Journal of Industrial and Engineering Chemistry 17 (2011) 367–376

Contents lists available at ScienceDirect

Journal of Industrial and Engineering Chemistry

journal homepage: www.e lsev ier .com/ locate / j iec

Review

Continuous production of carbon nanotubes – A review

Lim Siang Ying a, Mohamad Amran bin Mohd Salleh a,b,*, Hamdan b. Mohamed Yusoff b,Suraya Binti Abdul Rashid a,b, Jeefferie b. Abd. Razak c

a Universiti Putra Malaysia, Institute of Advanced Technology (ITMA), Serdang, 43400 Selangor, Malaysiab Universiti Putra Malaysia, Department of Chemical & Environmental Engineering, Serdang, 43400 Selangor, Malaysiac Universiti Teknikal Malaysia Melaka, Faculty of Manufacturing Engineering, Karung Berkunci No. 1752, Pejabat Pos Durian Tunggal, 76109, Durian Tunggal, Melaka, Malaysia

A R T I C L E I N F O

Article history:

Received 10 August 2010

Accepted 8 November 2010

Available online 14 May 2011

Keywords:

Carbon nanotubes

Fluidized bed

Chemical vapour deposition

Continuous CNT synthesis

A B S T R A C T

Carbon nanotubes (CNTs) up to now are the most researched materials of the 21st century with an

international intention of growing industrial quantities due to their superior properties for use in many

applications. Thus far large quantities of carbon nanotube scan be grown in a continuous manner by both

arc as well as chemical vapour deposition methods. In this paper, an innovative approach of feeding

gases, a carbon precursor (solid or gases) and a catalyst into the reaction zone is reviewed. This is

followed by a study of the reaction process concerning how the method is initiated, the effect of growth

environment and catalyst on CNTs as well as the discharging mechanism for the final carbon products. A

study of the arc method consists of a novel way of growing CNTs in a liquid solution from an arc discharge

generated by carbon rods, by growing CNTs in a plasma zone using carbon gases or solid carbon and a

more direct method of using carbon tape as the anode for the synthesized source are also reported. In the

case of the chemical vapour deposition (CVD) method, some use a horizontal reactor and some use a

vertical reactor with all having different installed devices for use in continuous feeding and discharging

of resources and products respectively. Additionally, problems regarding the CNT yield and some issues

that have not been taken into consideration by others, are discussed. At the end of the review, an

additional mechanism to integrate catalyst preparation and carbon nanotube purification into the

current research synthesizing process for future study is proposed for a highly productive continuous

CNT synthesis process.

� 2011 The Korean Society of Industrial and Engineering Chemistry. Published by Elsevier B.V. All rights

reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

2. Continuous synthesizing methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

2.1. Feeding mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368

2.2. Reaction process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

2.2.1. Approach of initiating CNTs formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369

2.2.2. Effect of growth environments and characteristic of catalyst on CNTs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371

2.3. Discharging mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372

2.4. Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374

3. Efficiency of the process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374

4. Current issues in continuous synthesis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374

5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375

* Corresponding author at: Universiti Putra Malaysia, Institute of Advanced

Technology (ITMA), Serdang, 43400 Selangor, Malaysia. Tel.: +60 389466286;

fax: +60 386567120.

E-mail address: asalleh@eng.upm.edu.my (Mohamad Amran bin Mohd Salleh).

1226-086X/$ – see front matter � 2011 The Korean Society of Industrial and Engineer

doi:10.1016/j.jiec.2011.05.007

1. Introduction

The discovery of CNTs in 1991 was primarily due to Iijima, andsince that day much research has been dedicated to studying CNTs

ing Chemistry. Published by Elsevier B.V. All rights reserved.

[(Fig._1)TD$FIG]

Fig. 1. Schematic of a reactor used for CNT synthesis. Major component for the continuous process are the syringe pump, mass flow controller and furnace [1].

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376368

[15]. CNTs have received much attention from scientific commu-nities up to this date mainly because of their superior propertiessuch as having a wide band gap, high melting point, high tensilestrength and high thermal conductivity [9]; these properties beingthe result of their small size and the covalently bonded carbonatoms. CNTs are normally categorized into three types, there aresingle-walled carbon nanotubes (SWCNT), double-walled carbonnanotubes (DWCNT) and multi-walled carbon nanotubes(MWCNT). To date, a great deal of papers state success in growingCNTs by using common synthesizing methods including thechemical vapour deposition method [19], the arc method [18] andthe laser method [20]. Only recently, the continuous synthesis ofCNTs has been a hot topic of discussion due to their ability to growlarge quantities of CNTs effortlessly. These techniques include butare not restricted to arc-jet plasma [5], simplified carbon-arc [16],floating catalyst [26] and fluidized-bed CVD [23]. The newtechniques have been developed due to the difficulty of currentmethods to grow high quantities of CNTs. Being the newestmethods, current research in these fields is very limited and itsprogress is somewhat slow as only a limited number of papers havebeen generated. Nevertheless this review discusses to what extent

current development in continuous CNTs synthesizing has beensuccessfully achieved.

