Cleanroom Energy Optimization Methods

Philip J. Naughton, P.E.

Co-author and presenter
Senior Staff Engineer
Motorola Semiconductor Products
7700 West Parmer Lane
Mail Drop: PL06
Austin, Texas 78729
(512) 996-6612

Martha Montemayor-Rapier, P.E.

Co-author
Austin ESIH Strategic Programs Manager
Motorola Semiconductor Products
3501 Ed Bluestein Boulevard
Mail Drop: F13
Austin, Texas78721
(512) 933-3430

ABSTRACT

Energy efficiency has not been a high priority for the semiconductor industry in the past,
since costs related to this issue have historically represented a relatively small
percentage of overall operating costs. As the semiconductor industry enters the next
millennium the need to review energy consumption and its absolute costs has been
renewed due to global resource conservation goals and the fact that conservation in
today’s megafabs can result in a significant positive impact to a company’s bottom line.
This paper will present some of the methods being used by a multi-national
semiconductor company to change the way they design and build cleanrooms with a
focus on resource conservation, energy conservation methods, and cost of ownership.
Various clean air and energy management scenarios will be compared with their
potential for energy savings.

INTRODUCTION

It was estimated that the US semiconductor industry would consume in excess of 8400
million kWh in energy in 1995 (Energy). The environmental impact of this energy is
874,000 tons of CO2, 17,000 tons of NOx, and 42,000 tons of SO2.emmissions
(Energy). In addition to the environmental impact the financial cost is over $1 billion.
Operating costs for wafer fabs have historically been considered inconsequential to the
overall cost of ownership. More recently, the environmental impact and the scrutiny of
all manufacturing costs have focused much attention on fab energy efficiency. There
are numerous studies and benchmark programs being conducted to measure energy
consumption in the semiconductor industry. Studies are also beginning to concentrate
more on where the energy is being used within a typical fab. This paper will focus on
the cleanroom air management systems and the HVAC support to the cleanroom. The
scope of this paper limits any analyses of the manufacturing tools since there is little
research of methods that reduce individual tool energy at this time.

ENERGY USE WITHIN FABS

Within the typical semiconductor fab the energy consumption can be divided into
several functional blocks: manufacturing equipment, cleanroom air management
equipment, chillers, makeup air systems, process exhaust, central plant utilities
(compressed air, process cooling water, nitrogen generator, etc.), Figure 1 (Sematech).
Also, in today's cleanrooms HVAC systems encompass much more than traditional
factors of heat load calculation, cooling and/or heating methodology, and environmental
control of temperature and humidity. HVAC systems are also used to control
particulate contamination, airflow pattern, sound, vibration, industrial engineering
aspects and manufacturing equipment layouts.

Figure 1: Fab Energy Divisions

Process Equipment
Cleanroom Recirculation Fans
Makeup Air Handling
Chillers
Central Plant Utilities
Process Exhaust

Cleanroom Characteristics

According to international consensus a Cleanroom is a specially constructed enclosed

area environmentally controlled with respect to airborne particulate, temperature,
humidity, air pressure, air pressure flow patterns, air motion, vibration, noise, and
lighting (ASHRAE). Control of these environmental parameters is critical and therefore
the primary mission of the cleanroom designer. In order for the designer to meet his
goal, it is imperative that he completely understand all of the factors effecting the
control of these critical parameters. Figure 2 is a schematic view of the energy
transferring components of a fab and their interconnectedness with each other and
HVAC systems.

Figure 2. Energy Flow Within Cleanrooms

The cleanroom designer must understand energy use, energy flow and appropriate
design of HVAC systems for the industry. The characteristics of semiconductor
cleanrooms are different from cleanrooms used in other industries such as
pharmaceutical, aerospace, and biotechnology. These differences lead to having
unique considerations when designing the cleanroom layout and the HVAC systems.
The following list include some of the characteristics of semiconductor manufacturing:

1. Within the semiconductor industry, cleanroom air cleanliness requirements, as

defined by Federal standard 209E, "Federal Standard Clean Room and Work
Station Requirements, Controlled Environment," including cleanliness classes M1 to
M6. This document is to be replaced by the new ISO 14644 documents. To
achieve the desired cleanliness level, unidirectional airflow cleanrooms with average
 air velocity ranges of 40 fpm (0.20 m/s) to 100 fpm (0.51 m/s) are typically used,
with 70-80 fpm (0.35-0.41 m/s) as the most common design velocity. Vertical
unidirectional airflow air management concepts are the design of choice for the
majority of semiconductor manufacturing facilities. The ramifications for choosing an
air management concept will be discussed later.
2. Within the cleanroom component load analysis, room sensible heat ratios of 0.99 or
higher are normal due to large concentrations of manufacturing equipment. High
2 2
sensible internal heat loads approaching 200 W/ft (2152 W/m ) are common, while
2 2
75 W/ft (807 W/m ) is a typical design criteria for many merchant, high volume,
semiconductor factories. The HVAC parameters restrict the cleanroom design’s
flexibility of heat transfer options. Representative environmental control tolerances
in many facilities today are ±0.2°F (±0.11°C) to ±0.5°F (±0.28°C) and ±1% RH to
±2.5% RH for many process areas. These arduous tolerances are required
throughout a specific process area and present difficult control methodology
requirements as the size of the process area increases.
3. The typical metal oxide semiconductor (MOS), bi-polar, or gallium arsenide facility
requires large quantities of conditioned fresh air to replace process exhaust and to
provide for cleanroom pressurization. Process exhaust requirements of up to 10
cfm/ft2 (51 L/s•m2) may be required for some processes areas, while typical
2
industry averages are 2 to 3 CFM per ft (10.2-15.3 L/s•m2). This fact also limits the
degree of energy reduction associated with cleanroom make-up and its interaction
with the cleanroom air management concept.
4. In addition to controlling the cleanroom air cleanliness level, control of cleanroom
temperature and relative humidity is crucial for a semiconductor factory to operate
successfully. Temperature and humidity control within the cleanroom will have a
direct influence on the quality of the products manufactured to the point that without
proper environmental control, many products will fail minimum functional tests and
thus cannot be sold or used. Providing the correct temperature and humidity control
as part of the air management concept must be fully understood when optimizing
energy consumption.

Cleanroom Performance

The cleanroom designer should balance the need to meet critical environmental factors,
cleanroom specs set by operator/owner, and maintain reasonable installation and
operating costs; while at the same time providing an efficient, functional, and flexible
cleanroom. The cleanroom designer must understand the aforementioned
characteristics and their interactive nature to begin an analysis of the energy
components that will achieve the given scope. With the proper analysis, the cleanroom
designer can optimize the performance of individual components and systems to
produce an effective and energy efficient total design. The final performance of the
cleanroom is judged by the quality of control of the critical environmental factors.

AIR MANAGEMENT CONCEPTS

During the evolution of cleanroom designs several air management concepts have
been developed as the layout for the semiconductor fabs. Whereas, in other industries
such as pharmaceutical, whose cleanroom design has changed little in the past 30
years, semiconductor cleanrooms have changed. From early semiconductor factories
with single level bay and chase designs to today’s multi-level megafabs, and tomorrows
300 mm fabs with fully automated wafer transport vehicles, the mission of the
cleanroom has remained relatively the same. The designs have matured with the
industry into a handful of common designs. Figures 3, 4 and 5 present schematic
representations of the three most common air management concepts in use today.

Figure 3. Fan Tower Air Management Concept 1

Figure 4. Recirculation Air Management Concept 2

Figure 5. Fan Filter Air Management Concept 3

In Concept 1, there are very large size recirculation zones, and the Fab is divided into
recirculation zones / rooms typically along process functional requirements. Each room
has its own temperature and humidity control system. In Concept 2, the large size
recirculation zone concept, a Fab is divided into recirculation zones. The zones contain
many rooms. Each room has its own temperature and humidity control system. In
Concept 3, the medium size recirculation zone concept, a Fab is divided into
recirculation zones. The zones contain many rooms. A zone has one temperature and
humidity control system.

