rus
Issue #1/2019
V.A.Kuznetsov, B.Ch.Kholkhoev, V.G.Makotchenko, A.N.Lavrov, Ye.N.Gorenskaya, A.S.Berdinsky, V.F.Burdukovskii, A.I.Romanenko, V.Ye.Fedorov
Composite tensoresistive materials based on polybenzimidazole matrix
Composite tensoresistive materials based on polybenzimidazole matrix
Studied are the film composite materials based on the polymer polybenzimidazole matrix with nano-structured carbon fillers. As a matrix the poly[2,2’-(p-oxydiphenylen)-5,5’-bisbenzimidazole] (OPBI) has been used. Graphite nanoplates and few-layered graphene were applied as the fillers. Piezoresistive effect in the film composite samples, strain gauge factor and endurance were studied experimentally. It is shown that the strain gauge factor does not depend on the concentration of filler and, on the average, equal 15 (for FLG composites) and 13 (for GNP composites) within the error. It was found that strain gauge factor of the samples does not change up to 100,000 alternating loading cycles. Temperature dependences of electrical resistance of composite samples have been measured and analyzed. It has been established that the main insulating polymer gaps between the carbon filler nanoparticles make the main contribution to the electrical resistance and conduction mechanism is tunneling between the particles.
Теги: composite films electron transport few-layered grapheme (flg) graphite nanoplates (gnp) polybenzimidazole (opbi) strain gauge factor графитовые нанопластины композиционные пленки коэффициент тензочувствительности малослойный графен полибензимидазол электронный транспорт
INTRODUCTION
Composite materials based on polymer matrix with the conducting fillers have many uses in the fields where established semiconductor technologies can not be applied, particularly in electronic devices on flexible substrates. Recently the polymer composites attract the ever increasing attention of researchers as the advanced materials for electronics. Application of new functional materials as sensitive elements of the detectors is one of the modern trends in measuring mechanical values. On the base of a large number of publications it can be suggested that such materials are either already used or will soon be used in manufacturing of strain gauges on the flexible substrates [1–3]. The operating principle of strain gauges is based on the piezoresistive effect, when the electrical resistance of a conductor is changed as a result of its deformation.
Nowadays a large number of polymers [4] is used as the polymer matrix for production of composites although the composites based on the polybenzimidazole matrix still have not been studied sufficiently. It should be noted that comparatively to other polymers the polybenzimidazole has the outstanding mechanical properties and stability at high temperatures [5, 6].
It has high creep resistance, high mechanical strength and fatigue strength together with the significant chemical resistance. So, the composites based on polybenzimidazole are prospective as the strain sensing elements. We proposed to use the carbon nanoparticles – graphite nanoplates (GNP) and few-layered graphene (FLG) as a conductive phase due to their similarity in nature but with the different aspect ratios. This paper presents the experimental results of piezoresistive effect and temperature dependences of electrical resistance for composites based on the polybenzimidazole matrix with two types of the fillers (GNP and FLG).