2. Continuous synthesizing methods

According to current research papers published, research in thecontinuous synthesizing process has made some progress in boththe arc method and the CVD method while other continuousmethods are currently not generally known. Nevertheless, each ofthese methods of operation is reviewed according to fourcategories comprising of feeding mechanism, reaction process,discharging techniques and instrumentation.

2.1. Feeding mechanism

In continuous arc methods, different feeding mechanisms areused to grow CNTs as reported by different authors. For anexample, Ishigami et al. (2000) uses a continuously fed graphitecathode with non-consumable copper as an anode connected to aDC power supply dipped into liquid nitrogen as depicted in Fig. 1[16]. In a similar study, Biro et al. (2003) uses two carbon

[(Fig._2)TD$FIG]

Fig. 2. Moving bed reactor made up of a quartz tube with major components

including the conveyor belt, gas inlet, collector, tube furnace and solid sampler [22].

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376 369

electrodes dipped into deionized water rather than expensiveliquid nitrogen presented in Fig. 2 [3]. Both methods consume theelectrodes to synthesize CNTs by forming an arc when theelectrodes are brought into contact, here the liquid is used as acooling agent and also to prevent oxidation. While another studydirected by Yusoff et al. (2006), uses a moving carbon tapesubstrate passed between two electrodes indicated in Fig. 3 togrow CNTs [36]. This process uses nitrogen gas as a buffer and acathode is slowly sublimed with the carbon tape substratesupported by the anode. Additionally, CNTs are also found togrow in a plasma jet reactor in a study conducted by bystrzejewskiet al. As illustrated in Fig. 5(a), a graphite rod is fed directly into thereaction chamber perpendicular to the plasma jet and close to anozzle exit [4]. Different carbon precursors to grow CNTs replacingthe electrodes are also found to be used in the plasma reactor. Asper Figs. 4 and 5, the study utilizes an injection system to introduceplasma gas, carbon powder/carbon gases and catalyst continuously

[(Fig._3)TD$FIG]

Fig. 3. Schematic of an advanced fluidized-bed reactor with legend indicating each

of the components [2].

from above the plasma into a vaporization chamber [8,12],whereas Choi et al. (2006) and Ohishi et al. (2008) use carbongases and a catalyst fed perpendicular to the plasma jet [27,5]. Theillustration is given in Fig. 6(b) and Fig. 7. This kind of synthesizingprocess consumes the carbon source and catalyst to form CNTswith the help of a plasma torch formed by introducing plasmagases into the electrical current path of non-consumable electro-des.

In the case of the CVD method, there are also limited papersexisting in the field for the purpose of growing CNTs continuously.In the floating catalyst technique for instance, to grow CNTscontinuously it uses a dry powder injector to constantly introducecatalyst powder into the reactor from above with the help of acarrier gas and a carbon precursor, Fig. 7 [32]. As for Liu et al.(2001), they grow CNTs using a modified fixed-bed CVD as shownin Fig. 8, this method continuously drops catalyst onto a movingconveyor belt with gases introduced horizontally along the belt toinitiate the reaction process [22]. A horizontal quartz tube reactorcapable of operating on a continuous or semi-continuous basis hasalso been reported as shown in Figs. 9 and 10 respectively. Thismethod introduces a carbon precursor in the form of a liquid mixedwith a catalyst by continuously injecting liquid carbon into thereactor [1,34]. Here inert gases are used to volatilize the liquid andassist the flow of the carbon mixture into the high temperaturezone for reaction [1]. There is also a CVD reactor made up of acontinuous mass production apparatus with several catalyst-containing ceramic boats preinstalled within the reaction cham-bers for continuous operation with gases fed directly from a gasdecomposer, as illustrated in Fig. 11 [30]. In the case of thefluidized-bed CVD, only two papers have been found thus far to berelated to the continuous production of CNTs using this techniquedespite its advantages to run continuously as stated by themajority of the published papers. Nevertheless, the first papers asdescribed by Arena et al. (2006) with a reactor picture presented inFig. 12, uses recycled polypropylene as a carbon source fedcontinuously using an automated feeder from the top of the reactorwhile fluidization gases flow in from bottom of the reactor [2].Another paper presented by Maghsoodi et al. (2009) uses a blendof floating catalyst and fluidization method in a modified fluidized-bed reactor. This method uses a ferrocene injection system placedon top of the reactor and an MgO transfer tube located at thebottom of the reactor [23]. In this system, reaction occurs withferrocene supported on MgO as both of them are injected to thepoint of contact using the floating catalyst and fluidization conceptrespectively. A detailed instrumentation of the method isdisplayed in Fig. 13.

