A rapid and simple method to draw polyethylene nanofibers with enhanced thermal conductivity

Thermal conductivity of polymers is an important thermal property for both polymer applications and processing. Bulk polymers are generally considered as thermal insulators, and typically have thermal conductivities on the order of 0.1 W m⁻¹ K⁻¹ near room temperature.^1–6 The low thermal conductivity of polymers comes from their structure, with numerous molecular chains coiled up and entangled together. However, in applications, such as in electronic packaging and encapsulations, satellite devices, and in areas where good heat dissipation, low thermal expansion, and light weight are needed, polymers with high thermal conductivity are preferred.

A common method to enhance a polymer's thermal conductivity is to compose it with high thermal conductivity materials such as metal,⁷ ceramic particles,⁸ or other polymer.² It has been reported that with the help of fillers, the thermal conductivity of polymers can be increased to 1 to 10 W m⁻¹ K⁻¹.² However, adding a large amount of fillers to polymers not only significantly increase the material cost but also may deteriorate other properties of the polymer, such as electrical and optical properties.

Drawing polymeric fibers to increase chain alignment and crystallinity is another common method to increase thermal conductivity of polymer. The drawing process can be implemented via both electrospinning^9,10 and mechanical stretching.^3,11,12 Electrospinning is a favorable route for producing nanoscale polymeric fibers for applications in various fields, such as tissue engineering, drug delivery systems, optoelectronics, and filtration techniques.¹³ A recent study showed that the thermal conductivity of single Nylon-11 electrospun fibers could be as high as 1.6 W m⁻¹ K⁻¹, nearly one order of magnitude higher than the typical Nylon-11 bulk value of around 0.2 W m⁻¹ K⁻¹.⁹ Canetta et al. measured the thermal conductivity of individual polystyrene nanofibers electrospun at 7–10 kV,¹⁰ and the thermal conductivity of the measured nanofibers ranges from 6.6 to 14.4 W m⁻¹ K⁻¹, a significant increase compared to the typical bulk value for polystyrene. In our previous work, polyethylene (PE) nanofibers electrospun at 45 kV and 150 mm needle-collector distance could have a thermal conductivity up to 9.3 W m⁻¹ K⁻¹, which is over 20 times higher than the typical bulk value.¹⁴ Although electrospinning has been successfully used in preparing many kinds of polymer fibers with high thermal conductivity, it has some disadvantages. For example, relatively high electrical conductivity of polymer solvent is required and electrospinning can only be operated at high voltage of several kV. Also, there are many polymer materials cannot be used for electrospinning or the electrospinning setup is too complicated.

Mechanical stretching^3,11,12 is another drawing process to prepare polymer fibers. Shen et al. reported a method of mechanical stretching to fabricate high-quality ultra-drawn polyethylene nanofibers from the heated polyethylene gel.³ In their drawing process, the fibers were drawn from the heated gel using either a sharp tungsten tip (diameter ∼100 nm) or a tipless atomic force microscope (AFM) cantilever. The largest overdraw ratio was 800. The thermal conductivity of these ultra-drawn nanofibers is as high as ∼104 W m⁻¹ K⁻¹, larger than the conductivities about half of the pure metals and all the electrospun polymer fibers reported. Shen et al. attributed the unexpectedly high thermal conductivity to the restructuring of the polymer chains by stretching, which improved the fiber quality toward an “ideal” single crystalline fiber. Compared to electrospinning method, the mechanical stretching method proposed by Shen et al. needs a more skillful and sensitive treatment.

In this paper, we report on a rapid and simple stretching method of directly drawing PE nanofiber from PE solution. We measured the thermal conductivity of drawn PE nanofibers with different diameters. The highest thermal conductivity of the drawn PE nanofiber is comparable to the electrospun PE nanofiber.¹⁴ We used micro-Raman spectroscopy to characterize the microstructure change during the PE nanofiber drawing process. Raman results suggest that the simple directly drawing process can improve the molecular orientation and crystallinity of the drawn PE nanofibers to the level of that of electrospun fibers.

The direct drawing technique used to fabricate PE nanofibers is described in Figure 1. First, PE powder of high molecular weight (3–6 × 10⁶) is dissolved in p-xylene solvent (all materials were bought from Sigma-Aldrich, Inc.). P-xylene is a well-known solvent for PE.^15–17 The boiling temperature of P-xylene solvent is about 411 K while the dissolution temperature for PE powder is about 393 K. Therefore, we chose to prepare PE solution at around 400 K by stirring it until the PE powder completely dissolved and kept it around 400 K during the nanofiber drawing. To draw a nanofiber, a 28 G metal syringe needle was inserted into the PE solution (about 2 mm inserted) and draw out from the solution quickly. During the drawing process, the solvent evaporated and a PE nanofiber extend between the tip of the needle and the solution. The PE nanofibers were collected on a piece of PDMS for further transfer.

