APPLIED PHYSICS LETTERS 88, 083119 �2006�

Enhancement of strength and stiffness of Nylon 6 filamentsthrough carbon nanotubes reinforcement

Hassan Mahfuza�

Department of Ocean Engineering, Florida Atlantic University, Boca Raton, Florida 33431

Ashfaq AdnanSchool of Aeronautics and Astronautics, Purdue University, West Lafayette, Indiana 47907

Vijay K. Rangari, Mohammad M. Hasan, and Shaik JeelaniTuskegee University’s Center for Advanced Materials (T-CAM), Tuskegee, Alabama 36088

Wendelin J. WrightMaterials Science and Engineering Department, Stanford University, Stanford, California 94305

Steven J. DeTeresaUniversity of California, Lawrence Livermore National Laboratory, 7000 East Ave., Livermore,California 94550

�Received 7 July 2005; accepted 1 February 2006; published online 24 February 2006�

We report a method to fabricate carbon nanotube reinforced Nylon filaments through an extrusionprocess. In this process, Nylon 6 and multiwalled carbon nanotubes �MWCNT� are first dry mixedand then extruded in the form of continuous filaments by a single screw extrusion method. Thermogravimetric analysis �TGA� and differential scanning calorimetry �DSC� studies have indicated thatthere is a moderate increase in Tg without a discernible shift in the melting endotherm. Tensile testson single filaments have demonstrated that Young’s modulus and strength of the nanophasedfilaments have increased by 220% and 164%, respectively with the addition of only 1 wt. %MWCNTs. SEM studies and micromechanics based calculations have shown that the alignment ofMWCNTs in the filaments, and high interfacial shear strength between the matrix and the nanotubereinforcement was responsible for such a dramatic improvement in properties. © 2006 AmericanInstitute of Physics. �DOI: 10.1063/1.2179132�

In order to utilize the extraordinary strength and stiffnessof carbon nanotubes �usually in the range of 200–900 MPaand 200–1000 GPa, respectively� in bulk materials, severalresearchers have attempted to align CNTs along preferreddirections.1–6 Almost all of these attempts were made withthe infusion of single or multiwalled CNTs, or vapor growncarbon nanofibers �CNF� into thermoplastic polymers. Infu-sion was carried out either through a liquid route using soni-cation, or a dry route followed by melt mixing in an extruder.Alignment of CNTs or CNFs in the composite was enforcedby extrusion or spinning, and was accompanied by subse-quent stretching. The resulting composites either in consoli-dated or filament form have no doubt demonstrated im-proved mechanical and thermal properties. However, thefullest potential of the nanoparticle reinforcements could notbe harnessed primarily because of the lack of alignment orfailure to develop strong interfacial bonding between nano-tubes and the polymer.

In the present investigation, we dry mixed MWCNTswith Nylon 6 powder. The amount of nanoparticle loadingwas restricted to 1 wt. % beyond which agglomeration pro-hibited the continuous drawing of filaments. The dry-mixedpowder was then melted in a single screw extruder whichwas followed by distributive mixing, extrusion, stretching,and heat stabilization to continuously draw MWCNT-reinforced filaments. In parallel, control filaments were alsoextruded following identical procedures. The dry mixing ofNylon powder with MWCNTs was performed in a mechani-

a�Author to whom correspondence should be addressed; electronic mail:

hmahfuz@oe.fau.edu

0003-6951/2006/88�8�/083119/3/$23.00 88, 08311Downloaded 05 Mar 2009 to 128.100.48.236. Redistribution subject to

