Cw

LS

a

ARRAA

KCSSS

1

rmtsmour

lSottmthrfimtm

0h

Applied Surface Science 261 (2012) 415– 421

Contents lists available at SciVerse ScienceDirect

Applied Surface Science

j our nal ho me p age: www.elsev ier .com/ loc ate /apsusc

omparison studies of surface cleaning methods for PAN-based carbon fibersith acetone, supercritical acetone and subcritical alkali aqueous solutions

inghui Meng, Dapeng Fan, Yudong Huang ∗, Zaixing Jiang, Chunhua Zhangchool of Chemical Engineering and Technology, Harbin Institute of Technology, P.O. Box 410, Harbin 150001, PR China

r t i c l e i n f o

rticle history:eceived 11 June 2012eceived in revised form 6 August 2012ccepted 6 August 2012

a b s t r a c t

Four kinds of polyacrylonitrile-based carbon fibers were cleaned by three methods and were char-acterized by X-ray photoelectron spectroscopy, monofilament tensile strength test and atomic forcemicroscopy (AFM). Experimental results of these tests reveal that the method using supercritical ace-tone or subcritical potassium hydroxide aqueous solution act as the processing medium shows a better

vailable online 14 August 2012

eywords:arbon fiberurface cleaning

cleaning effect compared to the traditional method, Soxhlet extraction with acetone. The method usingsupercritical acetone is more appropriate to wipe off the oxygenated contaminants on carbon fibers’ sur-faces and causes a relatively smaller damage to the bulk strength of each carbon fiber. As far as treatingmethod using the subcritical alkali aqueous solution, it can thoroughly remove silicious contaminants onthe surfaces of treated fibers.

upercritical acetone

ubcritical aqueous

. Introduction

As a kind of important reinforced material and catalyst car-ier, carbon fiber has gained increasing attention from more andore researchers and was investigated extensively in labs all over

he world [1]. When used as experimental subjects, especially forurface modification, manufacture of non-resin matrix compositeaterial and applied as catalyst carrier, epoxy resin coating layer

n the surface of carbon fiber, which formed in commercial man-facture, becomes an interference factor and is intractable to beemoved [2–4].

Presently, prevailing methods to remove epoxy resin coatingayers on the surfaces of carbon fibers include three categories:oxhlet extraction with organic solvent, ultrasonic cleaning inrganic solvent and thermolysis in an inert atmosphere. Each ofhese three methods has its disadvantages, therefore, cannot meethe requirements of experiment research satisfactorily [5]. Nor-

ally, the Soxhlet extraction process cannot remove residual layershoroughly within a reasonable period [6,7]. Because carbon fiberas a significant brittleness, ultrasonic cleaning usually leads to aelatively extreme injury to monofilament tensile strength of thebers. For another traditional cleaning method, the nature of ther-olysis is an elimination process of heteroatom on the surfaces of

reated carbon fibers [8,9]. In fact, it is not a cleaning process anday change surface appearances of the cleaned fibers irretrievably.

∗ Corresponding author. Tel.: +86 451 86414806; fax: +86 451 86221048.E-mail addresses: ydhuang.hit1@yahoo.com.cn, s seele@163.com (Y. Huang).

169-4332/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.apsusc.2012.08.025

© 2012 Elsevier B.V. All rights reserved.

Supercritical fluid has both liquid-like and gas-like characteris-tics and possesses a complete solvency for most gases and organiccompounds. Moreover, a lot of reactions were usually catalyzed byprotons or hydroxyl ions occur in near-critical water without theaddition of acids or bases, simply because of the high ionic prod-uct of water under this condition. Jiang et al. [10] recycled threedifferent PAN based carbon fibers from epoxy resin/carbon fibercomposites using supercritical n-propanol. The tensile strengthand modulus of the recycled carbon fiber was very similar to thecorresponding as-received carbon fibers and the surface oxygenconcentration decreased significantly. Hernanza et al. [11] inves-tigated chemical recycling of carbon fiber reinforced compositesusing subcritical and supercritical alcohols as reactive-extractionmedia. After recycling the produced fibers that retained 85–99% ofthe strength of the virgin fibers.

In this study, we designed two novel methods to clean the sur-faces of polyacrylonitrile-based (PAN-based) carbon fibers, basedon the properties of supercritical and subcritical fluid, such as highdiffusivity and special catalytic action [12,13]. In comparison withthe traditional Soxhlet extraction process with acetone, the twocleaning methods with supercritical acetone and subcritical potas-sium hydroxide solution are proven to be more feasible and showa better cleaning efficiency.