Each air management concept is capable of meeting the facility design criteria
established by the process requirements with limitations, advantages and
disadvantages, based on cleanroom configuration. The concepts have varying degrees
of flexibility to process changes, and their cost of construction can also be significantly
different. Energy efficiency is also quite different. Table 2. shows advantages and
disadvantages of the air management concepts. To choose the appropriate concept for
final design, the cleanroom designer should take into consideration different
applications and factors effecting design.

Table 2. Summary of Advantages and Disadvantages of Air Management Concepts

Attribute Concept 1 Concept 2 Concept 3

Flexibility to process Fair Good Very
changes Good
Flexibility to facility changes Good Good Good
Expandability Fair Good Good
Operating Cost Very Good Good Fair
First Cost Very Good Good Fair
Reliability features Very Good Very Good Good
Maintainability Very Good Very Good Fair
Ease and cost of Very Good Good Poor
Automation
FACTORS EFFECTING FAB CLEANROOM ENERGY USE

The cleanroom HVAC system designer provides an efficient design for achieving the
owner's critical environmental parameters, thus providing a suitable manufacturing
environment with minimized levels of contamination. When deviations from the critical
environmental parameters are considered potential contamination may occur. The
cleanroom designer must minimize and if possible eliminate the sources of
contamination under his control.

Airborne Cleanliness

Controlling airborne cleanliness is a fundamental requirement of a cleanroom.

Maintaining cleanliness requires the removal of internally generated contaminants, the
filtration of external contaminants from the fresh air introduced to the cleanroom, and
the prevention of external contaminants by means of positive pressurization. To
accomplish these goals, large air change rates of 300 to 600 air changes per hour for
air cleanliness classes of class M1 (ISO Class 2) or better will be necessary, see Table
1. The control of internally generated particles is a function of the cleanroom protocol
established by the owner's contamination control specialist, while the need for sufficient
airflow to quickly remove the contaminants from the workspace is a function of the
cleanroom design. The ability to maintain room cleanliness is proportional to the
volume of recirculation air. The challenge facing cleanroom designers is to provide the
minimum amount of recirculation air needed to maintain room cleanliness.

Table 1. Cleanroom Classifications

Temperature and Humidity Control Zones

The method of temperature and humidity control can produce one of the largest
variations in energy consumption. During the past years a thorough understanding the
heat transfer dynamics in semiconductor cleanrooms has resulted in several methods
for temperature and humidity control. The air management concept will also impact the
available choices for control, especially humidity control. Temperature and humidity
control can be one of the largest users of energy over which the designer has the most
amount of influence.

Process Exhaust

While process exhaust does not have a direct impact on cleanroom energy
consumption the method of exhaust replacement or makeup air does. How the makeup
air is configured into the overall air management concept can have significant impact on
overall cleanroom energy consumption. Depending upon local climatic conditions, the
treatment of makeup air is also a significant user of energy in maintaining the
cleanroom environment

Process Equipment
As previously, noted the process equipment within the cleanroom is typically the single
largest energy user in today’s semiconductor factories. The heat generated by this
equipment must be efficiently removed in order to maintain a proper manufacturing
environment. The air management concept of the cleanroom can have a significant
role in the removal of this heat. Manufacturing equipment heat load is removed either
with process cooling water, process exhaust, or cleanroom air circulation. For, most
typical cleanrooms the air circulation systems removes 85% of the tool heat load. Heat
transfer via air is not very efficient due to the low heat transfer coefficient of air versus a
liquid. Fortunately, the large air movement required for particle control provides a
significant mass of heat transfer fluid, or in plain language the “bigger hammer”
approach is used. As air circulation rates have been reduced to reduce energy
consumption and/or for capital savings, the reduced mass has presented some new
challenges to the cleanroom designer.

Humidity Control Method

The control of relative and absolute humidity is of extreme importance to the cleanroom
designer. Many processes require precise humidity control to insure product quality;
and the control of humidity is also important for the control of electro-static discharge.
Corrosion of thin films during metal deposition and etch processes may occur,
depending on the process. Humidity is also controlled to provide for operator comfort.
Figure 6 below shows the typical psychrometric operating zone. Due to the high
sensible heat ratios of semiconductor cleanrooms, humidity excursions are typically
caused by localized wet processes or moisture migration from adjacent spaces.

