When organic chemists submit a paper detailing the synthesis of a new molecule to the likes of the Journal of Organic Chemistry, they must comply with a set of well-established requirements.
The researchers have to include nuclear magnetic resonance and infrared spectra, submit mass spectrometry data, and report melting points at each stage of the molecule-making process. Yield needs to be calculated, and for chiral compounds, optical rotation measured.
Chemists use this battery of analytical tests to prove that they’ve made what they say they’ve made—and to confirm that they’ve removed by-products and impurities along the way. The characterization data serve as part of a blueprint others can follow to reproduce the work.
No such standards have yet been established for the synthesis of nanomaterials, however.
Papers explaining how to make nanospheres, -rods, and -cubes typically display simple electron microscopy images of the freshly synthesized particles. Synthetic procedures list centrifugation steps to remove residual reaction reagents from nanoparticles and, perhaps, to isolate particles of a desired size. But yield is rarely mentioned.
During the early days of nanotechnology discovery, when nanomaterials were just something scientists tinkered with in the lab, this lack of detailed characterization probably didn’t raise eyebrows. But today, researchers are using nanomaterials as catalysts as well as incorporating them into electronic devices and sensors. They’re also developing the tiny materials to be injected into patients’ bodies as cancer-fighting medicines and imaging agents for disease.
Scientists working in this area are becoming more concerned about nanoparticle characterization, and they’re debating what levels of purity are needed for particular applications. Some are establishing new methods to characterize and purify the nanomaterials that labs around the world are pumping out with increasing vigor. Others are discussing whether manuscript reviewers and journals should set minimum characterization requirements for authors seeking to publish syntheses.
An established set of required characterizations would help nanoscience mature, argues Mary Beth Williams, a chemist at Pennsylvania State University, University Park. That’s because it would help researchers reproduce nanoparticle syntheses from the literature more consistently, she says.
Williams, whose lab synthesizes both supramolecular inorganic compounds and magnetic nanoparticles, says the lack of standards didn’t bother her when she first got into nanoscience. But then her students began having problems reproducing published procedures.
“Even stuff we’d worked out and reported, my students would say that it might work only 50% of the time” when they later returned to it, she says. “They would be really frustrated.”
According to the scientists C&EN interviewed, nanoparticle synthesis is exquisitely sensitive to reaction conditions—the purity of the precursors or surfactants used, temperature, pressure, and stir rate, to name just a few. Sometimes, Williams quips, “it seems like the phase of the moon is involved.”
Catherine J. Murphy is no stranger to the reproducibility issue in nanomaterial synthesis either. In 2001, she and her group, then at the University of South Carolina, Columbia, reported the controlled synthesis of gold nanorods (J. Phys. Chem. B, DOI: 10.1021/jp0107964). Afterward, “people would e-mail me and say, ‘Hey, I’m trying to reproduce your synthesis, and I’m not getting the same stuff,’ ” says Murphy, a chemist now at the University of Illinois, Urbana-Champaign.
In some cases, she says, the problem was lab temperature. “Our lab was warm” during the initial experiments, she explains. Because one of the reaction reagents gets added to the synthetic pot at saturating concentrations, researchers in cooler labs were precipitating it out of solution.
In other cases, the purity of cetyltrimethylammonium bromide (CTAB) was to blame. CTAB, a surfactant that stabilizes Murphy’s nanorods as they grow, was causing other shapes to form depending on the supplier from which it was sourced. In 2009, a team led by Brian A. Korgel of the University of Texas, Austin, figured out that some suppliers’ CTAB contained parts-per-million levels of iodide impurities that inhibited nanorod growth by sticking to the particles’ crystalline facets (Langmuir, DOI: 10.1021/la900757s).
“That study was a huge eye-opener for me,” Korgel says. “I hadn’t anticipated that parts-per-million levels of anything in a surfactant could completely change a reaction product.”
As a result of these and other discoveries about the nanorod synthesis, Murphy says she’s come to have the view that “nanoparticle synthesis is extremely kinetically controlled.” Tiny changes in how the reaction is carried out can influence the end result. “It’s almost like cooking,” she adds. “You could give a chocolate-chip cookie recipe to 100 people,” and because of differences in ingredient quality, oven temperature, and so forth, “the cookies you get are not going to be precisely the same.”
