Characterization of Dielectric Properties of Nanocellulose from Wood

Article pubs.acs.org/JPCB

Characterization of Dielectric Properties of Nanocellulose from Wood and Algae for Electrical Insulator Applications David Le Bras, Maria Strømme, and Albert Mihranyan* Nanotechnology and Functional Materials, Department of Engineering Sciences, Box 534, Uppsala University, 75121 Uppsala, Sweden S Supporting Information *

ABSTRACT: Cellulose is one of the oldest electrically insulating materials used in oil-filled high-power transformers and cables. However, reports on the dielectric properties of nanocellulose for electrical insulator applications are scarce. The aim of this study was to characterize the dielectric properties of two nanocellulose types from wood, viz., nanofibrillated cellulose (NFC), and algae, viz., Cladophora cellulose, for electrical insulator applications. The cellulose materials were characterized with X-ray diffraction, nitrogen gas and moisture sorption isotherms, helium pycnometry, mechanical testing, and dielectric spectroscopy at various relative humidities. The algae nanocellulose sample was more crystalline and had a lower moisture sorption capacity at low and moderate relative humidities, compared to NFC. On the other hand, it was much more porous, which resulted in lower strength and higher dielectric loss than for NFC. It is concluded that the solid-state properties of nanocellulose may have a substantial impact on the dielectric properties of electrical insulator applications.

1. INTRODUCTION Cellulose is the most abundant polymer on earth and has enormous industrial importance. One of the common applications for native cellulose includes electrical insulators, mainly in oil-filled power transformers and paper-insulated power cables. The traditional use of oil-impregnated cellulose as a reliable electrically insulating material is justified by the combination of its dielectric and mechanical properties such as high resistivity (∼1010 Ω cm−1), high electrical strength (∼180 kV cm−1 in oil), chemical stability, flexibility, nonthermoplastic properties, availability in nature, and low cost.1 Therefore, it is often said that the use of cellulose as an insulator in the electrical industry is as old as the industry itself2 because cotton rags, hemp, paper clippings and various textiles have been used even before the 1920s. In a book published in 1935, Whitehead3 states the following: “In spite of its limitations, paper has proved itself by far the superior material for the insulation of high-voltage cables. Both history and product are truly remarkable. The impregnated paper cable is unique, it cannot be even approached in performance by any other type and without it we would still be at a very early stage in the remarkable expansion in transmission and distribution of electric power.” Some of the earliest comprehensive overviews of cellulose use in electrical insulators are those made by Whitehead3 and Kohman2 in the 1930s. More recent reviews include those by Schaible4 and Prevost and Oomen.5,6 The dielectric properties of cellulose depend on many factors including impurities, hemicellulose and lignin content, degree of polymerization (DP), fiber length, density, and moisture content. The resistivity of cellulose was shown to be highly dependent on moisture content. The first study on the effect of © XXXX American Chemical Society

moisture on insulating materials, including cotton, was done in 1914 by Evershed,7 who observed significant current leakage due to moisture present in the material. Fundamental contributions to understanding the effect of moisture on the dielectric properties of cellulose were made by Murphy.8−12 Murphy proposed a phenomenological power-law model in which the electrical conductivity (σ′) of cellulose increases with increasing water content to the power of 9, which was explained by the formation of a pathway of adsorbed water molecules in cellulose on which ions could move.12 The assumptions behind this model were later developed in further detail and used to explain the conduction mechanism in microcrystalline cellulose at varying relative humidity.13 Also, refined models have been developed to include percolation effects and hopping transport mechanisms.14,13,15 The major hurdle for cellulose’s use in electrical insulator applications is the hygroscopicity of cellulose.5,6 The moisture level in cellulose for insulator applications should preferably be about 0.5% or lower,5 whereas the typical moisture content of tree-derived cellulose varies in the range of 4 to 8% for RH between 30 and 70%, i.e., the typical range of humidity on factory floors throughout the year. Therefore, prior to its use the paper needs to be processed and kept dry, which is an elaborate, time-consuming, and costly process. If moisture is not properly removed, then partial discharge inception becomes significant (above 3% moisture content) and may result in gas (hydrogen) evolution.3,5 Furthermore, paper Received: January 23, 2015 Revised: March 28, 2015

A

DOI: 10.1021/acs.jpcb.5b00715 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B degradation and aging will be accelerated. In this respect, the lifetime of regular Kraft paper is reduced by half for every doubling of moisture content.5 The mechanical strength of Kraft paper is also decreased by half every 24 °C at 1% moisture content whereas at 8% moisture content the doubling interval is only 8 °C.5 The water sorption capacity of cellulose can be moderated by oil impregnation, which also removes air, another factor detrimental to insulator applications.3 Furthermore, mineral oil dissipates heat from the device, and cellulose acts as mechanical reinforcement. Nonetheless, it is not desirable to expose the impregnated insulator paper over long periods to moisture because once moisture is reabsorbed into cellulose it will be even harder to remove by heating due to the presence of oil.6 Therefore, intrinsically nonhygroscopic mechanically strong cellulose materials are in high demand in the electrical insulator industry. Previous studies with pharmaceutical cellulose grades16−20 showed that H2O sorption by cellulose as well as water availability to partake in chemical reactions is highly dependent on the source of the material and especially the cellulose degree of crystallinity, viz., the higher the degree of crystallinity, the less H2O absorbed. In this respect, nanocellulose materials of various origins are interesting because they have shown great potential as mechanical strength reinforcement in various composites. Nanocellulose, which can be produced from higher plants, bacteria, algae, or tunnicates, is the product of defibrillation of cellulose fibers into nanoscale units except for bacterial cellulose, which is directly produced by microorganisms. Nanocellulose has been widely considered to be a suitable substrate for paper electronics applications.21 However, reports on the dielectric properties of nanocellulose and especially those specifically targeting insulator applications are so far scarce.22 The aim of this study was to investigate the dielectric and mechanical properties of two nanocellulose types differing in their solid-state properties such as the degree of crystallinity, water sorption capacity, specific surface area, and porosity.