2.2. Reaction process

2.2.1. Approach of initiating CNTs formation

Regardless of the methods used, the process of synthesizinggoes back to the root of chemical reaction at a high temperature.Even so, each method has different techniques of initiating thereaction process. In arc methods, current research proves to beable to grow CNTs in a gas environment and a liquidenvironment as described earlier. The difference in the reactionprocess depends on the use of a carbon source by either using agas environment to act as a buffer [36] or a liquid environmentto provide oxygen free environments [35]; given the graphiteelectrodes, the reaction process takes place with electricallyconnected electrodes initially brought into contact to generatean arc discharge [3,16,35,36]. As for the carbon gases, powdersand consumable carbon electrodes, the reaction occurs in a hightemperature plasma generated by the non-consumable electro-des. In detail, the reaction process of the carbon gas as aprecursor is initiated by forming plasma with the simultaneous

[(Fig._4)TD$FIG]

Fig. 4. Schematic of arc-jet plasma using a non-consumable cathode with (a) continuously fed graphite rod [4] and (b) gases carbon precursor with a CNT collection tube [5].

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376370

introduction of inert gases and followed the by feeding of thecatalyst into the reaction region [5,12,27]. The same goes for thepowders and consumable carbon electrodes with plasmagenerated in an inert environment with simultaneous feedingof the carbon powder/carbon electrodes into the reactor [4] andwith the additional introduction of a catalyst in [8].

In CVD methods, all the reaction processes take place in a hightemperature CVD reactor which requires the introduction of acatalyst, inert gas and a carbon precursor in the form of a liquid orgas. Unlike the reaction process in arc methods which have muchsimilarity in the work of different authors, the step by step processof initiating a reaction in CVD methods is very different in thedifferent work and it varies according to different authors.Methods used by [1] and [34] show an example of similarity in

[(Fig._5)TD$FIG]

Fig. 5. Schematic of underwater a

devices but the reaction process take place differently. In the workof Andrews et al. (1999), the mixed carbon liquid is firstly passedthrough a preheater through a capillary tube and is volatized andswept into the furnace by a flow of argon mixed with hydrogengases once the liquid leaves the capillary [1]. In the case of Su et al.(2006), the reactor was initially flushed with argon gas to eliminateoxygen inside the reactor followed by heating the reactor and anargon flow was introduced to dissolve the supplied alcoholsolution mixed with ferrocene [34]. Fig. 8 based on the study ofLiu et al. (2001) preceded the experiment by heating the reactor tothe reaction temperature in a flow of nitrogen gas. A pre-reducedcatalyst is later dropped onto a conveyor belt moving into thereaction chamber followed by the introduction of methane [22]. Inpapers that state the use of a customized CVD reactor, such as the

rc discharge equipment [3].

[(Fig._6)TD$FIG]

Fig. 6. 3-phase AC plasma reactor with powder injector, vaporization chamber, cooling system, filter, circulation pump and a three phase electrical source [47].

[(Fig._7)TD$FIG]

Fig. 7. Thermal plasma reactor with helical extension for continuous CNT

production [12].

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376 371

paper by So et al. (2009), the reaction takes place whenhydrocarbon gases, essential for CNT synthesis, is simultaneouslysupplied and decomposed via the use of a high-temperature gasdecomposer directly connected to the reaction chamber where theCNTs are to grow [32]. According to So et al. (2009), the use of thistechnique is to ensure the decomposition of hydrocarbon gas and[(Fig._8)TD$FIG]

Fig. 8. Arc method with continuously fed graphite rod and non-consumable cat

the reaction between the decomposed gas and the catalystproceeds separately. Lastly in the case of the fluidized-bed reactor,the process is started with the reaction chamber firstly heated to areaction temperature according to [2], [26] and [23]. This isfollowed by the injection of a catalyst at a predetermined ratewhich is carried by the flow of methane and inert gases into thereaction zone [23,26] or it is carried just by the inert gas [2].

2.2.2. Effect of growth environments and characteristic of catalyst on

CNTs

It is widely known that CNT formations in any of the abovementioned methods are very much depended on its growingenvironment and characteristic of the catalyst. Here an overview ofthe literatures is given to provide a brief understanding for theroles of growing environment and catalyst have on CNT. Ananalysis of the literatures thus far indicates that, growing CNT in aspecial environment is a must regardless of their synthesizingmethods [1–45]. The role of the growing environment is particularimportant because it could affect the outcome of CNT selectivity

hode dipped in a liquid nitrogen reaction chamber for CNT synthesis [16].

