Literature on Carbon Nanotube Research

Title: CNT Literature
	About: Moderator: [Markus Landgraf] Created: March 13th, 2009 Modified: April 3rd, 2009	Tags: This is a collaborative article Discipline(s): Wiki, Engineering, Chemistry

I have hijacked this page to write down my views on the literature on Carbon Nanotube (CNT) growths and processing, a procedure that should give us the cable/ribbon we desire for the space elevator.

If anyone has something to add, please do not hesitate!

Summary and Way Forward

For now I have finished the review of current literature on carbon nanotube (CNT) research, at least the part that is concerned with the creation of strong fibers/yarns for application in the space elevator. The following nine papers are interesting to read and give quite a good overview of the state of the art.

The state of the art as of early 2009 appears to be that it is impossible to catch the impressive specific tensile strength of sub-millimetre size single-wall nanotubes (SWNTs) for infinitely long wires or yarns. The current limited understanding of the CNT growth process and the inter-fiber forces in a spun yarn does not allow us to build a sufficiently strong wire for the space elevator from CNTs. The maximum reported breaking strength is in the order of 10 GPa for a 1 mm long spun yarn. However, in the respective paper by Koziol et al. (2007) it is very clear that this strength is lost when going to longer yarns. Realistically speaking, we are still at around 3 GPa of breaking strength.

On the positive side: there is progress in the understanding of the molecular dynamics of CNT growth in chemical vapour deposition (CVD) processes. In a recent paper actual observation of the growth of CNTs on a catalyst are presented.

One idea to produce the strong wires we need could be to optimise the manufacturing of the catalytic growth substrate in the CVD process, so that the CNTs grow infinitely long. This could be achieved by using the aluminium oxide buffer layer on top of the silicon base as it was done by G. Zhang et al. (2005). The catalyst should be applied on top of the buffer layer using lithographic techniques. The catalyst layout created by the lithographic process would have to be optimised in order to guarantee a long lifetime of the catalyst as well as good accessibility of the growth site (interface between catalyst and underlying layers) to the feedstock compounds.

Overall, this process will not be able to grow the space elevator wire in one piece, because the growth rate is about <math>100 \mu m min.^{-1}</math> or so, meaning for 100,000 km to grow we would need 2 million years! Now the yarn spinning explained by M. Zhang et al. (2004) comes into play. If we are able to transfer the full tensile strength from one fiber to another we would not need the full 100,000 km as a single fiber. I am quite optimistic that if we have 1 m-long CNT fibers, we can transfer the full tensile strength of the fiber to a neighbouring one. In this context "transfer the full tensile strength" means that in a pull test of a yarn spun of two fibers one of the fibers would break before they can be separated from each other. If the 1 m fibers are enough, we are in good shape as they take one week to grow in the CVD chamber at <math>100 \mu m min.^{-1}</math>.

Whether or not this proposed approach works or whether there are show-stoppers remains to be seen. I propose to contact the CNT industry in order to find out.

A new paper by Wang et al. (2009) discussed below describes a promising new development: CNTs 18.5cm in length! If spun to a yarn, perhaps the van der Waals forces between those long fibers can be strong enough to transmit the impressive mechanical properties of the fibers to the macroscopic yarn, which could make our ribbon.

Direct mechanical measurement of the tensile strength and elastic modulus of multiwalled carbon nanotubes

B. G. Demczyk et al., Materials and Engineering, A334, 173-178, 2002
The paper by Demczyk et al. (2002) is the basic reference for the experimental determination of the tensile strengths of individual multi-wall nanotube (MWNT) fibers. The experiments are performed with a microfabricated piezo-electric device. On this device CNTs in the length range of tens of microns are mounted. The tensile measurements are obseverd by transmission electron microscopy (TEM) and videotaped. Measurements of the tensile strength (tension vs. strain) were performed as well as Young modulus and bending stiffness. Breaking tension is reached for the SWNT at 150 GPa and between 3.5% and 5% of strain. During the measurements 'telescoping' extension of the MWNTs is observed, indicating that single-wall nanotubes (SWNT) could be even stronger. However, 150 GPa remains the value for the tensile strength that was experimentally observed for carbon nanotubes.

Direct Spinning of Carbon Nanotube Fibers from Chemical Vapour Deposition Synthesis

Y.-L. Li, I. A. Kinloch, and A. H. Windle, Science, 304,276-278, 2004
The work described in the paper by Y.-L. Li et al. is a follow-on of the famous paper by Zhu et al. (2002), which was cited extensively in Brad's book. This article goes a little more into the details of the process. If you use a mixture of ethene (as the source of carbon), ferrocene, and theophene (both as catalysts, I suppose) into a furnace (1050 to 1200 deg C) using hydrogen as carrier gas, you apparently get an 'aerogel' or 'elastic smoke' forming in the furnace cavity, which comprises the CNTs. Here's an interesting excerpt:

Under these synthesis conditions, the nanotubes in the hot zone formed an aerogel, which appeared rather like “elastic smoke,” because there was sufficient association between the nanotubes to give some degree of mechanical integrity. The aerogel, viewed with a mirror placed at the bottom of the furnace, appeared very soon after the introduction of the precursors (Fig. 2). It was then stretched by the gas flow into the form of a sock, elongating downwards along the furnace axis. The sock did not attach to the furnace walls in the hot zone, which accordingly remained clean throughout the process.... The aerogel could be continuously drawn from the hot zone by winding it onto a rotating rod. In this way, the material was concentrated near the furnace axis and kept clear of the cooler furnace walls,...

The elasticity of the aerogel is interpreted to come from the forces between the individual CNTs. The authors describe the procedure to extract the aerogel and start spinning a yarn from it as it is continuously drawn out of the furnace. In terms of mechanical properties of the produced yarns, the authors found a wide range from 0.05 to 0.5 GPa/g/ccm. That's still not enough for the SE, but the process appears to be interesting as it allows to draw the yarn directly from the reaction chamber without mechanical contact and secondary processing, which could affect purity and alignment. Also, a discussion of the roles of the catalysts as well as hydrogen and oxygen is given, which can be compared to the discussion in G. Zhang, et al. (2005, see below).

Multifunctional Carbon Nanotube Yarns by Downsizing an Ancient Technology

M. Zhang, K. R. Atkinson, and R. H. Baughman, Science, 306, 1358-1361, 2004
In the research article by M. Zhang et al. (2004) the procedure of spinning long yarns from forests of MWNTs is described in detail. The maximum breaking strength achieved is only 0.46 GPa based on the 30micron-long CNTs. The initial CNT forest is grown by chemical vapour deposition (CNT) on a catalytic substrate, as usual. A very intersting formula for the tensile strength of a yarn relative to the tensile strength of the fibers (in our case the MWNTs) is given:

<math> \frac{\sigma_{\rm yarn}}{\sigma_{\rm fiber}} = \cos^2 \alpha \left(1 - \frac{k}{\sin \alpha} \right) </math>

where <math>\alpha</math> is the helix angle of the spun yarn, i.e. fiber direction relative to yarn axis. The constant <math>k=\sqrt(dQ/\mu)/3L</math> is given by the fiber diameter d=1nm, the fiber migration length Q (distance along the yarn over which a fiber shifts from the yarn surface to the deep interior and back again), the quantity <math>\mu=0.13</math> is the friction coefficient of CNTs (the friction coefficent is the ratio of maximum along-fiber force divided by lateral force pressing the fibers together), <math>L=30{\rm \mu m}</math> is the fiber length. A critical review of this formula is given here.

In the paper interesting transmission electron microscope (TEM) pictures are shown, which give insight into how the yarn is assembled from the CNT forest. The authors describe other characteristics of the yarn, like how knots can be introduced and how the yarn performs when knitted, apparently in preparation for application in the textile industry.

Ultra-high-yield growth of vertical single-walled carbon nanotubes - Hidden roles of hydrogen and oxygen

G. Zhang et al.,PNAS, 102, no. 45, 16141-16145, 2005

Important aspects of the production of CNTs that are suitable for the SE is the efficiency of the growth and the purity (i.e. lack of embedded amorphous carbon and imperfections in the Carbon bounds in the CNT walls). In their article G. Zhang et al. go into detail about the roles of oxygen and hydrogen during the chemical vapour deposition (CVD) growth of CNT forests from hydrocarbon sources on catalytic substrates. In earlier publications the role of oxygen was believed to be to remove amorphous carbon by oxidation into CO. The authors show, however, that, at least for this CNT growth technique, oxygen is important, because it removes hydrogen from the reaction. Hydrogen has apparently a very detrimental effect on the growth of CNTs, it even destroys existing CNTs as shown in the paper. Since hydrogen radicals are released during the dissociation of the hydrocarbon source compount, it is important to have a removal mechanism. Oxygen provides this mechanism, because its chemical affinity towards hydrogen is bigger than towards carbon.

In summary, if you want to efficiently grow pure CNT forests on a catalyst substrate from a hydrocarbon CVD reaction, you need a few percent oxygen in the source gas mixture. An additional interesting information in the paper is that you can design the places on the substrate, on which CNTs grow by placing the the catalyst only in certain areas of the substrate using lithography. In this way you can grow grids and ribbons. Figures are shown in the paper.

In the paper no information is given on the reason why the CNT growth stops at some point. The growth rate is given with 1 micron per minute. Of course for us it would be interesting to eliminate the mechanism that stops the growth so we could grow infinitely long CNTs.

This article can be found in our archive.

Sustained Growth of Ultralong Carbon Nanotube Arrays for Fiber Spinning

Q. Li et al., Advanced Materials, 18,3160-3163,2006

Q. Li et al. have published a paper on a subject that is very close to our hearts: growing long CNTs. The longer the fibers, which we hope have a couple of 100 GPa of tensile strength, can hopefully be spun into the yarns that will make our SE ribbon. In the paper the method of chemical vapour deposition (CVD) onto a catalyst-covered silicon substrate is described, which appears to be the leading method in the publications after 2004. This way a CNT "forest" is grown on top of the catalyst particles. The goal of the authors was to grow CNTs that are as long as possible. They found that the growth was terminated in earlier attempts by the iron catalyst particles interdiffusing with the substrate. This can apparently be avoided by putting an aluminium oxide layer of 10 nm thickness between the catalyst and the substrate. With this method the CNTs grow to an impressive 4.7 mm! Also, in a range from 0.5 to 1.5 mm fiber length the forests grown with this method can be spun into yarns.

The growth rate with this method was initially <math>60{\rm \mu m\ min.^{-1}}</math> and could be sustained for 90 min. This is very different from the <math>1{\rm \mu m\ min.^{-1}}</math> reported by G. Zhang, et al. (2005), which shows that the growth is very dependent on the method and materials used. The growth was prolonged by the introduction of water vapour into the mixture, which achieved the 4.7 mm after 2 h of growth. By introducing periods of restricted carbon supply, the authors produced CNT forests with growth marks. This allowed to determine that the forest grew from the base. This is in line with the in situ observations by S. Hofmann, et al. (2007).

Overall the paper is somewhat short on the details of the process, but the results are very interesting. Perhaps the 5 mm CNTs are long enough to be spun into a usable yarn.

In situ Observations of Catalyst Dynamics during Surface-Bound Carbon Nanotube Nucleation

S. Hofmann et al., Nano Letters, 7, no. 3, 602-608, 2007.

The paper by S. Hofmann et al. (2007) is a key publication for understanding the microscropic processes of growing CNTs. The authors describe an experiment in which they observe in situ the growth of CNTs from chemical vapour deposition (CVD) onto metallic catalyst particles. The observations are made in time-lapse transmission electron microscopy (TEM) and in x-ray photo-electron spectroscopy. Since I am not an expert on spectroscopy, I stick to the images and movies produced by the time-lapse TEM. In the observations it can be seen that the catalysts are covered by a graphite sheet, which forms the initial cap of the CNT. The formation of that cap apparently deforms the catalyst particle due to its inherent shape as it tries to form a minimum-energy configuration. Since the graphite sheet does not extend under the catalyst particle, which is prevented by the catalyst sitting on the silicon substrate, the graphite sheet cannot close itself. The deformation of the catalyst due to the cap forming leads to a retoring force exerted by the crystaline stracture of the catalyst particle. As a consequence the carbon cap lifts off the catalyst particle. On the base of the catalyst particle more carbon atoms attach to the initial cap starting the formation of the tube. The process continues to grow a CNT as long as there is enough carbon supply to the base of the catalyst particle and as long as the particle cannot be enclosed by the carbon compounds. During the growth of the CNT the catalyst particle breathes so drives so the growth process mechanically.

Of course for us SE community the most interesting part in this paper is the question: can we grow CNTs that are long enough so we can spin them in a yarn that would hold the 100GPa/g/ccm? In this regard the question is about the termination mechanism of the growth. The authors point to a very important player in CNT growth: the catalyst. If we can make a catalyst that does not break off from its substrate and does not wear off, the growth could be sustained as long as the catalyst/substrate interface is accessible to enough carbon from the feedstock.

If you are interested, get the paper from our archive, including the supporting material, in which you'll find the movies of the CNTs growing.

High-Performance Carbon Nanotube Fiber

K. Koziol et al., Science, 318, 1892, 2007.
The paper "High-Performance Carbon Nanotube Fiber" by K. Koziol et al. is a research paper on the production of macroscopic fibers out of an aerogel (low-density, porous, solid material) of SWNT and MWNT that has been formed by carbon vapor deposition. They present an analysis of the mechanical performance figures (tensile strength and stiffness) of their samples. The samples are fibers of 1, 2, and 20 mm length and have been extracted from the aerogel with high winding rates (20 metres per minute). Indeed higher winding rates appear to be desirable, but the authors have not been able to achieve higher values as the limit of extraction speed from the aerogel was reached, and higher speeds led to breakage of the aerogel.

They show in their results plot (Figure 3A) that typically the fibers split in two performance classes: low-performance fibers with a few GPa and high-performance fibers with around 6.5 GPa. It should be noted that all tensile strengths are given in the paper as GPa/SG, where SG is the specific gravity, which is the density of the material divided by the density of water. Normally SG was around 1 for most samples discussed in the paper. The two performance classes have been interpreted by the authors as the typical result of the process of producing high-strength fibers: since fibers break at the weakest point, you will find some fibers in the sample, which have no weak point, and some, which have one or more, provided the length of the fibers is in the order of the frequency of occurrence of weak points. This can be seen by the fact that for the 20 mm fibers there are no high-performance fibers left, as the likelihood to encounter a weak point on a 20 mm long fiber is 20 times higher than encountering one on a 1 mm long fiber.

As a conclusion the paper is bad news for the SE, since the difficulty of producing a flawless composite with a length of 100,000 km and a tensile strength of better than 3 GPa using the proposed method is enormous. This comes back to the ribbon design proposed on the Wiki: using just cm-long fibers and interconnect them with load-bearing structures (perhaps also CNT threads). Now we have shifted the problem from finding a strong enough material to finding a process that produces the required interwoven ribbon. In my opinion the race to come up with a fiber of better than Kevlar is still open.

Tensile and Electrical Properties of Carbon Nanotube Yarns and Knitted Tubes in Pure or Composite Form

S. Hutton, C. Skourtis, and K. Atkinson, Int. J. Technology Transfer and Commercialization, 7, no. 2/3, 258-264, 2008

The paper by S. Hutton, et al, is the latest on yarns spun out of CNTs. The core of the paper is concerned with the effect the different amounts of twist has on the tensile strength and on the electrical conductivity of the yarn. The bad news for us is that they arrive only at 1 GPa/ccm for the optimum tensile strength of the yarn. However, some insight is given in the spinning process and in the different methods of processing the CNTs. They use relatively short CNTs (0.2 to 0.3 mm) grown into a MWNT forest by chemical vapour deposition (CVD) onto a silicon substrate covered with metal catalyst. The latter method appears to have become standard recently.

Strong and Ductile Colossal Carbon Tubes with Walls of Rectangular Macropores

H. Peng et al., Phys. Rev. Lett., 101,145501,2008

This paper does not actually fit here, because it is not about CNTs. This is about "collosal Carbon Tubes" (CCTs), which are rolled-up sandwiches of porous sheets of amorpohous carbon. I came across the paper from the article on the elevator on the main wiki. The two obvious questions are: do CCTs actually exist, and what can CCTs do for the space elevator?

