Sensors and Instrumentation Technical Advisory Meeting (SITAC)

MINUTES OF MEETING

 

Committee Title: ISTAC meeting
Date: May 8-9, 2012
Time: May 8th from 9:10 AM to 3:30 PM 
May 9th from 9:05 AM to 3:20 PM
Location: Herbert C. Hoover Building

Agenda Item Presentations/Discussions:

PUBLIC SESSION (May 8, 2012)

The meeting opened at 9:10 AM. Approximately 32 people were in attendance.

• Opening: Jonathan Wise opened the meeting with introductions and request for comments from the public. There were no public comments.

 

• E-beam Lithography: Martyn Ansell and Rainer Schmid of Vistec Lithography made a presentation on e-beam lithography, addressing these four points: 1) What is e-beam lithography; 2) Who is Vistec; 3) Who are Vistec’s customers and competitors; 4) Export control issues for Vistec’s e-beam lithography.

Overview of e-beam lithography

• The essence of e-beam lithography is to generate and control an electron beam, and use it to write patterns, analogous to writing with a pencil. The physics of e-beams allows them to be focused as small as 3 nm spot size, which provides the opportunity to produce correspondingly small features. However, the beam must be rastered sequentially across the substrate, and thus it is very slow (especially when compared to optical lithography, in which the patterning occurs in parallel).
• David Lindsay asked how the processing speeds of e-beam and optical lithography compare. Rainer responded that optical lithography can process on the order of 100 300-mm wafers per hour, whereas e-beam lithography can process a single 300-mm wafer in one month.
• E-beam lithography equipment is complex, and has rigorous requirements for tolerances of the mechanical stages (otherwise the 3 nm spot smears/blurs).
• David Lindsay asked what types of patterns are typically written by e-beam lithography. Rainer responded that e-beam lithography is most commonly used in R&D applications, where the flexibility of direct-write is an important advantage over the use of masks in optical lithography, because masks are difficult to changes quickly.
• For e-beam lithography, there is a trade-off between spot size and throughput, and the optimum operating point depends on what one is trying to accomplish. Vistec machines operate in the regime of Gaussian spot size, which allows the smallest spot sizes, because this is most useful for the target market, which is R&D.
• E-beam lithography history dates from the 1920s, when Bush proposed the lens properties of coils and verified that the lens formula for glass optical lenses is valid for magnetic lenses. In the 1950s, SEM technology was widely investigated, and contamination layers in SEM are well known. In 1958, Buck proposed to use the contamination effect for 100nm etch masking. In 1961-62, Speidel/Tuebingen did the first e-beam lithography patterning. In the 1960s-70s, there was much work with e-beam lithography worldwide. In the 1970s-80s, e-beam lithography has been used for mask and wafer manufacturing applications.
• David Lindsay asked how deep a hole is created by the e-beam. Rainer responded that it depends on the interaction with the resist, and there are no resists that are specifically designed for e-beam lithography; rather, people use UV resists and do the best that they can.

About Vistec

• Vistec’s history begins with Phillips, Cambridge Instruments and Carl Zeiss Jena. Though a series of mergers, portions of these predecessor companies have become Vistec.
• Vistec’s technology sites are in US, Germany and the Netherland, with sales offices worldwide.
• Vistec has three main products, EBPG5000, EBPG5200 and VB300. The two EBGP5xxx products are the major products, with many installations worldwide, principally in university research settings. The VB300 is a specialized instrument for 300-mm wafers; to date, only four VB300 instruments have been installed.
• Key technical parameters of the EBGP5200 are: Gaussian beam shape, 100 kV acceleration voltage, lambda/1024 laser interferometer, 50 MHz clock rate, up to 200 mm wafer size, step-and-repeat writing, 20-bit main field resolution, 2.2 nm minimum theoretical spot size, <8 nm feature size.
• Customer applications include; nanotechnology, micromechanics, Si direct write, micro optics, telecommunications.
• Vistec’s strategy is to hold high-value IP and core-competencies in-house.