What’s encouraging, says Christopher B. Murray, a chemist at the University of Pennsylvania, is that people have realized this reproducibility problem exists. “Now, within a few weeks—or certainly within a few months—many of the better recipes for nanomaterials get reproduced by multiple groups.” Researchers are starting to have a feel for which reaction parameters can be sticking points on the basis of well-publicized cautionary tales such as the nanorod/CTAB story. And they’re beginning to report more of the details in the methods sections of their papers. “So there’s a natural peer process bringing some of this under control,” Murray notes.
“But again,” Williams says, “because there’s no real confirmation of purity reported in the literature, it’s a problem that’s been compounded over time.” And scientists are still working through it.
“Nanomaterial purity” can mean a number of different things, says Vincent A. Hackley, a research chemist at the National Institute of Standards & Technology, in Gaithersburg, Md. There’s purity in terms of physical properties—all the particles having the same size, shape, and degree of aggregation, he says. And then there’s purity in terms of having minimal by-products and generating particles with the same chemical composition (for example, identical amounts of cadmium and selenide in a CdSe quantum dot).
Residual reagents from nanomaterial synthesis also stick around by adsorbing to the freshly made particles’ surfaces, says Scott E. McNeil, director of the Nanotechnology Characterization Laboratory, in Frederick, Md. NCL helps cancer researchers test the toxicity and preclinical efficiency of their lab-designed nanoparticles. Some researchers try to remove these reaction remnants with methods such as filtration and buffer exchange, but those steps are often not 100% effective, he adds.
The degree to which each of these purity parameters matters depends entirely on the intended use of the nanoparticle, scientists in the community say.
For example, incomplete removal of leftover synthesis reagents is a particularly large problem for nanomaterials meant for biomedical use. Case in point: In 2009, Murphy and her group realized that even though they washed their nanorods and centrifuged them, minuscule amounts of the toxic surfactant CTAB hung around. So when the researchers added the nanorods to human cells to test their safety, the tiny particles seemed to kill the cells.
“But then my students did a clever experiment,” Murphy says. “They took our particle solutions, centrifuged off the particles, and added the supernatant to the cells.” The liquid, which contained free-floating CTAB molecules at a micromolar level, also killed the cells (Small, DOI: 10.1002/smll.200801546). “We purified the nanorods,” she explains, “but it wasn’t good enough as far as the cells were concerned.”
Because they will eventually be injected into the human body, nanomaterials intended for biomedical applications, in general, should be purified and well characterized, according to the scientists interviewed by C&EN. At this point, however, the Food & Drug Administration has no specific policies regarding the materials’ regulation. “Nanomaterial-containing products,” including those in clinical trials, “are dealt with on a case-by-case basis,” says Lisa C. Kubaska, an FDA spokeswoman.
Clean nanoparticle surfaces and characterization of their composition are also vital when scientists use the materials as catalysts, says Keith J. Stevenson, a chemist at UT Austin. Stevenson, who develops multimetal nanostructures for electrocatalysis, thinks it would be especially beneficial for the nanoscience community to establish a set of purity benchmarks that take into consideration the class and intended use of nanomaterials. But setting standards for ensuring particle purity, he says, “has to be motivated by a very important problem that people care about investing resources in.”
That’s because generating a batch of particles that are free of contaminants and that are identical in size and shape—otherwise known as monodisperse—is a feat that takes time, money, and energy, says Nicholas A. Kotov, a chemical engineer at the University of Michigan, Ann Arbor. “I wholeheartedly agree that monodispersity is convenient and good,” Kotov says. “But it comes at a price, so we really need to find a boundary between where polydispersity is acceptable and where it is not.”
For instance, applications in which nanoparticles are being encapsulated in a composite or coating may not require monodispersity, says Raymond E. Schaak, a chemist at Penn State’s University Park campus.
And, Murphy explains, “if you want particles that broadly absorb or emit over a wide range of light, a mixture of sizes or shapes is actually desirable.”
Kotov encourages scientists in the field to embrace nanomaterial polydispersity in situations where it makes sense. Taking a cue from natural supraparticles such as viruses, he and his group recently demonstrated that, given enough time, a group of CdSe nanoparticles with a broad size distribution will self-assemble into larger, more uniformly sized superstructures (Nat. Nanotechnol., DOI: 10.1038/nnano.2011.121). Using these assemblies rather than smaller, individual particles for certain applications, he argues, might obviate the need for extensive nanomaterial purification.