Figure 6.

Pressurization Control

Good pressure control is needed for many reasons in today’s cleanrooms. Pressure
control of the clean spaces can have an impact on atmospheric processes such as film
deposition. Pressure control is used to provide barriers to contamination. Positive
cleanroom pressure is maintained to keep external contamination (particles and
moisture) out of the cleanroom, and negative pressurization is used when trying to
contain hazardous materials inside the cleanroom from exiting (metal ions like gold,
boron, potassium, etc. must be carefully contained). The method of pressure control
will have an impact on the makeup air quantity and quality, and the method that
makeup is introduced into the air management concept will also impact the total energy
consumption.

AIR MANAGEMENT CONCEPT COMPONENT ANALYSIS

During the preliminary or programming stages of the cleanroom design, the

cleanroom designer must factor many different requirements into the overall
selection of the air management concept. The energy consumption or losses of
components within each air management concept are included in this evaluation.
The designer should also consider the total system dynamics when performing
optimization evaluations. Total system dynamics represents the interrela-
tionships of all elements of the cleanroom design. Within each air management
concept there are similar components or processes involved in maintaining the
critical environmental parameters. Upon review of the air management concepts
the major energy-using or energy loss components of the system include:

1. Fans

2. Coils

3. Filters

4. Ductwork

5. Sound attenuators

6. Power conveyances (motors, VFDs)

7. Miscellaneous conditioning apparatus

Items 2-5 compose the cleanroom air path and many aspects of the design
optimization can be applied equally to the different elements of the cleanroom air
path. When the design engineer begins an analysis of a cleanroom design, he
must understand the obstacles to providing energy-efficient components. Are
there restraints upon the cleanroom configuration that will eliminate one or more
of the air management concepts or impact traditional optimization methods such
as those being discussed. Once the designer is aware of any constraints he
may begin applying optimization principles beginning with the recirculation fan.

Recirculation Fan Evaluation

The cleanroom recirculation fans are used to provide a uniform airflow of ultra-clean air
over the product workspace. In a large semiconductor cleanroom with 100,000 ft2
(9290 m2) of class M1 work space (i.e., less than 10 particles per cubic foot of particles
less than 0.1µm in size), airflow rates of 6,000,000 to 9,000,000 CFM (2,832,000 to
4,248,000 L/s) may be involved. Large amounts of energy are used to transport the
cleanroom air and remove the fan heat. Many significant energy savings are possible
when high-efficiency components are used for large quantities of air circulation.

Improving fan system efficiencies will reduce the recirculation air handler brake
horsepower requirements. Changes in fan system efficiencies are achieved through:

-- Fan Drive efficiency

-- Fan mechanical efficiency
-- Motor efficiency

Proper selection of cleanroom fans involves evaluating fan types, drive mechanisms,
maintenance requirements, sound power levels, etc. Concepts 1 and 2 may use
several fan options: open and closed scroll centrifugal fans, either forward curve,
backward inclined, or airfoil centrifugal fans, or vane and tube axial fans. When vane-
axial fans are selected, the higher fan sound power levels may require additional sound
attenuation with correspondingly higher pressure drops, while some centrifugal fans
have lower efficiencies but operate at lower sound power levels. The lower static
pressure system combined with low speed large diameter vane-axial fans will have
substantial reduction to sound power thus reducing the need for some sound
attenuation. An analysis of the factors effecting total power can be show by the
following:

Airflow x Static Pressure (1)

Bhp =
6356 x ηm x ηf
Rearranging terms and solving for static pressure

Airflow x Static Pressure(v) Airflow x Static Pressure(c) (2)

=
6356 x ηm x ηv 6356 x ηm x ηc

Combining similar terms

Static Pressure(v) Static Pressure(c) (3)