Vistec’s Customer Base and Competitors

• Over 80% of Vistec customers are universities, 60% of which are in US and/or EU. The overall installed based is 42% in EU, 44% in US and 14% in Asia/Pacific.
• Key competitors are Nanobeam (UK); Raith (Germany); and JEOL, Elionix and Crestec (Japan). Vistec commented that although e-beam lithography is a small and specialized market, there are nevertheless many competitors.
• The emerging market is at universities in China.
• Vistec currently ships 6-8 systems per year. System lifetime is typically >10 yrs, and often 15-20 years. The equipment lasts a long time, and the service lifetime depends on the noise level in the electronics and mechanics: Noise adversely impacts the performance, and when performance is degraded to the point that it is no longer useful, the instrument may be retired.

Vistec’s export control issues for e-beam lithography

• e-beam lithography instruments are currently considered by BIS to be 3B001.f.3.
• Vistec has customer interest in China, India and Russia, but is discouraged from pursuing this based on withdrawal of two and denial of one prior export license applications. They observed that EU countries have typically processed licenses for e-beam equipment in 4-6 weeks, while the US licensing process extended for six months or more, and has little or no transparency.
• Data was presented indicating that competitors JEOL, Nanobeam and Raith have all exported systems to China.
• There was some discussion about whether the Vistec e-beam lithography machines are properly classified under 3B001.f, because it is not clear that they are “specially designed for mask making or semiconductor device processing”; they might in fact be general-purpose research instruments.
• Jonathan Wise asked whether linear writer speed might be a relevant control parameter, but this was not discussed or evaluated.
• Action: The matter was referred to the 3B Working Group

• Working Group Reports: Key points from the Working Group reports were:

 

Cat 3A (reported by Jonathan Wise): This group has awaiting feedback from Wassenaar on the proposals currently in progress. Additionally, this group held a telecon to discuss “specially designed” as it is used in 3A; a summary of this will be presented later today in the session on “specially designed”.

Cat 3B (reported by Jeff Rogers): This group is awaiting feedback from Wassenaar on the proposals currently in progress.

Cat 4 (reported by Mark Renfeld): This group is looking ahead and trying to assess technology directions, to try to understand Cat 4 might look like in 10 years.

Cat 5p1 (reported by Frank Quick): This group held a telecon to discuss “specially designed” as it is used in Cat 5p1; not much progress was made, and that will be presented later today in the session on “specially designed”. The issue of the ENC control threshold for satellite modems remains open, and ViaSat will speak on this later today.

Cat 5p2 (reported by Roz Thomsen): This year has seen the largest number of proposals in Cat 5p2 in recent years. In particular, this group is discussing with BIS the proposal for mass-market components, which ran into difficulty last year due to the question of software components.