2. Experiment

2.1. Materials

The sources of experimental materials engaged in this study aregiven in Table 1.

416 L. Meng et al. / Applied Surface Science 261 (2012) 415– 421

Table 1The source of experimental materials.

Materials Source of materials

No. 1 carbon fiber T700, Japan TorayNo. 2 carbon fiber T300, Japan TorayNo. 3 carbon fiber GuangWei Group, Weihai, PR ChinaNo. 4 carbon fiber Jilin Carbon Co. Jilin, PR ChinaAcetone The First Factory of Chemical Agents, Tianjin, PR China

2

ba

aafvowf

aibwa

2

EwerasfS

3

3

bas

Table 3The mechanical data and the results of Weibul analysis.

Specimens Ra m �0 Expectation (GPa)

No. 1-I 0.992 −6.23 3.79 4.68No. 1-II 0.993 −6.15 3.81 4.54No. 1-III 0.987 −5.74 3.56 4.52No. 2-I 0.997 −8.29 6.21 3.53No. 2-II 0.998 −7.77 5.73 3.60No. 2-III 0.997 −8.41 6.36 3.50No. 3-I 0.994 −6.48 4.17 4.31No. 3-II 0.994 −7.49 4.76 4.43No. 3-III 0.993 −8.08 5.24 4.30No. 4-I 0.995 −5.03 3.59 3.66

TS

Potassium hydroxide Shanghai Chemical Reagent Co.Deionized water Made in lab

.2. Surface cleaning processes

Method I: each kind of PAN-based carbon fibers were extractedy Soxhlet with acetone for 24 h, then they were dried in a dry ovent 393 K for 24 h.

Method II: each PAN-based carbon fiber was placed into a batchutoclave (95 ml, with 25 ml acetone and 5 ml deionized water)nd sealed. The sealed autoclave was heated by salt-base at 623 Kor 20 min. The purpose of mixing water in this system is to pro-ide a protons environment and augment the dissolving capacityf supercritical acetone for polar groups. After cleaned, the fibersere washed with acetone and then dried in a dry oven at 393 K

or 24 h.Method III: each kind of PAN-based carbon fiber was placed into

batch autoclave (95 ml, with 80 ml 0.1 mol/L potassium hydrox-de aqueous solution) and sealed. The sealed autoclave was heatedy salt-base at 613 K for 20 min. After treated, carbon fibers wereished by deionized water and acetone respectively, then dried in

dry oven at 393 K for 24 h.

.3. Performance testing

The surface composition analysis was performed on a ScientaSCA 300 X-ray photoelectron spectroscopy (XPS) system equippedith a monochromatic Al K� X-ray source (1486.60 eV). The pass

nergy was set at 187.83 and 29.35 eV for the survey and the highesolution spectra, respectively. The instrument was operated withn analyzer chamber pressure of 2.6 × 10−7 Pa at 25 ◦C. Single fibertrength tests were carried out according to ASTM-D3379. Atomicorce microscopy (AFM) images were obtained using an NT-MDTolver P47H system.

. Results and discussion

.1. Surface chemical composition analysis by XPS

Survey spectrograms of twelve specimens (four kinds of PAN-ased carbon fibers cleaned by three methods) are given in Fig. 1,nd the detailed analysis results are shown in Table 2. As can beeen that after cleaning by Soxhlet extraction with acetone for 24 h

able 2urface composition of the treated carbon fibers.

Specimens Composition of C (%) Composition of O (

No. 1-I 64.41 26.42

No. 1-II 87.89 4.94

No. 1-III 86.58 8.80

No. 2-I 82.18 9.73

No. 2-II 88.14 4.30

No. 2-III 88.44 7.18

No. 3-I 59.41 25.79

No. 3-II 79.36 8.97

No. 3-III 79.51 15.86

No. 4-I 82.70 12.00

No. 4-II 92.79 3.50

No. 4-III 94.65 2.69

No. 4-II 0.992 −5.30 3.67 3.82No. 4-III 0.982 −4.96 3.62 3.55

the oxygen content on the surfaces of No. 1 carbon fiber and No.2 carbon fiber are 26.42% and 9.73% respectively. When cleanedwith supercritical acetone act as treated medium, the oxygen con-tent on the surfaces of No. 1 carbon fiber and No. 2 carbon fiber are4.94% and 4.30% respectively. After cleaned by subcritical potas-sium hydroxide aqueous solution such values are 8.80% and 7.18%respectively. These results show that the cleaning method usingsupercritical acetone and subcritical potassium hydroxide aqueousis preferred to Soxhlet extraction with acetone for these two kindsof carbon fibers.

