1 Introduction

Carbon nanotubes (CNTs) have drawn tremendous attentions from various areas of science and engineering arising from their unique structure, unprecedented mechanical, electrical, and thermal performance [1,2,3]. One significant application for CNTs is the fabrication of conductive polymer nanocomposites (CPCs) [4, 5], in which polymers act as the matrix and CNTs act as the conductive fillers. CPCs have gained considerable interest for decades in lots of fields (e.g., antistatic materials, conductors, transparent conductive material, sensors, electromagnetic interference (EMI) shielding) [6,7,8,9], because of their easy manufacturing, low cost, controllable shapes and adjustable electrical properties [10].

The ultimate goal of developing CPCs is to achieve a wide range of practical applications. However, until now, deploying CPCs in daily life has been hugely challenging. On one hand, an important reason for this is that the presence of CNTs generally weakens the mechanical properties of CPCs [11, 12]. In other words, electrical conductivity is generally opposite to the robust mechanical performance [13]. Such case can be easily understood as follows: (i) for the CNTs filled polymer composites, although high CNTs content in CPCs can bring excellent electrical properties; this can lead to inferior mechanical properties because of the CNTs aggregations [14, 15]; (ii) with respect to segregated CPCs (s-CPCs), the segregated structures in s-CPC materials can ensure the formation of conductive networks at relatively low filler contents. However, the segregation of the conductive filler at the interface significantly prevents the molecular movement between the isolated domains. In addition, the agglomerations of fillers at the domain interface are also unavoidable. Both these factors notably harm the mechanical performance of s-CPCs [16, 17]; (iii) for the CPCs with double-percolated structures, conductive fillers are added to the immiscible polymer blend during the processing, and they are selectively located in one continuous component of the two-polymer matrices [18, 19]. However, CPCs exhibit the reduced toughness and lower ductility due to the incompatibility between the polymers [20, 21].

On the other hand, the methods to endow composites with enhanced mechanical properties always have an adverse effect on electrical performance. The mechanical properties of samples can be improved due to the significant orientation of CNTs and polymer chains under the condition of high shear/elongational flow [22]. However, high orientation of CNTs can steadily decrease the CNTs contact, resulting in much lower electrical conductivity, and therefore, higher filler content is required to construct effective conductive network in the composites [23, 24]. Another is focused on improving the dispersion state of conductive fillers to achieve the mechanical enhancements of CPCs [25]. The functionalization of CNTs is one of the most effective methods to achieve homogeneous dispersion of CNTs in polymer matrix. However, heterogeneous atoms introduced by functionalization on the fillers surface can degrade the natural electrical properties of fillers. Meanwhile, the acids used to functionalize CNTs may shorten and damage the CNTs and tend to reduce the aspect ratio. In a word, the aforementioned two aspects are harmful for the formation of effective conductive networks in CPCs [26].

The mechanical properties are the foundation for CPCs application. Moreover, enhanced electrical conductivity can not only further improve their functionalities but also expand their applications. Therefore, it will be very promising if both enhanced mechanical and electrical properties for CPCs applications can be simultaneously achieved. In view of this, one of the ideal strategies of CPCs preparation is to endow CPCs with excellent conductivity at low filler contents without sacrificing mechanical properties. However, simultaneously enhanced mechanical and electrical properties might damage other performances (such as optical performance [27]). In a word, it must be interesting to prepare the CPCs maintaining their initial performance of each domain or even exhibiting synergistic effect.

In this study, a facile method is reported for coating MWNTs on linear low-density polyethylene (LLDPE) substrates through swelling and ultrasound to form an effective conductive network on the surface of the LLDPE film. Subsequently, the MWNTs are further pressed onto the surface layer of the LLDPE film via compression molding. In this way, crystallization induced by MWNTs with good dispersion enhances the mechanical properties of LLDPE film apparently without decreasing the electrical conductivity. Meanwhile, the transparency is not significantly reduced. The properties of LLDPE/MWNTs composite films are studied, and the application potential to serve as organic vapor sensing detector is evaluated as well.

