To be submitted to Atmospheric Environment

while simultaneously controlling indoor particle concentrations with improved air filtration.
Time- and size-resolved particle concentration data were gathered at an operating northern
California data center while using air filtration of varying levels of efficiency. Along with
measuring the size- and time-resolved indoor and outdoor particle concentrations, chemical
analysis of particulate sulfate, nitrate, chloride, and carbon was performed under each filter and
economizer configuration. Metered electricity data were also gathered to compare overall
operational energy use for each condition. Fan power requirements were calculated, using the
mechanical design specifications of the data center, to disaggregate the potential increase in fan
energy associated with improved filtration from the energy use associated with compressor-
based cooling. The measured particle characteristics and energy use are evaluated to explore the
potential energy savings associated with a shift from conventional cooling and filtration practices
in data centers to a system with economizer use, to save energy, combined with improved
filtration, to ensure protection of equipment from particle contamination.
2. Methods
2.1 Experimental Design
Particle concentrations were measured at a data center in Sunnyvale, CA during August,
2008. Real-time measurements were made and two- or three-day integrated filter-based samples
were collected during 8-29 August for subsequent analysis. The mechanical system at this data
center is designed with an economizer, allowing the amount of outside air entering the zone to be
adjusted depending on outside temperature and humidity conditions. The amount of outside air
entering the data center is controlled by an energy management and control system (EMCS).
This data center was evaluated in a previous study (Shehabi et al., 2008), which presents details
of the building layout and mechanical design. Briefly, economizer dampers within the air-

3