cal blender for 3 h. The MWCNTs were obtained from MERCorporation �7960 South Kolb Road, Tucson, Arizona85706�. The nanotubes were 10–15 nm in diameter and ap-proximately 5 �m long, with 5–20 graphitic layers. TheMWCNTs had a tube-rich core surrounded by a fused carbonshell. The core volume ranged from 10% to 25%. The distri-bution of MWCNTs in the as-received materials was around40%. Commercial grade Nylon 6 was procured from NymaxPolyone. A mechanical crusher was used to produce micron-sized powders of Nylon which were then used in the me-chanical blender for mixing with MWCNT. The density ofNylon 6 was 1.14 g/cm3 with a melting point of 215 °C. Inorder to eliminate moisture, the Nylon and CNT mixture wasplaced into a cylindrical drying chamber. Hot air was sup-plied to the chamber through an insulated flexible tube usinga vortex blower. The dryer was operated for 24 h with tem-perature set at 90 °C. Prolonged heating accompanied with avortex flow broke up large agglomerates of CNTs if any leftafter the mechanical blending. Once the mixture was dried, itwas extruded through a Wayne Yellow Label Table Top Ex-truder. Five thermostatically controlled heating zones wereused to melt the mixture prior to extrusion. The die zoneconsisted of a circular plate, a 10 cm long steel tubing withan inner diameter of 4 mm, and the die itself. A distributivemixing of the CNTs within the Nylon was enforced throughthe use of the circular plate with multiple orifices. A speciallydesigned die was used in the process. The die configurationgenerated two distinct flow regimes that significantly af-fected the nanotube alignment. After extrusion, filamentswere solidified by passing them through chilled water main-

tained at approximately 10 °C. In the next step, filaments

© 2006 American Institute of Physics9-1 AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

083119-2 Mahfuz et al. Appl. Phys. Lett. 88, 083119 �2006�

were stabilized by using a Wayne Yellow Jacket Stabilizingunit. The heater temperature was set at 110 °C, and the fila-ment travel per minute �FPM� was adjusted in the Godetstations to allow continuous drawing of filaments. Finally,the filaments were wound on a spool using a filament winderat a winding speed of 70 rpm. Several of these spools wereproduced.

The thermal stability of the neat and nanophased fila-ments was characterized by TGA. The weight loss curves asa function of temperature is shown in Fig. 1. In the presentstudy, 50% of the total weight loss was considered as thethermal decomposition temperature of the system which usu-ally coincides with the peak of the derivative curves indi-cated in Fig. 1. It is observed in Fig. 1 that the thermaldecomposition temperature of MWCNT-Nylon is around493 °C, whereas that of neat Nylon is only 459 °C. Thisclearly demonstrates that the MWCNT-Nylon system is ther-mally more stable than the corresponding neat Nylon system.We have also observed similar trends with othernanoparticle-reinforced polymers.7–11

Results from DSC are tabulated in Table I. DSC studies

FIG. 1. TGA curves of �a� 1 wt. % MWCNT-Nylon and �b� neat Nylon.

TABLE I. DSC results.

Type of sampleTg

�°C�Tm

�°C��Hf

�J/g�Crystallinity

�%�

Neat Nylon 6 43.92 221.36 54.27 23.601% MWCNT Nylon 6 49.11 221.67 59.142 25.71

Downloaded 05 Mar 2009 to 128.100.48.236. Redistribution subject to

allowed determination of glass transition temperatures �Tg�,melting temperatures �Tm�, and heat of fusion ��Hf�. It isobserved in Table I that the glass transition temperature ofextruded Nylon �neat� has increased from 44 °C to 49 °Cwith the infusion of MWCNTs. This increase in Tg is evi-dently due to the presence of MWCNTs which may haveimposed restrictions on molecular mobility at an earlierstage. This effect can also be understood in terms of decreas-ing free volume of polymer. From the concept of free vol-ume, it is known that in the liquid state when free space ishigh, molecular motion is relatively easy because of the un-occupied volume. As the temperature of the melt is lowered,the free volume would be reduced until there would not beenough free space to allow molecular motion or translation.With the infusion of MWCNTs, this free space is evidentlyfurther reduced.