2 Experimental

2.1 Materials

Commercially available LLDPE (trade-marked as 7042) was provided by Maoming Petrochemical Corp., China. Its average molecular weight (\(\overline{{\text{M} }_{\text{w}}}\)) and melt index (MFI) were ca. 14.0 × 104 g/mol and 2.0 g/10 min (190 °C, 21.6 N), respectively. The multi-walled carbon nanotubes (MWNTs) used in this work were prepared by chemical vapor deposition (CVD). They were purchased from the Organic Chemical Limited Company, Chengdu, China. Their outer diameter was in the range of 20–40 nm, and length was about 10–30 μm, and their purity was about 95%. MWNTs were used as received without any pre-treatment. The xylene (DMB) was purchased from Zhiyuan Reagent Co., Ltd. Tianjin, China.

2.2 Sample preparation

Firstly, LLDPE particles were hot pressed at 150 °C for 10 min under a pressure of 5 MPa, followed by quenching in ice water (FM-11, Beijing Future Material Sci-Tech Co., Ltd., China). The thickness of the films was 0.2 mm. Secondly, different contents of MWNTs were dispersed in DMB to obtain the dispersions with various concentrations (in ratio of 1, 0.5, 0.25, 0.1, 0.05, and 0.025 mg/mL to DMB). Ultrasonication was then applied to the MWNT dispersion for 10 min in a cold-water bath using ultrasonication instrument (SCIENTZ-II, 315W, Ningbo Scientz Biotechnology Co., Ltd., China) to achieve a homogenous MWNTs/DMB solution. Thirdly, the hot-pressed films were cut into film strips (100 mm × 10 mm × 0.2 mm) and swelled for 10 min at 60 °C in DMB. After being swelled, the films were put into the MWNTs dispersion and sonicated for 20 min in a cold-water bath using ultrasonication instrument. Finally, each film was washed by ultrasonic treatment in ethanol for 1 min and dried in air.

The as-prepared films were transferred into a compression molding machine and hot pressed at 150 °C for 2 min under a pressure of 2 MPa. Following this, the temperature was gradually decreased to 105 °C (near the temperature of maximum crystallization rate [28]) with the help of cooling water and maintained for 10 min to form the stable and perfect crystals. Finally, the temperature was decreased to 50 °C. The films were taken out and cut into rectangular sheets (20 mm × 5 mm × 0.2 mm) for the following measurements. For comparison, pure LLDPE was also prepared following the same procedures without adding MWNTs. For convenience, the composite films are defined as LCNTx-U and LCNTx-C. Herein, LCNT represents MWNTs/LLDPE composite films, and x represents the content of MWNTs (wt %) in composite films, which is determined by TGA. U and C mean that the MWNTs/LLDPE composite films were respectively prepared by ultrasonic treatment and that ultrasonic treated followed by compression molding. For example, LCNT0.15-C represents that the MWNTs/LLDPE composite films were prepared by compression molding and the content of MWNTs was 0.15 wt %. LCNT0-U represents the pure LLDPE prepared by ultrasonic treatment.

2.3 Characterizations

2.3.1 Thermogravimetric analysis

Thermogravimetric analysis was performed under a nitrogen atmosphere on a TGA/TA Q5000IR (TA Instruments Co., USA). A sample of about 10 mg was heated from ambient temperature to 600 °C at a constant heating rate of 10 °C/min in a nitrogen atmosphere. The nitrogen flow rate was 25 mL/min.

Differential scanning calorimetry (DSC) analysis was carried out on a DSC/MDSC2920 (TA Instruments Co., USA). Approximately 6 mg sample was encapsulated in an aluminum pan and heated from 40 to 160 °C at a heating rate of 10 °C/min. All the tests were performed in nitrogen atmosphere at a flow rate of 20 mL/min.

2.3.2 Electrically conductive property test

The volume resistance of the specimens was measured by a high resistance meter (TH2683, Changzhou Tonghui Electronics Co., Ltd.). Before measurement, both ends of the samples were covered with copper electrodes to eliminate the effect of the contact resistance. The resistivity (ρ) was calculated by the equation ρ = RS/L, where R is the volume resistance, S is the cross-section area of the strip, and L is the length between the two electrodes. Five samples were tested to ensure the measurement accuracy for each composite.

