Modle de prsentation EDF Gris

1
Potential risks of using new electronic component
technologies in IC systems for nuclear power
plants G. Simon, EDF RD, guillaume.simon_at_edf.fr
2
1. Research activities in EDF RD in the field of
electronic

• A Research Group which is in charge of the
  following tasks
• Identify the aging risks of old electronic
  systems still in use in nuclear power plants
• To avoid generic failures
• To get data to decide how to modernize
• Identify the risks linked to the use of new
  electronic technologies
• To check the knowledge of our manufacturers
• To help them to get this knowledge
• With a large expertise in the field of
  Reliability of electronic components and Electro
  Magnetic Compatibility

3
2. Content of the presentation

• Introduction
• Technological risks
• Regulatory risks
• Industrial risks
• Recommendation
• A European initiative

4
3.1 Introduction

• Electronic technology is driven by large
  diffusion products (PC, mobile phone, toys,
  automotive) characterized by
• High performance, miniaturization, short term
  needs (no aging and obsolescence issues),
  continuous innovation
• On the contrary, our industrial needs for new or
  modernized IC systems are
• High reliability on a long term period (more than
  20 years), long term availability, stability of
  the technologies
• The industrial world has to fit its needs with
  electronic technologies which are not designed
  for it !

5
3.2 Introduction

• This contrast between the available technology
  and the industrial needs may lead to potential
  issues that have to be
• Identified and assessed
• Managed by industrial world
• These potential issues are of three types
• Technological risks
• Regulatory risks
• Industrial risks
• The goal of this presentation is to briefly
  describe part of these potential issues and to
  propose anticipative actions to cope with them.

6
4.1 Technological risks

• A technology trend toward miniaturization driven
  by the Moore Law
• In 2018 gate oxide thickness of 0.5 nm, ASICs
  with 6,000 connection pads, a cost of a
  manufacturing plant of about 10 billions dollars
• A technology breakdown is forseen around
  2015-2020
• One can expect a constant reliability level, but
• Risks of new aging phenomena (not interesting for
  component manufacturers which look at life time
  shorter than 10 years)
• A higher sensitivity to environmental conditions
• An increase of quality issues or design issues
• An increase of maintenance cost
• A high difficulty to qualify complex logical
  systems
• In the following we focus on two technological
  risks.

7
4.2 Technological risks Aging of Thin Gate Oxide

• Miniaturization leads to a decrease of the supply
  voltage and therefore to a decrease of the
  thickness of the gate oxide in CMOS technology
  (up to 0.5 nm in 2018).
• In current technology, aging of thicker oxide (gt
  3 nm) is well managed in the design of the
  circuits. The Hard Breakdown failure is
  definitive.
• In thinner oxides, new aging phenomena will
  occur
• Soft breakdown slow increase of the leakage
  current in transistors, without immediate
  failure this aging mechanism may be a concern
  after 10 years of operation, depending on
  temperature and circuit design
• Proton migration in the thinnest oxides
• New oxides are needed with high kin progress but
  very few data on the aging mechanism of these
  oxides.

8
4.3 Technological risks Single Event Upset (SEU)

• Silicon interacts with radiation and charged
  particles (from the process but also from the
  natural radiation).
• With decreasing size, the current needed to
  switch integrated transistors decreases. The
  parasitic current created after an interaction
  between a neutron and silicon may switch
  transistors. This phenomenon may be destructive
  or reversible (SEU).
• Some industries have already report failures due
  to SEU on ground systems (more usual in aeroplane
  and space).
• An experiment on SRAM showed that increased
  miniaturization leads to higher SEU risks.
• This potential issue needs to be specifically
  adressed when designing electronic boards
  (redundancies, qualification of components, use
  of software verification).

9
5.1 Regulatory risks

• A current example the interdiction of lead and
  halogen flame retardants in electronic systems
  (European 2002/95/CE in july 2006).
• Control systems are out of the scope of this
  rule, but
• It will accelerate obsolescence
• Our suppliers need to get knowledge and practice
  with this new technology and electronic board
  manufacture process a risk of low quality during
  a few years
• It will increase the manufacturing costs
• A confusion issue when repairing the boards
  (quality of repair actions)
• A risk of increase of tin whiskers growth
• The challenge for utilities is to help our
  suppliers to get adequate skills to be sure that
  their systems are reliable.

10
5.2 Regulatory risks
Lead of the diode
Tin whisker
A tin whisker on a diode lead
11
6. Industrial risks

• High production cost means concentration of
  manufacturing tools and plants.
• This fact may accelerate obsolesence
• Low performance components (standard ones for
  industrial needs) will probably be fabricated by
  minor manufacturers.
• A risk of quality loss, which has already been
  reported by system manufacturers (ex ceramic
  capacitors).

12
7. To conclude our recommendation

• What would happen if utilities would not involve
  in the management of these risks ?
• Most of the system manufacturers wont have
  ressources to survey and solve these risks
  (especially for aging issues).
• Increased maintenance cost and availability
  losses which are not acceptable for utilities.
• The good approach may be the following, within a
  collaboration between end-user, system
  manufacturer, component manufacturer and
  scientific world
• Identify the risks through scientific studies and
  a more pragmatic collect work of failure datas
  shared between the industrial actors,
• Write and use audit guides to check wether the
  system manufacturer manage or not these risks.

13
8. A European initiative

• EDF RD intends to create a network of interested
  partners to share data about generic failures
  (aging, quality or design issues for old and new
  technologies) of IC systems.
• A Coordination Action is proposed within the
  EURATOM Call of the 11th of october 2005 (called
  MAGIC Management of AGing of IC systems)
• This network should provide a list of generic
  failures and a guide to audit manufacturers
  (system or component)
• 100 funded by European Community, planned for
  two years
• We are still looking for partners !