For No. 4 carbon fiber, all of these three methods displayedfavorable cleaning effects. But for No. 3 carbon fiber, cleaning effectof each method is dissatisfactory, but the latter two methods arestill far better than the former one. Further, we hold the opin-ion that the advantage of using supercritical acetone as cleaningmedia rather than using acetone lies in its high diffusivity and goodheat-transporting properties. As to subcritical potassium hydroxideaqueous, this cleaning treatment is more like a pyrogenic decompo-sition and hydrolyzation process, therefore can remove the layerson the surfaces of carbon fibers more thoroughly.

A further contrast can be performed between cleaning processthrough supercritical acetone and subcritical potassium hydroxideaqueous solution. The former is more appropriate to remove oxyge-nous groups on the surfaces of the cleaned carbon fibers. The latterpossesses a better cleaning effect for silicious groups. (After clean-ing by subcritical potassium hydroxide aqueous solution, there isno Si be detected on the surfaces of carbon fibers.)

3.2. Single filament strength tests

Since Soxhlet extraction with acetone has widespreadly beenseen as a feasible cleaning method leading an acceptable strength

loss, another two methods are compared with it. The results ofWeibul analyses of single tensile tests are presented in Fig. 2 andTable 3. It can be seen that the regularity of monofilament ten-sile strength distributions of the carbon fibers basically remains

%) Composition of N (%) Composition of Si (%)

2.97 6.206.01 1.164.61 –3.52 4.575.25 2.314.38 –4.95 9.866.59 5.084.63 –2.94 2.363.36 1.352.67 –

L. Meng et al. / Applied Surface Science 261 (2012) 415– 421 417

1400 120 0 10 00 800 600 40 0 20 0 0

0

500 0

1000 0

1500 0

2000 0

2500 0

3000 0

Inte

nsity

/ cp

s

Energy

0

1000 0

2000 0

3000 0

4000 0

5000 0

Fiber 1 clea ned by method I

Fiber 1 clea ned by me thod II

Inte

nsity

/ cp

s0

500 0

1000 0

1500 0

2000 0

2500 0

Inte

nsity

/ cp

s

Fiber 1 clea ned by me thod II I

1400 1200 1000 800 600 400 200 0

0

5000

10000

15000

20000

25000

30000

Fiber 2 cleaned by metho d I

Fiber 2 cleaned by method II

Fiber 2 cleaned by method III

Inte

nsity

/ cp

sIn

tens

ity /

cps

Inte

nsity

/ cp

sEne rgy

0

10000

20000

30000

40000

50000-50 00

05000

10000150002000025000300003500040000

1400 1200 1000 800 600 400 200 0-500 0

05000

10000150002000025000300003500040000

Fiber 3 cleaned by method I

Fiber 3 cleaned by method II

Fiber 3 clean ed by method II I

Inte

nsity

/ cp

sIn

tens

ity /

cps

Inte

nsity

/ cp

s

0

10000

20000

30000

40000

50000

60000-500 0

05000

100001500020000250003000035000

1400 120 0 100 0 80 0 60 0 40 0 20 0 0

0

100 00

200 00

300 00

400 00

500 00

Inte

nsity

/ cp

sIn

tens

ity /

cps

Inte

nsity

/ cp

s

0

200 00

400 00

600 00

800 00

-100000

100 00200 00300 00400 00500 00600 00700 00

Fiber 4 cleane d by method I

Fiber 4 clea ned by met hod II

Fiber 4 clea ned by metho d II I

s of t

lisad

mckpcs2s

Ene rgy

Fig. 1. Survey spectrogram

evel when treated by various cleaning methods. Such phenomenonndicates that these three methods have not caused serious bulktrength losses. (If serious bulk strength losses taken place, monofil-ment tensile strength distributions should show a significantifference between each others [14].)

According to expectation strengths represented in Table 3, theonofilament tensile strength of each experiment carbon fiber

leaned with supercritical acetone are higher than that of the sameind fibers cleaned by Soxhlet extraction process. Such resultsrove that the former lead to lesser damage to the bulk of the

leaned fibers. As to the method cleaning with subcritical potas-ium hydroxide aqueous solution, such values decrease by about% compared to cleaning by Soxhlet extraction, but this degree oftrength loss is acceptable for the most experimental researches.

Energy

he cleaned carbon fibers.

3.3. Surface appearance analysis

The AFM images of No. 1 carbon fibers cleaned by various meth-ods are given in Fig. 3. There is more and clearer grain perpendicularto the carbon fiber’s axis can be seen on the surfaces of the fibersthat were cleaned by supercritical acetone or subcritical potassiumhydroxide aqueous solution. This phenomenon indicates that thesetwo methods have better cleaning effects for No. 1 carbon fiber.

The AFM images of No. 2 carbon fiber cleaned by various meth-ods are given in Fig. 4. It can be seen that the grooves on the surfaces

of No. 2 carbon fiber are covered by residual epoxy resin layerseven after cleaning by Soxhlet extraction with acetone. Therefore,such grooves clearly appear in Fig. 4(b). Such phenomenon indi-cates that it is appropriate for No. 2 carbon fibers to be cleaned with

418 L. Meng et al. / Applied Surface Science 261 (2012) 415– 421

2.42.01.61.20.80.40.0

-6

-4

-2

0

2

ln ( sf )

ln ln

[ 1

/ ( 1

- F

( sf )

]

Fiber 1 cleaned by method I

Fiber 1 cleaned by method II

Fiber 1 cleaned by method III

2.42.01.61.20.80.40.0

-6

-4

-2

0

2

ln ( sf )

ln ln

[ 1

/ ( 1

- F

( sf )

]

Fiber 2 cleaned by method I

Fiber 2 cleaned by method II

Fiber 2 cleaned by method III

2.42.01.61.20.80.40.0

-6

-4

-2

0

2

ln ( sf )

ln ln

[ 1

/ ( 1

- F

( sf )

]

Fiber 3 cleaned by method I

Fiber 3 cleaned by method II

Fiber 3 cleaned by method III

2.42.01.61.20.80.40.0

-6

-4

-2

0

2

ln ( sf )

ln ln

[ 1

/ ( 1

- F

( sf )

]

Fiber 4 cleaned by method I

Fiber 4 cleaned by method II

Fiber 4 cleaned by method III

Fig. 2. Weibul analyses of the cleaned carbon fibers’ single filament strength.

Fig. 3. The AFM images of No. 1 fibers: (a) cleaned by Soxhlet extraction with acetone; (b) cleaned by supercritical acetone; (c) cleaned by subcritical KOH solution.

L. Meng et al. / Applied Surface Science 261 (2012) 415– 421 419

Fig. 4. The AFM images of No. 2 fibers: (a) cleaned by Soxhlet extraction with acetone; (b) cleaned by supercritical acetone;(c) cleaned by subcritical KOH solution.

Fig. 5. The AFM images of No. 3 fibers: (a) cleaned by Soxhlet extraction with acetone; (b) cleaned by supercritical acetone; (c) cleaned by subcritical KOH solution.

420 L. Meng et al. / Applied Surface Science 261 (2012) 415– 421

ceton

sohfi

oabt

omt

4

rseukachfirfi

A

N

[

Fig. 6. The AFM images of No. 4 fibers: (a) cleaned by Soxhlet extraction with a

upercritical acetone. In Fig. 4(c), there are many irregular residuesn the surfaces of the fibers which cleaned by subcritical potassiumydroxide aqueous solution. So, it is unsatisfactory for No. 2 carbonbers cleaned with subcritical potassium hydroxide aqueous.

The AFM images of No. 3 carbon fiber cleaned by various meth-ds are given in Fig. 5. According to the images, using supercriticalcetone and subcritical potassium hydroxide aqueous solution isetter than using Soxhlet extraction with acetone. However, all ofhem are unsatisfactory.

The AFM images of No. 4 carbon fiber cleaned by various meth-ds are given in Fig. 6. According to the images, all of these threeethods can clean the surfaces of No. 4 carbon fiber thoroughly, and

hese results are accordance with the information of XPS analyses.