SYNTHESIS AND EXPERIMENTAL TECHNIQUE
A representative of polybenzimidazoles, poly[2,2’-(p-oxydiphenylen)-5,5’-bisbenzimidazole] (OPBI) was prepared by high-temperature polycondensation of 3,3',4,4'-tetraaminodiphenyl oxide and 4,4'-oxybis (benzoic acid) in Eton's reagent according to the method described in [7]. To prepare the composite films, the dispersions (colloidal systems) consisting of graphite nanoplates particles or few-layered graphene distributed regularly in OPBI solution have been synthesized. The natural commercial graphite with particles lower than 200 µm produced by Merck was used as a precursor of graphite nanoplates. Dispersions were obtained by mixing GNP portions with N-methyl-2-pyrrolidone (MP) and two-stage ultrasonic treatment (by ultrasonic technological disperser "Volna" UZTA-0.4/22-ОМ type, power 400 W, frequency 22 kHz). Supernatant (supernatant of a dispersion, i.e., without particles precipitated while centrifuging) approximately constituting of 70% of the dispersion was used. In that case the GNP particle prepared by this method is nearby 10 nm in thickness with linear sizes from tens to hundreds nm [8]. The synthesis technique to prepare few-layered graphene was described in [9, 10]. The average thickness of FLG particles is 3–4 nm and corresponds to 9–12 structural layers [10] of graphite particles with interlayer distances of 0.347 nm. Linear dimensions of FLG particles were in the range from tens to hundreds of nanometers. FLG dispersions were synthesized by sonication of the necessary quantity of FLG in 2% OPBI solution in MP. It was applied the supernate in quantity of 90% of the dispersion to obtain the composite films. Composite films with GNP and FLG were prepared with one and the same conditions by the method called "flow coating", when the dispersions were flowed on glass substrates and dried at 70–80 °C during 24 hours. As a result, the films of about 50 µm in thickness were prepared. Afterwards the films were removed from the substrates with a tweezers and have been dried additionally in vacuum at 100 °C for 24 hours and in air at 200 °C for 2 hours in order to remove the solvent remained. In this paper the films with mass concentrations FLG and GNP from 0.25 to 2.00% and from 17 to 45% were studied. Hereinafter we designate the composites as OPBI-FLG-x and OPI-GNP-x, where x is a mass concentration of filler in percentage. One of the most important parameters of the piezoresistive effect – the strain gauge factor, which is the relation of relative change in electrical resistance to the strain that caused this change [11]:
K=ΔR/(Rεmin·ε), (1)
where ΔR=Rεmax–Rεmin, Rεmax – electrical resistance of a sample at strain of εmax, Rεmin – electrical resistance of a sample at deformation of εmin, ε = εmax – εmin, ε – strain of a sample (relative changes of its length). To study the piezoresistive effect, the experimental samples were bonded with ВЛ-931 glue to the beams of uniform strength (in bending) (which are elastic elements of trapezoidal shape) in order to achieve the uniform distribution of strain along the length of the sample. To insulate the samples, the beams were covered previously with the ВЛ-931 glue layer. To ensure deformation of the samples, one end of the beam was rigidly fixed and the free end of the beam was deflected with the device described in detail in [12]. Deflection of the free end of the beam in opposite directions to the balance position makes it possible, in the frame of one experiment, to measure the electrical resistance of the samples both at their tension and compression. Samples were made as strips of 7 mm long and 1 to 2 mm wide. Fig.1 presents the image of the sample bonded on a beam of uniform strength (in bending). Electrical contacts were made with silver paste and leads conductors – of thin copper wires. Electrical resistance was measured using four-contact method with power supply operating in the voltage stabilizing mode.
Temperature dependencies of electrical resistance of the samples were measured to study the electronic transport properties. Temperature was set up by an insert to immerse into the Dewar vessel, helium gas was used as a coolant. Temperature was measured with iron-rhodium resistance thermometer in the range of temperatures studied. The cooling rate at low temperatures did not exceed 0.5 К/min. Geometry and electrical contacts were the same as in case of when the piezoresistive effect was investigated. Keithley 2,000 voltage meter with the 2,000-SCAN internal scanning plate was used to measure the voltage on the potential contacts of the sample and resistance coils.
RESULTS AND DISCUSSION
The piezoresistive was investigated in the OPBI-FLG composite samples of mass concentration FLG 0.25, 0.75 and 2.00%, and OPBI -GNP composites of mass concentration of GNP 17, 30, 40 and 45%. Fig.2 indicates the dependence of electrical resistance and the relative change of the electrical resistance of OPBI-FLG-2,00 composite on strain of the sample. It was shown that the experimental data may be fitted by linear function within the error accuracy in all studied range of strain (from –0.14% to +0.14%). All studied composites present the same shape of the curves.
The strain gauge factor was determined by the formula (1). Coefficient values for OPBI-FLG and OPBI-GNP composites were from 14.4 to 16.6 and from 12.1 to 14.5, correspondingly. The obtained factor did not depend on the concentrations within the error. Fig.3 presents the dependence of the strain gauge factor for OPBI-FLG-2,00 composite sample on the number of alternating loading cycles with ±0.14% strain. It is shown that the strain gauge factor is stable up to 100,000 cycles. The same stability was observed for all studied samples.