=
ηv ηc
Solving for centrifugal fan efficiency

Static Pressure(v) = Static Pressure(c) x

ηv (4)
ηc
Where:
Airflow is in ft 3 per minute
Static pressure is in inches WC
6356 is a conversion factor
ηm is motor efficiency

ηf is fan efficiency

ηc is centrifugal fan efficiency

ηv is vane-axial fan efficiency

Using Equations 1 through 4, one can see that the change in fan efficiencies results in
a linear offset in static pressure. Thus, when comparing an 75% efficient (total
efficiency) vane-axial fan with a 65% efficient (total efficiency) centrifugal fan, the ratio
of 75/65 (1.154) times the static pressure of the centrifugal fan will allow a 15.4% higher
static pressure for the vane-axial fan for equal brake horsepower requirements. When
high-efficiency fans are selected, considerable energy savings will result.

The trend toward smaller local fans for cleanroom recirculation should be evaluated
against the use of larger, more efficient central fans. When localized fan filter units
(FFU) (Gerbig, Grout, O’Halloran) are used to reduce the system's static pressure,
there is a trade-off between decreasing efficiency and static pressure decreases. FFUs
cannot develop significant static pressure so the cleanroom designer must be careful in
designing the cleanroom air path. Careful selection of very low velocity, low pressure
cooling coils is needed. After air passes through any internal sound attenuators and the
discharge HEPA filters, there is comparatively little pressure left to offset external static
pressure losses. For many older small FFUs with motor sizes less than 1 horsepower
(0.75 kW) and in-line centrifugal forward-curved fans, the total efficiency drops to 50%
or less. An 85% mechanically efficient, direct-drive vane-axial fan has a motor
efficiency of 94% and total efficiency of (0.85 x 0.94) x 100% = 79.9%. Compare this
with an FFU operating at 1.5 in. WC (373.5 Pa) of static pressure, 55% mechanical
efficiency, 70% motor efficiency, and total efficiency of (0.55 x 0.70) x 100% = 38.5%.
FFUs designs are now available with the fan, the motor and the motor controller
optimized for efficiency. These designs include a brushless, electronically commutated
dc motor with an external rotor. The fan impeller is fitted directly onto the rotor. Either
embedded microprocessors or remote microprocessor control is used to adjust the
motor voltage to match the torque requirement of the fan, thereby minimizing
inefficiencies due to slip. Overall, the resultant motor efficiency is 75% to 80%,
compared to less than 40% for phased split capacitor or shaded pole motor designs.
With this improved efficiency also comes the byproduct of quieter operation. New
backward inclined versus forward curve centrifugal fans are used with a mechanical
efficiency of 60%, this allows for a potential total efficiency of (0.6 x 0.8) x 100% = 48%,
still significantly less than the 80% achieved by vane-axial fan systems.

Good fan selection provides an optimum mechanical design; the choice of drive
mechanism and motor type will also yield considerable energy-savings potential.
2 2
Additional energy savings of 2-5 W/ft of cleanroom space (21.6-54 W/m ) may be
achieved with the substitution of high-efficiency motors for standard efficiency motors.
The choice to use fans with motors outside the primary air stream will save 1.0-1.5
W/ft2 (10.7-16.15 W/m2) due to motor inefficiency, but removal of the motor from the
airstream by using belt-driven fans with lower mechanical efficiencies will reduce the
energy savings. High-efficiency motors have an inefficiency of 6-8%, whereas belt drive
systems have inefficiencies of 4-6%. The use of special direct-drive fans with motors
outside the airstream (bifurcated fan housings) may result in higher system effects and
lower total efficiency.

To summarize cleanroom fan energy: the energy associated with the cleanroom fan is a
function of the type of fan used, the fan arrangement (i.e. direct drive, belt driven),
motor type, and the total pressure (static pressure + velocity pressure = total pressure)
of the fan system. As seen in equations 1-4, static pressure is also a contributor to
energy consumption. Significant air-side saving can be achieved by lowering the
system static pressure: 1) Reduce the cleanroom airflow static path pressure losses. 2)
Reduce the cleanroom airflow in all areas or in selected areas with mixed cleanroom
HEPA filter velocities. Many manufacturers use cleanroom unidirectional velocities of
less than the historical average of 90 fpm (0.457 m/s). The potential air-side energy
savings for cleanroom velocity reductions only are summarized in Figure 7.