It is also seen in Table I that there is no change in themelting temperature of the two systems. This is confirmed bythe close crystallinity of the neat and nanophased systems.The degrees of crystallinity as calculated from melting en-dotherms and heats of fusion are 24% and 26%, respectively.This is somewhat different than what was found with linearlow density polyethylene �LLDPE� infused with MWCNTs.5

Tensile tests of individual filaments were carried out toassess the increase in strength and stiffness of thenanophased Nylon. Tests were performed in a regular MTSmachine by taking machine compliance into account as perASTM standard D3379-75.12 The Young’s modulii of neatand nanophased filaments determined in this manner werefound to be 1.1 GPa and 3.6 GPa, respectively.

Figure 2 shows the stress-strain curves for samples ofneat Nylon 6 and MWCNT-Nylon with 102 mm gagelengths. The 2% offset yield strength of the neat Nylon 6 was35 MPa. The ultimate tensile strength was 125 MPa, and thestrain to failure varied from 300% for the longest samples�102 mm gage length� to 800% for the shortest samples�13 mm gage length�. With the limited displacement travel ofthe testing system, it was not possible to fail samples with203 mm gage lengths.

The 2% offset yield strength of Nylon 6 with 1 wt. %MWCNTs was 95 MPa. The ultimate tensile strength andstrain to failure ranged from 330 MPa at 19% strain for thelongest samples �102 mm gage length� to 410 MPa at 48%strain for the shortest samples �13 mm gage length�. Presum-

FIG. 2. Engineering stress vs engineering strain for neat Nylon, and Nylonwith 1 wt. % multiwalled carbon nanotubes with 102 mm gage length.

ably the trend of increasing ultimate tensile strength with AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp

083119-3 Mahfuz et al. Appl. Phys. Lett. 88, 083119 �2006�

decreasing gage length exhibited by both materials is due toa statistical distribution of flaws. The probability of samplinga flaw of sufficient size to precipitate failure increases withgage length, thereby decreasing the ultimate tensile strengthand strain to failure. Note that for an addition of only1 wt. % MWCNTs, the strength of the composite material ismore than twice that of the unreinforced Nylon which indi-cates the exceptional load-bearing capability of CNTs andtheir potential as structural materials.

The high value of Young’s modulus reported here for theMWCNT reinforced Nylon 6 �3.6 GPa� is expected given thehigh degree of alignment of the MWCNT as shown in Fig. 3;however, the percent increase in strength of the compositematerial relative to the neat Nylon 6 ��150% � is surprisinggiven previously reported results for nanotube reinforcedpolymers and the low percentage of nanotube addition.13–16,4

This large increase in strength suggests that this particularmaterial has exceptionally high interfacial shear strength be-tween the matrix and the nanotube reinforcement, causing anefficient transfer of stress from the matrix to the fiber. Usingthe method of Cox,17,18 the interfacial shear stress �i along afiber is given by

�i =n�1

2Efiber sinh�nx

r�sech�ns� , �1�

where �1 is the strain in the matrix which is approximated asthe overall composite strain, x is the position along the lengthof the fiber relative to the midpoint, r is the fiber radius, s isthe fiber aspect ratio, and n is a dimensionless constant givenby

n = � 2Ematrix

Efiber�1 + �matrix�ln�1/f��1/2

, �2�

where �matrix is the Poisson’s ratio of the matrix material.One can estimate the effective nanotube volume fraction �f�by considering the density of graphite �2.25 gm/cm3�, 10%–25% core volume, and 40% purity of MWCNT in the1 wt. % loading fraction. With such calculated values of �f�,and nanotube Young’s modulus �Efiber1.0 TPa�, the shearstress sustained at the interface between the matrix and fiberat the yield point of the composite �as determined by a de-viation from linearity rather than an arbitrary offset strain�

FIG. 3. FE-SEM micrograph showing alignment of MWCNTs in ananophased filament.