Fig.4 shows the temperature dependences of resistivity for all studied composite samples. In the composite materials that present a insulating matrix with the conductive phase particles the electron transport can be effected by the particles contacting each other or particles separated with polymer gaps. Due to stability of the dispersions of the FLG and GNP particles in the OPBI solution for a long time (up to a month and longer without sedimentation), can be concluded that FLG and GNP particles are not coagulated and are surrounded by the polymer solution. After the films of such dispersion have been formed, the particles remain separated by the insulating polymer gaps.
The fact that the particles of the conducting phase are surrounded by insulating gaps is supported by the fact that the value of resistivity of graphite (or a sample consisting only of particles of GNP – bulk sample of FLG), differ from the value of resistivity of the composite samples by more than four orders of magnitude. This difference cannot be explained by the percolation theory without taking into account the influence of the insulating gaps. Otherwise, with an increase in concentration of the conducting phase, a sharp decrease in resistivity of the composite near the percolation threshold would have been firstly observed, and then a slight decrease in resistivity with a further increase in the concentration. However, this type of behavior is not observed.
Fillers of FLG and GNP composites are the nano-scale particles of semimetallic graphite. Electrical current in a sample flows through the nanoparticles separated by the insulating polymer gaps. There are two possible mechanisms of overcoming the potential barriers involved by the gaps: above-barrier transitions of the charge carriers and tunneling. It is possible to estimate the effective activation energy from temperature dependences of electrical resistance of the samples. In case of the only above-barrier transitions it would have been equal to the height of the potential barrier involved by the insulating gaps.
Fig.5 presents temperature dependences of Δeff ≈ d(lnR)/d(1/T) (effective conductivity activation energy expressed in units of temperature, K). As one can see, at any temperatures the values of Δeff are lower than the corresponding temperature. This fact means that were the above-barrier transitions of charge carriers the only electron transport mechanism, the termal activation would have allowed the charge carriers overcoming the barriers at any temperatures (T > Δeff) (see Fig.5). In this case electrical resistance of the samples would have been determined by electrical resistance of conductive phase particles and percolation network topology. This model can not explain the increase of composite resistivity by many orders of magnitude.
Hence, it was proposed to explain the behavior of electronic transport by the tunneling of charge carriers through the insulating polymer gaps between the semimetallic conducting phase particles.
It is well known that tunneling is a temperature-independent mechanism of electrical conductivity, however, Fig.4 shows increasing electrical resistance of the samples while the temperature decreases. We explain this fact as follows. Because the FLG and GNP fillers are the nanoparticles of semimetallic graphite with defects and with relatively small density of states, the allowed states spectrum in the particles will be discrete. Accordingly, the electron transport will be carried out by the tunneling transitions (hoppings) between the localized states of different energy. The lower the temperature, the smaller quantity of unoccupied states with available energy in the system, and, as a result, the tunneling transitions become longer. That is decreasing of the temperature leads to an increase of the electrical resistance of the composite. This mechanism of electron transport has a character of the variable range hopping conduction mechanism. As the temperature rises, the hopping length decreases down to the thickness of insulating gaps.
Changes of the conductive phase concentration do not lead to the changes in the electron transportmechanism. It means that the physical nature of the piezoresistive effect is in changes of electrical resistance of tunneling contacts at deformation of the samples.
CONCLUSIONS
This paper presents the research of composite materials based on polybenzimidazole matrix with nano-structured fillers, few-layered graphen and graphite nanoplates as the strain sensing elements. It was shown that the strain gauge factor of OPBI-FLG and OPBI-GNP composites did not depend on conductive phase particle sizes and their concentration in all studied range. It was observed that the insulating gaps made main contribution to the composite electrical resistance, and the electron transport is carried out by tunneling between FLG or GNP particles through the gaps. Changes in the conductive phase concentration in the wide range do not lead to change of the mechanism of electronic transport. The studied composite samples are resistant to long-term cyclic loadings and their strain gauge factor does not change up to 100,000 compression-tension cycles. The obtained experimental data confirm that the strain sensing elements based on polybenzimidazole matrix are promising for creation of the mechanical quantities sensors, strain gauges in particular.
