resources in-house to ensure its development over the long term.

Thanks to its know-how, Riber has also diversified into the field of evaporators for the OLED (organic LEDs) and photovoltaic (PV) industries.

VK: Though a range of companies are using MBE in a production environment, there is still a popular misconception that MBE is only suitable for research and pilot production. What is your response to that?

PL: We identified two main reasons for this misconception. First, during interviews with industrials, it emerged that most of them had experimented with MBE in research laboratories as a tool for understanding the fundamental growth mechanisms, which is great! But based on this experience, MBE remains a research tool for them. The second reason is historical: the two growth processes (MBE and MOCVD), which are fundamentally very different, arose and matured at nearly the same time. Both have relative advantages and disadvantages depending on the process.

This is made evident in that the development of either process has always been accelerated by advances in the other.

Common misconceptions about MBE are primarily linked to maintenance, usability, and safety. Riber has addressed these 3 items:

  • The run time exceeds one year; maintenance operation takes from 4-6 weeks compared to MOCVD’s a few days per month plus the necessary recalibration each time.
  • MBE uses pure material, and no safety handling procedures as with MOCVD are required.
  • MBE can be seen as a complicated technique but relies on a simple physical evaporation process. And Riber actively works to increasingly automate growth steps through its monitoring tools and exclusive process control software.

Riber MBE6000: production MBE system – Courtesy of Riber, 2022

VK: What current applications are quasi-exclusive to MBE?

PL: As MBE is often perceived as a niche technology, it is used in many less prominent areas, but there are some key applications where MBE is leading. Mention may be made of IR sensors or Sb-based lasers, since the corresponding MO precursor is not as optimized. We can also note interesting trends in some applications where MBE’s UHV conditions are likely to be game-changing compared to usual practices, for example, in the passivation market.

VK: We see that MBE is being deployed for upcoming disruptive technologies like microLED and quantum computing. In your opinion, what are the key new markets that have big growth potential for MBE?

PL: Any market which is not yet fully mature represents an opportunity. Even though MBE may not appear to be the technique of choice at any given time, addressing Moore’s Law’s limitations or increasing device performance to satisfy the end-user may require shifting to ultra-high quality epitaxy techniques (low defects, low contamination, lower temperature…). This is precisely one of the main characteristics of MBE. While microLEDs and quantum computing are indeed markets of interest to grow our installed base, we also want to be a recognized actor in silicon photonics (for electro-optic modulators, filters, etc.), as we have identified this as a great opportunity for growth.

VK:  If we understand it correctly, MOCVD and MBE are in the same arena when it comes to material capabilities. According to you, how is MBE differentiating itself and keeping its edge over MOCVD?

PL: MBE is a technology requiring UHV conditions, with base vacuum less than 5.10-11Torr, and pure material, Gallium purity 7N5. No hydrogen incorporation comes from the metal-organic source or carrier gas. On this basis, MBE offers a low residual background of less than 7.10-14 cm-3 @ 300K for GaAs.

The process temperature (400-600°C) is also lower than for MOCVD (500-700°C), and so is the growth rate at around 1-2 µm/h. As a result, MBE has the advantage of producing a structure with more abrupt interfaces compared to other growth techniques. This is particularly attractive for high-performance VCSEL structures.

MBE also offers the capability of reaching higher doping concentration levels – i.e., 20cm-3 p-type doping, interesting for devices with tunnel junctions. Uniformity, the key criterion, is also very stable during a full-year campaign.

Finally, with our MBE 8000 system, the world’s largest MBE machine, Riber offers the same reactor capacity as the MOCVD tools, 8×6’’ wafers per run.

On this basis, MBE is a suitable tool for mass production, providing very stable, reproducible, and uniform performance.

VK: GAFAM seems to be increasing their presence in the hardware side of things in their quest for vertical integration. We feel that they are staying in the R&D and IP development blocks and not showing any signs of moving into volume manufacturing. What is your take on this?

PL: We cannot provide information on this topic at our level. But of course, we share the same observation as you regarding GAFAM activity in R&D and IP development.

VK: How is the geopolitical situation shaping equipment sales? Are European players like Riber better positioned than their American counterparts to serve the Chinese market?

PL: As an advanced manufacturing technique for complex components, MBE is a key technology in several fields, some being quite sensitive to the geopolitical situation, indeed. Riber machines are classified as dual-use equipment, thus subject to export licensing rules. American players have the same obligation as Europeans regarding this export license policy. Nevertheless, granting an export license relies on the respective national agency’s appreciation. Depending on the project, this appreciation may differ slightly on a case-by-case basis, but it is not necessarily in favor of European players when it comes to Chinese market opportunities. Actually, Riber recently lost a number of opportunities due to the national export license policy, with a direct impact on our overall revenue.

VK: Equipment suppliers are experiencing their best years in 2020 and 2021. In your opinion, is this an actual demand, or do you expect to see a market correction in the following years?

PL: We are confident in our development strategy focused on consolidating our market share for systems, achieving sustained growth in our service-based activities, and expanding our portfolio of technologies and applications. The demand is durably strong, and the MBE market is expected to continue to grow in 2022.

VK: If you would like to add one final message to our semiconductor readers, what would it be?

PL: MBE is often considered a niche technique compared to more widely used technologies like MOCVD. This situation is partly inherited from the initial applications of these technologies, probably less demanding in terms of performance, such as large-scale LED manufacturing, for example. However, now that the markets are changing, higher component performance is required, the situation is more balanced, and MBE is one of the answers to beat Moore’s law and increase component performance. Over the years, we have built long-term relationships with some of our production customers who early understood that MBE could play a key role in high-end production applications. Overall, this is not a battle between MBE and MOCVD; the future will be to take advantage of both techniques.

Headline image – Title: Riber cracker and effusion cells – Courtesy of Riber, 2022.

Interviewee

Philippe Ley, 51, is an Ecole Nationale Supérieur d’Arts et Métiers (ENSAM) engineer. After starting his career with ASSYSTEM in 1994, he held various executive positions with RENAULT Automation (1997 to 2001) then COMAU France (2001 to 2007). From 2007 to 2015, he was Operations Director and a Management Board member with RIBER. From 2015 to 2018, he was Managing Director and a corporate officer at ERCA, a subsidiary of the IMA industrial group. He has been CEO of Riber since 2018.

Interviewer

Vishnu-KUMARESAN-rounded-ronde-photo

Vishnu Kumaresan, Ph.D., is a Technology & Market analyst in the Semiconductor Manufacturing Team, part of the Semiconductor, Memory & Computing division of Yole Développement, France. Vishnu focuses on the semiconductor manufacturing domain, covering both equipment and material segments. His scope includes mainstream microelectronic applications as well as More-than-Moore applications. Having lived and worked in four countries, he has more than 11 years of international experience in the electronics industry, covering semiconductor, display, and software technologies. Prior to joining Yole, Vishnu worked as an epitaxy engineer at Aledia, an advanced startup in the microLED display industry, and has previously gained corporate experience at IMEC, CNRS, Saint-Gobain, and Infosys.
Vishnu obtained his Ph.D. in Epitaxy, Material Physics & Chemistry from Sorbonne University, France, and his Masters in Microelectronics from the National University of Singapore and the Technical University of Munich, Germany.

Related reports

Epitaxy Equipment for More than Moore 2021 - cover

The market for epitaxy equipment, such as MOCVD, HTCVD, and MBE, crucial in Power and Optoelectronic applications, is expected to reach around $1.1B in 2026.

Related webcast

Epitaxy and ALD equipment are crucial enablers for the surging compound semiconductor markets such as GaN, SiC, MicroLED and VCSEL market.

up