Purity and Dispersion Measurement Issues Workshop on …

   Purity and Dispersion Measurement Issues Workshop on
          Single Wall Carbon Nanotubes (SWCNTs)
                                Workshop Summary
                                          NIST
                                 Gaithersburg, Maryland
                                     May 27-29, 2003

A workshop organized jointly by the National Aeronautics and Space Administration, Lyndon B.
Johnson Space Center (JSC) and the National Institute of Standards and Technology (NIST) was
held May 27-29, 2003 at NIST in Gaithersburg, MD. In attendance were 68 participants,
representing 11 private corporations, 19 universities, and 9 government agencies. The primary
purpose of the workshop was to bring together leading researchers in the field of single wall
carbon nanotubes (SWCNTs) to discuss and prioritize measurement needs relative to nanotube
purity and dispersion.

The topics of purity and dispersion were chosen because it was recognized that the ability to
accurately measure and describe the purity of nanotube-containing materials and their dispersion
in liquids or polymers is crucial for future development and use of SWCNTs. Currently, a
variety of measurement techniques are utilized for purity and dispersion; significant differences
in both methodology and interpretation exist from one laboratory to another. For this reason,
comparison of different SWCNT materials is extremely difficult.

To address these challenges, the organizing committee invited 23 speakers, and developed an
agenda that encouraged active participation from attendees. Breakout sessions addressing both
workshop topics were held to foster open discussion and to invite consensus regarding best
techniques and measurement methods. A final panel discussion led to recommendations for
future work and to plans for developing documentation of existing techniques. The agenda, as
well as speaker and poster session abstracts, is included in the appendices.

The following sections provide details of the issues relative to measurement of purity and
dispersion.

Purity
The purity of single walled carbon nanotubes can be loosely defined as the quantity of SWCNTs
relative to the metal catalysts and other carbon-like materials present (amorphous, graphitic, and
C60 carbons). A number of measurement techniques used for purity determination were
discussed. The most extensively utilized techniques are thermogravimetric analysis (TGA),
scanning electron microscopy (SEM), transmission electron microscopy (TEM), Raman
spectroscopy, and infrared spectroscopy, but others are employed as well. The comment was
made that no single technique can describe the quality of a sample of nanotubes.
TGA is commonly used to determine the presence of both carbon and non-carbon impurities,
e.g., metal. A draft TGA protocol was proposed by Nikolaev et. al. (JSC) that specified
parameters for pre-desiccation, maximum heating rates, moisture content in the ambient
atmosphere, minimum sample mass, and minimum number of samples. The authors stated that
the residual mass, Mr, after burnout can be used to determine the fraction of residual metal. The
maximum temperature is a measure of thermal stability. The standard deviation of both of the
above parameters gives a picture of the homogeneity of the sample. Issues of detecting non-
metallic impurities, e.g., metal carbide, through TGA analysis were also discussed. Various
limitations to the use of TGA were identified, e.g., the presence of metal carbides or oxide,
temperatures spikes at high heating rates, etc.

Both TEM and SEM are used extensively for qualitative analysis of a sample containing
SWCNTs. There was general accord that unless the TEM image demonstrates the existence
of a significant quantity of SWNTs, no one will agree on the quality of the material. A draft
protocol for TEM specimen preparation and observation was proposed by Nikolaev et. al. (JSC).
Discussions at the breakout sessions led to a modification of this protocol. Nikolaev et. al. also
pointed out that electron dispersive spectroscopy (EDS) can be used in combination with
observation in the TEM to yield information on the chemistry of impurities. There was
consensus that the methodology by which TEM and SEM images are selected should always be
specified.

Haddon (U. Ca., Riverside) suggested that Near Infrared Absorption (NIR) can be utilized for
carbon characterization and to distinguish SWCNT from "schmutz". At present, such analysis is
limited to nanotubes produced by the arc process. Haddon also pointed out that NIR cannot
identify non-SWCNT carbons. Samples must be in the solution phase or be in film form for
rapid analysis. The NIR technique must be normalized with a "best available" material standard;
other needs include a correction for metal, an absolute scale of purity, an extension to SWCNTs
of all diameters and production methods, as well as a correction for diameter distribution.

Raman spectroscopy was discussed as an important technique--an excellent quick test for the
presence of SWCNTs. Dresselhaus (MIT) pointed out that Raman is also valuable as a tool to
distinguish metallic-type nanotube samples from semiconducting-type nanotube samples, which
is critical for electronic applications. The interpretation of Raman spectra is complex, and not
uniformly applied. The minimum detectable level of nanotubes is one percent. It was pointed
out that Raman spectroscopy cannot be used to detect the presence of metallic impurities.

During the panel discussion, it was pointed out that magnetization measurements can also be
used to obtain quantitative metal analysis.

Dispersion
Dispersion can be defined as the distribution of nanotube bundles, the splitting of the bundles
into individual tubes, and the agglomeration of SWCNTs in solvents or polymers. A
fundamental question is that of dispersability, i.e., the degree and the ease of placing the
nanotubes in suspension. In macrodispersion, the focus is on eliminating agglomerates. In
nanodispersion, the focus is on eliminating SWCNT ropes. The question of dispersion stability
over time is also important.
The first step is to disperse the nanotube-containing sample in a solvent or surfactant. It was
decided that the community should agree upon a dispersive agent--a standard solvent or
surfactant to be used at the concentration 0.1 mg per ml, for the purpose of measurement
characterization only. It is understood that various applications will require various solvents.
Useful organic solvents include dimethyl formamide (DMF) (which, unfortunately, interacts with
samples); tetrahydrofuran (THF); and tetrachloroethylene (TCE). Organic solvents for the
dispersive medium are necessary for subsequent dispersion into organic polymer. It was
suggested that an impartial laboratory should develop a stable suspension in an appropriate
solvent or surfactant and then coordinate a round robin to characterize the sample.