We tried three different concentrations of PE solution: 0.1%, 0.5%, and 1% (wt. %). At 0.1% concentration, the viscosity of the PE solution is too low to form continuous fibers. When the concentration is increased to 0.5%, the viscosity becomes higher and nanofibers can be drawn. The diameter of nanofibers drawn varies from 40 nm to 500 nm. At 1% concentration, the viscosity of PE solution is very high. Although it is easy to draw nanofibers from this concentration, the diameters of the drawn fibers are usually on the order of a few hundred nanometers. The sample we measured in this study was drawn from 0.5% PE concentration. During the drawing process, the rate was controlled by hand. We found if the fiber was drawn immediately (about 200 mm/ms), the morphology of the fiber would be uniform and tinny. While if it was drawn slowly (about 200 mm/s), the morphology of the fiber would be rough, thick, and the length of the fiber usually was short.

The thermal conductivity of individual nanofibers^18,19 was measured with a suspended microdevice consisting of two adjacent suspended SiN_x membranes integrated with Pt resistance heaters/thermometers for transport property measurements of individual one-dimensional (1D) nanostructures. Each resistance heater/thermometer is electrically connected to four contact pads by Pt lines on the suspended beams, enabling electrical joule heating and four-point measurement of its electrical resistance to determine the temperatures of the membranes. With the help of a home-built micromanipulator, the drawn fibers collected on PDMS were cut and transferred to a suspended microdevice, as shown in Figure 2(a). The microdevice was placed in a high vacuum cryostat (<10⁻⁶ mbar) to eliminate contribution of convective heat transfer to the measured conductance. As we reported in a previous study,¹⁴ solvent may remain within the drawn fibers, and be evaporated quickly in such a high vacuum. In the vacuum chamber, we installed two radiation shields with one directly mounted on the high temperature stage to minimize the radiation loss from the suspended membranes to surroundings.²⁰ The thermal conductivity of each fiber was measured in a temperature range of 100–320 K. This upper temperature limit (320 K) was chosen to avoid potential annealing effects. It should be pointed out that no contact treatment was done to reduce the contact thermal resistance between the measured fiber and the heat source/sink. Thus, the results presented here represent a lower limit for the true thermal conductivity. However, it has been demonstrated in our previous work¹⁴ that the PE nanofiber can attach onto the membranes steadily, the contact thermal resistance is negligible compared to the intrinsic thermal resistance of the measured PE fiber.

The quality of the measured sample was examined with Raman spectroscopy. The Raman spectra were collected at room temperature with the use of 10 mW of radiation at 532 nm (from LaserQuantum) at the sample as shown in Figure 2(b). The spectra were accumulated for 1 min and taken with a slit width equivalent to 1.5 cm⁻¹ resolution with LabRam HR800 system. Raman spectroscopy has been widely used to characterize PE and the assignments of the main Raman bands are well known,^21–27 and are given in detail in Table I. The band between 1000–1600 cm⁻¹ are frequently used to study the morphological structure of PE and it can be divided in the following vibrational areas:²⁸ the C-C stretching between 1000 and 1200 cm⁻¹, which is sensitive to molecular orientation, stress, and conformation; the -CH₂- twisting vibrations around 1295 cm⁻¹, which can be used as an internal standard; and the -CH₂- bending modes between 1400 and 1470 cm⁻¹, which is sensitive to chain packing (the 1415 cm⁻¹ band is assigned to orthorhombic crystallinity). We used this technique to characterize the ultra-high molecular weight PE powder and nanofibers of different diameters.

Figure 3(a) shows the measured thermal conductivity of drawn PE nanofibers. The thermal conductivity shows a diameter dependence, and increases as the diameter decreases. At room temperature, the thermal conductivity of 148 nm diameter PE nanofiber is only 0.8 W m⁻¹ K⁻¹, and the thermal conductivity of 135 nm diameter sample is about 1.1 W m⁻¹ K⁻¹. As the diameter decreased to 120 nm, its thermal conductivity increased to 3.7 W m⁻¹ K⁻¹, and the thermal conductivity of 100 nm diameter is 3.8 W m⁻¹ K⁻¹. These two nanofibers of similar diameters give similar thermal conductivities, which indicates that the quality of the nanofibers directly drawn from solution is stable. The thermal conductivity of 40 nm diameter sample is 8.8 W m⁻¹ K⁻¹ at room temperature, which is about 20 times higher than the typical bulk value. In our previous work,¹⁴ we have investigated the thermal conductivity of PE nanofibers electrospun from 9 kV to 45 kV. The highest thermal conductivity we obtained from electrospun PE nanofibers is also given in Figure 3(a) (Hollow symbol). This fiber was electrospun at 45 kV and has a diameter of 53 nm. It can be seen that the highest thermal conductivity of drawn PE nanofiber is almost the same as the highest thermal conductivity of electrospun PE nanofiber. It should be noted that thermal conductivity of electrospun nanofiber does not show a clear diameter dependence. These fibers were electrospun at a large voltage range. The higher electrospinning voltage leads to stronger “whipping instability,” and hence large variations in the actual electric field the jet experiences during the fabrication process, which likely overshadow any dependence on the fiber diameter.¹⁴ From the comparison, we can see that this rapid and simple drawing method also can effectively enhance the thermal conductivity of PE nanofiber.

