As a graduate student and postdoc, AMAX Career Development Assistant Professor in Materials Science and Engineering Robert J. Macfarlane conducted groundbreaking research on how to make ordered crystal nanostructures by grafting oligonucleotides — short strands of artificial DNA that are chemically synthesized to have a specific, targeted structure — onto nanoparticles.
These oligonucleotides are biologically inactive but nevertheless form the same double helical structure common in DNA of living organisms, which makes them rigid. They also terminate in “sticky ends” — short sections of DNA that act like tiny locks and keys to bind nanoparticles to each other.
“With the DNA-based system, we can actually program in the structure with nanometer-scale precision in where we put particles in three dimensions,” Macfarlane says. His published papers show more than 20 different crystallographic symmetries, with hundreds of different crystal structures, including seven or eight geometries that have never been synthesized in either nanoparticle or atom-based crystals. The nanoparticles can consist of any material composition that can be grafted with these oligonucleotides, and Macfarlane’s work demonstrated that this process works with particles made of metals, oxides, semiconductors, and even organic materials.
“What happens is I have two particles with DNA strands attached to the surface. They have these long rigid sections and they terminate in just small little tags of DNA strands that are unpaired,” Macfarlane explains. “The tags — those DNA sequences — on adjacent particles are complementary to each other, which means that the sticky ends of two different particles can hybridize to one another and form a linkage. You can think of it almost like a smart Velcro. You have these really weak interactions between each individual pair of DNA sticky ends, but because you have a large number of them on the DNA particle surface, you get a large number of DNA connections coming together and forming a very strong particle-to-particle bond.”
“If you know what DNA structure you’re working with and you know what particle structures you’re working with, you can predict what crystal structure is actually going to form from that given set of building blocks and vice versa,” he says.
Overcoming limits of DNA
While this work demonstrates control over molecular structure that currently cannot be replicated with any other synthetic method, the high-cost of these chemically synthesized DNA strands and their narrow range of processing conditions limit their potential use as scalable, functional materials. Although DNA is easily synthesized in our cells, nanoparticle assembly requires that the oligonucleotides be synthesized in the laboratory, so that they contain the correct base sequence to direct the assembly process. Because they are challenging to synthesize in high yield, Macfarlane says a gram of these chemically synthesized oligonucleotides can cost up to thousands of dollars.
“While DNA is the best nanoscale assembly tool we have, hands down, it’s a terrible building block for functional materials at the bulk scale because you simply can’t make enough of it,” he explains. Macfarlane, who joined the MIT faculty in summer 2015, wants to transcend these limits by designing polymer-based materials that mimic the organizational power of DNA. “Polymers have the advantage that, although they are not as precise as DNA, they can be made at the gram to kilogram scale. So they’re much more applicable to functional and useful bulk materials, if you’re talking about building a structure or device.”
The materials that the Macfarlane lab aims to make are categorized as nanocomposites — materials that consist of a polymer matrix with embedded inorganic particles. The particle components of a polymer nanocomposite are typically added to improve the mechanical properties of a polymer, but can also be used to alter their electrical, optical, magnetic, or other chemical and physical properties in ways that would be impossible for polymers alone. A common example of enhanced polymers is automotive tire rubber. “People throw carbon black into tire rubber; it changes the mechanical properties, making them tougher and more resilient and therefore better for use as car tires,” he notes. In these materials, the particles are typically randomly dispersed throughout the polymer. “What we’re planning on doing, and we actually have the first examples of this, is using polymers that are grafted to a nanoparticle to precisely engineer where the particles are located within the polymer matrix,” Macfarlane explains. “So you basically have a particle with a bunch of polymer chains grafted to it. You can think of it like a Koosh ball — a rigid core with a bunch of flexible strands coming off of it — and those strands terminate in functional groups that engage in some sort of weak bonding interaction like the ‘sticky end’ interactions between the DNA strands in our previous work.” When these particles come together, the interactions between the polymer strands link them together into a networked structure.
Hoping to mimic the programmability of DNA-directed molecular assembly, Macfarlane is exploring hydrogen bonding as a method for programming crystal structure. “The idea is we have particle A with, for example, hydrogen bond donors, [and] particle B, with hydrogen bond acceptors. You put them together in solution, and they form a linkage,” he says.
“What we’re investigating is ways to actually induce nanoparticle ordering in polymer composites via self-assembly routes,” Macfarlane says. “It’s those sticky ends and where they’re positioned relative to the particles that dictates these bonding interactions, and so this is why it’s potentially [an] incredibly powerful crystallization technique — the chemical composition of the building blocks doesn’t matter in terms of dictating structure. What that means is you can separate the atomic level structure from the nanoscale ordering and macroscale shape of the material; you can control all three independently of one another.”
“If you think about the way chemists build materials, they have atoms that have electrons that occupy electron orbitals and bonds are formed when electron orbitals overlap with one another. ... You’ll get electrons being shared or swapped, depending on the type of crystal structure you’re talking about,” Macfarlane explains. “So if we’re analogizing these nanoparticles to atoms as building blocks, now these polymers and these sticky ends, they are kind of like our electron orbitals that actually dictate bonding interactions between particles.”
Programmable atom projects
Working on this research in Macfarlane Lab is a team consisting of postdoc Jianyuan Zhang; MIT graduate students Peter J. Santos, Sangho Lee, Diana Lewis, and Paul Gabrys; and sophomore Caroline Liu. “We grew way faster than I expected,” Macfarlane says. “The students are fantastic so I’m certainly happy with growing quickly, though — it’s just a matter of keeping the funding going to allow these great students to keep doing their work.”
Macfarlane’s team hopes to find a key to unlock the secrets of a “universal” nanocomposite assembly strategy that will work with multiple polymer and nanoparticle compositions. Such control of crystal structure formation from the bottom up could offer an alternative to current top-down methods of forming electronic devices, Zhang suggests. “Leading techniques in industry can now make the circuits in commercial chips with the FET [field effect transistors] about 14 nanometers in size, but we are approaching the limit of the smallest size that lithographic techniques can achieve,” Zhang says. “If we build the materials and circuits or devices from bottom up, that is, we throw the nanoparticles onto some surface, and then they automatically arrange to the shape or size that we want, that would be a totally revolutionary approach to that.”
One approach is the group’s Universal Ligand project, which aims to attach polymers onto nanoparticles for crystallization in an organic solvent. “Universal means it can be modified in many ways because it’s a synthetic polymer, so one set of design principles can potentially fit various complex systems,” Zhang explains. The polymers can be adjusted by changing their chemical composition and size, altering their polarity [electrical charge state] or modifying their attraction to, or repulsion of, water. Other characteristics of the polymer system that can be changed include the relative sizes of the attached polymers and their “sticky ends” and the number of attached polymers per nanoparticle. Compared to the DNA system, “we have more handles,” Zhang says. Macfarlane says, “The idea here is if we have this universal self-assembly methodology, we can target many different material types simply by changing out the chemical composition of the building blocks we’re working with.” Future projects may explore how this bottom-up assembly can be controlled by electrostatic or metal coordination interactions.
Lee, whose background is in computational simulation, is focusing on developing computational models that give information about the particle and polymer behavior at the level of individual polymer-grafted particle building blocks. “This is the first time [for me] to try experimental techniques or these soft materials. So it is totally new to me, but it is very interesting to learn many new things and combine the two aspects, theory and experimental techniques,” Lee explains. Liu joined the lab this summer through MIT’s Undergraduate Research Opportunities Program (UROP) after taking a chemistry class with Macfarlane in the spring. She worked on her project 40 hours a week during the summer but will work fewer hours during the semester. “I’m interested in soft materials, so not metals and not electronics. This is kind of just my way of exploring the field and trying to figure out what I want to do with my future,” Liu says.
Besides the Universal Ligand project, Macfarlane’s team also is continuing DNA-enabled self-assembly work; studying self-assembly in polymer melts; and designing hybrid inorganic/polymer hydrogels. Zhang, who got his PhD at Virginia Tech, has a background in synthetic chemistry, particularly the cage-like carbon structure known as a fullerene. He previously served as a postdoc at the University of Washington, where he worked on fullerene-based solar cells and electronic devices. “Rob has this goal of, or has this idea of, adding polymers to these gold nanoparticles, so that’s a really, really good step forward to establish a bridge between inorganic nanomaterials and organic nanomaterials,” Zhang explains.
Graduate student Santos is working on the goal of taking the self-assembly processes the Macfarlane lab is developing, and performing them directly within a polymer solid, instead of a liquid solvent [how all current assembly processes are performed]. “Polymers and glasses have a property where they go from something that’s a liquid that can move around and behave like liquids do, to something that behaves like a solid,” Santos explains. This transition from a liquid-like to a solid state in polymers is called its glass transition temperature. Santos’s project aims to develop a process in which nanoparticles can self-assemble in the heated polymer in its liquid-like phase, then be cooled to its solid phase. “Now you have something you can actually make devices out of, which would be beneficial,” he says. “This way you would have something that’s a solid, that you could make into things that people could use, but also it’s easily processable.”
For his lab work, Santos starts with solution-dispersed nanoparticles and grafts polystyrene chains with “sticky ends” capable of forming hydrogen-bonded pairs to their surfaces. This allows the particles to be easily embedded within a polystyrene matrix. “What P.J. is investigating is how to get these things to actually move around once they’re embedded and tangled in the polymer chains that make up the matrix material,” Macfarlane says.
Macfarlane notes self-assembly within a polymer melt could eliminate the separate step of assembling nanostructures first and then adding them to polymers. “It also allows us to think about interesting things like actually making large-scale materials via additive manufacturing or roll-to-roll processing. You simply embed the particles within the polymer, form it into the three-dimensional structure you want, and then direct the particles to assemble either while the polymer is being molded into its macroscopic shape, or afterwards,” he says. “You get your three-dimensional macroscopic objects, but you also get your nanoscale structure embedded within it. That currently does not exist.”
In a separate project, assisted by graduate students Paul Gabrys and Diana Lewis (a Draper Fellow), Macfarlane is continuing to work with DNA by pairing DNA-programmed self-assembly, which controls crystal structure at the nanoscale, with top-down lithography that determines structure across a larger footprint. “We use a process called e-beam lithography to pattern a surface so that it looks like the top layer of a nanoparticle superlattice,” Macfarlane explains. “If we then functionalize that pattern with DNA, in terms of the way the particles see that substrate, they see it just as another arrangement of particles that they can bind to,” he adds. “What we do is essentially pattern a substrate with an array of posts that mimic the size and shape of the superlattice we are trying to create, and then build up the superlattice layer-by-layer, so this allows us to control both the individual particle arrangements as well as the overall macroscopic size and shape of the superlattice itself.” The Macfarlane lab is using DNA programmed particle assembly because of the advantages it provides in precisely programming crystal structure, but ultimately aims to pair these strategies with the polymer-based particle assembly approaches they are developing. Macfarlane is collaborating on this work with his former doctoral adviser at Northwestern University, Chad Mirkin.
“We have some small funding that we came in with from the Air Force that allows us to do this DNA-linked assembly work. That will be ending at the end of the year, so we’re looking for other sources of funding to keep this work going,” Macfarlane says.
Advantages of gold
Macfarlane is conducting his self-assembly research with gold nanoparticles because attaching polymers or DNA to them is easy and because they give visual cues when their state transforms. “Gold nanoparticles have what’s called a surface plasmon resonance, basically a collective oscillation of electrons, where the energy associated with that oscillation is the same as the energy of the wavelengths of visible light,” Macfarlane explains. “What that means is [that] gold nanoparticles absorb very strongly in the visible regime. This surface plasmon resonance is very sensitive to the size and the shape of the nanoparticle, but also its local environment. So if I take two particles that have surface plasmon resonances that make them red in color, I bring those particles very close together, the surface plasmons can actually couple, the energy of the oscillation changes, and now the particles absorb different wavelengths of light, and appear purple. So that means for doing self-assembly that’s a very nice optical tag. So if they’re in the free, discrete state, where they are not bound to each other, they’re red, as soon as everything comes together, the color shifts, and they’re purple. So we can stick things in a UV-vis spectrophotometer and over time measure how these things are actually assembling, which makes it a very easy way to track them without sophisticated or complicated equipment.”
“Gold obviously is not the best material to make things out of because it’s expensive, unless you want to use the specific chemical properties of gold, but it’s a very nice testbed material to do all this work with, because it’s very easy to characterize and see what’s going on,” he says.
Once the basic design strategies are outlined, the next step will be to utilize particles with other desirable optical, electrical, magnetic, or mechanical properties. In particular, “nanocomposites containing silica are an attractive goal that we are pursuing next, since they are inexpensive and have good mechanical properties for controlling the strength and toughness of composites. However, once the design rules have been worked out, any polymer or particle combinations are fair game; this method could be applicable to any current polymer or particle composition used in academia or industry,” Macfarlane says. “We think this may be the next big revolution in controlling nanocomposites for a whole host of applications.”
Macfarlane currently teaches Course 3.034 (Organic and Biomaterials Chemistry). He will co-teach 3.091 (Solid-State Chemistry) with fellow MIT faculty member Niels Holten-Andersen in the spring. In both classes, he uses the technology he is developing in his research lab as a tool to educate undergraduates on modern applications of materials synthesis and design, and to reinforce basic concepts in materials science. “Ultimately, I have plans to replicate some of the simpler experiments we are now performing in the 3.034 lab; this would give students a taste of what cutting-edge research is actually like, instead of canned experiments based on decades-old science,” Macfarlane says.
A native of Palmer, Alaska, Macfarlane graduated from Willamette University in Salem, Oregon, where he minored in art. He got a master’s at Yale University before completing his PhD at Northwestern in the lab of Chad Mirkin. From there, he completed a postdoc at Caltech in the labs of Bob Grubbs and Harry Atwater, working on synthetic polymers with novel optical properties, such as photonic crystals for IR-reflective paints and surface coatings. He began in the MIT Department of Materials Science and Engineering in 2015. His lab conducted all of its initial work in temporary space in building E34 but recently moved to newly renovated space in Building 13.