All dispersion characterization methods have in common the critical need for careful sample
preparation. Many researchers use sonication, although some use mechanical stirring, because it
was determined that chemical reactions can occur during sonication. It was postulated that there
is a power floor above which sonochemistry occurs. All agreed that sonication depends on
ultrasonic frequency. Vessel specifications and displacement are also thought to be important.
Sonication is broadly used yet conditions vary considerably among researchers. Workshop
participants unanimously agreed there was an urgent need for sonication methods research that
will study the effects of time, frequency, power, and concentration of nanotubes in dispersion.

The technique of choice to determine the degree of macrodispersion appears to be optical
microscopy. However, there is lack of agreement on what constitutes good versus poor
dispersion. It was suggested that Raman mapping techniques and SEM were useful as
complements to optical microscopy.

In the case of nanodispersion, the best method seems to be UV/Vis Spectroscopy, for which
researchers did not agree on standard sample preparation. However, there was consensus that the
conditions for running the UV/Vis need to be standard and consistent and the method for
quantification should be standard and needs further development. Other techniques that may be
of value include light scattering, Raman, AFM, dynamic light scattering, and electrophoretic
spectrophotometry.

A method for determining dispersion stability was suggested by Nikolaev (JSC). There was a
general consensus that this method would be a useful starting point for developing a standard
protocol. Researchers were encouraged to use centrifugation to speed the settling and then
measure dispersion again. The measurements seem to be very dependent on the sample
preparation. It was agreed that further studies are needed on the effects of one, two, and four
hours of sonication.

SEM studies can also yield valuable information, but sample preparation is absolutely critical.
The two sample preparation techniques that were suggested were freeze-drying of liquids and
spin coating. Unfortunately, pullout of nanotubes often occurs during cryofracture. In spin
coating, researchers can make thicker coatings and selectively etch away the matrix to expose
particles, thereby enabling the examination of particle distribution. A choice must be made
between selective etches: either oxygen sources, which have the disadvantage of possibly
attacking nanotubes, or hydrogen sources, which can possibly attack the matrix. SEM has
limitations, particularly when the dispersion is so fine that small single bundles are created.
Atomic Force Microscopy (AFM) can eliminate some of the problems with artifacts that are
inherent in electron microscopy. Compared with SEM, AFM provides higher magnification and
better depth of field. Protocols are needed to standardize AFM methods. Key issues are the
development of methods by which to hold a single nanotube onto the surface, as well as methods
to eliminate charging. One limitation is that the AFM does not give a true representation of the
nanotube because the tip is so large in comparison with the nanotube. It is therefore necessary to
deconvolute the AFM tip shape to give an accurate representation of the tip.

It was noted that the nanotube community may be able to adapt some measurement methods
from the industries that produce carbon black and carbon fiber, where there exist very well
established characterization methods. In those industries, commercial standards are available to
measure the dispersion of particles of carbon down to 2 nm. There are also standards for TEM
imaging and sample preparation.

Summary
This workshop focused on present-day techniques for characterizing the purity and dispersion of
single wall carbon nanotubes. A number of the attendees commented that they appreciated the
opportunity to openly discuss their measurement concerns with their colleagues, and to learn
about different approaches to the various techniques. A questionnaire filled out by the attendees
confirmed the value of this workshop.

Participants agreed to support a special issue of the Journal of Nanoscience and Nanotechnology.
The issue will feature a workshop summary, multi-authored review papers on specific
techniques, and articles on research findings presented during the workshop. Possible topics are
TGA, TEM, SEM, EDS, optical microscopy, UV/IR, light scattering, SANS, surface
energy/area/tension, surfactants/dispersants, sonication, and magnetic techniques. In some
instances NIST Recommended Practice Guides will also be written in order to provide more
technical detail.

On the topic of purity, the strengths, limitations, and research needs for most of the commonly
used techniques, e.g., TGA, TEM, SEM, Raman, and NIR were discussed. It was agreed that
TEM observations are of primary importance. While reference materials will ultimately be
useful, it was noted that the nanotube community must first agree on measurement and
characterization methods. There was consensus that protocols for measurement techniques
would be valuable even if they were incomplete.

In the area of dispersion, the most prevalent need was for information on sonication methods,
e.g., power levels, frequency, and time. There was a suggestion that a standard dispersed liquid
and a standard dispersed solid be distributed to researchers willing to perform characterization so
that methods can be compared. Variables could include composition, particle size, and
dispersablity. An agreement on a solvent for dispersion is also needed. Optical microscopy was
conceded to be the primary technique for determining dispersion, but other techniques of value
include SEM, UV/Vis Spectroscopy, and the AFM.

Potential future workshop topics were discussed. These include: 1) diameter, chirality, length,
and types of nanotubes; 2) defects 3) surface chemistry and functionalization; 4) functional, e.g.
electronic properties; 5) applications and performance measures; and 6) health and safety issues.
There was consensus that identification of measurement procedures relative to nanotube
size and chirality would be particularly welcomed. NIST and NASA/JSC agreed to plan
for such a workshop.