The bulk polyethylene is of low thermal conductivity, mainly due to its randomly oriented molecular chains. It is widely accepted that the enhanced thermal conductivity of stretched polymer films and fibers can be attributed to the improvement in chain alignment and crystallinity.^{3,9,10,15,29–31} In PE powders, some PE molecular chains are folded locally in the form of small crystallites, usually termed as “lamellae,”^15,29 while other molecular chains are randomly distributed as shown in Figure 3(b). Figure 3(c) shows, when the PE powder completely dissolved in the solvent, both lamellae and entangled molecular chains extended and oriented randomly. During the drawing process, the random molecular chains will be aligned in an ordered fashion to form a nanofiber as shown in Figure 3(d).

A qualitative index of the molecular orientation in PE is the ratio of the 1130 cm⁻¹ and 1060 cm⁻¹ Raman bands (I₁₁₃₀/I₁₀₆₀).^27,28,32 The bands at 1130 and 1060 cm⁻¹ have different vibrational symmetries. The 1130 cm⁻¹ Raman band is thought to arise from the C-C symmetric stretching of the all-trans PE chain segments while the 1060 cm⁻¹ band is due to the C-C antisymmetric stretching. If the molecules are oriented in a preferred direction, the 1130 cm⁻¹ band becomes stronger with respect to the 1060 cm⁻¹ band.

Raman spectra study of the measured nanofibers was performed using a Horiba LabRam HR800 system, as shown in Figure 4(a). It can be seen that as the diameter decreases, the band intensity ratio of the 1130/1060 bands increases. I₁₁₃₀/I₁₀₆₀ is 1.16 for PE powder, and increased to 5.14 for the 40 nm diameter PE nanofiber. These two bands have different vibrational symmetries. The 1130 cm⁻¹ Raman band is thought to arise from the C-C symmetric stretching of the all-trans PE chain segments while the 1060 cm⁻¹ band is due to the two degenerated C-C antisymmetric stretching. If the molecules are oriented in the preferred direction, the 1130 cm⁻¹ band has been reported to become stronger with respect to the 1060 cm⁻¹ band.¹⁴ As we mentioned above, thermal conductivity of the 40 nm diameter PE nanofiber prepared from direct drawing is very close to that of the 53 nm diameter PE nanofiber prepared from electrospun reported in Ref. 14, as shown in Figure 3(a). It is interesting to see that the values of I₁₁₃₀/I₁₀₆₀ obtained on these two samples are also very close, and this confirms that thermal conductivity enhancement of polymer nanofibers is indeed due to the improvement molecular chain alignment. From Figure 4(b), we can see the thermal conductivity growth dramatically as the orientation improved.

The crystalline phase of PE is primarily orthorhombic.^26–28,33 The band at 1416 cm⁻¹ has been unanimously assigned to the orthorhombic crystalline phase. According to the Raman spectra of PE nanofibers (Figure 4(a)), it is clear that the 1416 cm⁻¹ band, or the orthorhombic crystallinity, becomes stronger compared with the PE powder. The structural characterization confirmed that the drawing process could lead to better aligned molecular chains and improved crystallinity, which contributes to the enhanced thermal conductivity along the axial direction of the nanofibers. Meanwhile, smaller fibers tend to possess better alignment and crystallinity than the ones of larger diameter.

In this study, PE nanofibers with diameter down to 40 nm have been drawn directly from PE solution in a rapid and simple manner. Results show that drawn PE nanofibers exhibit thermal conductivity significantly higher than the bulk values (∼0.4 W m⁻¹ K⁻¹), with the highest value measured as 8.8 W m⁻¹ K⁻¹, over 20 times higher than the typical bulk value. This thermal conductivity enhancement is attributed to the higher degree of molecular orientation and enhanced level of crystallinity, as evidenced by the micro-Raman spectroscopy characterization. Although this direct drawing method currently have not achieved ultra-high thermal conductivity in PE nanofibers, it expects to be applicable to enhance the thermal conductivity of many other materials fibers like polyvinylidene fluoride (PVDF), polyethylene oxide (PEO), polytrimethylene terephthalate (PTT), etc.

The authors thank the Natural Science Foundation of China (Nos. 51176032 and 51375089) and Natural Science Foundation of Jiangsu Province (BK20150636) for the financial support. Qian Zhang appreciates the financial support from the U.S. National Science Foundation (Grant No. CMMI-1462866).