Feature Story: Nanoscale Motors
by Thomas E. Mallouk and Ayusman Sen
In the 1966 movie Fantastic Voyage, a team of scientists, physicians, and secret agents are shrunk to molecule size and travel through the body of the shrinking ray’s inventor, Jan Benes, to destroy a blood clot in his brain. Their micron-size submarine, the Proteus, navigates hazards in the patient’s lungs, heart, inner ear, and circulatory system before finally completing its mission. From our point of view, the most fascinating thing about this story is not Racquel Welch in her nano-wetsuit, but the idea of a tiny, controllable, motorized craft that could perform such a useful medical procedure. In addition to their obvious applications in microsurgery and targeted drug delivery, autonomous micro-machines and nano-machines could act as the “assemblers” of a molecularly engineered world of new devices, from electronic circuits and solar cells to head-of-a-pin-size diagnostic laboratories.
Molecular Machines
Credit: (top) Paul Weiss, Penn State; (bottom) James Tour and Kevin Kelly, Rice University
(top) A double-decker molecular rotor pokes out from within an array of single-decker molecules, which isolate the double-decker molecules from each other and thereby give them to room to rotate. (bottom) Four-wheeled nanocars move back and forth on a gold surface along the direction defined by their axles. In contrast, three-wheeled nanotripods simply spin in place.
Unfortunately, we cannot shrink atoms as Dr. Benes does. To do so, we would need to change such fundamental quantities as the speed of light, Planck’s constant, and the mass of an electron. However, we can use ordinary atoms and the tools of chemical synthesis to design and make molecules and nanoscale objects that look—and in some ways act—like simple machines. Work on this problem is part of a team effort involving faculty and students in the Eberly College of Science and the College of Engineering at Penn State, as part of the Center for Nanoscale Science supported by the National Science Foundation.
Figure 1 (above) is an illustration of one of these nanoscale molecules. It shows two molecular machines, which have been studied by Paul Weiss and his group at Penn State, and by collaborators Jim Tour and Kevin Kelly at Rice University. The top image shows a molecule that consists of disk-shaped segments at the top and bottom joined by a single metal atom. These disks are similar chemically to the molecules found in certain pigments and in hemoglobin—the proteins in blood that carry oxygen—and they easily join together to make double-decker molecules. The image at the upper left, of a double-decker molecule inserted into a two-dimensional crystal of individual disk molecules, was acquired using a scanning tunneling microscope (STM). In this image, in which the false color scale corresponds to the height of the molecule’s electron cloud, the bright yellow cross marks the highest point in the double-decker molecule. The STM is able to resolve individual molecules and some internal features of the molecules, such as the rings at the four corners of the disks, but it cannot resolve individual atoms. Nevertheless, one can determine by STM that the bottom disk (the stationary “stator” component that functions as a pivot point) is fixed, and the top disk (the revolving rotor component) can turn freely on its “axle.”
The second molecule—which is more of a synthetic tour de force than the rotor—is a nanoscale “car” that drives on buckminsterfullerene (C60) wheels. With its 3-nanometer wheelbase, this car is only slightly wider than the cross section of the DNA double helix. STM images of these molecules on a terraced surface of gold atoms, shown in the lower right of Figure 1 behind the space-filling drawings, resolve only the outline of the molecule. However, they do reveal something interesting about its motion. Because of the way the STM images are obtained, one can track the movement of the car on the gold surface by comparing its position and orientation from frame to frame. This kind of analysis shows that the cars move back and forth in the direction perpendicular to their axles (i.e, in “forward” and “reverse”) more easily than they move from side to side. These observations are consistent with rolling rather than sliding C60 wheels. In a control experiment, Tour and Kelley made and tested an analogous three-wheeled triangular vehicle that was able only to spin in place, and this experiment confirmed the rolling-versus-sliding hypothesis.
credit: Will Hancock and Tom Jackson, Penn State
A capped circular channel fitted with kinesin biomotors loads microtubules from the left and steadily accumulates a large population of circulating fluorescent tubules, as shown in superimposed before- and after-images.
The nanocar is far more easily mass-produced than the miniature Proteus submarine in Fantastic Voyage or, for that matter, any other kind of real or toy car. The Proteus was made by shrinking a full-size submarine—no doubt built with government funding—and it was, unfortunately, gobbled up by a white blood cell at the end of its mission. Although the synthesis of nanocars requires several steps, one milligram of the product contains 1017 cars. By comparison, a human body consists of about 1013 cells, so it is not unreasonable to think of molecular vehicles as disposable reagents for carrying cargo into and within cells.
Why can’t the nanocar do what the Proteus can? The answer is, it does about as well as the Proteus would without a steering wheel or an engine. We will return to the steering wheel question in the last part of this article. The engine problem is particularly vexing for all kinds of nanoscale machines because none of our macroscopic designs for powering movement—internal combustion engines or electromagnetic motors, for example—can be scaled down to that level. Control and power are thus the two most essential challenges today in the nano-machine business.