[(Fig._9)TD$FIG]

Fig. 9. Schematic of fluidized-bed reactor with apparatus to provide continuous CNT

production using a floating catalyst and fluidization concept [23].

[(Fig._11)TD$FIG]

Fig. 11. Schematic of arc-jet plasma with a non-consumable cathode and a substrate

holder [27].

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376372

and structural quality [46]. In an arc discharged, CNT is grown outdirectly from a graphite rod without the need for catalyst. Due tothe absence of catalyst in arc discharged method, CNT structure ismainly affected by their growing environment. For instance, CNTsgrown in deionise water [35] tend to have a better structuralquality than CNTs grown in liquid nitrogen [16] that yield adistorted morphology and degraded structure [35]. Of course, CNTssynthesis from arc discharged is also affected by other parameterincluding electrical current [3] and graphite rod [21]. For a systemthat requires the presence of metal catalyst such as CVD [10] andplasma arc [12] method. It is generally considered that differentcarrier gases will have a different effect on CNT structures. As anexample, the use of argon gas will usually yield MWCNTs whileSWCNTs were formed with the use of hydrogen gases [46]. Theroles of catalyst in these systems is as important as its environmentand is said to influence CNT selectivity with alloy catalyst (Fe–Co,Ni–Co, Fe–Mo, Co–Ni–Fe, Fe–Ni and Mg–NiO) being very effectiveon CNT selectivity [6,38–43] compared to metal catalyst (Fe, Co, Ni)that yield both SWCNT and MWCNT respectively, indicating poor[(Fig._10)TD$FIG]

Fig. 10. Floating catalyst reactor with a dry powder injector. In addition, a vertical

shaker is included to maintain homogeneity of gases and catalyst mixture and two

collectors as well as a trap are installed at the end of the reactor [26].

selectivity [8]. CNT growth rate is another parameter influence bythe catalyst and it is said to be dependent on catalyst size [45] alsoon catalyst types (i.e. Fe catalyst yields more CNTs) [44]. Perhaps,the most interesting part of the catalyst is its ability to influencethe structural quality of CNT [29]. According to Kukovecz et al.(2000), high quality CNT is obtainable using Co and alloy catalystCo–Fe whereas using Fe tends to produce lower quality CNT as atrade-off for higher CNT yields [44].

2.3. Discharging mechanism

In terms of collecting the final products, not much informationis reported and most of it is identified through the figures ofsubmitted papers. In the continuous arc method, the authors of[36] grow CNTs on top of a moving carbon tape and as the tapemoves, synthesized CNTs move along with it and at the same time anew region of tape moves into the reaction zone of the previouslygrown CNTs to form fresh CNTs. This process is continuous with

[(Fig._12)TD$FIG]

Fig. 12. Quartz tube reactor with a continuous flowing carbon source provided by a

quartz capillary and Ar gas. Carbon product is collected at the exit of the reactor

[34].

[(Fig._13)TD$FIG]

Fig. 13. A continuous mass production apparatus consisting of a catalyst-containing

ceramic boat installed in the reaction chambers, a moving heater and a gas

decomposer [32].

[(Fig._14)TD$FIG]

Fig. 14. Schematic of a continuous arc method with moving carbon tape as an

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376 373

CNTs constantly growing on the moving tape and transported outof the reaction zone. A detailed visualisation of the moving tapeand the growing spot between the anode and the cathode ispresented in Fig. 3. In a plasma reactor however, the study in [8]collected the final products when the high temperature gasesmixed with the CNTs is carried into the quenching system and thefiltered CNTs are discharged at the bottom of the filter while thegases are released at the top as shown in Fig. 4. A different studyconducted by Hahn et al. (2004) using a thermal jet plasma reactorcollected the carbon products by flowing gases through a helicalextension reactor into the collector with a small exhaust aspresented in Fig. 5 [12]. In a similar way, Choi et al. (2004) droppedCNTs into a collection tube as they are taken by gas flowing to thevacuum pump placed directly below the plasma jet as depicted inFig. 6(b) [5]. On the other hand some researchers, for example theauthors of [26] did not discharge the synthesized carbon productsbut instead they are collected inside the reaction chamber on top ofa substrate holder according to Fig. 14. In case of an arc created in aliquid environment, the experimental setup depicted in Fig. 1shows that Ishigami et al. (2000) collected the CNTs at the bottomof a bucket full of liquid nitrogen. This bucket is sealed with a valvethat automatically opens from time to time to flush CNTs from thevessel [16]. Another similar study according to Fig. 2 and [3] isspeculated to leave the CNTs within the water without discharging.

ode. (a) Anode support and (b) side view of the experimental set-up [36].

[(Fig._15)TD$FIG]

Fig. 15. Roles of current to weight of carbon produced (*), electrodes weight loss

(&) and relative product yield (*) [3].

