Project Overview

What if you could capture and convert sunlight into electricity with a material as cheap and as versatile as a plastic bag? What if the material could be produced on a massive scale, with easily accessible technology? What if other versions of the material could be coated, painted, or sprayed on building surfaces for solar energy production? What if these materials were ultra-thin and ultra-light for portable devices? And finally, what if they were inexpensive and could provide electricity to people in the developing world? 

Solar Energy and the Energy Supply of the Future

The fossil fuel based lifestyle and economy of the present must give way to one based on renewable and sustainable energy. Getting there is one of humanity’s greatest challenges. Organic solar cells offer the potential to realize this goal. The Harvard Clean Energy Project was conceived in 2008 to help find materials to make this exciting technology a reality!

The sun is an abundant source of energy and Earth receives enough solar energy per hour to meet mankind’s annual demand. Each year enough sunlight irradiates the Earth’s surface to provide as much as twice the energy of all fossil fuel and nuclear energy reserves. Solar energy is thus an obvious alternative to these non-renewable sources and will play an important role in sustainably supplying rising energy demands.

Silicon-Based Solar Cells

The solar cells of today are typically made of silicon, an inorganic semiconductor material. Unfortunately, these devices require complicated and energy-intensive manufacturing processes leading to high production costs. Research shows that commercially-available solar cells require two years to generate the amount of energy used to make them! As a result, the cost of electricity from silicon-based photovoltaics is notably higher than that of nonrenewable grid-scale electrical power generation. Silicon based solar cells also tend to be heavy, bulky, rigid, and fragile.

Carbon-Based Solar Cells

Carbon-based (i.e., organic) solar cells have emerged as one of the interesting alternatives to this silicon-based technology. They consist of crystalline (ordered) small molecules or amorphous (disordered) polymers such as plastics. Organic photovoltaics (OPVs) have great potential for several reasons. They promise simple, low-cost, high-volume production and the prospect of merging the flexibility and versatility of plastics with electronics. OPVs can be processed via roll-to-roll printing, and there is active research into sprayable or paintable materials. They can be semi-transparent, are light-weight, and can be molded into  any shape. These properties make OPVs promising materials for harvesting solar energy. Particularly interesting target applications include integration into buildings and wearable devices.

Challenges for Carbon-Based Solar Cells

Despite their promising outlook, there are still significant issues to overcome in order to make plastic solar cells a viable technology. The biggest problems are their relatively low performance and limited lifetime. A solar cell’s performance is measured by its power conversion efficiency (PCE) - the percentage of incident solar energy it captures. Single-junction OPVs have PCEs as high as 11.1% in the lab, but production scale cells are still only about 4-5% efficient. Current state-of-the-art materials degrade when exposed to the environment. Improving production materials by increasing their efficiency to 10-15% in combination with extending their lifetimes to over 10 years will push the power generation costs of OPVs below those of other energy sources such as coal. The key parameters for the improvement of OPVs are essentially known; however, engineering materials which combine all these features still remains a challenge because suitable compounds are difficult to identify.

Traditional Development of New OPV Materials

Traditional experimental development of new materials is largely based on empirical intuition or experience with a certain family of chemical compounds. Using this approach, an experimental group can only study a few materials per year due to the lengthy synthesis and characterization process.

The CEP’s Virtual High-Throughput Screening Framework

The CEP is a theory-driven research effort that uses a virtual high-throughput screening approach to identify potential high-performance material candidates on an unprecedented scale. The CEP uses an automated computational framework to study candidate structures. The current project phase is primarily concerned with the characterization of millions of compounds using first-principles molecular quantum mechanics. A computational study of molecular materials combining the scale and level of theory found in the CEP is unprecedented.

The CEP stands out from other computational materials science approaches as it combines conventional molecular modeling with strategies from modern drug discovery. It also adopts techniques from cheminformatics, materials informatics, and machine learning to scale the process of developing structure-property relationships and improve existing efficiency models.

Distributed Volunteer Computing via the IBM World Community Grid

The scale of this study requires a correspondingly large computational resource, which is provided by distributed volunteer computing on the World Community Grid (WCG). This virtual supercomputing platform was founded by IBM to harvest idle computing time and utilize it for philanthropic research. Participants can donate their spare computing time by running the supported science applications on their personal computers, either on low priority in the background or in screensaver mode during idle periods. In the CEP, we employ a custom version of the Q-Chem 3.2 package, which was ported to the Berkeley Open Infrastructure for Network Computing (BOINC) environment. From the user perspective, participation in a WCG project is fully automated and usually does not require any input or maintenance beyond the initial setup. Presently, more than 600,000 users have signed up 2,300,000 computers to the various WCG projects (as of June 2013). You can contribute to this effort by joining the WCG and signing up for this project!

The WCG is a powerful resource and provides us with the means for our high-throughput investigation. It is, however, also unusual due to the non-specialized hardware and host demands. These limitations have to be taken into consideration in the design of suitable tasks. In addition to the WCG, CEP utilizes the Harvard FAS Odyssey cluster and has accounts at NERSC and XSEDE (formerly TeraGrid) for problems which are outside the scope of volunteer grid computing due to their size and computational complexity. The CEP currently characterizes about 20,000 compounds per day, and so far we have collected results equivalent to more than 17,000 years of CPU time.

The Harvard Clean Energy Project Database

The results from this large-scale study are compiled and analyzed in a massive reference database – the Harvard Clean Energy Project Database (CEPDB) – that is available for public use. It contains data on 2.3 million molecular structures in 22 million geometries. This took 150 million density functional theory calculations and comprises 400TB of data. For the data storage we have custom-built racks of hard drives which currently have a capacity of 750TB. We nicknamed these pods “Jabba”s after Jabba the Hut.


An encouraging early result of the CEP was the discovery of DA2T, which emerged from our original proof-of-principle study. DA2T is a very powerful organic semiconductor, with a hole mobility of 12.3 cm²/Vs. To our knowledge, this is the second best organic semiconductor reported in the literature.

The CEP as a Joint Venture Between Theory and Experiment

While the centerpiece of the CEP is the computational study of OPV candidates, we point out that an overarching theme of our project is also the tight integration of experimental and theoretical work. The project is in many instances guided by input from experimentalist collaborators, in particular the Bao Group at Stanford University. Our most promising candidates are studied in-depth in their laboratories. The CEP is ultimately a community tool and we invite and welcome collaborations.

CEP Publications

We have published a number of articles in scientific journals about this work. Direct PDF download links are available from this website.


The above text was adapted from our 2011 paper “The Harvard Clean Energy Project: Large-Scale Computational Screening and Design of Organic Photovoltaics on the World Community Grid” in The Journal of Physical Chemistry Letters. Please also consider all the work by others cited therein.

© 2007-2013 The Clean Energy Project | Terms & Privacy | Credits
Follow us: