Applications and Mass Fabrication of Carbon Nanotubes€¦ · I.5. Carbon Fibers Carbon fibers are closely linked to carbon nanotubes and are already used in commercial applications

Applications and Mass Fabrication of Carbon Nanotubes

Ivar Dijck - S2306301

Rijksuniversiteit Groningen, Maria Antonietta Loi

March, 2016

Contents

I Introduction 1I.1 Single versus multiwalled . . . . . . . 2I.2 Structural Properties . . . . . . . . . . 2I.3 Electrical Properties . . . . . . . . . . . 3I.4 Thermal Properties . . . . . . . . . . . 3I.5 Carbon Fibers . . . . . . . . . . . . . . 3I.6 Health Risks . . . . . . . . . . . . . . . 4

II Applications 4II.1 Transistors . . . . . . . . . . . . . . . . 4II.2 Conductor . . . . . . . . . . . . . . . . 5II.3 Composites . . . . . . . . . . . . . . . 5II.4 Aerogels . . . . . . . . . . . . . . . . . 6

III Production 6III.1 Aggregates versus Aligned . . . . . . 6III.2 Laser Ablation . . . . . . . . . . . . . . 6III.3 Arc Discharge . . . . . . . . . . . . . . 7III.4 Chemical Vapour Deposition . . . . . 7III.5 Catalysts . . . . . . . . . . . . . . . . . 8III.6 Fixed Bed Reactor . . . . . . . . . . . . 8III.7 Fluidized bed Reactor . . . . . . . . . 8III.8 HiPCo . . . . . . . . . . . . . . . . . . 8

IV Discussion 9IV.1 Research . . . . . . . . . . . . . . . . . 9IV.2 Microscopic versus Macroscopic . . . 10IV.3 Costs . . . . . . . . . . . . . . . . . . . 10IV.4 Top Down Approach . . . . . . . . . . 10IV.5 Recipes . . . . . . . . . . . . . . . . . . 10IV.6 Health . . . . . . . . . . . . . . . . . . 11

V Conclusion 11

VI Acknowledgments 11

Abstract

Carbon nanotubes promise to be a powerful mate-rial for many applications. Especially structural isinteresting, but also electrical properties of carbonnanotubes (CNTs) are unique. This paper will lookcritically at the synthesis of CNTs and some of theirpossible applications commercially. CNTs can beused for their exceptional structural, thermal andelectrical properties. They are already in use on asmall scale for high-end applications where cost isless of a concern. Carbon nanotubes are anotherform of carbon together with better known formslike carbon fibers and graphite. The structure of thenanotubes gives it very different material propertiesthan other carbon forms. To get these benefits ineveryday applications the cost of high quality CNTshas to go down a lot. To this end the only realcontender is production based on chemical vapourdeposition (CVD). As opposed to laser ablation andthe arc discharge method, chemical vapour depo-sition scales easily. Moreover CVD can not onlyproduce agglomerated multiwalled nanotubes but,depending on the exact geometry and catalyst used,can even make single walled aligned CNTs in ascaleable fashion.

I. Introduction

First an overview of the properties of CNTs is given,that is what’s important after all. Before discussingsome of the production methods we will look at themost prominent applications of CNTs currently.

Carbon is easily the most versatile element inchemistry. In recent years more allotropes of carbonhave been found apart from diamond and graphite.Due to its valency, carbon can form many allotropes,including so called carbon nanotubes.[2] CNTs arehollow cylinders made out of solely carbon atoms.[2]These carbon atoms bind exclusively by sp2 bonds tocreate a honeycomb-like structure.[1] The sp2 bondsin graphene are the strongest bonds in chemistry,but individual sheets have little interaction betweenthem making the material weak as a whole.[7] Car-bon nanotubes do not suffer from this and individ-

1

ual CNTs can have very strong van der waals inter-actions between them.[7] CNTs embody some of thegreatest promises in nanotechnology; stronger thanany material, best thermal conductor, best electricalconductor and more.

The carbon nanotube structure can have extremeaspect ratio’s, meaning it can be much longer thanit is wide. CNTs on the order of tens of centimetershave been confirmed, with aspect ratios of 108.[4]This is akin to a 1000 km long fiber optic cable.[3]CNTs are therefor strongly anisotropic, propertiesalong the tube (axial) can be vastly different than inthe radial direction.[6]

I.1. Single versus multiwalled

There can be differences between individual CNTssuch as diameter and length. Carbon nanotubes areoften further subdivided into single walled carbonnanotubes (SWCNT) and multi walled carbon nan-otubes (MWCNT). SWCNTs are usually about onenanometer in diameter and can be either conduct-ing or semi-conducting, depending on their chiralityand size.[1] For this reason SWCNTs are interest-ing in electrical applications. SWCNTs are oftenmore difficult to fabricate than MWCNTs as the con-ditions have to be better controlled. Multi walledcarbon nanotubes are obtained from wrapping mul-tiple layers of carbon atoms on top of one another.This can be done either with concentric larger car-bon tubes, or in rarer cases like a rolled up papersheet as one continuous layer.[12] MWCNTs are of-ten stronger and metallic compared to single SWC-NTs. The amount of layers a MWCNT can haveis essentially limited by fabrication only and verylarge diameters compared to SWCNT can be made.For most applications the SWCNTs are consideredbetter since their properties are well defined (anddependant on chirality[1]). Per nanotube MWCNTsare usually stronger but not per weight, howevergiven the low weight in general this is not such aproblem.

I.2. Structural Properties

The main interest into CNTs are their unusual prop-erties compared to other materials. To build any-thing of consequence will mean large volumes of

Figure 1: The different dimensions for the different kinds ofcarbon nanotubes. The distance between layers in aWMCNT are always very close to 0.36 nm. [13]

CNT material, this is then the most important whenlooking at industrial scale production of CNTs.

To quantify strength there are several constantsto denote the properties of materials, most oftenin Young’s modulus and ultimate tensile strength.These relate how a material reacts when beingpulled on at both ends, called tensile strain. For com-pressive strain there are similar numbers, howeverthere have been very few studies to verify the com-pressive strength of CNTs. Since material strengthgoes with the size this is denoted in Pascal (Newtonper area). Young’s modulus or elastic modulus is theslope of force versus deformation curve for elasticdeformations. This means high modulus materialsdon’t deform much with force, while they keep theirintegrity. Tensile strength is how much force mustbe applied before the cable will break. For the val-ues given below you could hang a 15000 kg weighton a 1 mm2 carbon nanotube cable while it onlyelongates by 16% (in theory). Clearly the structuralproperties of carbon nanotubes are vastly superiorto either steel or Kevlar:[6]

2

Material Young’s Modulus Tensile StrengthMWCNT 900 GPa 150 GPaSteel 200 GPa 1.5 GPaKevlar 150 GPa 3.7 GPaCarbon fiber 180 GPa 4.0 GPa

Table 1: For the MWCNTs these values were measured usingonly a single perfect nanotube, in the axial direction.Theoretical models show that this is only slightlyless than the the upper limit for CNTs as it shouldbe.[7] CNTs are much less stiff in the radial direction,especially SWCNTs. The values for steel and Kevlarare for the highest grade materials available.

Also carbon fibers are mentioned, carbon fiberis a composite made from carbon fiber are epoxy(see below). The high values for CNTs are quiteamazing in itself but coupled with the low atomicweight of carbon this means the specific propertiesof CNTs are (by relative standards) even better, over300 times that of steel.[6] Under excessive tensionCNTs will undergo plastic deformation instead ofelastic like any other material. Plastic deformation isthe kind that is not reversible when the load is takenoff, as opposed to elastic deformations which are re-versible. Materials that only or mostly undergo elas-tic deformation are normally preferred. For CNTsthis plastic deformation is thought to be mediatedby the so called Stones-Wales deformation; here 4hexagons deform into 2 pentagons and 2 heptagons,releasing some tensile strain.[12] Such deformationsare not before 50 GPa however.[12] Naturally un-der more extreme loads higher order deformationswill occur too.[12] This very high strength is themain reason for using carbon nanotubes as a newstructural material.

Figure 2: Stone-Wales deformations are involved in the plas-tic deformation of CNTs. After such a deformationthe lattice has more freedom.[12]

I.3. Electrical Properties

Not only structural but also electrical properties ofCNTs are unlike any other material. Others have de-scribed the electrical properties of carbon nanotubesin depth, here we only look at it only superficially.

The carbon tubes can carry current incredibly well,some studies reported as high as 109 A/cm2, higherthan most superconductors.[8] CNTs do not exhibitthe Meissner effect however, and are not supercon-ductors in name. Small scale CNTs are good con-ductors because of ballistic transport.[8] Ballistictransport is when the mean free path for electronsis comparable to the dimensions of the medium.Meaning on average it will simply follow a ballisticpath from one end to the other. The very few scatter-ing centers in the hollow carbon nanotubes meansit has excellent electrical conductivity in the axialdirection.

I.4. Thermal Properties

Added to the special electrical and structural prop-erties is also excellent thermal conductivity. Thethermal properties stem from phonons, for muchthe same reason in regard electrons and ballistictransport.[9] For a single carbon nanotube exper-imental values as high as 3500 W/mK have beenreported, almost ten times that of copper.[9] Suchhigh thermal conductivity means that temperaturedifferences will equilibrate very quickly and unequalthermal expansion is not a big concern. Interestinglyin the radial direction it has only 1.5 W/mK, ratherpoor conductivity.[9] In a bit of irony we will see thatone of the applications mentioned (aerogels) are infact thermal insulators, not conductors. This seem-ing contradiction is mostly due to simply the verylow conducting volume and density in aerogels.

I.5. Carbon Fibers

Carbon fibers are closely linked to carbon nanotubesand are already used in commercial applicationssuch as in high-end aerospace engineering solu-tions. Such products could also be made with thesuperior carbon nanotube versions to reduce weightand increase strength. Carbon fiber is a compositematerial, this means that it is made out of a ma-trix, usually epoxy materials, and the carbon fibers

3

themselves.[10] The matrix acts as the glue to holdthe fibers together, while the fiber provides the ac-tual strength of the whole material.[10] CNTs can beadded to a matrix material to make composites outof. Some of this has already been done, where CNTswere added to the epoxy matrix of carbon fibers.[14]Although still prohibitively expensive and relativelylittle benefits yet. Pure CNTs based composites havenot yet been used in any commercial application.

Figure 3: Shows how CNTs can be added to the epoxy matrixof already existing carbon fiber composites for addedstrength. This technology was used to make a lighterbicycle that went on to win the Tour de France.[14]

I.6. Health Risks

One can’t talk about CNTs without mentioning atleast some of the risk involved. One of the biggestproblems with carbon nanotubes is the possible dan-ger to human health. Just this possibility is alreadycumbersome as research with dry carbon nanotubeshas to be performed in controlled areas.

The carbon nanotubes are somewhat similar toother high aspect ratio fibers such as asbestos, whichare known to cause cancer. Research has shown thatCNTs can also be the cause of cancer through thesame mechanism of strained phagocytosis.[11] Thiscan be especially scary if CNTs are airborne and

breathed in where the white blood cells in the lungstry to break down the CNTs.[11] However luckilyCNTs strongly tend to aggregate, this clumping canmake particles much larger than cells and thus nostrained phagocytosis occurs.[24] This is still a riskof course, statistics dictate some percent will enterthe lungs as a single tube. Free multi walled nan-otubes have been found to be more carcinogenicthan asbestos fibers.[11] Asbestos has been bannedfor its danger, the hope is such drastic measures donot have to be implemented for CNTs as well.

II. Applications

Here we will describe only the most prominent ap-plications for CNTs. With all these apparent amaz-ing properties for pristine CNTs there are a plethoraof possible uses for it. Trying to sum up all possibil-ities would be a waste of time.

Most of these different applications have thus faronly seen limited research however and it remainsto be seen if most of these will become actual com-mercial products. Furthermore with other advancesin the field of nanotechnology and science in generalit’s very difficult to know if the CNTs will pan out inthe end as the best solution. Especially in electronicdevices, such as transistors, there is a lot of parallelresearch. Besides the applications below there areeven more possible uses for CNTs, including butnot limited to: hydrogen storage, terahertz polar-ization, near-ideal black body absorption, artificialmuscles.[23] In a bit of irony, research is also donefor using CNTs against cancer, because of its strongabsorption in the infrared part of the spectrum.[22]

II.1. Transistors

Transistors are the building blocks for all micro elec-tronics. Transistors effectively are small switches,either letting current through or not. To make agood transistor you need a semi-conductor material,for carbon nanotubes this means chiral SWCNTs.[1]Single transistors made with SWCNTs easily outper-form any silicon based transistor to date because ofhigher electron mobility, high current density andsmaller size.[1] However to create fully functioningintegrated circuits a much higher level of controlover the chirality and deposition of individual SWC-

4

NTs is needed than has been attained thus far. Oneupshot here is that in weight only a small amountof CNTs are needed. In the most ideal case of onlyone SWCNT per transistor, even with a trillion tran-sistors per square centimeter, this is still only onthe order of a microgram per CPU. Thus for thisapplication no mass production would be needed.

Single walled carbon nanotubes have also beenproposed for usage in new organic solar panels.Here it can have more than one function: simplyas a good flexible conductor, or even as part of theoptically active component.[23]

II.2. Conductor

Another promising electrical application for CNTs issimply as conductor. There are hundreds of kilome-ters of electrical wires everywhere. Potentially thiscould be replaced by wire made from carbon nan-otubes. For this application large scale productionwould be needed. Especially the cost per kilome-ter is important for high-voltage power cables. Forsmall scale and low-voltage the advantages of usinga CNT wire over a copper wire are much smaller.

Bulk metallic SWCNT can have electrical conduc-tivity superior to copper with orders of magnitudehigher specific current carrying capacity.[4] Currenthigh-voltage power cables have diameters of 20 cmor so, these heavy cables could be replaced by pa-per thin (0.1 mm) CNT cables. Currently there arealready efforts for a copper-carbon-nanotube alloy,although being an alloy it will never give the bigbenefits of pure CNT cables.[15] Bulk pure singlewalled carbon nanotubes cables with electrical prop-erties akin to the values of single SWCNTs havenot been made to date. As a conductor it will alsobe an excellent electromagnetic (EM) shielding ma-terial, especially where weight and thickness is aconcern.[23] This has already been used for EM-shielding in satellites.[23]

II.3. Composites

The main use for carbon nanotubes will likely be incomposites (also see 3). Composites form a largefamily of materials, most of which are fairly new.For this reason the possibilities are very excitingin this area of research, even more so with the in-

Figure 4: The CNT array is a mat of aligned carbon nan-otubes that were made by CVD process. The matis simply drawn and a super thin sheet is pulledoff and rolled. In this way sheets and yarns can bemade. The deposition state of the CNT array is char-acterized by so called "forest-like growth".[19][25]

troduction of trying to combine already existingcomposites with CNTs.

Already CNTs composites have been used withcarbon fibers for the strongest lightest material inthe world.[14] Given the price however it will stayfor use only in high-end applications for the fore-seeable future, just like carbon fiber itself (still isat least). Currently airplanes and space rockets areslowly switching from aluminum to carbon fibercomposites (The largest single composite is the Vegafirst stage).[33] Mass is one of the most importantcharacteristics for these craft, material costs are of-ten secondary. Especially in these aerospace appli-cations one would expect to see CNT compositesincorporated in the near future. This will meana large increase in producing CNTs, on the scalecomparable to that of carbon fiber composites now.Structurally speaking CNT enhanced carbon fibercomposites are some of the toughest materials wehave, but singular MWCNTs have very much bet-ter mechanical properties than any bulk materialproduced thus far. Carbon fiber composite are lessstrong than individual carbon fibers.[19] The lossesin material properties from pure MWCNTs to a bulkcomposite with epoxy as matrix materials remainsto be seen.

5

Instead of making composites purely from CNTs,often just a few percent CNTs added to existing ma-terials already gives a significant boost in structure.One area where CNTs might be considered is in gasturbines. Gas turbines power much of our modernworld and the blades have to be made from "superal-loys" operating at extreme temperature and tension,perfect for CNT materials.[32] The efficiency of thesegas turbines (as well as most jet engines) are limitedby the rotational speed and temperature at the tipof the blades.[32]

II.4. Aerogels

Aerogels have been made out of carbon nanotubes.These are the lightest solids in existence, muchlighter than even air.[15] However since it is com-pletely porous it is normally filled with air and doesnot float. These carbon aerogels are like a foam,as opposed to some other (silica) aerogels that aremore like brittle bricks. Cool as it may be, therehaven’t actually been many real world uses for theseaerogels yet. Given their porousness they are verygood thermal insulators and have been implied asmaterial for heat shields and the like.

III. Production

The Production of the carbon nanotubes will natu-rally be one of the most important things for the fu-ture of CNTs. The focus will lie on chemical vapourdeposition as it is simply much more promisingthan other alternatives.

For most materials the production costs determinehow much it is used in practice. Given that the rawmaterial cost of carbon is extremely low, the ease ofproduction and thus the method is easily the mostimportant for the widespread adoption of CNTs.This is often true for not just the adoption in realworld applications, but also concerning research.Cheap materials are simply easier to research. Ofcourse for research the quality is very important andmust be strictly controlled, while less so for mostapplications. This means that the scientific methodsfor producing CNTs don’t line up perfectly with thecommercial methods. It is clear that for any commer-cial application the CVD approach will be the bestin cost and production, since it is scaleable. While

especially in research the arc discharge method isstill used a lot, and to a lesser extend laser ablationas mentioned below.

III.1. Aggregates versus Aligned

There are a variety of ways for making carbon nan-otubes of various quality. Apart from differencesin making single walled versus multiwalled there’salso the issue of making aggregates as opposed toaligned nanotubes. Sometimes aggregates versusaligned CNTs are referred to as cooked and un-cooked spaghetti. Since CNTs are much weaker incompressive stress than tensile stress the aggregatesfall short of the aligned CNTs in strength. Given acertain uniform catalyst density and temperature,the forming tubes will coalesce and self-assembleinto the aligned state.[20] Since we usually want thecarbon nanotubes to align, this puts some restric-tions on the geometries that can be used.

Figure 5: Shows the microscopic difference of aggregates ofCNTs against aligned CNTs. Getting the alignedstate was done for (b) by having a uniform spread ofiron particle catalyst instead of randomly sputteredfor (a).[19]

III.2. Laser Ablation

Laser ablation was one of the first successful meth-ods of making CNTs. Because the parameters forlaser ablation are controlled very well this gives agood reproduceability for separate batches. Becauseof the physical nature of this method the CNTs pro-duced are almost purely single walled CNTs withsmall (1-1.6 nm) diameters.[18] Rather little progressin the method has been made since it was first dis-covered, and is only used for high quality lab CNTs.

6

Figure 6: The reaction chamber is heated and an inert gas flowis added to carry the particles. The graphite targetalso contains some catalyst (cobalt and nickel). TheCNTs form on the water cooled collector.[17]

III.3. Arc Discharge

The arc discharge method for making CNTs worksby passing a high current through a pair of carbonelectrodes that evaporate and deposit in the reac-tion chamber.[30] The reaction chamber is evacuatedand usually a catalyst gas such as iron or cobalt isadded.[30] Because of the simplicity, and the muchhigher efficiency compare to laser ablation this isused quite often for labs still. Concerning commer-cial production this still doesn’t produce nearly asmuch as CVD though.

Figure 7: In arc discharge deposition the carbon nanotubesform on the cathode and the space is filled with aninert gas as well as some (iron, nickel or cobalt)catalyst.[30]

III.4. Chemical Vapour Deposition

Chemical vapour deposition is the most promisingmethod for commercial production. Here we de-scribe 3 distinct different methods within CVD: fixedbed, fluidized bed, and floating catalyst (HiPCo).CVD is a broad term and the individual executionscan vary a lot.

Chemical vapour deposition is a process of lettinga gas undergo a reaction of some sort and deposit asa solid. This is done inside a furnace with catalystparticles to actually grow the CNTs. There are manyvariations on CVD of different kinds of geometryand reactions that are all slightly different.[17] Thevarious CVD techniques are by far the most promis-ing in making large amounts of CNTs, both singlewalled and multi walled. This is mostly becauseit is scaleable, the reaction chamber can simply bemade as large as needed with essentially the sameefficiency. This makes it a very powerful manufactur-ing technique and has long been used in the siliconand other thin films industry.

In the simplest setup (8) a hydrocarbon gas (usu-ally acetylene) is thermally decomposed near thecatalyst metal nano particles that reside there onthe substrate.[17] These nano particles then absorbthe freed carbon atoms which form the nanotube.The gas that didn’t react can be recycled, but forlab conditions is often contaminated too much. Thecatalyst is of course not directly consumed, but it isslowly passivated by CNTs and other processes.[17]The CNTs that are scraped off the walls have to bechemically treated to get rid of this excess catalystfor purer CNTs.

Figure 8: Simplest CVD setup using only thermal decompo-sition in a fixed bed geometry.[30]

7

III.5. Catalysts

Besides the temperature the catalyst is the most im-portant parameter for CVD grown CNTs. The mostactive metals for CNTs are transition metals.[21] Es-pecially iron, nickel and cobalt are interesting giventheir low material price. In the catalytic CVD pro-cess the catalyst is "used up" as the CNTs grow onthe metal particles, thus new catalyst needs to beadded to replace the used.[21] This means there is anefficiency yield of produced carbon nanotube massdivided by catalyst mass. For lowest cost this wouldbe as high as possible, not just for least catalyst ma-terial needed but also for less catalyst remnants inthe final product (thus higher quality). In practicethis will mean growing the longest possible CNTsper catalyst particle.

III.6. Fixed Bed Reactor

In a fixed bed reactor the CNTs have two distinctgrowth modes, it can either happen at the base or atthe top. The difference is the adhesion of the metalwith the substrate versus the carbon nanotube. i.e.more spread out over the substrate (lower contactangle) means more adhesion, and will more likelyhave the base growth mode as illustrated in 10. Inboth cases the growth is thought to stop because ofpassivation of the catalytic surface.[17]

Figure 9: Tip growth mode as it is currently thought to go,the hydrocarbons are absorbed by the metal catalystparticle and diffuse down to for part of the growingCNT.[17]

Figure 10: Base growth mode, here the hydrocarbon are ab-sorbed into the metal catalyst particle that formsthe carbon nanotube.[17]

III.7. Fluidized bed Reactor

In the fixed reactor bed configuration the catalystparticles are simply on a substrate inside the furnacearea of the CVD setup (as in 8). However there is an-other setup that is preferred by large scale chemicalindustries called fluidized bed reactor. In this setupthe reactor is turned ninety degrees and the gasinlet is done via a distributor. The gas is pumpedin, causing the solid nano particles of catalyst tobehave like a liquid at high enough pressure.[27]This mixing makes sure the temperature is uniformand the mass transfer is maximal.[27] Once the gashas reacted it is pumped off again and possibly re-cycled. In the case of CNT production the catalystneeds to be replaced continuously as it is used upand the product taken out. The output productionof fluidized bed reactors scales easily and alreadya commercial scale pilot plant has been made witha claimed total production of 15 kg/h.[29] As withthe fixed bed type the product has to be purified toget rid of some of the catalyst.

Figure 11: At higher gas injection velocity the solid particlesfluidize more and the catalyst mixes better withthe reactant. Aggregate fluidization causes nonuniform mixes and vibrations.[27]

III.8. HiPCo

Currently a different method is popular for pro-duced reasonable amounts of pure SWCNTs, theHiPCo process (High Pressure Carbon monoxide).

8

As the name suggests this involves CVD of CO athigh pressure (30 atm) and medium temperatures(1050 ◦C) with an iron catalyst.[24] This process isdistinct from bed reactors for it makes the catalystparticles in situ during the CVD process from gasphase Fe(CO)5. This is the biggest upshot of thismethod that the iron particles are formed in situthus making it a continuous process. The produc-tion rate of this process is about 0.5 g/h, fairly low,and sells for more than 1000$ per gram.[24] One ofthe biggest problems is purifying the CVD result totake out the iron catalyst. Although the temperatureis relatively low compared to some other techniquesthe pressure is not. High pressure vessels are veryexpensive, driving up the cost a lot. The maximumlength the SWCNTs attain is limited by passivation(by carbon) of the metal catalyst clusters.[24] Sincethe clusters are created in situ from a precursor gasthere is very little control over the catalyst parame-ters. Because of the small size of the metal clusters,and the consistency given the right parameters, onlysingle walled nano tubes are created.[24]

IV. Discussion

In this discussion section some of the issues with thecurrent situation will be highlighted. Especially withthe future in mind to move the field forwards. Espe-cially the metal-free catalyst that will be mentionedis interesting for it might give a very significantreduction in cost for CNTs commercially.

IV.1. Research

Perhaps the biggest gripe with carbon nanotubesis the fact that although we know how to producethem in various ways, we are still not capable of pre-dicting why those certain parameters would work.This means that most productions methods thusfar have really been by trial and error, instead ofknowing the best solution and implementing that.Doing specific research into making a batch of car-bon nanotubes with very precise defined parameters(diameter, length etc) by two separate, different, pro-duction methods may be a good way to understandbetter how CNTs are grown. In this way you wouldexpect that both those methods would use the samekind of reaction for creating the CNTs, and can be

compared to each other to find the common ground.Such a thing would obviously require being ableto produce not only nanotubes with very preciseparameters but also the same precise parameters bytwo methods, two methods that are beforehand notthat well understood. e.g. reliably making SWCNTswith 10 nm diameter and 200 nm length by CVD andby say laser ablation. This would suggest the situa-tion for both of those methods are the same duringthe CNT growth period, and can be compared.

Somewhat recently metal-free catalysts like nanoparticle diamond have been successfully used.[17]This suggests there is indeed something deeperabout carbon atoms forming nanotubes on parti-cles as the nano diamond was proven to be in solidstate the whole way through (meaning no diamondcarbon ended up in the CNT).[17] Exactly how thisworks is unknown, but metal-free catalysts in CVDmight not need the chemical treatments afters depo-sition anymore, making it purer and cheaper.

Figure 12: A model for the starting mechanism of growingCNTs by carbon atoms adsorbed on the surface ofa metal catalyst particle. First the carbon atomstry to form a graphene like structure whereby theypush themselves up off the metal particle to formthe end cap of the CNT.[17]

One of the biggest problems is the passivation ofthe (metal) catalyst particles. For most purposes thelonger the CNTs are the better. Therefore the exactgeometry of the nano particles that are used to growthe CNTs can have a large impact on the averagelength a nanotube grows before the catalyst particleis passivated. It’s very difficult to research becausethat means making a large amount of nano particleswith a very specific geometry, say perfectly sphericalversus star-like. Particles that would not suffer frompassivation could in theory make CNTs as long as isdesired.

9

IV.2. Microscopic versus Macroscopic

Of course all of this would mean little if the mi-croscopic amazing properties of CNTs don’t trans-late well into actual macroscopic properties. Thepromises of CNTs are from atomically perfect con-taminate free specimens, instead of an average overcarbon defects of various sorts. Although materi-als have been made with superior bulk propertiesalready they aren’t really worth the effort or priceyet. Presumably this will get fixed by making bet-ter cross links between the tubes as well as havinglonger pristine tubes. A very promising avenue inthis is radiating the tubes with electrons to formcross links.[31] In this method high energy electronsforce the aligned (MW)CNTs to form covalent bondsbetween the tubes, strengthening the matrix.[31] Ifwe are to make use of CNTs materials in "everyday"life it needs to not only be cheap enough to manufac-ture but also have very clear advantage over alreadyexisting materials.

IV.3. Costs

In the ideal case a great amount of CNTs are madeper hour per catalyst at low temperature with a highquality end product. But even in the very best ofcases the price of CNTs will simply never be ableto compete with conventional metal production andmanufacturing. This would mean that we shouldn’texpect CNTs products at every corner in the future.Although theoretically possible to build bricks fromCNT composite materials far superior to either steelor concrete, such a building would be many timesmore expensive than using current techniques. Sim-ilar arguments count concerning everyday clothingmade from ultra strong carbon nanotube fibers. Thatsaid, it has been shown that even small amounts ofCNTs added into other materials can have significantmaterial advantages (see [14] and [15]).

IV.4. Top Down Approach

When speaking of CNTs, essentially always a bottomup approach to making them is assumed. Bottom upmeaning the CNTs are made from carbon atoms oneby one as described earlier. On the other hand car-bon fibers have been made with smaller and smallerfiber thickness by new techniques. One could imag-

ine a pathway of making carbon nanotubes froma top down approach instead. Historically topdown approaches have often been cheaper thanbottom-up, while giving less control over the processin general. Carbon nanofibers have already beenmade with diameters in the nanometer range byelectrospinning.[22] In electrospinning the materialis suspended in a fluid and extracted by high voltage,creating a noodle-like extremely thin wire.[22] Asexpected from their size the nanofibers have prop-erties somewhere between carbon nanotubes andcarbon fibers, essentially an intermediate. Carbonnanofibers by electrospinning are roughly speakingtwo orders of magnitude cheaper than CNTs perweight currently.[26] A down side here is that thesenanofibers are made as aggregates and not in thealigned state.

Figure 13: Shows the exact geometry of the carbon nanofibers,essentially the fibers are made up of a stack ofconcentric hollow cones of graphene.[26]

IV.5. Recipes

Whether we understand it fundamentally or not, atsome point we’ll have a pretty good list of exactparameter/catalyst combinations for making carbonnanotubes with certain properties. Currently it’sstill a big mess of what exact method is the bestin making what exact CNT for a given application.And on top of that often where CNTs are used in theliterature it is not stated precisely what kind withwhat distribution, more transparency in this wouldhelp. Needless to say this is an on going processthat is getting better as CNTs are studied more.

10

IV.6. Health

A mayor concern with carbon nanotubes has beenits similarity to other tough high aspect ratio fibers.Not only are biological reactions to intruder parti-cles very complex, it is also difficult to do researchon the harm to humans and animals without need-lessly harming them in the process. Knowing itscarcinogenic effect we should look for methods toneutralize its effects.[11] When a certain technologyis deemed dangerous to the public eye it is almostimpossible to get rid of that reputation, whethertrue or not becomes irrelevant. For the most partthe CNTs are contained inside some other materialand the danger effect will be minimal, but for purematerials, the ones with the most promising proper-ties, this will need real attention before they can beput on the market.

V. Conclusion

Nanotechnology is still in its infancy. Easily one ofthe most promising materials is the carbon nanotube(CNT). Although the strength promised by theoret-ical arguments thus far has not yet been achievedin bulk form. The main application where indus-trial scale CNTs are required will be for structuralpurposes, but there are also plenty of electrical ap-plications. CVD (chemical vapour deposition) isa process for making CNTs as either aggregatesor aligned. Industry already has a lot of experi-ence using chemical vapour deposition. CVD is theonly scaleable process known for making CNTs, it istherefore favored for commercial production. Nev-ertheless the process is not well understood at allas can be seen from the fact that metal-free catalystwork as well. There are no obvious reasons whywe wouldn’t be able to get commercial scale CNTproduction in the near future however, some of thishas already been done (see [29], [24] and [33]). Thisimplies we will see many places where CNTs areused soon, while at the same time they will likelystay as an expensive high-end material. We shouldstay weary of any negative consequences it has, es-pecially to health, as damage is harder to fix thanprevent.

VI. Acknowledgments

Gratitude goes out to professor M.A. Loi as supervi-sor.

References

[1] W. Gomulya, M.A. Loi, "Conjugated polymer-wrapped carbon nanotubes: physical proper-ties and device applications" (2013, The Euro-pean Physical Journal B)

[2] Iijima S., Ichihashi T, "Single-shell carbon nan-otubes of 1-nm diameter" (1993, Nature)

[3] Lee, Byoungho, "Review of the present statusof optical fiber sensors." (2003, Optical FiberTechnology)

[4] Wang, X., Fan, Shoushan "Fabrication of Ul-tralong and Electrically Uniform Single-WalledCarbon Nanotubes on Clean Substrates" (2009,Nano Letters)

[5] Demczyk, B.G. Ritchie "Direct mechanical mea-surement of the tensile strength and elasticmodulus of multiwalled carbon nanotubes"(2002, Materials Science and Engineering)

[6] Yang Y., Li W. Z. "Radial elasticity of single-walled carbon nanotube measured by atomicforce microscopy" (2011, Applied Physics Let-ters)

[7] C.A. Coulson "Valence" (1952, Oxford Univer-sity Press)

[8] Hong Seunghun, Myung, S. "Nanotube Elec-tronics: A flexible approach to mobility" (2007,Nature Nanotechnology)

[9] Pop Eric et. al. "Thermal conductance of anindividual single-wall carbon nanotube aboveroom temperature" (2005, Nano Letters)

[10] J. M., Corum, M. B. Ruggles "Basic Properties ofReference Crossply Carbon-Fiber Composite"(2000, Oakridge National Laboratories)

[11] Boyles M.S., Duschl A. "The Significance andInsignificance of Carbon Nanotube-Induced In-flammation" (2014, Fibers)