Table 1CNT yield and C2H4 conversion in 2.5 cm diameter reactor of 0.5 g catalytic mass,

60 min duration and flow rate of C2H4, N2 and H2 at 60 sccm, 160 sccm and

120 sccm. Adapted from [28].

T (8C) CNT produced (g) C2H4 conversion (%) gCNT/gcatalyst

500 0.28 7.8 0.56

550 0.55 15.3 1.10

600 0.95 26.4 1.90

650 1.35 37.6 2.70

700 1.28 35.5 2.55

750 1.23 34.2 2.46

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376374

A range of ways to collect CNTs also exists in continuous CVDmethods. As reported in [22] and Fig. 8 using a moving bed reactor,CNT products are collected when a moving conveyor belt drops theCNTs vertically into the collector at the end of the reactor. Adifferent approach is adopted in [26] using the cyclone separatorconcept with the carbon product being carried away by gases andcollected in two consecutive collectors and a trap which aresituated at the bottom of a floating catalyst reactor as shown inFig. 7. Similar to [2] and [23], CNTs are continuously withdrawn atthe bottom of the reactor using a withdrawal device shown inFig. 12 (no information is given for the type of device used) and aneffluent tube shown in Fig. 13 respectively. It is noted that [1] and[34] effectively discharge all carbon products horizontally into acollector at the end of the reactor by manipulating the flow of gas(Fig. 9 and Fig. 10). One study conducted by [32] using a massproduction apparatus did not clearly report the dischargingmethod even with a graphical presentation as given in Fig. 11. Itis possible that after the reaction, the carbon products inside thereaction chamber are taken out upon releasing the sealed door.

2.4. Instrumentation

With all the different approaches used in each research fordeveloping a continuous process, it is difficult to obtain a clearpicture of how each process works and thus a presentation of theinstrumentation in terms of figures is available to give a betterpicture of the working equipment in each case. Each of the feedingand discharging mechanism has been described in detail above.

3. Efficiency of the process

Despite the advantages of continuous synthesis to constantlygrow CNTs compared to batch synthesis, the process efficiency interms of CNT yield mostly has been wrongly reported. This is trueeven for both arc and CVD methods in batch processes. Theproblem associated with this is the difficulty to differentiate CNTsfrom soot experimentally [30]. In one example of arc discharge,according to Biro et al. (2003) as per Fig. 15, the data provided onlyshows electrode conversion into carbon products with themaximum obtainable conversion at 85 A (85% purity) while itdid not exactly indicate the CNT yield in the conversion [3].Similarly, the yield also has not been reported for the continuousCVD method. Even in batch processes, only a limited amount ofpapers state the efficiency of the process and from these it ispossible to have some insight of the CNT yield which is supposed to

be nearly similar to the continuous processes. To make thingseasier, a comparison made by See et al. (2008) indicatesconsistencies in yield for the fluidized bed and horizontal CVD[30]. Thus, a list of experimental data compiled by Philippe et al.(2007) for the fluidized bed provides adequate yield informationfor the both techniques with a CNT yield (gCNT/gcatalyst) shown torange from 0.05 to 27.9 [28]. However, the yield measured in termsof gCNT/gcatalyst does not actually reflect the true efficiency of theprocess for a given carbon source. Intelligently, Philippe et al.(2007) supported the information with additional data for carbonconversion to CNT as presented in Table 1 for his own experiment[28]. Although the method seems promising in determining theeffectiveness of the process but the CNT yield per mole of carbonsource (gCNT/molcarbon source) can be more informative in the sensethat the efficiency of carbon conversion to CNT can be directlyobtained. Yet, up till now it has not been studied by most authors.

More interestingly to noted, instead of measuring gCNT/gcatalyst

and gCNT/molcarbon source, a popular method of determining CNTyield directly from soot using TGA is used by many authors[8,26,31,33]. This method measures the purity of CNTs in the sootand is commonly considered by some as the CNT yield. Regrettably,some authors have even mistaken this as the soot yield from CNT[10,14,24]. We think that this method of calculating yield isentirely misleading as it only shows the yield of CNT from soot andnot the carbon source, not to mention others for falsely reportedsoot yield from CNTs. Provided here is an example, a method usedby Fabry et al. (2004) who were able to obtain a CNT yield rangingfrom 10 to 50% in a soot of production rate of 250 g/h using the arcmethod [8]. For the CVD method, the yield is reported over andover again. These authors claim to be able to synthesize CNTsranging from 50% [34], 80% [12] to over 90% carbon content [32].Some even yield between 5 and 15% for the production of 6 g/h ofSWCNTs [26].

