By Carol Clark
When most people think of a chemistry lab, they picture scientists in white coats mixing chemicals in beakers. But the lab of theoretical chemist Francesco Evangelista looks more like the office of a tech start-up. Graduate students in jeans and t-shirts sit around a large, round table chatting as they work on laptops.
“A ‘classical’ chemist is focused on getting a chemical reaction and creating new molecules,” explains Evangelista, assistant professor at Emory University. “As theoretical chemists, we want to understand how chemistry really works — how all the atoms involved interact with one another during a reaction.”
Working at the intersection of math, physics, chemistry and computer science, the theorists develop algorithms to serve as simulation models for the molecular behaviors of atomic nuclei and electrons. They also develop software that enables them to feed these algorithms into “super” computers — nearly a million times faster than a laptop — to study chemical processes.
The problem is, even super computers are taxed by the mind-boggling combinatorial complexity underlying reactions. That limits the pace of the research.
“Computers have hit a barrier in terms of speed,” Evangelista says. “One way to make them more powerful is to make transistors smaller, but you can’t make them smaller than the width of a couple of atoms — the limit imposed by quantum mechanics. That’s why there is a race right now to make breakthroughs in quantum computing.”
Evangelista and his graduate students have now joined that race.
The Department of Energy (DOE) awarded Evangelista $3.9 million to lead research into the development of software to run the first generation of quantum computers. He is the principal investigator for the project, encompassing scientists at seven universities, to develop new methods and algorithms for calculating problems in quantum chemistry. The tools the team develops will be open access, made available to other researchers for free.
While big-data leaders — such as IBM, Google, Intel and Rigetti — have developed prototypes of quantum computers, the field remains in its infancy. Many technological challenges remain before quantum computers can fulfill their promise of speeding up calculations to crack major mysteries of the natural world.
The federal government will play a strong supporting role in achieving this goal. President Trump recently signed a $1.2 billion law, the National Quantum Initiative Act, to fund advances in quantum technologies over the next five years.
“Right now, it’s a bit of a wild west, but eventually people working on this giant endeavor are going to work out some of the current technological problems,” Evangelista says. “When that happens, we need to have quantum software ready and a community trained to use it for theoretical chemistry. Our project is working on programming codes that will someday get quantum computers to do the calculations we want them to do.”
The project will pave the way for quantum computers to simulate chemical systems critical to the mission of the DOE, such as transition metal catalysts, high-temperature superconductors and novel materials that are beyond the realm of simulation on “classical” computers. The insights gained could speed up research into how to improve everything from solar power to nuclear energy.
Unlike objects in the “classical” world, that we can touch, see and experience around us, nature behaves much differently in the ultra-small quantum world of atoms and subatomic particles.
“One of the weird things about quantum mechanics is that you can’t say whether an electron is actually only here or there,” Evangelista says.
He takes a coin from his pocket. “In the classical world, we know that an object like this quarter is either in my pocket or in your pocket,” Evangelista says. “But if this was an electron, it could be in both our pockets. I cannot tell you exactly where it is, but I can use a wave function to describe the likelihood of whether it is here or there.”
To make things even more complicated, the behavior of electrons can be correlated, or entangled. When objects in our day-to-day lives, like strands of hair, become entangled they can be teased apart and separated again. That rule doesn’t apply at the quantum scale where entangled objects are somehow intimately connected even if they are apart in space.
“Three electrons moving in three separate orbitals can actually be interacting with one another,” Evangelista says. “Somehow they are talking together and their motion is correlated like ballerinas dancing and moving in a concerted way.”
Much of Evangelista’s work involves trying to predict the collective behavior of strongly correlated electrons. In order to understand how a drug interacts with a protein, for example, he needs to consider how it affects the hundreds of thousands of atoms in that protein, along with the millions of electrons within those atoms.
“The problem quickly explodes in complexity,” Evangelista says. “Computationally, it’s difficult to account for all the possible combinations of ways the electrons could be interacting. The computer soon runs out of memory.”
A classical computer stores memory in a line of “bits,” which are represented by either a “0” or a “1.” It operates on chunks of 64 bits of memory at a time, and each bit is either distinctly a 0 or a 1. If you add another bit to the line, you get just one more bit of memory.
A quantum computer stores memory in quantum bits, or qubits. A single qubit can be either a 0 or a 1 — or mostly a 0 and part of a 1 — or any other combination of the two. When you add a qubit to a quantum computer, it increases the memory by a factor of two. The fastest quantum computers now available contain around 70 qubits.