11

[12] Zhou L. G., Shi San-Qiang "Formation energyof Stone-Wales defects in carbon nanotubes"(2003, Applied Physics Letters)

[13] Andrew Penn http://cargocollective.com/AndrewPenn/Carbon-Nanotubes (2011)

[14] Ron George http://cozybeehive.blogspot.nl/2008/03/carbon-nano-tube-bicycle-technology.html (2008)

[15] http://ultrawire.eu/ (2014, UltraWire)

[16] Mecklenburg M., Schulte K. "Aerographite: Ul-tra Lightweight, Flexible Nanowall, Carbon Mi-crotube Material with Outstanding MechanicalPerformance" (2012, Advanced Materials)

[17] Mukul Kumar, Yoshinori Ando, "Chemical Va-por Deposition of Carbon Nanotubes: A Re-view on Growth Mechanism and Mass Produc-tion" (2010, Journal of Nanoscience and Nan-otechnology)

[18] T. Guo, R.E. Smalley, "Catalytic growth ofsingle-walled nanotubes by laser vaporization"(1995, Chemical Physics Letters)

[19] Huisheng Peng, Tao Chen, "Polymer Compos-ites with Carbon Nanotubes in Alignment"(2011, Fudan University)

[20] R. R. Schlittler, M. E. Welland, "Single Crystalsof Single-Walled Carbon Nanotubes Formed bySelf-Assembly" (2001, Science)

[21] Ding F., Bolton K. "The importance of strongcarbon-metal adhesion for catalytic nucleationof single-walled carbon nanotubes." (2008,Nano Letters)

[22] Michio Inagaki, Ying Yang, "Carbon NanofibersPrepared via Electrospinning" (2012, AdvancedMaterials)

[23] Michael F. L., De Volder, A. John Hart, "CarbonNanotubes: Present and Future CommercialApplications" (2013, Science)

[24] Michael Bronikowski, Richard Smalley "Gas-phase production of carbon single-walled nan-otubes from carbon monoxide via the HiPco

process: A parametric study" (2000, Rice Uni-versity)

[25] http://advancedtextilessource.com/category/featured/ (2013, Advanced Textiles Source)

[26] David Burton, Andrew Palmer "Propertiesand Applications of Carbon Nanofibers (CNFs)Synthesized using Vapor-grown Carbon Fiber(VGCF) Manufacturing Technology" (2011,Sigma-Aldrich inc.)

[27] http://encyclopedia.che.engin.umich.edu/Pages/Reactors/PBR/PBR.html (2011, Encyclo-pedia of Chemical Engineering Equipment)

[28] Xiaolin Zhu, Fei Wei "Review on the nanoparti-cle fluidization science and technology" (2015,Chinese Journal of Chemical Engineering)

[29] Fei Wei, Qiang Zhang "The mass productionof carbon nanotubes using a nano-agglomeratefluidized bed reactor: A multisca29ime analy-sis" (2008, China Particle Technology)

[30] Jay P. Gore, Anup Sane "Carbon Nanotubes- Synthesis, Characterization, Applications"(2011, InTech)

[31] Peng B, Espinoza HD, "Measurements of near-ultimate strength for multiwalled carbon nan-otubes and irradiation-induced crosslinking im-provements" (2008, Nature Nanotechnology)

[32] Mustafizur Rahman, Thamir K. Ibrahim "Ther-modynamic performance analysis of gas tur-bine power plant" (2011, International journalof physical sciences)

[33] http://www.esa.int/Our_Activities/Launchers/Launch_vehicles/Vega (2012,European Space Agency)

12

http://cargocollective.com/http://cozybeehive.blogspot.nl/2008/03/carbon-nano-tube-bicycle-technology.htmlhttp://cozybeehive.blogspot.nl/2008/03/carbon-nano-tube-bicycle-technology.htmlhttp://cozybeehive.blogspot.nl/2008/03/carbon-nano-tube-bicycle-technology.htmlhttp://cozybeehive.blogspot.nl/2008/03/carbon-nano-tube-bicycle-technology.htmlhttp://ultrawire.eu/http://encyclopedia.che.engin.umich.edu/Pages/Reactors/PBR/PBR.htmlhttp://encyclopedia.che.engin.umich.edu/Pages/Reactors/PBR/PBR.htmlhttp://www.esa.int/Our_Activities/Launchers/Launch_vehicles/Vegahttp://www.esa.int/Our_Activities/Launchers/Launch_vehicles/Vega

IntroductionSingle versus multiwalledStructural PropertiesElectrical PropertiesThermal PropertiesCarbon FibersHealth Risks

ApplicationsTransistorsConductorCompositesAerogels

ProductionAggregates versus AlignedLaser AblationArc DischargeChemical Vapour DepositionCatalystsFixed Bed ReactorFluidized bed ReactorHiPCo

DiscussionResearchMicroscopic versus MacroscopicCostsTop Down ApproachRecipesHealth

ConclusionAcknowledgments