Figure 7. Potential Air-Side Savings

Airflow Path Evaluation

For concepts 1 and 2 the cleanroom design involves complex air paths that may be
optimized. The primary source of energy saving’s potential is from reduced air-side
static pressure losses, using efficient ductwork fittings, lower coil and filter face
velocities, etc. The following air-side components typically are used within the standard
cleanroom air supply and return system:

-- Air-moving apparatus
-- Cooling/heating apparatus
-- Air filters
-- Sound attenuators
-- HEPA/ULPA filter plenums or supply ducts
-- Air balance devices
-- Supply/return ductwork and fittings
-- Fire and smoke protection devices
-- Cleanroom components including raised flooring, ionization grids, etc.

The design and selection of each component are interrelated, and the cleanroom
designer must balance the energy use of each component with its impact on cleanroom
performance. This interrelationship must be considered in order to assess the
operating cost benefits of one device vs. the first cost of another. When the optimized
system requires a threefold increase in capital cost with a ten-year simple payback, the
cleanroom owner may decide to distribute the investment dollars to other systems.

Most of the economical static pressure reductions can be accomplished with lower coil,
damper, and/or filter face velocities and efficient ductwork design. Removing or
redesigning air filtration and cleanroom components must be balanced against the
primary goals of the cleanroom, such as the desired air cleanliness level.

When hundreds of thousands of CFM are involved, the reduction in fan static pressure
2
of just 0.1 in. WC (24.9 Pa) can result in $7200 per year of savings for a 10,000 ft (929
2
m ) cleanroom (see Figure 7). In addition to fan horsepower savings, each 0.1 in. WC
(24.9 Pa) will also produce 3.9 tons (13.7 kW) of air conditioning savings due to the
reduced fan heat load (Naughton).

Lowering the air-side pressure drop associated with heat transfer coils and filters is a
readily available energy-saving option for designers. While the use of air filters is the
primary means to achieve the desired air cleanliness level, the quantity and quality of
filtration are normally established by the cleanroom owner's contamination control
specialist. Many semiconductor cleanrooms use only minor amounts of prefiltration,
typically one set of 30% efficient prefilters. Other semiconductor firms use two stages
of prefiltration, a 30% and a 90% efficient filter in series. The removal of pre-filters after
a cleanroom is operating is also quite common. The owner normally makes the choice
of one-stage filtration, two-stage filtration, or no prefiltration. Therefore, the extra
operating costs associated with the 0.75 in. WC (186.8 Pa) to 1.0 in. WC (249 Pa)
pressure drop required for prefiltration must outweigh the shorter HEPA filter life
expectancy. (Prefilter operating costs will vary from $5.00 to $8.00 per year per square
foot verse the HEPA filter cost of $12.00 to $15.00 per square foot.) The expected life
of the HEPA/ULPA filters is not the only issue. The potential down-time required to
replace the HEPA filters are typically an overriding factor in the decision.

A good HEPA filter ceiling system design can prevent external contaminants from
entering the cleanroom and provide the proper unidirectional airflow to remove the
internal contaminants. The cleanroom designer has many choices available for
potential energy savings within the HEPA filter system. The quality of HEPA filter
removal efficiency and the pressure drop of the HEPA filter must be discussed during
project planning. HEPA filters with 99.97% removal efficiency of all particles 0.3µm or
larger are available, as well as HEPA filters with removal efficiencies of 99.99999% for
particles 0.12µm or larger. The pressure drop associated with many HEPA filters will
depend on the type of filter media used and the quantity of media per square foot of
filter face area. Today, HEPA filter pressure drops vary from 0.2 in.WC (49.8 Pa) to 1.0
in. WC (249 Pa). Normally, the lower the pressure drop, the higher the cost of the filter,
but with higher dust holding capacity and lower operating costs. The choice of which
filter to use must be input from the cleanroom owner.