“Quantum computers are not just exponentially faster, they work in a radically different way from classical computers,” Evangelista says.
For instance, a classical computer can determine all the consequences of a chess move by working one at a time through the chain of possible next moves. A quantum computer, however, could potentially determine all these possible moves in one go, without having to work through each step.
While quantum computers are powerful, they are also somewhat delicate.
“They’re extremely sensitive,” Evangelista says. “They have to be kept at low temperatures to maintain their coherence. In a typical setup, you also need a second computer kept at very low temperatures to drive the quantum computer, otherwise the heat from the wires coming out will destroy entanglement.”
The potential error rate is one of the challenges of the DOE project to develop quantum software. The researchers need to determine the range of errors that can still yield a practical solution to a calculation. They will also develop standard benchmarks for testing the accuracy and computing power of new quantum hardware and they will validate prototypes of quantum computers in collaborations with industry partners Google and Rigetti.
Just as they develop algorithms to simulate chemical processes, Evangelista and his graduate students are now developing algorithms to simulate quantum software so they can run tests and adapt the design based on the results.
Evangelista pulled together researchers from other universities with a range of expertise for the project, including some who are new to quantum computing and others who are already experts in the field. The team includes scientists from Rice University, Northwestern, the University of Michigan, CalTech, the University of Toronto and Dartmouth.
The long-range goal is to spur the development of more efficient energy sources, including solar power, by providing detailed data on phenomena such as the ways electrons in a molecule are affected when that molecule absorbs light.
“Ultimately, such theoretical insights could provide a rational path to efforts like making solar cells more efficient, saving the time and money needed to conduct trial-and-error experiments in a lab,” Evangelista says.
Evangelista also has ongoing collaborations with Emory chemistry professor Tim Lian, studying ways to harvest and convert solar energy into chemical fuels. In 2017, Evangelista won the Dirac Medal, one of the world’s most prestigious awards for theoretical and computational chemists under 40.
Related:
$2 million NSF grant funds physicists' quest for optical transistors
Chemists find new way to do light-driven reactions
Physicists devise method to reveal how light affects materials
from eScienceCommons http://bit.ly/2Bnc4F1
By Carol Clark
When most people think of a chemistry lab, they picture scientists in white coats mixing chemicals in beakers. But the lab of theoretical chemist Francesco Evangelista looks more like the office of a tech start-up. Graduate students in jeans and t-shirts sit around a large, round table chatting as they work on laptops.
“A ‘classical’ chemist is focused on getting a chemical reaction and creating new molecules,” explains Evangelista, assistant professor at Emory University. “As theoretical chemists, we want to understand how chemistry really works — how all the atoms involved interact with one another during a reaction.”
Working at the intersection of math, physics, chemistry and computer science, the theorists develop algorithms to serve as simulation models for the molecular behaviors of atomic nuclei and electrons. They also develop software that enables them to feed these algorithms into “super” computers — nearly a million times faster than a laptop — to study chemical processes.
The problem is, even super computers are taxed by the mind-boggling combinatorial complexity underlying reactions. That limits the pace of the research.
“Computers have hit a barrier in terms of speed,” Evangelista says. “One way to make them more powerful is to make transistors smaller, but you can’t make them smaller than the width of a couple of atoms — the limit imposed by quantum mechanics. That’s why there is a race right now to make breakthroughs in quantum computing.”
Evangelista and his graduate students have now joined that race.
The Department of Energy (DOE) awarded Evangelista $3.9 million to lead research into the development of software to run the first generation of quantum computers. He is the principal investigator for the project, encompassing scientists at seven universities, to develop new methods and algorithms for calculating problems in quantum chemistry. The tools the team develops will be open access, made available to other researchers for free.
While big-data leaders — such as IBM, Google, Intel and Rigetti — have developed prototypes of quantum computers, the field remains in its infancy. Many technological challenges remain before quantum computers can fulfill their promise of speeding up calculations to crack major mysteries of the natural world.
The federal government will play a strong supporting role in achieving this goal. President Trump recently signed a $1.2 billion law, the National Quantum Initiative Act, to fund advances in quantum technologies over the next five years.
“Right now, it’s a bit of a wild west, but eventually people working on this giant endeavor are going to work out some of the current technological problems,” Evangelista says. “When that happens, we need to have quantum software ready and a community trained to use it for theoretical chemistry. Our project is working on programming codes that will someday get quantum computers to do the calculations we want them to do.”