Part of the difficulty of addressing the importance of nanomaterial purity is that each scientist has a different system in mind, and “each class of nanomaterials is at a different stage of development,” says Paul S. Weiss, a chemist and materials scientist at the University of California, Los Angeles, and editor-in-chief of the journal ACS Nano. Simple metal clusters can be more easily characterized than a complex particle made of thousands of atoms and functionalized with multiple ligands. For the more complicated systems, Weiss says, “metrology tools are lagging pretty far behind” the complexity of what researchers are currently concocting in the lab.
Today, nanoscience researchers typically examine batches of nanoparticles with electron microscopy to assess an average particle size and shape. The problem with this practice is that sometimes researchers draw conclusions from a single transmission electron microscopy image without any corroborating information, Schaak says. And when the image includes only a small sampling of particles—50 or fewer—it can be deceiving and not at all reflective of the overall sample, he adds. “As chemists, we work on an Avogadro’s number-type scale,” he says. “Fifty is trivial.”
As in organic synthesis, nanoscience researchers should be using multiple characterization techniques to assess their materials, according to the scientists C&EN spoke with.
Part of the problem with requiring researchers to use a particular set of characterization techniques is that, at this stage, analytical tools for nanomaterials are specialized and expensive, says Schaak, who is also an associate editor for ACS Nano. “Most research-intensive chemistry departments around the country have the required tools for characterizing molecules,” he says, “but the same can’t be said of characterizing nanoparticles.”As a result, for journals, setting standards is complicated, Schaak notes.
Not only that, UCLA’s Weiss says, but nanoparticle synthesis is still not understood in a lot of detail. “We’re just not yet at a point where, as in crystallography, we can expect to have X-ray structures that can be put in databases,” he adds.
Some in the nanoscience community have tried to get around doing extensive characterization by focusing on perfecting their protocols to generate monodisperse samples, experts say. This strategy, however, is hard to get away with as nanomaterials become more and more complex, Williams says.
“It’s like multistep organic synthesis,” she adds. “If you don’t purify along the way,” the resulting batch of particles can be a mess. So there’s a growing need for developing ways to isolate desired products in nanoscience, she argues.
So far, though, “folks working in nanoscience haven’t been attuned to developing analytical separation tools,” Williams says. There’s not yet a nanoparticle purification equivalent for doing high-performance liquid chromatography on organic molecules, she says.
Many scientists who synthesize nanomaterials currently use gradient ultracentrifugation, filtration, or dialysis methods to purify their nanomaterials. But other techniques that are more sensitive to elemental composition are needed, Williams contends.
For this reason, she and others are developing separation methods to fill the void. Collaborating with Schaak in Penn State’s Center for NanoAnalytics, Williams reported a magnet-assisted chromatography method last year that the researchers used to separate hybrid magnetic nanoparticles (Angew. Chem. Int. Ed., DOI: 10.1002/anie.201104829). These complex structures contain two, three, and four particles linked to form a larger particle with multiple functionalities.
A team led by Bruce K. Gale, a mechanical engineer at the University of Utah, also published a paper this year about the development of an electrical field-flow fractionation method for separating metal nanoparticles by size (Anal. Chem., DOI: 10.1021/ac300662b). And Hackley says he and others at NIST are working on joining methods like this one to inductively coupled plasma mass spectrometry to combine size separation and composition analysis.
Analytical tools for nanomaterial characterization are evolving, but their development is often hard to get funded and published, scientists in the community say.
In nanoscience, “you typically have to do application-driven work,” UT Austin’s Stevenson says. “When you make a new nanoparticle, you immediately have to show that it’s good for something.”
Still, Williams hopes that as nanoscience matures, more researchers in the field will tackle the challenge and collaborate with analytical scientists to develop characterization tools and set standards.
From a fundamental perspective, Schaak says, if nothing else, such tools should help researchers better understand nanomaterial synthesis by identifying by-products and figuring out what led to their formation. That knowledge, he adds, will eventually lead to improved synthetic protocols and nanomaterials of higher purity.