2.3.3 Mechanical property test

A universal tensile testing machine (UTM2203, Sun technology Stock Co., Ltd.) was used to measure the tensile property at a crosshead rate of 50 mm/min according to GB/T 1040.3–2006 standard. The measured temperature was around 20 °C. At least five specimens had been measured and the average values were derived.

2.3.4 Morphology investigation

The morphological characterization was carried out with thermal field emission scanning electronic microscope (JSM-7500F). The morphologies of LCNTx-U (such as LCNT0-U, LCNT0.1-U, and LCNT0.15-U) were observed to investigate the distribution and dispersion of MWNTs. In order to investigate the interface interaction between MWNTs and LLDPE film, the cross-sectional morphology of MWNTs/LLDPE composite films (such as LCNT0.15-U and LCNT0.15-C) was observed. Before SEM observation, the composites were fractured by immersing in liquid nitrogen for 1 h and breaking quickly. To evaluate the electrically conductive networks of MWNTs after compression molding process, the morphologies of LCNTx-C (such as LCNT0-C, LCNT0.1-C, and LCNT0.15-C) were observed. Before SEM observation, the specimens after compressing molding were etched in a mixture of 1:1 volume of concentrated sulfuric acid and nitric acid by solving 0.7 wt % of potassium permanganate [29, 30]. All scanning surfaces were then coated with a thin layer of platinum for better imaging.

2.3.5 Optical transmittance test

LCNTx-C was characterized for transmission in the wavelength range from 400 to 1100 nm using dual beam, ultraviolet, visible light, and near-infrared spectrophotometer (Agilent Cary 5000 UV–Vis-NIR spectrophotometer) on a transparent quartz substrate.

2.3.6 Organic vapor sensing measurement

The vapor sensing behaviors of the composite films were evaluated by recording their resistance variation towards two representative organic solvent vapors of acetone and xylene. In order to study the vapor sensing responsivity, the composite films were immersed in acetone and xylene for 300 s, respectively; subsequently, the composite films were dried in air for at least 300 s to remove the organic vapor completely. A complete immersion-drying runs (IDR) included 10-s immersion in vapors and 50-s drying in air. Resistances of the composite films were recorded by a high resistivity meter (model TH2683, supplied by Changzhou Tonghui Electronics Co., Ltd.). To make a visualized comparison, the collected resistance values were all calculated via a normalization treatment. The responsivity (Rres) was defined as the ratio of (Rt − R0)/R0, where Rt represents the resistance of a sample during test and R0 is the original resistance of the sample. All the samples were tested at 40 °C.

3 Results and discussion

3.1 Thermostability of composites

To understand the thermostability of the composite films and the exact MWNTs contents in the final products, thermogravimetric analysis was first performed under nitrogen atmosphere. Figure 1a shows the TGA curves, and Fig. 1b shows the corresponding derivative thermogravimetric (DTG) curves. Table 1 shows the onset decomposition temperature (Tonset), the temperature of maximum DTG (Tmax), and the content of MWNTs in MWNTs/LLDPE composite films. As can be seen in Fig. 1a, the mass loss of MWNTs is as low as 1.6 wt % at 540 °C. The LLDPE phase in the as-prepared composite films becomes unstable when the heat temperature is above 440 °C. The weight of MWNTs/LLDPE composite films decreases rapidly in the temperature range 440–540 °C, indicating that the combustion of LLDPE matrix occurs in this temperature range.

Fig. 1
figure 1

a The TGA and b DTG curves of MWNTs, pure LLDPE, and MWNTs/LLDPE composite films containing different MWNTs contents. DSC thermograms of c LCNTx-U and d LCNTx-C

Full size image
Table 1 TGA results of MWNTs/LLDPE composite films prepared with different concentrations of MWNTs fillers
Full size table

Moreover, it is noted that Tonset and Tmax for MWNTs/LLDPE composite films generally shift toward higher temperature with increasing the concentration of MWNTs dispersion (see Fig. 1a and b), indicating that the addition of the nanoscale MWNTs greatly improves the thermal stability of MWNTs/LLDPE composite films. For example, as shown in Table 1, Tonset and Tmax of the pure LLDPE are 439.4 and 482.5 °C, respectively. However, Tonset and Tmax of the MWNTs/LLDPE composite films are observed to increase with increasing the MWNTs loading, and the increases of Tonset and Tmax are as high as 15 and 6 °C even when the MWCNT loading is only 0.15 wt %.

