Advanced Search

Please click here to take a brief survey

Reader Report: The Carbon Footprint of Research at Caltech

This piece is by Asa S. Hopkins, a graduate student studying physics at Caltech. We're happy to share his ideas about how to make this major research institution more sustainable.

We encourage submissions from members of Worldchanging's global audience who volunteer to write up their notes from conferences, workshops and other worldchanging happenings they participate in. If you'd like to contribute your own report, please email

CAltech.jpg Over the years, the California Institute of Technology (Caltech) has earned a name for itself as one of the world’s foremost scientific research universities. In its more than 100 buildings, a small community of just over 3,000 faculty, students, and researchers study everything from chemical engineering to planetary sciences. But, as a member of Caltech’s Sustainability Council, I learned that the facilities and machines used to conduct this research are responsible for generating more than 100,000 tons of greenhouse gas emissions per year.

Prior to 2006, students felt as if our pleas to increase sustainability efforts fell on deaf ears. But that changed almost overnight once Caltech hired its current President, Jean-Lou Chameau, who brought with him a continuing commitment to improving institutional sustainability and implementing sustainable technology.

Caltech has, in fact, been making incremental progress toward sustainability, including reducing its use of energy intensive lighting and power options, as well as arranging a "power purchase agreement" for solar power from a photovoltaic arrays to be installed atop several buildings.

But despite these efforts, we still face big challenges in terms of carbon emissions. In many ways, these challenges have their roots in Caltech's defining purpose: research comes first. The operations and administration of the Institute are not in a position to challenge the faculty on energy use, and research energy use is intense. Laboratory buildings operate 24/7 -- many experiments simply can't be shut off -- and Caltech's academic buildings are almost all laboratory buildings.

I'd like to highlight two research energy uses in particular: fume hoods and high performance computing. Fume hoods are used for chemical and biological research, with an air flow away from the researcher, then filtered and out a chimney, in order to allow the use of dangerous chemicals. One fume hood uses about as much energy as 3-4 average California homes -- and Caltech has about 800 fume hoods, almost all of them older, constant flow models. (It's not the fan itself that's the energy hog -- it's the fact that the air is cooled and blown into the building through the HVAC system, only to be sucked right back out again.) Measurements at the Jet Propulsion Laboratory indicated that replacing a constant flow fume hood with a variable flow hood (one which pulls less air when the glass sash is pulled lower) would save at least 15 MWh of electricity per year (15,000 kWh). For Caltech, the energy savings from simply upgrading fume hoods would cut our total energy use by about 10 percent.

In comparison with relatively well-understood fume hoods, which have an easy prescription, high performance computing is a puzzle yet to be cracked. According to Institute administrators, a computer cluster, and facilities to host it, is a component of almost every new faculty start-up package. Caltech expects that increases in energy use from computing will be the dominant driver of increased energy use for the Institute as a whole over the next 10 years. Rooms throughout Caltech have been refit for the added air conditioning load, and racks of servers run 24/7. The Institute is considering asking faculty to consolidate servers into shared (hopefully more efficient) facilities, but institutional deference to faculty desires (which is often amply rewarded with research success) may make this difficult.

A critical difference between fume hoods and computing is who makes the choices, and what motivates them. Fume hoods are part of the building infrastructure, and the same people who design and maintain the building (the Institute) pay the power bills. Computing, on the other hand, is bought by faculty, usually using federal grant money, while the electric bill is paid by the Institute. Faculty members have every incentive to buy the most powerful computers they can buy, within a fixed initial cost, ignoring how efficient they may be. The Institute charges a fixed overhead, to pay for electricity, bandwidth, support staff, etc., but is not allowed to pass the actual energy costs on to the faculty. Federal grant money is highly restrictive, and overhead must be a fixed quantity, paid at time of purchase. Energy costs for servers can exceed purchase costs in a matter of years, yet the Institute can't charge for this power.

While Caltech's administration can't control exactly how electricity is used, it does have some control over its input energy mix. The Institute gets about 70 percent of its total power from its co-generating natural gas plant, located on campus, and purchases the rest from the city utility, Pasadena Water and Power (PWP). PWP's energy mix is relatively coal-heavy for California, so cuts in purchased electricity will have the greatest impact on carbon emissions. In off-peak times, Caltech can meet all its on-campus needs from natural gas; we buy electricity during the workday, and especially during "peak use" times for air conditioning. Fume hood upgrades will help here, due to some reduced air conditioning load, but Caltech’s new megawatt of solar photovoltaic arrays will tackle this head-on. Solar photovoltaic power will peak just when we need it to: summer afternoons. A rebate program from the City of Pasadena, which expires at the end of 2008, is driving the construction schedule.