4. Current issues in continuous synthesis

One other issue which could be considered as important, otherthan the process efficiency described earlier, which has beenneglected by most authors, is the study of how the CNTs and thecatalyst interact with the reactor. To our knowledge so far this hasnot been reported in any indexed journal. In the CVD method, bulkmaterial handling is common for the process in which manyproblems encountered are attributed to the flow characteristic ofthe powders [37]. In truth, the scientific and engineeringknowledge that govern the above conditions has been very muchbased on past performance [17]. Data provided by Merrow, (1985)for handling liquids or gases shows the ratio of actual to thedesigned production rate on average is about 90%, whileparticulate handling is roughly 40% with a variance of 40% [25],and this data provides the proof of a lack of scientific research.Commonly affected problems are the mechanics of the particulatebed and the flow ability of the powders such as surface activation,solidifying bridge between particles, internal rupture of particles

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376 375

and caking [17]. These are attributed to particle size, shape,environmental conditions, cohesiveness, surface roughness, elec-tromagnetic force, orifice diameter, wall effects and the angle ofinclination, etc. [11,13]. For arc discharge synthesis, discussion ofthe problem associated with particle–reactor interaction is hardlyto be found and possibly there is none at all. For the plasma arc, it ispossible some of the studies have a certain similarity to CVD only interms of continuous synthesis, that is during the CNT dischargesuch as in the work of Choi et al. (2006) [5], Fabry et al. (2004) [8]and Hahn et al. (2004) [12].

It is not so certain what further possible issues are involvedsince hardly any problems associated with the reactor or duringcontinuous synthesis are reported. However it can be sure in eachand every research, there will be problems that are omitted orwhich most authors fail to identify.

5. Conclusion

In conclusion, all continuous processes using either arc or CVDmethods as judged by the reviewed papers have shown innovativeways of growing CNTs in large amounts. To ensure CNTs are growncontinuously, these authors have given priority to feeding anddischarging mechanisms by supplying a carbon precursor, gasesand a catalyst into the reaction zone as well as discharging the finalcarbon products at a constant rate. Most of the papers reviewed sofar show the continuous process is achieved mostly by the use ofconventional and non-continuous equipment by simply addingadditional feeding and discharging mechanisms capable ofoperating on a continuous basis and these methods have provento be successful. However, the process of synthesizing CNTscontinuously requiring the presence of a catalyst is still lacking interms of integrated approaches to preparing the catalyst and forthe purification of the CNTs. It is obvious from the reviewed studiesthat catalysts are prepared separately and no purification step istaken into consideration, as the carbon products are simplycollected in a collector, particularly in the case of the CVD method.In future research, the continuous synthesis process can be carriedforward by including an additional mechanism for the purificationof CNTs to provide a fully integrated one step synthesizing processfrom catalyst introduction to purified CNTs. This kind of integratedprocess could possibly be capable of growing very large quantitiesof purified CNTs in a shorter amount of time as compared to thecurrent processes reviewed due to the elimination of the need for aseparate purification process. After reviewing several papers, wefeel that in terms of efficiency of the process, some authorsreported the wrong CNT yield while others did not give anyconsideration to the carbon conversion process which is importantto know the true efficiency of the technique. Such authorsexpressed the yield as gCNT/gcatalyst instead of gCNT/molcarbon source.Finally, the current issues of continuous synthesis reveal that bulkparticle handling and its interaction with the reactor require muchattention, at least for the CVD method.

Acknowledgements

The authors gratefully acknowledge Universiti Putra MalaysiaEngineering Faculty for their support.

References

[1] R. Andrews, D. Jacques, A. Rao, F. Derbyshire, D. Qian, X. Fan, et al., Continuousproduction of aligned carbon nanotubes: a step closer to commercial realization,Chemical Physics Letters 303 (1999) 467–474.

[2] U. Arena, M. Mastellone, G. Camino, E. Boccaleri, An innovative process for massproduction of multi-wall carbon nanotubes by means of low-cost pyrolysis ofpolyolefins, Polymer Degradation and Stability 91 (4) (2006) 763–768.

[3] L. Biro, Z. Horvath, L. Szalmas, K. Kertesz, F. Weber, et al., Continuous carbonnanotube production in underwater AC electric arc, Chemical Physics Letters 372(3–4) (2003) 399–402.

[4] M. Bystrzejewski, A. Huczko, H. Lange, W. Płotczyk, R. Stankiewicz, T. Pichler,et al., A continuous synthesis of carbon nanotubes by dc thermal plasma jet,Applied Physics A 91 (2) (2008) 223–228. , doi:10.1007/s00339-008-4400-y.

[5] S. Choi, J. Nam, J. Kim, T. Hwang, et al., Continuous process of carbon nanotubessynthesis by decomposition of methane using an arc-jet plasma, Thin Solid Films506 (2006) 244–249.