Uniform dispersion and network formation of the MWNTs with unaffected superior thermal conductivity may help to transfer heat released and retard the decomposition of the LLDPE matrix [31]. And the polymorphism, morphology, and structure change of LLDPE matrix due to the addition of MWNTs can also contribute to the significantly improved thermal stability. Furthermore, the contents of MWNTs in the composite films could easily be calculated from the weight remainder after the samples being heated over 700 °C and shown in Table 1. The MWNTs contents in MWNTs/LLDPE composite films are determined to be 0.064, 0.088, 0.10, 0.12, 0.15, and 0.20 wt % for samples after the ultrasonication in various CNT fillers, respectively.

3.2 Calorimetric analysis

To deeply understand the crystallization-property relationship, DSC heating curves of LCNTx-U and LCNTx-C are, respectively, represented in Fig. 1c and d. A large peak (primary peak) locates at around 122 °C for both LCNTx-U and LCNTx-C and corresponds to the fusion of stable crystals with thick lamellae. However, for LCNTx-U, two small melting peaks appear near the range of 115–117 °C and 100–108 °C (see Fig. 1c). It is worth noting that these two melting peaks are lower than primary peak. This phenomenon can be considered as the melting of smaller or imperfect lamellae formed during the ultrasonic treatment process [32]. It is worth mentioning that the two small melting peaks, which are below the melting temperature of the main melting peak, have abruptly disappeared after the compression molding samples (see Fig. 1d). This phenomenon indicates that almost all the small and imperfect crystals formed in the ultrasonic treatment process will melt and recrystallize into stable crystals.

What is more, for all the samples experienced compression molding except LCNT0.064-C and LCNT0-C, there is a shoulder peak at between 125 and 127 °C, which is higher than the main melting peak by about 4–6 °C. Combined with the result shown in Fig. 1c, this phenomenon indicates that the compression molding process has facilitated the formation of crystalline fraction with high melting point. Several factors can be contributed to the upshift of the melting temperature, i.e., the formation of thicker lamellae, confined crystals in a nanoenvironment, or simply the favorable interaction between the LLDPE chains and the MWNTs surfaces [33]. For LLDPE in the case of crystallization experienced compression molding, the temperature of 4–6 °C higher than the main melting peak is usually considered to be the melting of extended-chain crystalline structure [34]. If it is the case, it is not difficult to speculate that compression molding process causes the formation of extended-chain LLDPE layer in the presence of MWNTs, which is wrapped around of MWNTs shish. This result will be well confirmed by the SEM observation and the statistical diameter of MWNTs below.

3.3 Electrical performances

The volume electrical conductivity of MWNTs/LLDPE composite films as a function of the MWNTs content is shown in Fig. 2a. As expected, the electrical conductivities of all the composite films strongly depend on the MWNTs loading content. At the low contents of MWNTs, the volume resistivity gradually decreases with increasing the MWNTs content. However, at concentration above 0.088 wt % MWNTs, the volume resistivity sharply decreases. For example, the volume resistivity decreases from 1 × 1013 Ω∙cm for pure LLDPE to 105 Ω∙cm for the MWNTs/LLDPE composite films with the MWNTs content of about 0.12 wt %. This stepwise change in resistivity is a result of the formation of an effective interconnected MWNTs networks and can be regarded as an electrical percolation threshold [35].

