| Comment Record |
|
Commentor |
Mr. Raj Jaisinghani |
Date/Time |
2003-04-11 15:30:48 |
|
Organization |
Technovation Systems, Inc. |
|
Category |
Company |
| Comments for FDA General |
| Questions |
|
1. General Comments
|
The interpretation of 21 CFR Part 210 usually is interpreted for validation purposes in the following manner with respect to the HEPA filtration and airflow rates or air changes per hour. We believe these guidelines or interpretations should be changed keeping in mind FDAs desire to base the interpretations based on good scientific and engineering principles. This is discussed in detail below.
Part A - Terminal HEPA filters velocity and Air changes per hour.
The interpretation of guidelines is usually done in a manner so as to require terminal HEPA filters to operate or be tested at 90 fpm. The 90 fpm is not based on any valid technical basis and is in fact detrimental due to the following reasons:
(i) Requiring 90 fpm through the filter means that for the same and reasonable air changes per hour in a Class 100 cleanroom there will be less points of discharge of air in the cleanroom. This means less ceiling coverage. In fact having the same air changes per hour with hundred percent ceiling HEPA supply will result in better and more reliable cleanroom performance by elimination of dead zones in the cleanroom. A consequence of this full ceiling coverage approach to eliminate the dead zones will be either a) lower filter velocity (which as we will see below is more desirable) OR b) higher air changes per hour than required from a cleanliness or air conditioning point of view. The consequences of b) are much higher operating costs and initial costs for cleanroom facilities. Up to 40% of the operating and initial costs of cleanrooms are associated with the air changes per hour (cf. Lawrence Berkeley Laboratories “Lawrence Berkeley National Laboratories High Tech Buildings Program – Cleanroom Benchmarking Plan”, Internal Report, LBL, Berkeley, CA, 2000).
(ii) Assuming approximately 90-100 percent ceiling coverage of HEPA filters is used along with 90 fpm filter velocity, then the average room velocity will be between 81 – 90 fpm. This is much higher than the average velocity used in even Class 1 (ISO Class 1 or 2) cleanrooms in the semiconductor industry. The reason why Class 1 semiconductor cleanrooms are designed at 75 fpm is because higher average room velocities result in turbulence and eddy formation, keeping in mind that there are in reality no real laminar cleanrooms – just unidirectional cleanrooms. Operating rooms/filters at 90 fpm results in much higher turbulence. This is borne out by studies conducted at MIT (Vazquez, M. (1999), “The Study of Altering Air Velocities in Operational Cleanrooms”, Proc. Cleanrooms West ’99, Santa Clara, CA.), Assyt (Tannous, G and K. Compton, (1998), “Studies Conclude Low Air Velocity Increases Effectiveness of Minienvironment Design”, Cleanrooms, March issue.) and Sematech (Unpublished work). Jaisinghani (Jaisinghani, R.A., “Air Handling Design for Energy Efficient and Low Bioburden Cleanrooms”, Published in A2C2, January 2002; Jaisinghani, R.A., “New Ways of Thinking About Cleanroom Airflow Design”, A2C2, p. 13, December 2000.) presents additional evidence that based on dilution analysis alone the point of diminishing returns is reached at about 65 fpm. Why should pharmaceutical Class 100 (ISO Class 5) rooms have to operate at a velocity higher than Class 1 rooms? Why should we penalize our industry in terms of operating costs and initial costs of cleanrooms when the evidence shows that cleanrooms operate at an optimum level around 65 –75 fpm?
(iii) Filters actually work better at lower velocity than at higher velocity (cf. Dhaniyala, S. and B.Y.H. Liu, (1999), “Investigations of Particle Penetration in Fibrous Filters”, J. IEST, v 42, #1 ) This is another reason to reduce the filter velocity.
Filter face velocity has no bearing to cleanroom performance. What is really important is that the room particulate concentrations and airborne bio burden levels are maintained during operation. How one gets there should be up to the cleanroom designer. BASICALLY IF SUCH ACCEPTABLE PARTICULATE AND BIO BURDEN LEVELS ARE ACHIEVED ALONG WITH GOOD AIRFLOW PATTERNS (SWEEPING ACTION) FOR CONTAMINATION TRANSPORT AS DEMONSTRATED VIA FLOW VISUALIZATION TECHNIQUES, THEN VALIDATION SHOULD BE BASED ON THE ESTABLISHED DESIGN PARAMETERS FOR A PARTICULAR CLEANROOM VIZ. THE DESIGN FILTER VELOCITY AND AIR CHANGES PER HOUR.
Jaisinghani (Jaisinghani, R.A., “Air Handling Design for Energy Efficient and Low Bioburden Cleanrooms”, Published in A2C2, January 2002; Jaisinghani, R.A. “ENERGY EFFICIENT LOW OPERATING COST CLEANROOM AIRFLOW DESIGN”, paper to be published at the IEST ESTECH Conference, May 19-22, Phoenix, AZ) has shown that the variables affecting cleanroom performance are many and include the process equipment, make up air, the filtration system (not just terminal filters), air flow distribution into the room and airflow patterns. In order to not shackle the designer so that efficient and reliable room performance can be obtained the FDA should ACTIVELY DISCOURAGE filter validation to be interpreted at 90 fpm. The validation process should instead be done at the design conditions which have been demonstrated to be effective under OQ and PQ procedures.
90 fpm originates in an accidental manner from the first cleanroom built at Los Alamos, when HEPA filters were any filter above 99% efficient at 0.5 micrometers. To continue to use this baseless value is laughable in this day and age. A more detailed discussion on the fallacy of 90 fpm and other matters can be obtained from a following power point presentation:
Energy Efficient Air Handling Design for Clearnrooms: A DETAILED Power Point Presentation illustrating new methods for energy efficient air handling for Cleanroom design.
This power point presentation and other published papers referenced above are viewable at: http://www.cleanroomsys.com/downloads.htm
Part B – Validation of In-duct HEPA filters
The FDA should encourage (but not require) usage of in-duct or first stage HEPA filters, especially for makeup airflow. This ensures room compliance even under adverse makeup air conditions. However, the validation requirements of such HEPA filters should be very different from that of terminal HEPA filters. This is necessary for the following reasons:
a) The in duct HEPA filters cannot be scanned in situ because they have ducts attached on the outlet side. Instead of scanning such filters the filter should simply be validated by means of sampling to measure an average value of the penetration of aerosol. Scanning is simply not possible for induct filters.
b) It should be recognized that in duct filters are only to be used as final filters for Class 1K and above cleanrooms and for such rooms scanning is not important.
c) For Class 100 and below rooms since the use of terminal filters is required, scanning of the terminal filters ensures the validity of the filters and the facility. The induct filter in that case is simply a prefilter.
|
|