• ENC Threshold for Satellite Modem: ViaSat, represented by Maria Contrestano (Global Trade), Erwin Hudson (CTO) and Maarten Sengers (Consultant), proposed changes to the ENC threshold for satellite modems (740.17(b)(2)(i)(A)(3)). This was a continuation from the Nov 2011 ISTAC meeting, when ViaSat first raised the issue. At that time, the plan was to address this through both a Wassenaar proposal and through modification to ENC; in the interim, a Wassenaar proposal was not initiated. Accordingly, the current presentation was intended to resume work towards changes to the ENC threshold.
  • Satellite internet has generally been thought of as “internet when nothing else is available”. However, satellite internet is beginning to be competitive with DSL or cable in urban environments (satellite now offers 12+ Mb/s for $49.99/month).
  • David Lindsay asked what equipment is needed for satellite internet connection. The response was a 75-cm dish, the modem and a few cables. Historically, there has been 200-ms latency time for satellite internet, but with acceleration servers most websites load in 5-6 sec, which is generally acceptable.
  • The basic architecture of satellite internet is subscriber terminals, which connect to the satellite (a microwave repeated, with 72 small beams), which connects to earth gateways stations, which connect via fiber to operations/businesses. ViaSat’s Ka band satellite has 3 GHz bandwidth in each of the up- and down-links, and it is re-used 20x, for a total effective bandwidth of 120 GHz.
  • ViaSat’s consumer high-capacity modem, the “SurfBeam 2” has a receive rate up to 40 Mbps and a transmit rate up to 20 Mbps. More than 1M units have shipped. However, as consumer expectations of bandwidth continue to increase, there will be a continuing trend to higher-speed modems.
  • David Lindsay asked how bandwidth requirements are determined, supposing for example that a large subset simultaneously seek to download streaming video. The response was that ViaSat does statistical usage modeling to support the business model, and the satellite is sized accordingly (it is sized to support 1M users, and currently supports 400k-450k users). They are adding 50k subscribers per month. Usage of the satellite will peak in 4-5 years, and then drop, as increasing bandwidth requirements per user overtakes the number of users as the driving factor.
  • The competitive landscape is that there are only a few provided: ViaSat, iDirect and Hughes (the Hughes EchoStar is ViaSat’s primary competitor). Internationally, there is also NewTec (Belgium), but the details of the NewTec product are not well understood by ViaSat.
  • Customers generally request air-interface encryption, usually AES-256 (there is a 250-mile wide beam, which is easy to intercept, thus it is important to encrypt the traffic on the satellite up- and down-links.
  • It is essential to understand that VisSat is only an internet service provider. ViaSat does not make direct connections between customers; to do that requires a separate service (e.g., Skype). Encryption functionality provided by ViaSat is limited to protection on the air-interface.
  • The ViaSat modem is analogous to the Cisco or Motorola cable modems, all of which provide consumer broadband internet. The key point is that ViaSat is an Internet Service Provider, and only incidentally a satellite company.
  • Turning to EAR regulatory matters, ViaSat asked whether 740.17(b)(2)(i)(A)(3), which refers to satellite, has implications for 741.17(b)(2)(i)(A)(4) and (A)(5). That is, it is not clear whether only paragraph (A)(3) covers satellite, or whether (A)(4) and (A)(5) also cover satellite. Aaron Amundson responded that he thinks that the intent was that only (A)(3) refers to satellite, and that (A)(4) and (A)(5) do not.
  • ViaSat continued by noting that the current 10 Mbps threshold in 710.17(b)(2)(i)(A)(3) will not support consumer demands; is difficult to understand in the context of foreign availability (NewTec); and results in similar products (e.g. cable and DSL modems) being treated differently from satellite modems, even though all of these products perform essentially the same function.
  • The foreseeable trend is to 50 Mpbs modems by 2015, and these are treated by the EAR as ENC-Restricted. This becomes a serious challenge for maintaining the supply chain (in contrast to merely being able to provide the modem to the consumer, which is easier to handle).
  • ViaSat offered three proposals: 1) Delete 740.17(b)(2)(i)(A)(3), and treat satellite modems the same as other modems. 2) Increase the control threshold in (A)(3) from 10 Mbps to 60 Mbps; 3) Add an exclusionary Note to the effect of “740.17(b)(2)(i)(A)(3) does not apply to commercial satellite terminals” (or, equivalently, modify (A)(3) to refer to “…transmission over satellite at >10 Mbps, except for provision of broadband access to the public internet via satellite.” Either way, the intent is to differentiate between mesh applications (non-consumer) and broadband consumer internet access).
  • LinNan Lee agreed that the consumer market is only hub-and-spoke geometry, but noted that point-to-point is still used in some enterprise applications. He is not sure of the dividing line between these two cases.
  • Joe Young asked how to differentiate between a consumer modem and a ground station/hub. ViaSat stated that a modem is a small device whereas a ground station/hub is a huge industrial item.
  • Frank Quick added that 740.17(b)(2)(i)(A) is actually a list of exclusions from mass-market treatment, and observed that ViaSat’s option #3 has the effect of adding an exclusionary Note for satellite modem is thus an exclusion to an exclusion. If this approach is pursued, Frank suggests that it is important to be careful with the wording, to avoid confusion.
  • Action: ViaSat will discuss directly with BIS (Joe Young, Aaron Amundson, Mike Pender)

 

• 3B1c Etch Equipment: Joel Cook and Nick Bougopoulos of LAM Research spoke on several issues pertaining to Wassenaar export controls on etch equipment. The three key topics were: 1) Reasons for control; 2) “specially designed” and possibility of a positive list of components; and 3) foreign availability.

The first topic was reasons for the controls on etch equipment. Where do these controls come from? Why do they exist? What do they accomplish?