Fig. 2
figure 2

a Comparison of the volume electrical conductivity of LCNTx-U and LCNTx-C. The tensile strength and elongation at break of pure LLDPE and different composite films b after ultrasonic treatment and c after compression molding. The typical SEM images of surface of composite films: d and e LCNT0.1-U with magnification of d 5 k and e 50 k; f and g LCNT0.15-U with magnification of f 5 k and g 50 k, the red arrows show the MWNTs aggregations

Full size image

According to the percolation theory [36], the electrical percolation threshold is the critical content that conductive networks are formed by the conductive fillers, above which the electrical conductivity of material will sharply increase. In this study, this simply means that, at the MWNTs contents ranging from 0.088 to 0.12 wt %, a very high percentage of electrons is permitted to flow through the specimen due to the creation of an interconnecting conductive network [37, 38]. The obtained results are compared with those of several other reported literatures [39,40,41], and the percolation threshold of our studies is extremely lower than other conductive polymer composites. At concentration above 0.12 wt % MWNTs, the volume resistivity is low and decreases marginally with increasing the MWNTs content, indicating the formation of perfect conductive networks at around 0.12 wt % for electron transport. On the whole, it is worth mentioning that the resistivity of MWNTs/LLDPE composite films after compression molding treatment is slightly higher than the films after simple swelling and ultrasonic treatments. This should be ascribed to the extended-chain LLDPE layer formed around the MWNTs surface, slightly increasing the contact resistance of neighboring MWNTs. This result will be further proved by the following statistical diameter of MWNTs.

3.4 Mechanical performances

Good mechanical properties of CPCs are essential for their applications in many fields. In order to quantitatively estimate the variations of mechanical properties (e.g., the values of tensile properties), a series of monotonic tensile tests are carried out. Tensile strength and tensile fracture strain of the composite films can be obtained from the stress–strain curves, and their dependence on the concentration of MWNTs dispersion is shown in Fig. 2b and c. Interestingly, with increasing the MWNTs concentration, the variation tendency of tensile strength for LCNTx-U and LCNTx-C is entirely different. For the samples after ultrasonic treatment (see Fig. 2b), tensile strength generally decreases with the increase of MWNTs content. Upon further increasing the MWNTs loading content, the tensile strength levels off at a slight increase. Reasonably, during the process of swelling and ultrasonic, the MWNTs are very difficult to enter into the surface of films or reach enough depth, which results in the poor interfacial interaction between MWNTs and LLDPE matrix. In addition, the MWNTs detected in LLDPE/MWNTs composite films act as stress concentrators and cause the decrease of the tensile strength. However, for the composite films after compression molding treatment (see Fig. 2c), the tensile strength is firstly enhanced and then slightly decreases with the increase of MWNTs content. The maximum value in 20 MPa for the sample with 0.10 wt % MWNTs loading is 22% higher than that of the pure LLDPE. All these results indicate that the incorporation of MWNTs is good for the improvement of tensile strength of the composite films, arising from the intrinsic high mechanical property of MWNTs.

In addition, the tensile fracture strain of composite film shows no obvious change with increasing the MWNTs concentration. This result can be explained as follows: by using the compression molding, MWNTs are further pressed into the surface layer of LLDPE matrix, which will apparently induce stronger interfacial interactions between LLDPE chains and MWNTs. In addition, these well-dispersed MWNTs may have an effect on the physical cross-linking points, thus increasing the tensile strength. However, with further increasing the MWNTs concentration beyond the critical concentration, the MWNTs could not be well dispersed in the LLDPE film and agglomerate to big clusters because of huge surface energy of MWNTs and cause the decrease of the tensile strength. This result will be well confirmed by the SEM observations.

As well known, the mechanical performance of MWNTs/LLDPE composite films is highly dependent on the content of MWNTs, dispersibility, and the interaction between the MWNTs fillers and LLDPE matrix. For the two kinds of MWNTs/LLDPE composite films described above, the tensile strength of the films after compression molding with the same MWNTs content has been respectively improved by up to 24.3%, 32.9%, 69.2%, 110.8%, 120.3%, and 105.8% compared with that of the composite films only by ultrasonic treatment (see Fig. 2b and c). The improvement in the tensile strength may be caused by the strong interactions between LLDPE matrix and MWNTs after compression molding process. That is, LLDPE extended-chain tightly decorated on MWNTs shish can significantly enhance the interfacial adhesion and the stress transfer capability between MWNTs and LLDPE matrix during compression molding process. This behavior is to be expected and will be further discussed below. Interestingly, the tensile fracture strain of LCNTx-C is not significantly reduced, in comparison to LCNTx-U (see Fig. 2b and c). This indicates that such composite films still show a certain degree of toughness.