[6] J. Colomer, J. Bister, I. Willems, Z. Konya, A. Fonseca, G. VanTendeloo, Synthesis ofsingle-wall carbon nanotubes by catalytic decomposition of hydrocarbons, Chem-ical Communications 14 (1999) 1343–1344.

[7] A. Dupuis, The catalyst in the CCVD of carbon nanotubes – a review, Progress inMaterials Science 50 (8) (2005) 929–961.

[8] F. Fabry, T. Gruenberger, J. Gonzalez-Aguilar, H. Okuno, E. Grivei, N. Probst, et al.,Continuous mass production of carbon nanotubes by 3-phase AC plasma proces-sing, In NSTI Nanotechnology Conference and Trade Show Vol. 3 (2004) 228–231.

[9] J. Fischer, in: Y. Gogotsi (Ed.), Carbon Nanotubes: Structure and Properties, CRCPress, Boca Raton, 2006.

[10] G. Gulino, R. Vieira, J. Amadou, P. Nguyen, M. Ledoux, et al., C2H6 as an activecarbon source for a large scale synthesis of carbon nanotubes by chemical vapourdeposition, Applied Catalysis A: General 279 (1–2) (2005) 89–97.

[11] A. Guo, J. Beddow, A. Vetter, A simple relationship between particle shape effectsand density, flow rate and Hausner ratio, Powder Technology 43 (3) (1985) 279–284.

[12] J. Hahn, J. Han, J. Yoo, H. Jung, J. Suh, New continuous gas-phase synthesis of highpurity carbon nanotubes by a thermal plasma jet, Carbon 42 (4) (2004) 877–883.

[13] H. Hausner, Powder characteristics and their effect on powder processing, PowderTechnology 30 (1) (1981) 3–8.

[14] K. Hernadi, Catalytic synthesis of multiwall carbon nanotubes from methylace-tylene, Chemical Physics Letters 363 (1–2) (2002) 169–174.

[15] S. Iijima, Helical microtubules of graphitic carbon, Nature 354 (1991) 56–58.[16] M. Ishigami, J. Cumings, A. Zettl, S. Chen, A simple method for the continuous

production of carbon nanotubes, Chemical Physics Letters 319 (5–6) (2000) 457–459.

[17] H. Kalman, G. Tardos, Elements of particle technology in the chemical industry,Particulate Science and Technology 23 (1) (2005) 1–19.

[18] M. Keidar, Factors affecting synthesis of single wall carbon nanotubes in arcdischarge, Journal of Physics D: Applied Physics 40 (8) (2007) 2388.

[19] Y. Kobayashi, H. Nakashima, D. Takagi, Y. Homma, CVD growth of single-walledcarbon nanotubes using size-controlled nanoparticle catalyst, Thin Solid Films464 (2004) 286–289.

[20] M. Kusaba, Y. Tsunawaki, Production of single-wall carbon nanotubes by aXeClexcimer laser ablation, Thin Solid Films 506 (2006) 255–258.

[21] H. Lange, M. Sioda, A. Huczko, Y. Zhu, H. Kroto, D. Walton, Nanocarbon productionby arc discharge in water, Carbon 41 (8) (2003) 1617–1623.

[22] B. Liu, M. Qu, Z. Yu, Continuous production of carbon nanotubes by using movingbed reactor, Chinese Chemical Letters 12 (12) (2001) 1135–1138.

[23] S. Maghsoodi, A. Khodadadi, Y. Mortazavi, A novel continuous process for syn-thesis of carbon nanotubes using iron floating catalyst and MgO particles for CVDof methane in a fluidized bed reactor, Applied Surface Science (2009),doi:10.1016/j.apsusc.2009.11.026.

[24] P. Mauron, C. Emmenegger, P. Sudan, P. Wenger, et al., Fluidised-bed CVDsynthesis of carbon nanotubes on Fe2O3/MgO, Diamond and Related Materials12 (3–7) (2003) 780–785.

[25] E.W Merrow, Linking R-and-D to problems experienced in solids processing,Chemical Engineering Progress 81 (5) (1985) 14–22.

[26] E. Mora, T. Tokune, A. Harutyunyan, Continuous production of single-walledcarbon nanotubes using a supported floating catalyst, Carbon 45 (5) (2007)971–977.

[27] T. Ohishi, Y. Yoshihara, O. Fukumasa, Continuous synthesis of carbon nanoclustersusing well-controlled thermal plasmas, Surface & Coatings Technology 202 (22–23) (2008) 5329–5332.

[28] R. Philippe, A. Morancais, M. Corrias, B. Caussat, et al., Catalytic production ofcarbon nanotubes by fluidized-bed CVD, Chemical Vapor Deposition 13 (9) (2007)447–457.