ACHNOWLEDGEMENTS
The research was supported by the Ministry of Science and Education of the Russian Federation, NIIC SB RAS (study of electrical properties of composite materials, synthesis of few-layered graphene) and the Baikal Institute of Nature Management, Siberian Branch of the Russian Academy of Sciences (production of functional polymer composites), with the support of the RFBR grant No. 18-33-00655 mol_a (production of stable dispersions of graphite nanoplates). ■
Composite materials based on polymer matrix with the conducting fillers have many uses in the fields where established semiconductor technologies can not be applied, particularly in electronic devices on flexible substrates. Recently the polymer composites attract the ever increasing attention of researchers as the advanced materials for electronics. Application of new functional materials as sensitive elements of the detectors is one of the modern trends in measuring mechanical values. On the base of a large number of publications it can be suggested that such materials are either already used or will soon be used in manufacturing of strain gauges on the flexible substrates [1–3]. The operating principle of strain gauges is based on the piezoresistive effect, when the electrical resistance of a conductor is changed as a result of its deformation.
Nowadays a large number of polymers [4] is used as the polymer matrix for production of composites although the composites based on the polybenzimidazole matrix still have not been studied sufficiently. It should be noted that comparatively to other polymers the polybenzimidazole has the outstanding mechanical properties and stability at high temperatures [5, 6].
It has high creep resistance, high mechanical strength and fatigue strength together with the significant chemical resistance. So, the composites based on polybenzimidazole are prospective as the strain sensing elements. We proposed to use the carbon nanoparticles – graphite nanoplates (GNP) and few-layered graphene (FLG) as a conductive phase due to their similarity in nature but with the different aspect ratios. This paper presents the experimental results of piezoresistive effect and temperature dependences of electrical resistance for composites based on the polybenzimidazole matrix with two types of the fillers (GNP and FLG).
SYNTHESIS AND EXPERIMENTAL TECHNIQUE
A representative of polybenzimidazoles, poly[2,2’-(p-oxydiphenylen)-5,5’-bisbenzimidazole] (OPBI) was prepared by high-temperature polycondensation of 3,3',4,4'-tetraaminodiphenyl oxide and 4,4'-oxybis (benzoic acid) in Eton's reagent according to the method described in [7]. To prepare the composite films, the dispersions (colloidal systems) consisting of graphite nanoplates particles or few-layered graphene distributed regularly in OPBI solution have been synthesized. The natural commercial graphite with particles lower than 200 µm produced by Merck was used as a precursor of graphite nanoplates. Dispersions were obtained by mixing GNP portions with N-methyl-2-pyrrolidone (MP) and two-stage ultrasonic treatment (by ultrasonic technological disperser "Volna" UZTA-0.4/22-ОМ type, power 400 W, frequency 22 kHz). Supernatant (supernatant of a dispersion, i.e., without particles precipitated while centrifuging) approximately constituting of 70% of the dispersion was used. In that case the GNP particle prepared by this method is nearby 10 nm in thickness with linear sizes from tens to hundreds nm [8]. The synthesis technique to prepare few-layered graphene was described in [9, 10]. The average thickness of FLG particles is 3–4 nm and corresponds to 9–12 structural layers [10] of graphite particles with interlayer distances of 0.347 nm. Linear dimensions of FLG particles were in the range from tens to hundreds of nanometers. FLG dispersions were synthesized by sonication of the necessary quantity of FLG in 2% OPBI solution in MP. It was applied the supernate in quantity of 90% of the dispersion to obtain the composite films. Composite films with GNP and FLG were prepared with one and the same conditions by the method called "flow coating", when the dispersions were flowed on glass substrates and dried at 70–80 °C during 24 hours. As a result, the films of about 50 µm in thickness were prepared. Afterwards the films were removed from the substrates with a tweezers and have been dried additionally in vacuum at 100 °C for 24 hours and in air at 200 °C for 2 hours in order to remove the solvent remained. In this paper the films with mass concentrations FLG and GNP from 0.25 to 2.00% and from 17 to 45% were studied. Hereinafter we designate the composites as OPBI-FLG-x and OPI-GNP-x, where x is a mass concentration of filler in percentage. One of the most important parameters of the piezoresistive effect – the strain gauge factor, which is the relation of relative change in electrical resistance to the strain that caused this change [11]:
K=ΔR/(Rεmin·ε), (1)
where ΔR=Rεmax–Rεmin, Rεmax – electrical resistance of a sample at strain of εmax, Rεmin – electrical resistance of a sample at deformation of εmin, ε = εmax – εmin, ε – strain of a sample (relative changes of its length). To study the piezoresistive effect, the experimental samples were bonded with ВЛ-931 glue to the beams of uniform strength (in bending) (which are elastic elements of trapezoidal shape) in order to achieve the uniform distribution of strain along the length of the sample. To insulate the samples, the beams were covered previously with the ВЛ-931 glue layer. To ensure deformation of the samples, one end of the beam was rigidly fixed and the free end of the beam was deflected with the device described in detail in [12]. Deflection of the free end of the beam in opposite directions to the balance position makes it possible, in the frame of one experiment, to measure the electrical resistance of the samples both at their tension and compression. Samples were made as strips of 7 mm long and 1 to 2 mm wide. Fig.1 presents the image of the sample bonded on a beam of uniform strength (in bending). Electrical contacts were made with silver paste and leads conductors – of thin copper wires. Electrical resistance was measured using four-contact method with power supply operating in the voltage stabilizing mode.
Temperature dependencies of electrical resistance of the samples were measured to study the electronic transport properties. Temperature was set up by an insert to immerse into the Dewar vessel, helium gas was used as a coolant. Temperature was measured with iron-rhodium resistance thermometer in the range of temperatures studied. The cooling rate at low temperatures did not exceed 0.5 К/min. Geometry and electrical contacts were the same as in case of when the piezoresistive effect was investigated. Keithley 2,000 voltage meter with the 2,000-SCAN internal scanning plate was used to measure the voltage on the potential contacts of the sample and resistance coils.
RESULTS AND DISCUSSION
The piezoresistive was investigated in the OPBI-FLG composite samples of mass concentration FLG 0.25, 0.75 and 2.00%, and OPBI -GNP composites of mass concentration of GNP 17, 30, 40 and 45%. Fig.2 indicates the dependence of electrical resistance and the relative change of the electrical resistance of OPBI-FLG-2,00 composite on strain of the sample. It was shown that the experimental data may be fitted by linear function within the error accuracy in all studied range of strain (from –0.14% to +0.14%). All studied composites present the same shape of the curves.
The strain gauge factor was determined by the formula (1). Coefficient values for OPBI-FLG and OPBI-GNP composites were from 14.4 to 16.6 and from 12.1 to 14.5, correspondingly. The obtained factor did not depend on the concentrations within the error. Fig.3 presents the dependence of the strain gauge factor for OPBI-FLG-2,00 composite sample on the number of alternating loading cycles with ±0.14% strain. It is shown that the strain gauge factor is stable up to 100,000 cycles. The same stability was observed for all studied samples.
Fig.4 shows the temperature dependences of resistivity for all studied composite samples. In the composite materials that present a insulating matrix with the conductive phase particles the electron transport can be effected by the particles contacting each other or particles separated with polymer gaps. Due to stability of the dispersions of the FLG and GNP particles in the OPBI solution for a long time (up to a month and longer without sedimentation), can be concluded that FLG and GNP particles are not coagulated and are surrounded by the polymer solution. After the films of such dispersion have been formed, the particles remain separated by the insulating polymer gaps.