3.5 Filler distribution and composites morphology

In order to further understand the high electrical and mechanical properties of MWNTs/LLDPE composite films, the distribution of MWNTs in the composite films after ultrasonic treatment is characterized by SEM. As examples, some typical SEM images of the LCNT0.1-U and LCNT0.15-U surfaces are observed by SEM, and the results are presented in Fig. 2d–g. Well-dispersed bright dots and lines, which are the embedded MWNTs, are observed to appear and increase with the MWNTs loading. The SEM micrograph (Fig. 2d) of LCNT0.1-U shows a good dispersion of the MWNTs in the surface layer. Another image of the same composite at higher magnification (Fig. 2e) proves the existence of a nanotube network in the MWNTs/LLDPE composite films when MWNTs concentrations are higher than 1.0 wt %. The MWNTs are partly connected with others, and the interconnecting MWNTs conducting networks are formed, above which the electrical conductivity of composites sharply increases. However, with further increasing the MWNTs concentrations beyond the critical concentration, the localized MWNTs aggregations are observed as shown by the arrow in Fig. 2f. This is consistent with the result of electrical property displayed above (see Fig. 2a). Furthermore, as shown in Fig. 2e and g, the same information can be obtained that the boundary between LLDPE matrix and MWNTs is ambiguous and hardly any complete MWNTs are observed on the surface of composite films, indicating the MWNTs surface is wrapped by the LLDPE matrix to a certain extent. The above phenomenon is consistent with the result of tensile strength displayed above (see Fig. 2a).

Figure 3a presents the SEM image of the cross section of LCNT0.15-U composite films. The MWNTs do not just bond the surface of films but subside slightly into the surface layer of LLDPE matrix. In other words, one end of the MWNTs enters the surface of LLDPE matrix, and the other end is suspended on the surface of LLDPE matrix. This phenomenon mainly arises from the thin polymer-nanotube composite layer formed by the diffusion of MWNTs into the pores created by the swelling of the LLDPE matrix in DMB under the ultrasonication anchoring technique.

Fig. 3
figure 3

SEM images of a the cross section of LCNT0.15-U and b the surface morphology of LCNT0.15-C. c and d LCNT0-C etched by mixing acid solution in c low and d high magnifications. e and f The etched LCNT0.15-C in e low and f high magnifications

Full size image

As well demonstrated above, the mechanical properties of MWNTs/LLDPE composite films were obviously enhanced, and the electrical conductivity was not significantly decreased during the compression molding process. In order to explore such phenomenon and further understand the relationship between microstructure and the macro-properties, the surface morphologies of LCNT0-C, LCNT0.1-C, and LCNT0.15-C are observed by SEM. For LCNT0.15-C without etching by using the mixing acid solution (see Fig. 3b), MWNTs were not observed on the surface of composite films, indicating that MWNTs are completely pressed into the surface layer of composite films during compression molding process and have good interphase interaction with LLDPE matrix.