. Conclusion

In order to clean the surfaces of PAN-based carbon fibers foresearches in labs, using supercritical acetone or subcritical potas-ium hydroxide aqueous solution act as cleaning medium is morefficient than Soxhlet extraction with acetone. For the methodsing supercritical acetone, it leads to smaller damage to eachind of carbon fibers when treated, and is particularly appropri-te for removing epoxy resin coating layers on the surfaces ofarbon fibers. Comparatively, cleaning with subcritical potassiumydroxide aqueous solution may cause more serious losses of singlelament strength to the cleaned fibers, but using this method canemove silicious contaminants thoroughly for the cleaned carbonbers.

cknowledgement

The authors gratefully acknowledge financial support from theational Natural Science Foundation of China (no. 21174034).

[

e; (b) cleaned by supercritical acetone; (c) cleaned by subcritical KOH solution.

References

[1] H.J. Li, H.B. Ouyang, L.H. Qi, Y.L. Zhang, Z.J. Li, J.F. Wei, Effect of temperature onthe synthesis of SiC coating on carbon fibers by the reaction of SiO with thedeposited pyrolytic carbon layer, Journal of Materials Science & Technology 26(2010) 211–216.

[2] H. Qian, A. Bismarck, E.S. Greenhalgh, M.S.P. Shaffer, Carbon nan-otube grafted carbon fibres: a study of wetting and fibre fragmenta-tion, Composites Part A: Applied Science and Manufacturing 41 (2010)1107–1114.

[3] T.D. Hapuarachchi, T. Peijs, Multiwalled carbon nanotubes and sepiolite nan-oclays as flame retardants for polylactide and its natural fibre reinforcedcomposites, Composites Part A: Applied Science and Manufacturing 41 (2010)954–963.

[4] C. Qiang, J. Xu, Z. Zhang, L. Tian, S. Xiao, Y. Liu, P. Xu, Magneticproperties and microwave absorption properties of carbon fibers coatedby Fe3O4 nanoparticles, Journal of Alloys and Compounds 506 (2010)93–97.

[5] H. Guo, Y. Huang, L. Liu, X. Shi, Effect of epoxy coatings on carbon fibers duringmanufacture of carbon fiber reinforced resin matrix composites, Materials 31(2010) 1186–1190.

[6] Z.W. Xu, Y.D. Huang, C.Y. Min, L. Chen, L. Chen, Effect of gamma-ray radiationon the polyacrylonitrile based carbon fibers, Radiation Physics and Chemistry79 (2010) 839–843.

[7] Y.P. Tang, L. Liu, W.W. Li, B. Shen, W.B. Hu, Interface characteristicsand mechanical properties of short carbon fibers/Al compos-ites with different coatings, Applied Surface Science 255 (2009)4393–4400.

[8] V. Gandikota, G.F. Jones, A.S. Fleischer, Thermal performance of a carbon fibercomposite material heat sink in an FC-72 thermosyphon, Experimental Thermaland Fluid Science 34 (2010) 554–561.

[9] C. Pradere, J.C. Batsale, J.M. Goyheneche, R. Pailler, S. Delhaire, Thermalproperties of carbon fibers at very high temperature, Carbon 47 (2009)737–743.

10] G. Jiang, S.J. Pickering, E.H. Lester, T.A. Turner, K.H. Wonga, N.A. Warrior, Char-acterisation of carbon fibres recycled from carbon fibre/epoxy resin compositesusing supercritical n-propanol, Composites Science and Technology 69 (2009)

192–198.

11] R.P. Hernanza, J.G. Serna, C. Dodds, J. Hyde, M. Poliako, et al., Chemi-cal recycling of carbon fibre composites using alcohols under subcriticaland supercritical conditions, The Journal of Supercritical Fluids 46 (2008)83–92.

ace Sc

[

L. Meng et al. / Applied Surf

12] T. Yanagida, T. Minowa, Y. Shimizu, Y. Matsumura, Y. Noda, Recovery of acti-vated carbon catalyst, calcium, nitrogen and phosphate from effluent followingsupercritical water gasification of poultry manure, Bioresource Technology 100(2009) 4884–4886.

[

[

ience 261 (2012) 415– 421 421

13] M. Goto, Chemical recycling of plastics using sub- and supercritical fluids, TheJournal of Supercritical Fluids 47 (2009) 500–507.

14] Y.X. Zhou, Y. Wang, Y.M. Xia, S. Jeelani, Tensile behavior of carbon fiber bundlesat different strain rates, Materials Letters 64 (2010) 246–248.