Let me start by summarising the paper: In a process that is identical to the growth process of CNTs, except that no catalysts are used, apparently CCTs grow in the furnace. They are imaged by transmission electron microscopy (TEM) and measured: <math>50 \mu m</math> in diameter and half a centimetre long. In the paper it is stated that it is still unclear how the CCTs form. It is speculated that a graphite sheet with embedded rectangular macropores grows in the chemical vapour deposition (CVD) process. The two sides of the sheet grow at different rates, which curls it up to form the tube.

Now the interesting part comes: while the CCTs are not much stronger than regular carbon composite wires (6.9 GPa breaking strength), they are much lighter. The density of the carrying structure (the wall of the CCTs) is given in the paper with 0.116 g/ccm. This might be not so imortant for applications like bullet-proof vests, but is critical to the SE. We get a specific tensile strength of 59 GPa/g/ccm, enough for the SE!

The open questions are:

• Do CCTs really exist?
• Can they be produced in quantity?
• Can the produced as a infinitely long wire?
• If not, how much is the specific breaking strength reduced if a yarn is spun out of them?

Pressure-induced Interlinking of Carbon Nanotubes

T. Yildirim and O. Gulseren, Pressure-Induced Interlinking of Carbon Nanotubes, NIST Center for Neutron Research, National Institure of Standards and Technology, Gaithersburg, MD 20899-8562

This offers a new way for nanotubes to be linked to each other laterally. Previously, they were held to each other by Van der Waals forces, much like static cling. This research put various type of CNTs under lateral pressure of many GPas, which caused deformation, then for the tubes to bond to each other. This has the potential to be stronger than VdW forces, as well as a way to bundle CNTs other than weaving and tape.

Fabrication of Ultralong and Electrically Uniform Single-Walled Carbon Nanotubes on Clean Substrates

Xueshen Wang, Qunqing Li, Jing Xie, Zhong Jin, Jinyong Wang, Yan Li, Kaili Jiang, and Shoushan Fan, Fabrication of Ultralong and Electrically Uniform Single-Walled Carbon Nanotubes on Clean Substrates, Nano Lett., 2009, 9 (9), pp 3137–3141

This paper shows how to produce really long CNTs using CVD. The maximum reached length was 18.5cm. This is two orders of magnitude longer than previous work. The trick appears to be to place the catalysts on a thin CNT film and not on an inert substrate like Aluminium oxide or Silicates. In the paper only electrical properties are discussed with an application in micro electronics (FET devices). Mechanical properties that could be interesting for the SE are not discussed in the paper. It is however mentioned that the ultra long CNTs are capable to bridge gaps between substrate plates. They do, however break if the different substrate plates are moved. This shows that, as expected individual CNTs are not very strong due to ther small diameter (few nanometres). Thos means, in order to use this new finding for strong ropes, ribbons, or yarns. An efficient method for post processing must me defined. For this the treatment with acetone and subsequent spinning is quite conceivable and I would be very interested in results from such experiments. We can expect that the strength of yarn spun from CNTs increases in strength with the length of the individual fibers as indicated by the formula given above. Can we get the required ~50 GPa/g/ccm out of those ultra-long CNTs? We'll see.

Creation of nanotube yarn

Scientists have spun carbon nanotube threads on industrial scale. The Register reported on 14th January 2013 that an international team of scientists has successfully found a way to spin tens of millions of carbon nanotubes into a flexible conductive thread that's a quarter of the thickness of human hair. The thread has ten times the tensile strength of steel and is as conductive as copper, but is flexible enough to be wound around a spool or woven. The team envisages it being used in "smart" clothing and the aerospace industry, and says that its properties will be of particular use to electronics manufacturers.

Weakening of CNT by Statistical Defects

N. M. Pugno, The role of defects in the design of space elevator cable: From nanotube to megatube, Acta Materialia, 55 (2007) 5269-5279

It is reported by N.M. Pugno in Acta Materialia that the frequency of defects in the modelcular structure of CNT will be too large for a cable of the required length. This would lead to a weakening of the cable, such that the elevator cable would no longer be feasible. While his model and conclusions appear to be correct, it is possible that he starts from an incorrect assumption: that the cable must be made of a continuous CNT fibre. If the CNTs can be made at sufficient strengths long enough (could be in the order of 1 metre) it is conceivable that other forces, like van-der-Waals forces between the mesoscopic fibers of a spun cable are strong enough to transmit the tension.