• The Wassenaar control objectives speak to major or key elements for military capabilities.
• For 3B1c, no clear objective for the control objective has been articulated, and there is no clear understanding of the origins of the control parameters. Industry assumes that the underlying reason for control is to limit access to fab tools for production of leading-edge integrated circuits, but industry feels that pertains to economic security, not directly to military security.
• For 3B1c1, for which the current control threshold is 65 nm, industry wonders what it protects or protects against. That is, why is 65 nm of greater national security concern than some other value? And, what technical factors are of concern and why? Joel pointed out that an older etch tool, one not designed for 65nm process node, can still produce 65nm devices, but not economically.
• For 3B1c2, which pertains to uniformity, this is not an inherent technical capability but rather a matter of productivity. Productivity of the tool pertains to commercial viability; it is not a military issue. It appears that 3B1c2 may be a historic artifact.
• A key consideration is that economic viability is confused/comingled with capability in this control. Etch tools transfer patterns but do not create patterns (in contrast to lithography, which does create patterns).
• Nick added that LAM’s experience is that export license applications for etch tools have never been denied, and that all of their customers are believed to be commercial. That etch tools remain controlled creates significant overhead for compliance. Accordingly, LAM would like etch tools to be decontrolled by Wassenaar and will continue to work on proposals to accomplish that. Meanwhile, LAM proposed that 3B1c etch tools be made eligible for License Exception CIV.

The second topic was “specially designed” and the question of whether the umbrella control in 3B1 of “…and specially designed components and accessories therefor” could be replaced with a positive list of components.

• A plasma etching tool is analogous to (the musical instruments comprising) an orchestra: in the orchestra, many instruments in concert product an output; in the plasma etching tool, many components retain then form and function, regardless of whether the tool is controlled or uncontrolled. Critically, there is no single component that can be said to be solely/peculiarly responsible for achieving the output (wafer/dice). Extension of the analogy is that semiconductor fabrication is hundreds of steps, and a lack of quality control in one step can often be compensated for in other steps
• The semiconductor fab tool industry supports the idea of replacing “and specially designed components therefor” with a positive list of components, and further thinks that any effort towards a one-size-fits-all definition of “specially designed” risks creating ambiguities and less clarity.
• For etch tools, Joel believes that a positive list can be created, and presented a strawman list of those components that have high IP content or are difficult to engineer; other components are trivial and do not warrant control.
• This approach requires a lot of “up-front” analysis, but ultimately is simpler and better.
• The components that are proposed for control are:
  • RF generators, matching networks, applicators or plasma discharge regions providing any of the following:
  • RF application with
  • Rapid waveform modulation (modulate plasma parameters, e.g. Te)
  • Spatially or temporally variable energy deposition
  • Separate production/application of ions and reactive neutrals
• Electrostatic chucks and their ancillary components that enable
• Temporal (<10 s) or spatial temperature uniformity variation during processing
• Temporal and spatial RF amplitude uniformity variation during processing
• Reagent injection systems having temporal (<5 s) or spatial flow variation during substrate processing
• Components for active reactor interior surface temperature control

The third topic was foreign availability of plasma etch tools.

• AMEC, a Chinese company, is developing and selling etch tools.
• The first installation of an AMEC tool occurred in the PRC at SMIC in March 2012. Pr AMEC, it is targeted for processing wafers at 22 nm and below.
• The fact of installation implies “capability” because of the cost and risk associated with installation and use in a production setting: it is very expensive to install, and very expensive to run because it requires much engineering analysis and support.
• Once a new tool is installed, it enters a race to displace current Production Tool Of Record (“PTOR”); if it becomes PTOR, it can compete for installation in other fabs.
• Martyn Ansell asked whether the positive list of components would need to be updated or maintained. Joel responded that updates/maintenance might be required, but not frequently because technology for etch does not change quickly. Additionally, because so much up-front work would be required to draft a positive list, the intent is that this process would be thorough and expansive
• Jim Ramsbotham commented that RF generators and matching networks seem to be general-purpose components and asked whether it is appropriate for them to be included in the positive list. Joel explained that it would be necessary to limit control of these devices with appropriate qualifiers and technical parameters.
• Roz Thomsen asked how to connect the observation that US export licenses typically have more terms and conditions than foreign export licenses with the possibility of providing CIV eligibility for 3B1c etch tools. Joel responded that industry believes that Japanese export licenses typically have only one condition, “no military end use”, and this is essentially the same as CIV. Thus, providing CIV for 3B1c would level the playing field for US exporters relative to foreign competitors.
• Action: Joel and Nick will continue to work on this, through the ISTAC 3B Working Group.
• Use of “Specially Designed” in Categories 3, 4 and 5: This portion of the meeting was reports by the Working Groups on approaches to handling “specially designed” within their categories.
• Cat 3A: Jonathan Wise summarized findings pertaining to Specially Designed in Cat 3A.
• There is infrequent usage of the umbrella control “and specially designed components therefor” within Cat 3A. The umbrella control appears only twice, in 3A1c “acoustic wave devices” and in 3A981 “polygraphs, fingerprint analyzers, etc”. For 3A1c, we lack technical expertise but note that acoustic wave devices are themselves components and wonder whether the control on “Specially designed components of acoustic wave devices” is necessary; perhaps it can simply be deleted with little or no impact to the scope of control. For 3A981, this is a CC1 control; see discussion in the following paragraph.
• The umbrella control appears in two CC1 controls, 3A981 and the related technology entry 3A980. We suspect that there are policy implications associated with these controls and that there might be little support within USG for changes to these controls. This leads us to wonder whether “specially designed” should be interpreted broadly, even as “any modification, no matter how trivial”.
• There is one case, 3A292, where there is a direct opportunity for substitution of specially designed with a positive list. In 3A292, there is already a Note that enumerates the specially designed components.
• There is one case, 3A003, that is exceptionally wordy, and there appears to be the opportunity to replace specially designed through editing and clarification, approximately as follows:

 

Existing 3A003: Spray cooling thermal management systems employing closed loop fluid handling and reconditioning equipment in a sealed enclosure where a dielectric fluid is sprayed onto electronic components using specially designed spray nozzles that are designed to maintain electronic components within their operating temperature range, and specially designed components therefor

Possible rewrite: 3A003 Spray cooling thermal management systems having all of the following characteristics [and specially designed components therefor]

• Closed-loop dielectric fluid handling and reconditioning;
• Sealed enclosure; and
• Incorporating control components or spray nozzles specially [designed] to maintain electronic components within a specified operating temperature range

 

• There is frequent usage of Specially Designed in decontrol notes (e.g., “…does not control MMICs if they are specially designed for other applications); this usage appears to be correct.
• There is some mis-usage of Specially Designed, in the context of other technical parameters (e.g., 3A1d Electronic devices and circuits containing components, manufactured from “superconductive” materials, specially designed for operation at temperatures below the “critical temperature” of at least one of the “superconductive” constituent). In this cases, “specified” can be substituted for “specially designed” with no change to the scope of control and, in fact, with increased accuracy.
• For the software and technology entries 3D002, 3D003, 3D004, 3D101, 3D980, 3D982, 3D991) we are unsure what Specially Designed was intended to cover. Further analysis is required.

Cat 3B: As reported by Joel Cook in the prior presentation, it appears that substitution on “and specially designed components therefor” with a positive list of components is feasible. There was no additional analysis to report.

Cat 4: Mark Renfeld explained that the general approach is to bridge “computers” with “components”. There is a feeling that there are or may be situations when “specially designed” is a helpful term (“Specially Designed can be your friend”). There is also a possibility that Specially Designed could be changed to “specified”.

• 4A3c: This entry (equipment for aggregation) is believed to cover a growing pool of items (a computer is two fundamental parts: the microprocessor, and all of the equipment around the processor). The question of whether a positive list could be created remains vexing. However, the microprocessor itself is clearly a Cat 3 item.
• 4A3e: This equipment appears to belong more properly in Cat 3. Moving it to Cat 3 would obviate the need to handle it in Cat 4.
• 4A3g: This entry (components for aggregation) is very similar to 4A3c, and the issues with Specially Designed are the same as described for 4A3c.
• 4A4: This entry controls certain kinds of computers that do not currently exist, and therefore we do not and cannot know what “components” or “specially designed components” might be. In this situation, arguments can be made both for retaining and for deleting the term “specially designed”.

 

• Cat 5p1: Frank Quick explained that a positive list of components might be possible but that it would probably be very long, and it is not clear what processes should be used to develop such a list. It had been suggested that a cell phone could be used a as a model of 5A1 equipment, and dissected to attempt to identify what components might be important in 5A1 equipment. However, he cautioned that this is tricky because cell phones are uncontrolled. It appears that the catch-all “and specially designed components therefor” is used in Cat 5p1 out of necessity.
  • Bill Root commented that with regard to use of Specially Designed in decontrol notes, any definition of Specially Designed tends to make the situation unworkable. Probably “specially designed” can simply be deleted from the decontrol notes. Bill suggests that the ISTAC should take the lead in handling the issue of Specially Designed Software.
• Cat 5p2: Roz Thomsen had departed for the day, so discussion of Specially Designed in Cat 5p2 was deferred.