Furthermore, to distinguish the crystalline structure, the amorphous phase of LLDPE is chemically etched and extracted out from the surface of composite films by using the mixing acid solution. The SEM image of Fig. 3c shows that there exist many small or imperfect crystals and randomly lamellae structure throughout the whole surface for LCNT0-C. In the SEM image of Fig. 3d with relatively high magnification, the thickness of the lamellae is more than 100 nm, and the distribution of the lamellae is extremely sparse. Figure 3e and f give a typical image for the LCNT0.15-C films. A reticular crystalline microstructure induced by MWNTs can be observed in the whole view field. Interestingly, compared with pure LLDPE film, the lamellae for LCNT0.15-C is thinner and denser. These results lead to the conclusion that MWNTs can act as the nucleating agents and induce crystallization of molten LLDPE during the compression molding process. The above phenomenon can explain the enhanced tensile strength displayed above (see Fig. 1d); that is, LLDPE molecular chains tightly decorated on MWNTs can significantly enhance the interfacial adhesion and the stress transfer capability between MWNTs and LLDPE matrix. What is more, the special reticular crystalline superstructure enhances the force transfer effect, so the mechanical performances of films after compression molding are increased significantly compared with that after ultrasonic treatment (see Fig. 2c). Figure 3f shows the fibrous crystals which are connected with others, forming a perfect crystalline network. The most noteworthy finding of the crystalline structure reveals a striking similarity to the MWNTs networks on the surface of films after ultrasonic treatment, as observed in Fig. 2g. This phenomenon indicates that during compression molding process, the interaction between MWNTs and LLDPE could improve obviously without decreasing its electrical conductivity remarkably. Because there is no macron flow during compression molding process, conductive MWNTs networks can be well preserved without any obvious damage, and therefore, the electrical performances almost keep a relatively constant level. Moreover, when the MWNTs/LLDPE composite films are pressed, the MWNTs would only be pressed together and immersed into the LLDPE matrix without breaking the coating layer of MWNTs.

A very interesting phenomenon is also observed from the image of SEM shown in Figs. 2g and 3f. The diameter of MWNTs in the surface of composite films seems to be larger than MWNTs as provided (20–40 nm). Therefore, it is interesting to perform statistical analysis on the diameter of pristine MWNTs (DMWNTs-P), MWNTs after ultrasonic treatment (DMWNTs-U), and MWNTs after compression molding (DMWNTs-C). The diameter distribution and average diameter of DMWNTs-P, DMWNTs-U, and DMWNTs-C are respectively shown in Fig. 4a and b. For each sample, 100 points were randomly selected from its SEM image obtained under the same mode. Interestingly, the diameters of these three kinds of MWNTs are all of the same order of magnitude. In addition, the largest diameter of the three kinds of structures is DMWNTs-C, successively followed by DMWNTs-U and DMWNTs-P. According to the SEM and DSC results, the phenomenon can be interpreted in detail as the follows: during the ultrasonic treatment process, some degrees of the swollen LLDPE molecular chain are wrapped around the MWNTs surface because LLDPE can be easily swelled in DMB. Therefore, it is understandable that DMWNTs-U is bigger compared with DMWNTs-P. What is more, the LLDPE chains possess high flexibility so that dismantlement can easily take place during compression molding process, but the MWNTs networks have a high rigidity and low flexibility. Under these circumstances, the local flow behavior of melt LLDPE takes place on the surface of CNTs, and LLDPE extended-chain can easily form and wrap around MWNTs shish. This can be explained by so-called shear amplification in that a great enhancement of local stress occurred in the MWNTs region of two adjacent layered tactoids with different velocities [39]. This phenomenon should be the reason of good interphase interaction between MWNTs and LLDPE matrix discussed above. It has been confirmed that the resistance to stress is greater when the adhesion ability of the two binders is stronger [40], which can explain the enhanced tensile strength in this study.

Fig. 4
figure 4

a The diameter distribution and b average diameter of MWNTs, MWNTs after ultrasonic treatment and MWNTs after compression molding. c Transmittance spectra in the visible range of LCNTx-C

Full size image

3.6 Optical transmittance

Figure 4c presents the transmittance spectra of LCNTx-C. Within the wavelength range from 400 to 1100 nm, the optical transmittance shows a monotonic increase in the visible region and becomes relatively slow in the near-infrared region. As expected, as the MWNTs mass fraction increased, the optical transmittance is slowly decreased, while the spectral shape remains unchanged. The transmittance of the composite films is calculated by using their absorption values at λ = 550 nm. For LCNTx-C (X = 0, 0.064, 0.088, 0.1, 0.12, 0.15, 0.20), the optical transmittances observed are 30.3, 28.1, 26.7, 19.2, 17, 18, and 17.3%, respectively (for \(\lambda\) = 550 nm). The additive property of MWNTs has no significant influence on the optical transmittance of the composite films after compression molding treatment compared with LCNT0-C in a wide range of transmittance. This might be examined as follows: considering the low optical transmittance of pure LLDPE film, transparency is not significantly reduced during compression molding. In addition, the conductive networks for electron transport have been formed at very low content of MWNTs (see Fig. 2a), and the MWNTs detected on the surface of LLDPE/MWNTs composite films have a good dispersion (see Fig. 3e and f). Therefore, the transmittance of various composites films is exciting.