CrI =

I002 − Iam × 100 I002

(1)

Here I002 is the overall intensity of the peak at 2θ = 22°, and Iam is the intensity of the baseline for 2θ = 18°. 2.4. Porosity. The samples were degassed under vacuum overnight at 70 °C (ASAP 2020, Micromeritics, USA). The true density of the samples was measured using a He pycnometer (AccuPyc 1340, Micromeritics, USA). The total porosity in %, φ %, was calculated as follows: ⎛ ρ ⎞ φ% = ⎜⎜1 − bulk ⎟⎟ × 100 ρtrue ⎠ ⎝

(2)

Here ρbulk is the bulk density and ρtrue is the true density of the paper sheet. The bulk density is calculated from the geometric dimensions of the sheet using a digital caliper (Mitutoyo, Japan) with 0.001 mm precision. The relative density is defined as the ratio of bulk density to true density. 2.5. Scanning Electron Microscopy (SEM). Micrographs of each sample were taken using a scanning electron microscope (Leo Gemini1550 FEG SEM, U.K.). The samples were mounted onto carbon tape over aluminum stubs and coated with Au/Pt to reduce charging effects. 2.6. Nitrogen (N2) Gas and Moisture (H2O) Sorption Isotherms. The sorption properties of the paper sheets were analyzed from N2 gas and H2O (moisture) sorption isotherms, recorded at 77 and 298 K, respectively, using an ASAP 2020 instrument (Micromeritics, USA). For moisture sorption analysis the instrument was equipped with an auxiliary chemisorption unit. The analysis of the obtained data was performed using the standard software supplied to the ASAP 2020 instrument (Micromeritics, USA). 2.7. Tensile Properties. The mechanical properties of the samples were tested under axial tension (Mini Mat 2000, Rheometric Scientific, USA). The specimens were clamped on a tensile stage and stretched at a 10 mm/min strain rate until rupture occurred. The yield stress was determined from the intersection of the tangents to the lines corresponding to elastic and plastic deformation. At a given stress, the sample fails and the maximum force needed for the fracture is defined as the tensile stress. 2.8. Dielectric Spectroscopy. A dielectric spectroscopy instrument (Alpha-A High Resolution, Novocontrol Technologies, Germany) was employed to evaluate the frequencydependent response of the samples at different relative humidities in the frequency range of 1 × 10−2 to 1 × 106 Hz. The measured response was interpreted in terms of several interrelated, complex, frequency-dependent parameters such as impedance, conductance, capacitance, and dielectric permittivity. The complex permittivity ε was obtained from the capacitance C

2. MATERIALS AND METHODS 2.1. Materials. Two types of cellulose were used: Cladophora cellulose (G3095-10 batch; FMC Biopolymer, USA) and nanofibrillated cellulose (NFC; Innventia AB, Sweden). The materials were kind gifts from the respective companies. The NFC sample was produced from commercial never-dried bleached sulfite softwood dissolving pulp (Domsjö Fabriker AB, Sweden) by enzymatic pretreatment as described earlier.23 The solid-state properties of both algae- and woodderived nanocellulose samples used in this work have been partially characterized elsewhere.24 2.2. Paper Preparation. The Cladophora cellulose and the NFC paper sheets were prepared by vacuum filtration. A total of 0.57 g of each material was dispersed in 150 mL of deionized water using a high-shear ultrasonicator (Vibracell, Sonics, USA; 600 W, 20 kHz) for 4 min at 75% amplitude. The dispersion was then filtered through a Büchner funnel on filter paper. The collected wet cake was then dried in air fixed between two discs with screws to produce a 90-mm-diameter circular sheet of paper. 2.3. X-ray Diffraction (XRD). An X-ray diffractometer with Bragg−Brentano geometry was used (Diffraktometer D5000, Siemens, Germany). Cu Kα radiation was used (λ = 1.54 Å), and the angle of 2θ was set between 10 and 30°. The crystallinity index in % was calculated as follows:

ε = ε′ − iε″ =

C C0

(3)

where C0 is the capacitance of the empty probe station and ε′ and ε″ are, respectively, the real and imaginary parts of the permittivity. For the parallel plate probe station used in the current work,25 C0 = ε0A/d, where A is the electrode area (circular with a diameter of 1 cm), d is the electrode separation distance, and ε0 is the permittivity of free space (8.854 × 10−12 F m−1). The real part of the complex conductivity, σ′, is related to the imaginary part, σ″, of the dielectric permittivity as B

DOI: 10.1021/acs.jpcb.5b00715 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B σ ′ = ωε0ε″

Table 2. Solid-State Properties of the Nanocellulose Samples

(4)

where ω is the angular frequency. The dielectric response of the samples was measured at different relative humidities in air at room temperature by placing a Petri dish containing the various saturated salt solutions (Table 1) at the bottom of the closed compartment enclosing the probe station and letting the sample equilibrate for at least 24 h prior to the measurement.

crystallinity index, % bulk density, g/cm3 true density, g/cm3 porosity, % N2 surface area, m2/g H2O surface area, m2/g pore volumea, cm3/g

Table 1. List of Salts Used to Produce the Different Relative Humidities in the Dielectric Spectroscopy Measurements26 saturated salt solution

relative humidity, %

LiCl K2CO3 NaCl KNO3

11 43 75 94

Cladophora cellulose

NFC

93 0.895 1.606 44.3 107.0 92.63 0.464

66 1.523 1.577 3.44