Nature Does it Already
In 1966 when the Fantastic Voyage film was released, the prevailing view of the biochemistry of a living cell was that of a bag of enzymes carrying out a chain of metabolic reactions. We now view the cell as a highly structured, compartmentalized network of interlocking molecular machines, more like the inner workings of a Swiss watch than a chemical reactions in a flask. Such basic intracellular functions as protein synthesis, ADP phosphorylation, and DNA replication are intimately coupled to the movement of biological motors. In addition to the protein motors that drive skeletal muscle movement and cell division, there are “conveyor belts” that carry cargo inward and outward within our cells, protein motors that drive flagellae and cilia, and protein scaffolding that assemble and disassemble to change the shape of the cell itself.
There are three essential features of all known biological motors. First, regardless of the size of the object that they ultimately move (a miotic spindle or an elephant’s trunk), biological motors consist of nanoscale moving parts. Second, they derive their power by catalyzing the reaction of a fuel, for example adenosine triphosphate (ATP). Third, because the “power stroke” of breaking or making bonds in the fuel molecule is intimately coupled to motion, biological nanomotors tend to be very energy-efficient.
In the late 1990s, scientists began to experiment with biological motors as power sources for engineered systems. In classic experiments of this kind, Hiroyuki Noji at Tokyo Institute of Technology and Carlo Montemagno at Cornell University showed that genetically engineered ATP synthase—the rotor-stator enzyme that synthesizes ATP from ADP in living systems—could propel much larger, synthetic objects attached to it. The rotor part of the enzyme—a complex of six proteins—is only about 12 nanometers in diameter, and is thus far too small to see in an optical microscope. A microscope could, however, easily track the movement of micron-size nickel rods and plastic beads attached to the invisible rotor. Montemagno observed an attached rod rotating for a period of hours in the presence of ATP. By measuring the number of rotations generated from a given amount of ATP and estimating the drag force on the nickel propeller, he calculated that the energy-conversion efficiency of a semi-synthetic rotor between 50 percent and 80 percent. This range is consistent with the previously measured high efficiency of ATP synthase in the cell.
At Penn State, Will Hancock (Bioengineering), Tom Jackson (Electrical Engineering), Mary Beth Williams (Chemistry), Jeff Catchmark (Engineering Science and Mechanics), and their students are engineering another kind of semisynthetic biomotor based on the “conveyor belt” proteins of eukaryotic cells. By anchoring the motor protein kinesin to the walls of engineered microchannels, these researchers are building a micromachine that is able to move microtubules—the same objects that form the miotic spindle in cell division—at speeds of microns per second along the channels. In the cell, the kinesin protein walks along microtubule tracks, carrying cargo through the cell or forcing chromosomes apart in cell division. The movements of these proteins are driven by the hydrolysis of ATP. In the microchannels, the immobilized kinesin molecules pass the microtubules along in a hand-over-hand motion, in much the same way as the Nittany Lion mascot is passed overhead between the outstretched hands of fans during a home football game.
Microtubules are long, tubular assemblies of two globular proteins, ?- and ?-tubulin. The assembled structure is polar, meaning that kinesin moves in only one direction along the microtubule axis. Once assembled, microtubules are relatively stiff, like a length of inflated garden hose, so they cannot turn around sharp corners in the microchannels. This characteristic enables the construction of one-way turnabouts and corrals for the moving microtubules, as shown in the time-lapse fluorescence micrograph in Figure 2. Microtubules enter the 70 micron-diameter storage ring from a reservoir at the lower left. Those that are pushed counterclockwise by the immobilized kinesin molecules are trapped in the ring, but those moving the opposite way quickly exit. Gradually, a population of the microtubules that were pushed counterclockwise builds up in the ring, from which they can be released later using an electric field to open a microfluidic gate. In related research, the Penn State scientists have trapped microtubules in similar channels — steered them magnetically by means of tiny ferrite particles attached to the microtubules — and have sorted them by electric fields into separate destination reservoirs.
Credit: Ayusman Sen and Tom Mallouk, Penn State
Catalyzed movement of template-grown striped nanorods in 2.5% aqueous hydrogen peroxide. Left: Darkfield image of 2 micron long platinum-gold nanorods. Center: Schematic illustration of catalyzed axial movement in the absence of a magnetic field, and directional movement in an applied field. Right: Catalytic platinum-gold nanorod containing magnetic nickel stripes.
The microtubules have some of the Proteus-like properties that the nanocar lacks. They glide along the channels in one direction, powered by a chemical fuel that, in principle, they can use very efficiently. And they move independently from each other, but can be directed by the channels and by external fields. Still, their utility as autonomously powered micro-vehicles is limited by the fact that they can glide only along kinesin-lined channels. Because the tubulin-amino-acid sequence is one of the most tightly conserved in all of biology, there is also relatively little wiggle room for designing, de novo, variants that might take on other shapes or functions.
Catalytic Nanomotors
The biological nanomotor examples described above are inspiring in that they suggest a solution to the problem of scaling an engine down to the size of molecules. Catalysis is a well understood and fundamentally nanoscale process. In the chemical industry, catalytic molecules and nanoparticles are used for a wide variety of chemical reactions, including petroleum processing, the manufacture of plastics, and the combustion of gases in automotive exhaust. Catalytic reactions can release large amounts of energy, and we have many fuels at our disposal that are “better,” in terms of their energy content, cost, and stability, than ATP.
Our work on catalytic nanomotors began in 2003 with the idea that two things would be needed to convert chemical energy into mechanical energy on the nanoscale. First, we needed an energetic reaction that could run only on the surface of a catalyst. Second, we would need asymmetric catalyst particles, so that the forces generated in the reaction would not cancel each other out. For the fuel, we chose hydrogen peroxide, an unstable molecule that decomposes catalytically to oxygen and water on metal surfaces. It was a convenient choice because only one compound was needed. Later, Nicolas Mano and Adam Heller at the University of Texas at Austin demonstrated chemical locomotion in a similar system using a fuel-oxidant mixture of glucose and oxygen.
Credit: Jeffrey Catchmark, Penn State
100 micron diameter gold crogearsith platinum ethan rotate ~360o/sec in aqueous hydrogen peroxide systems.
The asymmetric particles we used were cylindrical nanorods made by sequentially electroplating gold and platinum in the pores of alumina filter membranes. This method provided a convenient supply of nanorods that were between 50 and 350 nanometers in diameter and several microns long. The pattern of metals along the length of the rods could be controlled by sequential electroplating. This level of synthesis control meant that complex axial patterns, including magnetic stripes for steering, could be built in.
Viewed under the optical microscope, the movement of chemically powered nanorods bears an eerie resemblance to that of live multiflagellar bacteria, such as E. coli. The catalytic “engine” drives the rods in the axial direction at speeds of tens of microns per second, but like the bacteria, the rods also are subject to random twists and turns from Brownian forces in the solution. A detailed analysis of the forces involved by Penn State theorists Paul Lammert and Vincent Crespi suggested several possible mechanisms by which some of the energy of the reaction might be converted to motion. Experiments by Penn State graduate students Walter Paxton, Tim Kline, and Yang Wang established that the dominant mechanism was electrochemical: hydrogen peroxide is oxidized and reduced at the platinum and gold surfaces, respectively. These reactions generate an imbalance in the proton concentration along the axis of the rod. It is actually the protons, which must move from platinum to gold to maintain charge balance, that push water past the rod and (by Galilean invariance) propel the rod through the fluid. In simple terms, the proton flux is like a paddle that pushes water along the side of a canoe. Once the researchers established this principle, several other catalytic motor designs followed. These designs included bimetallic rotary motors as well as immobilized catalyst patterns on surfaces that act as fluid pumps.
The study of catalytic motors has taught us several important lessons about the scaling laws for micro-engines and nano-engines, as well as about the challenges that lie ahead. One such lesson is the increasing dominance of Brownian forces—the random jolts from energetic solvent molecules—at very small length scales. Because smaller motors generate smaller forces, they must be increasingly efficient to work effectively against Brownian motion. They are additionally more effective if their random motion is constrained; for example, by immobilizing the motor to a membrane, or by confining it to a one-dimensional track.
Unconstrained motors smaller than bacteria (1 to 2 microns in length) are rare in biology. Indeed, bacteria exist in a very interesting size regime, where they are large enough to move as autonomous organisms but small enough to be constantly re-oriented by Brownian forces. Bacteria swim toward things they like, such as food and signals from other organisms, in a process known as chemotaxis. Interestingly, the mechanism of bacterial chemotaxis is uniquely suited to their length scale. Bacterial motion consists of a series of straight runs and random turns. When a bacterium is moving toward its target, it extends the length of its straight runs and therefore “triangulates” toward its goal. In the motion of catalytic nanorods, an analogous situation exists. That is, in a region of higher fuel concentration, the motion of the rod is faster and, therefore, it travels farther between random turns. Analysis of the movement of platinum-gold rods shows that they move up a gradient of hydrogen peroxide concentration, just as bacteria move towards chemical signals. This discovery is potentially important in the design of “smart,” autonomous nano-robots, which could move independently in the direction they are needed, perhaps by harvesting energy from glucose or other abundant fuels in biological systems.
Conclusion
In conclusion, the young field of nanomotors already has made significant progress toward the realization of autonomously moving, controllable nanomachines. In this research, biological motors provide inspiration for the synthetic systems, illustrating many of the physical principles and possible designs. Synthetic nanomotors have yet to exploit any of the reactions (especially catalytic hydrolysis and polymerization) that biology uses so effectively. By doing so, we hope to provide nanocars and related machines with engines and control mechanisms that will enable a range of unimagined and useful devices.
Thomas E. Mallouk and Ayusman Sen