The project will pave the way for quantum computers to simulate chemical systems critical to the mission of the DOE, such as transition metal catalysts, high-temperature superconductors and novel materials that are beyond the realm of simulation on “classical” computers. The insights gained could speed up research into how to improve everything from solar power to nuclear energy.
Unlike objects in the “classical” world, that we can touch, see and experience around us, nature behaves much differently in the ultra-small quantum world of atoms and subatomic particles.
“One of the weird things about quantum mechanics is that you can’t say whether an electron is actually only here or there,” Evangelista says.
He takes a coin from his pocket. “In the classical world, we know that an object like this quarter is either in my pocket or in your pocket,” Evangelista says. “But if this was an electron, it could be in both our pockets. I cannot tell you exactly where it is, but I can use a wave function to describe the likelihood of whether it is here or there.”
To make things even more complicated, the behavior of electrons can be correlated, or entangled. When objects in our day-to-day lives, like strands of hair, become entangled they can be teased apart and separated again. That rule doesn’t apply at the quantum scale where entangled objects are somehow intimately connected even if they are apart in space.
“Three electrons moving in three separate orbitals can actually be interacting with one another,” Evangelista says. “Somehow they are talking together and their motion is correlated like ballerinas dancing and moving in a concerted way.”
Much of Evangelista’s work involves trying to predict the collective behavior of strongly correlated electrons. In order to understand how a drug interacts with a protein, for example, he needs to consider how it affects the hundreds of thousands of atoms in that protein, along with the millions of electrons within those atoms.
“The problem quickly explodes in complexity,” Evangelista says. “Computationally, it’s difficult to account for all the possible combinations of ways the electrons could be interacting. The computer soon runs out of memory.”
A classical computer stores memory in a line of “bits,” which are represented by either a “0” or a “1.” It operates on chunks of 64 bits of memory at a time, and each bit is either distinctly a 0 or a 1. If you add another bit to the line, you get just one more bit of memory.
A quantum computer stores memory in quantum bits, or qubits. A single qubit can be either a 0 or a 1 — or mostly a 0 and part of a 1 — or any other combination of the two. When you add a qubit to a quantum computer, it increases the memory by a factor of two. The fastest quantum computers now available contain around 70 qubits.
“Quantum computers are not just exponentially faster, they work in a radically different way from classical computers,” Evangelista says.
For instance, a classical computer can determine all the consequences of a chess move by working one at a time through the chain of possible next moves. A quantum computer, however, could potentially determine all these possible moves in one go, without having to work through each step.
While quantum computers are powerful, they are also somewhat delicate.
“They’re extremely sensitive,” Evangelista says. “They have to be kept at low temperatures to maintain their coherence. In a typical setup, you also need a second computer kept at very low temperatures to drive the quantum computer, otherwise the heat from the wires coming out will destroy entanglement.”
The potential error rate is one of the challenges of the DOE project to develop quantum software. The researchers need to determine the range of errors that can still yield a practical solution to a calculation. They will also develop standard benchmarks for testing the accuracy and computing power of new quantum hardware and they will validate prototypes of quantum computers in collaborations with industry partners Google and Rigetti.
Just as they develop algorithms to simulate chemical processes, Evangelista and his graduate students are now developing algorithms to simulate quantum software so they can run tests and adapt the design based on the results.
Evangelista pulled together researchers from other universities with a range of expertise for the project, including some who are new to quantum computing and others who are already experts in the field. The team includes scientists from Rice University, Northwestern, the University of Michigan, CalTech, the University of Toronto and Dartmouth.
The long-range goal is to spur the development of more efficient energy sources, including solar power, by providing detailed data on phenomena such as the ways electrons in a molecule are affected when that molecule absorbs light.
“Ultimately, such theoretical insights could provide a rational path to efforts like making solar cells more efficient, saving the time and money needed to conduct trial-and-error experiments in a lab,” Evangelista says.
Evangelista also has ongoing collaborations with Emory chemistry professor Tim Lian, studying ways to harvest and convert solar energy into chemical fuels. In 2017, Evangelista won the Dirac Medal, one of the world’s most prestigious awards for theoretical and computational chemists under 40.
Related:
$2 million NSF grant funds physicists' quest for optical transistors
Chemists find new way to do light-driven reactions
Physicists devise method to reveal how light affects materials
from eScienceCommons http://bit.ly/2Bnc4F1
Aucun commentaire:
Enregistrer un commentaire