Another method of reducing the system's static pressure is to use a low-velocity duct
design. Low-velocity systems occupy more facility space and have higher capital costs.
Therefore, alternative designs must be used that compensate for higher velocity
distribution system pressure losses. A common design to solve this predicament is the
use of a pressurized plenum for air supply to the HEPA/ULPA filter ceiling.
Pressurized plenum designs may reduce system static requirements up to 1.0 in. WC
(249 Pa) (Aloi, Naughton) when compared to ducted HEPA filter systems. The primary
benefit of ducted HEPA filters is the precise balance of the cleanroom unidirectional
velocity profile or parallelism. Precise balance also provides the flexibility to have mixed
cleanroom velocities. For example, the cleanroom airflow velocity may provide 90 fpm
(0.457 m/s) over the production equipment and the remainder of the cleanroom may
operate at 60 fpm (0.305 m/s) which results in a lower overall cleanroom velocity
proportional to the ratio of mixed velocity areas. Semiconductor factories using
minienvironments are based upon this principle.

For concept 3, FFUs, the ability to reduce the air path static losses is quite limited. As
can be seen in figure 5, the cleanroom air path for FFUs is limited to the cleanroom
itself and this may seem to indicate a significant benefit to the FFU concept it also one
of the drawbacks. FFU arrangements cannot make changes to the air path without
impacting the manufacturing environment or manufacturing equipment arrangement.
Therefore, the range of system static pressure is small.
Figure 8 shows the static pressure range for each the common fan types used. Also,
indicated is typical power consumption per unit of airflow. Figure 8 also presents the
ranges for high efficiency fan systems and typical system efficiencies are also shown.

Figure 8. Static Pressure Range for Fan Types

The individual system components discussed so far can greatly impact the total system
operating costs. Within the total semiconductor cleanroom system, smaller subsystems
may also be optimized for improved total system operation. Methods to improve energy
efficiency within the direct cleanroom air management system have been presented.
The HVAC process also includes another large energy user, the make-up air treatment
system. One of the largest subsystem energy users is the make-up air-handling system
when both its fan energy and the chiller operating costs are considered. The heat
transfer load of the makeup is a very large portion (typically greater than 25%) of the
total heat transfer.

Make-up Air Treatment Evaluation

As previously mentioned, the treatment of the cleanroom make-up air is a significant

user of energy. Although the make-up treatment is not within the cleanroom envelop
per se’ its airflow is filtered using HEPA filters and the control of the temperature and
moisture content (dewpoint control) is very critical. This HVAC process affords the
cleanroom designer several opportunities to optimize the energy consumption.

The water temperatures selected for chilled/glycol water dehumidification systems will
directly affect the conditioned make-up air leaving temperature as well as the operating
costs. Assuming a typical semiconductor cleanroom space condition of 68°F (20°C) at
40%RH and the resultant 42.8°F (6°C) dewpoint, a chilled/glycol water temperature of
35°F (1.7°C) to 37°F (2.8°C) normally would be used. A minimum heat exchanger
approach of 5°F to 7°F (2.8°C-3.9°C) is recommended for good controllability. The cost
of generating the chilled water is a large portion of the total operating costs of the
make-up air conditioning system. Selecting chillers with low kilowatt per ton (kW/ton)
ratios will prove cost-effective for many large semiconductor facilities. The designer
should evaluate the chilled-water temperature differential design to minimize the
kW/Ton ratio.

When traditional chilled water dehumidification is used for conditioning make-up air, the
operating costs may be reduced with a number of innovative approaches, including pre-
cooling the make-up air with heat recovery modules, using evaporative cooling, using
closed-circuit cooling tower water, and using less expensive (i.e. lower kW/ton) chillers.
The use of evaporative cooling is not very feasible in humid climates. Closed-circuit
cooling towers may function in humid climates, but the additional capital costs may
prove this method unattractive. Using dual water temperature systems is an effective
and economical means of reducing operating costs. Further discussion of the HVAC
process is beyond the scope of this paper, but the possible range of energy
consumption and operating cost can vary by 70% is some cases and several times
higher in extreme cases (Naughton)
In summary, when the chilled-water system and the cleanroom air-management system
are examined, the professional designer should not underestimate any potential energy
savings with semiconductor cleanrooms.