Interestingly, while the city of Pasadena is driving our increased adoption of renewable energy sources (purchased from another source), the city also needs Caltech to continue purchasing electricity from them in order to balance the city's budget. If we cut our purchase of power from PWP, we directly affect the city's bottom line, resulting in either higher rates for the remaining power that we do buy from them (and that others pay throughout the city), or cuts in city services.

Caltech will be in Pasadena for decades to come -- and this interdependence requires the Institute and the City to embrace a broad definition of "sustainable" in order to build a bright green future with a budget that balances. Caltech can afford to look ahead many years when making its decisions, so it has the flexibility to try novel and potentially powerful actions to change the way it uses resources. My personal favorite idea is for Caltech to partner with Pasadena to construct a solar thermal power plant on Caltech-owned land in the California desert, and agree to buy power from that plant for an extended period to guarantee the financial sustainability of that investment.

Caltech is a world leader in its scientific research. With such impressive minds at work, can the institution pair existing solutions with innovation to lead the charge for its own greener future?

Photo credit: Flickr/taiwanjim, Creative Commons license.

Bookmark and Share


Very interesting article. I like the way it delves into some of the administrative challenges faced at universities in trying to cut energy consumption: Faculty are rewarded for good research and cannot legally be made to pay for their energy use. Not surprisingly, they choose to use a lot of energy to do research!

This is the disadvantage of faculty freedom. Sometimes the dictatorial nature of companies has a good outcome: If the CEO says to cut energy use by 20%, people will actually do it. Professors.... well, it's a slower process. Many simply aren't interested, and those who do care about "greening" their labs will probably face disadvantages! I wonder how we can change the incentives and rewards for having greener research labs/computers such that there are real benefits?

Posted by: Kim on 14 Sep 08

Thought provoking article!

As you mentioned, reducing energy use by installing more efficient fume hoods is a much clearer solution than lowering energy use in computer servers. Last week, the Economist took a more in depth look into this question and provides some interesting insight. The summary goes: new technology that uses water to cool processing chips could provide enough heat to warm 70 nearby homes.

Still five years off, but a good example of thinking outside of the box. Check out the article here:

Posted by: Mason on 14 Sep 08

@Kim: Caltech is looking into ways of charging for computing infrastructure using methods other than straight overhead. I don't pretend to be an expert n such things, but they hope to make some progress by changing incentives. They're also hoping that the same machines, hosted in a single facility, might be less energy intense, and are considering moving as many as possible into one place (with an offer to take over hardware admin tasks as a bonus). Of course, some faculty want to be able to physically touch their computers at any moment, and might not move....

@Mason: Good points. Unfortunately for Caltech, the co-generation plant already generates more hot water/steam than the school can use, so there's no "in-house" place to use the warm water generated by a server farm. And home heating isn't a big issue in southern California. :) Integrating Caltech's hot water with off-campus homes would be a massive undertaking -- interesting if building from scratch, hard to retrofit.

Posted by: Asa Hopkins on 14 Sep 08

A few solutions that seem fit for Caltech's campus are adsorption cooling, heat-recovery ventilation and cloud computing.

If they already generate more hot water/steam than they need, it's more efficient to use it to run an adsorption chiller (or a chiller at each building) than it is to run compressor-driven AC with electricity. Depending on the layout of the buildings and available space leftover from the PV, small solar thermal arrays (MicroCSP) could economically supplement their steam production (see Sopogy, a Hawaii-based company with offices in California).

Also, it seems obvious to me to install heat-recovery ventilation systems in those buildings with fume hoods, so the incoming air can be pre-cooled by the exhaust air. Surely someone at Caltech thought of this one...?

Cloud computing is possible for applications that don't require a ton of computing power, so it may or may not be applicable for some parts of their campus.

Posted by: greensolutions on 17 Sep 08

A few months ago on TED, they had Adam Grosser talking about a device he's working on called a sustainable fridge, which converts heat into cold. I know it gets hot in Pasadena, and computers and equipment make it doubly so. Could Caltech scientists work with this concept to use waste heat to actually cool the work spaces? The device is so small, it looks like a car could be air conditioned off its waste exhaust heat from the engine. It would make AC units today obsolete. Obsolete. And a third time for emphasis: Obsolete.

Posted by: Oysters and Trifle on 22 Sep 08



MESSAGE (optional):

Search Worldchanging

Worldchanging Newsletter Get good news for a change —
Click here to sign up!


Website Design by Eben Design | Logo Design by Egg Hosting | Hosted by Amazon AWS | Problems with the site? Send email to tech /at/
Architecture for Humanity - all rights reserved except where otherwise indicated.

Find_us_on_facebook_badge.gif twitter-logo.jpg