3.7 Organic vapor sensing behavior

It is well known that the electrical performance of CPCs has a great dependency on the chemical composite of polymer, nanofillers content, and interaction between polymer and organic solvents [41]. In the composite films, MWNTs particles are mainly localized on the surface of LLDPE matrix. These MWNTs particles interconnect with each other, constructing favorable conductive paths through the composite films. The conductive networks are the key to the subsequent vapor sensing capacity. In the present work, two representative organic solvents of acetone and xylene are selected for vapor sensing tests.

Figure 5a and b show the typical responsivity-time curve of the composite films (LCNT0.15-U and LCNT0.15-C) exposed to acetone and xylene organic vapors, respectively. It is found that the responsivity of the two composite films increases obviously when exposed to organic solvents. In general, the polymer matrix absorbs organic vapor molecules and starts to swell. This swelling polymer gradually changes the position of the conductive fillers, followed by the damage of conductive networks [42]. With further increasing the exposure time, different responsive behaviors are observed: The responsivity of the two composite films in xylene continues to increase apparently, while it changes much weakly in acetone. It is worth mentioning that the films after compression molding possess higher responsivity than the films after ultrasonic treatment. This phenomenon can be explained by the improvement of interaction between LLDPE and MWNTs and the conductive networks caused by compression molding treatment, which increase the synergistic effect of deformation for LLDPE and MWNTs conductive networks. In other words, when the deformation of LLDPE matrix occurs with the penetration of organic solvent vapor, the conductive networks wrapped by LLDPE matrix would be damaged more obviously, and higher vapor sensing responsivity is thus achieved.

Fig. 5
figure 5

Vapor sensing behavior of LCNT0.15-U and LCNT0.15-C towards acetone (a, c) and xylene (b, d) at 40 °C

Full size image

The responsivities of the two composite films towards organic vapors in ten successive immersion (10 s)-drying (50 s) runs (IDRs) are investigated to evaluate their stability and reproducibility (Fig. 5c and d). Among them, towards the vapors of acetone and xylene, the two composite films exhibit evidently different stimuli-response behaviors. With regard to acetone (see Fig. 5c), the two composite films show a similar sensing behavior, and the resistance almost decreases to the initial value after being exposed to air in each IDR, showing a reversible and reproducible behavior. However, towards the vapors of xylene (see Fig. 5d), the two composite films reach an equilibrium state just after the first run, showing a nice reproducibility. The equilibrium state of the remaining solvent volume can be achieved after one IDR, which was the probable reason for this phenomenon. Importantly, towards the same kind of organic vapor, the maximum responsivity of LCNT0.15-C was higher than that of LCNT0.15-U in IDR, indicating a more vulnerable conductive network formed during compression molding process.

Based on the results and discussions, it is proposed that compression molding is an effective approach to improve the vapor sensor behaviors of composite films. This phenomenon can be mainly ascribed to two aspects, i.e., on one hand, the strong interaction between the LLDPE matrix and the MWNTs caused by compression molding treatment, which increases the synergistic effect of deformation for the LLDPE matrix and MWNTs conductive networks. On the other hand, after compression molding treatment, the LLDPE extended-chains can be easily formed and wrap around MWNTs shish, which subsequently causes the damage of conductive networks.

4 Conclusion

The MWNTs/LLDPE composite films with good electrical, mechanical, optic performances, and vapor sensing behaviors were prepared by ultrasonic treatment and compression molding. After compression molding process, the crystallization induced by MWNTs enhances the mechanical properties of LLDPE film obviously without decreasing the electrical conductivity and optic performances. Moreover, vapor sensing responsivities of the MWNTs/LLDPE film in acetone and xylene vapors during ten IDRs are studied in details. The sensitivity of vapor sensors is remarkably improved after compression molding process. The present work suggests that the MWNTs/LLDPE composite film can be promising materials to prepare a nice vapor sensor.