 

• Action: No specific actions were decided; the ISTAC will discuss this with BIS.

 

• Ideas for 2013 Wassenaar Proposals: Jonathan Wise presented an overview of ideas for possible instrumentation-related Cat 3/4/5 Wassenaar proposals for 2013.
• 3A1b2: Evaluate whether the combination of frequency, power and fractional bandwidth is appropriate for all frequencies. If appropriate, replace power and fractional bandwidth with other control parameter(s).
• 3A1b4: From the perspective of microwave instrumentation manufacturers, there is a desire to obtain decontrol of instrumentation amplifiers, which are used as microwave test accessories. Because 3A2c/d/e instruments are decontrolled (at least in some form) up to 70 GHz, it would be useful for instrumentation amplifiers to be equally decontrolled. A particularly difficult issue, however, is the definition of “instrumentation amplifier” and determination of how it differs from other amplifiers. Most likely, no action will be taken on 3A1b4 unless action is taken first on 3A1b2; perhaps changes to 3a1b2 would propagate naturally to 3A1b4, and perhaps such changes would result in decontrol of instrumentation amplifiers. If so, then this issue would be moot.
• 3.A.1.b.7: This entry controls converters and harmonic mixers for extending the frequency range of other instruments described in 3A2c/d/3/f “beyond the limits stated therein”. It is ambiguous whether this is intended to refer to frequency only or to “operating range/envelope”, as a combination of frequency and other parameters. Before a proposal can be formulated, more study is needed to understand the underlying technical details.
• 3A1a2-3a1a4: Delete the controls on various types of magnetic tape instrumentation recorders, because these are a sunset technology, being replaced by hard-drive recorders.
• 3A2c1: Change or clarify the control on Resolution Bandwidth (RBW). This parameter was established as the inverse of 100 ns pulse in 3a2d1, but it is not clear that this reasoning is technically sound.
• 3B1: No proposals at this time.
• 3B2: Propose to delete the remaining portions (3B2a and 3B2c) of this entry. There are two key rationales for this: 1) the rationale for deleting 3B2b in 2004 was that testing increased throughput but does not hinder production of the devices; and 2) to the extent that it is necessary to control anything in 3B2, the entry is redundant: the equipment that would be controlled in 3B2a/c is already controlled in 3A2.
• 3D3: Delete this entry, as the subject software is believed not to exist and unlikely to exist in the future.
• 4A3e: Notionally move this control from Cat 4 to Cat 3, because this instrumentation is more commonly thought of as electronic test & measurement equipment than as computer equipment.
• 5E1c4b: Propose to delete this entry (radio technology >31.8 GHz), because 1) it is not clear how this differs from radio technology below 31.8 GHz, and 2) emerging commercial uses of broadband wireless at higher frequencies, principally 37.5-43.5 GHz and ~60 GHz.
• General understanding that all controls refer to specifications (i.e., to specifications in data sheets). In particular, clarify that although some control paragraphs contain the word “specified” and others do not, all entries should be understood to implicitly refer to specifications.

 

• Dave Robertson added that he will consider a proposal for relaxation to the control thresholds for 3A1a5 ADC; these were last changed in 2010.

 

• Mark Renfeld had no specific proposals for Cat 4, and is currently working to define the outlook for the next decade. As work progresses on that, proposal(s) might be developed later this year. The concern is that APP will cease to be meaningful within 10 years, and we need to start planning how to adapt export controls to that. David Lindsay notes that the aggregation formula will be the key problem.

 

• Henry Brandt is working on a proposal to delete 4A3g, principally because it appears redundant with 4A3c. Looking ahead, technology developments may lead to a future proposal to delete 4A3c.

 

• No proposals were suggested for Cat 5p1 or Cat 5p2.

 

• Action: The several Working Groups will continue to review their respective proposals, and the matter will be discussed again at the next ISTAC meeting, in July.

 

• Bill Root offered a comment regarding the upcoming USML-to-CCL effort, cautioning the USG about unilateral implementation of 600-series controls by the US without first consulting WA. In particular, he is concerned that the US may wind up with some more restrictive controls than those in the WA ML, thereby weakening the US position at WA.

 

• The open session was adjourned at 3:30pm.