Downloaded 05 Mar 2009 to 128.100.48.236. Redistribution subject to

ranges from 70 to 115 MPa. Since yielding of the compositeis likely due to yielding of the matrix material and not due tofiber pull-out �given the relatively low yield strength of neatNylon 6�, this range of stress values represents a lowerbound estimation of the interfacial shear strength. These highinterfacial stress values indicate the high strength of thebonding between the multiwalled carbon nanotubes and theNylon 6 matrix.14 The interface strength and the elasticmodus values of the composite shown in the paper are alsoconsistent with the predictions given in Ref. 19. In this ref-erence it is shown that at the lowest volume fraction��1% � of CNTs, and at an interface strength of around100 MPa, the elastic modulus of the composite is approxi-mately 3.2 GPa. In our case, the effective MWCNT volumefraction �f� was around 0.2%, calculated interface strengthwas between 70–115 MPa, and the corresponding compositeYoung’s modulus obtained was 3.6 GPa.

In conclusion, the methodology described here to pro-duce MWCNT infused Nylon filaments is vastly differentfrom what is described elsewhere in the literature in a varietyof ways, including melt extrusion and thermal stabilizationprocedures. The ultimate effect was in the spectacular en-hancement of the mechanical properties of the nanophasedfilaments. The source of this improvement is attributed to thesuccessful alignment of nanotubes, and the interfacial shearstrength developed during the fabrication process.

The authors would like to thank the Office of NavalResearch �Grant No. N00014-90-J-11995� and the NationalScience Foundation �Grant No. HRD-976871� for supportingthis research.

1L. Jin, C. Bower, and O. Zhou, Appl. Phys. Lett. 73, 1197 �1998�.2C. A. Cooper, D. Ravich, D. Lips, J. Mayer, and H. D. Wagner, Compos.Sci. Technol. 62, 1105 �2002�.

3E. T. Thostenson and T. W. Chou, J. Phys. D 35, L77 �2002�.4J. Gao, M. E. Itkis, A. Yu, E. Bekyarova, B. Zhao, and R. C. Haddon, J.Am. Chem. Soc. 127, 3847 �2005�.

5H. Mahfuz, A. Adnan, V. Rangari, and S. Jeelani, Int. J. Nanosci. 4, 55�2005�.

6R. Haggenmueller, H. H. Gommans, A. G. Rinzler, J. E. Fisher, and K. I.Winey, Chem. Phys. Lett. 330, 219 �2002�.

7H. Mahfuz, M. S. Islam, V. K. Rangari, and S. Jeelani, Composites, Part B35, 543 �2004�.

8H. Mahfuz, V. K. Rangari, M. S. Islam, and S. Jeelani, Composites, Part B35, 453 �2004�.

9N. Chisholm, H. Mahfuz, V. K. Rangari, A. Adnan, and S. Jeelani, Com-pos. Struct. 67, 115 �2005�.

10R. Rodgers, H. Mahfuz, V. K. Rangari, N. Chisholm, and S. Jeelani, Mac-romol. Mater. Eng. 290, 423 �2005�.

11H. Mahfuz, M. A. Baseer, V. K. Rangari, and S. Jeelani, SAMPE J. 41, 40�2005�.

12American Society for Testing Materials �ASTM� Standard D3379-75,“Tensile Strength and Young’s Modulus for High-Modulus Single-Filament Materials.”

13O. Lourie, D. M. Cox, and H. D. Wagner, Phys. Rev. Lett. 81, 1638�1998�.

14H. D. Wagner, O. Lourie, Y. Feldman, and R. Tenne, Appl. Phys. Lett. 72,188 �1998�.

15D. Qian, E. C. Dickey, R. Andrews, and T. Rantell, Appl. Phys. Lett. 76,2868 �2000�.

16L. S. Schadler, S. C. Giannaris, and P. M. Ajayan, Appl. Phys. Lett. 73,3842 �1998�.

17H. L. Cox, Br. J. Appl. Phys. 3, 72 �1952�.18D. Hull and T. W. Clyne, An Introduction to Composite Materials �Cam-

bridge University Press, Cambridge, 1996�.19S. Namilae and N. Chandra, ASME J. Eng. Mater. Technol. 127, 222

�2005�.

AIP license or copyright; see http://apl.aip.org/apl/copyright.jsp