Total System Dynamics

Several optimization methods have been reviewed. When the entire cleanroom energy
flow is considered and the interactions of air flow and heat transfer are optimized as a
total system this can be referred to as optimization of the total system dynamics. The
potential savings of the complete system dynamics are sometimes difficult to envision.
It is obvious to most design engineers that the use of high efficiency motors, fans,
pumps, etc., will provide for higher total system efficiency. The key is to foresee the
interaction of the individual components with the whole system, as demonstrated when
comparing reduced system static pressure vs. high fan efficiency. By using identical
high-efficiency equipment, many different psychometric processes can be analyzed.

After the consideration of air management concept is made, the system dynamics are
considered when designing the temperature and humidity control system. Within each
air management concept the factors affecting cleanroom energy are different. When a
total system dynamics analysis is performed the cleanroom designer can begin an
HVAC optimization process. This HVAC optimization process involves a thorough
knowledge of all heat transfer and particle transfer mechanisms occurring. It is beyond
the scope of this paper to derive or expand the concepts involved in a detailed
optimization or case study, but there are several sources available to those wishing to
explore this topic in more detail (O’Halloran, Grout, Naughton)

CONCLUSION

This paper presented some of the methods being used by multi-national semiconductor
companies to change the way cleanrooms are being designed and built. Considering
resource conservation, energy conservation methods, and cost of ownership, allows for
unique opportunities for employing innovative designs to reduce system operating costs
and help the customer compete in a world economy. Various clean air and energy
management optimization methods were presented along with their potential for energy
savings. This paper also introduced a holistic design approach, total system dynamics,
to evaluate the energy efficiency of cleanroom designs. The benefits of individual
system component efficiency and the optimization of the total system dynamics will
result in cost and energy savings for the operator and for the customer. Implementation
of these principles will produce better designs for the client and a rewarding satisfaction
for the designer.

BIBLIOGRAPHY AND REFERENCES

Aloi, C. W. 1976. "What you always wanted to know about centrifugal refrigeration
systems." 1976 Proceedings of Chilled Water Conference, Purdue University.
ASHRAE Handbook of Fundamentals, American Society of Heating, Refrigeration, and
Air-conditioning Engineers, Atlanta, Georgia
Energy Information Administration, U.S. Department of Energy
Gerbig, F. T. 1984. "Energy consumption in. vertical laminar-flow cleanrooms."
Microcontamination, April/May.
Grout, Richard V., 1998 “Strategic Selection of Air Management Technologies to
Reduce Costs and Optimize Fab Performance”
International Sematech, 1999, “Worldwide Fab Energy Survey Report” International
Sematech Technology Transfer document #990236698B-ENG
ISO 14644-1, Cleanrooms and associated controlled environments Part 1 Cleanliness
classification by airborne particles
ISO/DIS 14644-4, Cleanrooms and associated controlled environments Part 4 Design,
Construction, and Startup
Marsh, R.C. and Quinn, T.B. 1981. "Energy savings in laminar flow clean air systems."
Solid State Technology, July.
Naughton, Philip, 1991 “HVAC Systems for Semiconductor Cleanrooms — Part 1,
System Components”, 1991 ASHRAE Transactions
Naughton, Philip, 1991 “HVAC Systems for Semiconductor Cleanrooms — Part 2, Total
System Dynamics”, 1991 ASHRAE Transactions
O’Halloran, Michael, 1998. “Improving Wafer Fab Resource Reduction Using A
Systems Synergy Approach”
Schneider, R. K 1984. "Energy conservation in clean room design." 1984 Proceedings
of the Institute of Environmental Sciences, pp. 112-116.
Takenami, T. and Ohmi, T. 1988. "Air conditioning and particle filtration systems for
energy savings." 7th International Symposium on ULSI Ultra Clean Technologies,
July 21-23, Tokyo, Japan.