The fact that the particles of the conducting phase are surrounded by insulating gaps is supported by the fact that the value of resistivity of graphite (or a sample consisting only of particles of GNP – bulk sample of FLG), differ from the value of resistivity of the composite samples by more than four orders of magnitude. This difference cannot be explained by the percolation theory without taking into account the influence of the insulating gaps. Otherwise, with an increase in concentration of the conducting phase, a sharp decrease in resistivity of the composite near the percolation threshold would have been firstly observed, and then a slight decrease in resistivity with a further increase in the concentration. However, this type of behavior is not observed.
Fillers of FLG and GNP composites are the nano-scale particles of semimetallic graphite. Electrical current in a sample flows through the nanoparticles separated by the insulating polymer gaps. There are two possible mechanisms of overcoming the potential barriers involved by the gaps: above-barrier transitions of the charge carriers and tunneling. It is possible to estimate the effective activation energy from temperature dependences of electrical resistance of the samples. In case of the only above-barrier transitions it would have been equal to the height of the potential barrier involved by the insulating gaps.
Fig.5 presents temperature dependences of Δeff ≈ d(lnR)/d(1/T) (effective conductivity activation energy expressed in units of temperature, K). As one can see, at any temperatures the values of Δeff are lower than the corresponding temperature. This fact means that were the above-barrier transitions of charge carriers the only electron transport mechanism, the termal activation would have allowed the charge carriers overcoming the barriers at any temperatures (T > Δeff) (see Fig.5). In this case electrical resistance of the samples would have been determined by electrical resistance of conductive phase particles and percolation network topology. This model can not explain the increase of composite resistivity by many orders of magnitude.
Hence, it was proposed to explain the behavior of electronic transport by the tunneling of charge carriers through the insulating polymer gaps between the semimetallic conducting phase particles.
It is well known that tunneling is a temperature-independent mechanism of electrical conductivity, however, Fig.4 shows increasing electrical resistance of the samples while the temperature decreases. We explain this fact as follows. Because the FLG and GNP fillers are the nanoparticles of semimetallic graphite with defects and with relatively small density of states, the allowed states spectrum in the particles will be discrete. Accordingly, the electron transport will be carried out by the tunneling transitions (hoppings) between the localized states of different energy. The lower the temperature, the smaller quantity of unoccupied states with available energy in the system, and, as a result, the tunneling transitions become longer. That is decreasing of the temperature leads to an increase of the electrical resistance of the composite. This mechanism of electron transport has a character of the variable range hopping conduction mechanism. As the temperature rises, the hopping length decreases down to the thickness of insulating gaps.
Changes of the conductive phase concentration do not lead to the changes in the electron transportmechanism. It means that the physical nature of the piezoresistive effect is in changes of electrical resistance of tunneling contacts at deformation of the samples.
CONCLUSIONS
This paper presents the research of composite materials based on polybenzimidazole matrix with nano-structured fillers, few-layered graphen and graphite nanoplates as the strain sensing elements. It was shown that the strain gauge factor of OPBI-FLG and OPBI-GNP composites did not depend on conductive phase particle sizes and their concentration in all studied range. It was observed that the insulating gaps made main contribution to the composite electrical resistance, and the electron transport is carried out by tunneling between FLG or GNP particles through the gaps. Changes in the conductive phase concentration in the wide range do not lead to change of the mechanism of electronic transport. The studied composite samples are resistant to long-term cyclic loadings and their strain gauge factor does not change up to 100,000 compression-tension cycles. The obtained experimental data confirm that the strain sensing elements based on polybenzimidazole matrix are promising for creation of the mechanical quantities sensors, strain gauges in particular.
ACHNOWLEDGEMENTS
The research was supported by the Ministry of Science and Education of the Russian Federation, NIIC SB RAS (study of electrical properties of composite materials, synthesis of few-layered graphene) and the Baikal Institute of Nature Management, Siberian Branch of the Russian Academy of Sciences (production of functional polymer composites), with the support of the RFBR grant No. 18-33-00655 mol_a (production of stable dispersions of graphite nanoplates). ■
Readers feedback
