SEMI



Ballot Review Summary

2012 Cycle 4

REGION: North America

COMMITTEE: EHS

EVENT: SEMICON West 2012

DATE OF MEETING: July 12, 2012

PLACE OF MEETING: San Francisco Marriott Marquis Hotel in San Francisco, California

COMMITTEE CO-CHAIRS: Chris Evanston/Salus, Sean Larsen/Lam Research AG, Eric Sklar, Safety Guru, LLC

SEMI STAFF: Paul Trio

I. Document Number & Title

|Document 4683A |Line Item Revisions to SEMI S2-0310e, Environmental, Health, and Safety Guideline for Semiconductor |

| |Manufacturing Equipment. Delayed Revisions Related to Chemical Exposure |

II-1 Line item 3

|Line Item 3 |Change to applied OEL percentage for normal, maintenance and fault conditions |

1. Tally (Staff to fill in)

Voting Tally: As-cast tally after the close of the voting period

A minimum of 60% of the voting interests that have voting members within the technical committee must return votes. (Regulations ¶9.6.1)

2. Rejects

Reject 1 (IBM – Ernest Timlin)

Negative 2 of Reject 1

|Nega|Referenced Section |*TF/Committee to fill in if necessary |

|tive| | |

| | | |

| |Reason |*Original negative comment and as well as justification should be included. |

| | |2.)I have attached a set of three spreadsheets developed by the American Industrial Hygiene Association's |

| | |Exposure Assessment Committee members to help illustrate the statistics of air sampling datasets. |

| | |Environmental data is best described by use of Lognormal parametric statistics, because environmental data |

| | |is most typically lognormally distributed. I have attached three spreadsheets of my own making: a.) |

| | |"....Example of all data 1% or less" (Arithmetic Mean= 0.38 (%of the OEL)); b.) ....Example of all data 10% |

| | |or less." (Arithmetic Mean =2.1(%of the OEL)); and c.)"....Example B of all data 10% or less" Arithmetic |

| | |Mean=6.9(%of the OEL)) . In these examples, one can see that in all cases the data are not well described |

| | |by normal parametric statistics, but are well fitted to Lognormal distributions and therefore best described|

| | |by lognormal parametric statistics. What is also illustrated in the examples (a. and b.)is that a set of |

| | |18 datapoints, the first 14 of which are exactly the same and all under 1% or 10%OEL, can show remarkably |

| | |different results in terms of the percent of actual exposures (0% for the ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download