[29] C. See, A. Harris, A review of carbon nanotube synthesis via fluidized-bedchemical vapour deposition, Industrial and Engineering Chemistry Research 46(4) (2007) 997–1012.

[30] C. See, A. Harris, A comparison of carbon nanotube synthesis in fixed and fluidisedbed reactors, Chemical Engineering Journal 144 (2) (2008) 267–269.

[31] T. Setoguchi, M. Nozaki, H. Hashimoto, T. Fujii, Development of FabricationTechnology of Carbon Nanotube by Fluidized-Bed Reactor, 43, Mitsubishi HeavyIndustries, Ltd., 2006 1-3.

[32] D. So, H. Ham, W. Kim, S. Bang, B. Choi, K. Shim, et al., Continuous mass productionsystem of carbon nanotubes synthesis and processing, Ceramic Processing Re-search 10 (1) (2009) 105–108.

[33] S. Son, Y. Lee, S. Won, D. Lee, S. Kim, S. Sung, High-quality multiwalled carbonnanotubes from catalytic decomposition of carboneous materials in gas-solidfluidized beds, Industrial and Engineering Chemistry Research 47 (7) (2008)2166–2175.

[34] L. Su, J. Wang, F. Yu, Z. Sheng, H. Chang, C. Pak, Continuous production of single-wall carbon nanotubes by spray pyrolysis of alcohol with dissolved ferrocene,Chemical Physics Letters 420 (4–6) (2006) 421–425.

[35] A. Vittori, R. Marazzi, R. Krsmanovic, Synthesis of multiwall carbon nanotubes byelectric arc discharge in liquid environments, Carbon 41 (12) (2003) 2393–2401.

L.S. Ying et al. / Journal of Industrial and Engineering Chemistry 17 (2011) 367–376376

[36] H. Yusoff, R. Shastry, T. Querrioux, J. Abrahamson, Nanotube deposition in acontinuous arc reactor for varying arc gap and substrate temperature, CurrentApplied Physics 6 (3) (2006) 422–426.

[37] M. Yusoff, N. Page, Particle material, morphology and load effects on internalfriction in powders, Powder Technology 76 (2) (1993) 155–164.

[38] S. Lyu, C. Liu, S. Lee, C. Park, H. Kang, C. Yang, C. Lee, et al., Large-scale synthesis ofhigh-quality single-walled carbon nanotubes by catalytic decomposition of eth-ylene, Journal of Physical Chemistry B 108 (5) (2004) 1613–1616.

[39] A. Cassel, G. McCool, N. Tee, J. Koehne, B. Chen, J. Li, et al., Carbon nanotubenetworks by chemical vapor deposition, Applied Physics Letters 82 (2003) 817–819.

[40] H. Dai, A. Rinzler, P. Nikolaev, A. Thess, D. Colbert, R. Smalley, Single-wallnanotubes produced by metal-catalyzed disproportionation of carbon monoxide,Chemical Physics Letters 260 (1996) 471–475.

[41] Y. Baek, S. Honda, T. Ikuno, S. Ohkura, M. Katayama, T. Hirao, et al., Formation ofgraphite layers during carbon nanotubes growth, Japanese Journal of AppliedPhysics 42 (2003) 579–581.

[42] Y. Yang, Z. Hu, Y. Tian, Y. Lu, X. Wang, Y. Chen, High-yield production of quasi-aligned carbon nanotubes by catalytic decomposition of benzene, Nanotechnol-ogy 14 (2003) 733–737.

[43] P. Chen, H. Zhang, G. Lin, Q. Hong, K. Tsai, Growth of carbon nanotubes by catalyticdecomposition of CH4 or CO on a Ni–MgO catalyst, Carbon 35 (1997) 1495–1501.

[44] A. Kukovecz, Z. Konya, N. Nagaraju, I. Willems, A. Tamasi, A. Fonseca, et al.,Catalytic synthesis of carbon nanotubes over Co, Fe and Ni containing conven-tional and sol-gel silica-alumina, Physical Chemistry Chemical Physics 2 (3)(2000) 3071–3076.

[45] Z Yu, D. Chen, B. Totdal, A. Holmen, Effect of catalyst preparation on the carbonnanotube growth rate, Catalysis Today 100 (3–4) (2005) 261–267.

[46] F. Danafar, A. Fakhru’l-Razi, M. Salleh, D. Biak, Fluidized bed catalytic chemicalvapor deposition synthesis of carbon nanotubes – a review, Chemical EngineeringJournal 155 (1–2) (2009) 37–48.

[47] L. Fulcheri, et al. Production of Carbon Nanostructures Ranging from Carbon BlackOver Fullerenes to Nanotubes by Thermal Plasma, 16th International Symposiumon Plasma Chemistry (ISPC 16), Taormina (Italy), 2003.