Full Length ArticleComputational investigation of hydrogen-induced phonon changes in carbon fiber☆
Graphical abstract
Introduction
Carbon fiber (CF) is used in a variety of applications that require the advantages of lightweight and strong materials, such as the aeronautical and automotive industries. However, the mechanical properties of CF vary significantly under different manufacturing conditions [1], [2], [3]. Polyacrylonitrile, or PAN, is the most widely used CF precursor and generally yields fibers with high–tensile strength. It is often combined with acidic comonomers to increase drawability and promote polymerization of the fiber [4]. Mesopitch-based fibers are also produced and usually have a high modulus [5]. The manufacturing steps with the largest influence on the final material strength is the carbonization and graphitization temperatures, which are the temperatures to which the CF is heated to complete the cyclization of the precursor material, remove noncarbon components from the fiber matrix and increase the graphitic orientation of the fiber core. The tensile strength typically increases with increasing treatment temperatures up to 1600 °C; however, fibers are often heated up to 3000 °C to increase fiber modulus at the cost of reduced tensile strength [5]. Many credit the high-temperature increase of the tensile modulus to a decrease in defect density at higher treatment temperatures [5], [6], but this does not explain the abrupt reduction in tensile strength above 1600 °C. As such, the exact relationship between CF processing conditions, fiber microstructure, and tensile properties remains an open question.
To prevent the critical failure of CF during use, it is valuable to monitor the properties of CF materials using a noninvasive technique. Optical vibrational spectroscopy has become the primary tool for such investigations, and Raman spectroscopy in particular has been used extensively to study the vibrational properties of carbonaceous materials due to the perceived simplicity of the spectra [7], [8], [9], [10], [11], [12], [13], [14], [15], [16]. The most prominent peaks are the G peak near 1580 cm−1, the D peak around 1350 cm−1, and the D' peak around 1620 cm−1, though additional peaks have been reported [17], [18]. Despite the presence of only a few peaks, Raman spectroscopy is subject to selection rules, and interpretation of the spectra can become complex, especially when selection rules begin to break down because of defects [19], [20].
As Brubaker et al. showed, extracted information from a Raman spectrum, such as peak locations and widths, is extremely sensitive to the fitting procedure used. Their recommended fitting procedure varies slightly for different fiber modulus classes [20]. Inelastic neutron scattering (INS) experiments are a valuable supplement to Raman spectroscopy because INS has no selection rules. Because of this, Brubaker et al. performed INS experiments to probe the entire vibrational spectra of several commercially available CF [21]. Owing to the large incoherent neutron scattering cross section of hydrogen compared to other prevalent elements in CF, Brubaker et al. identified a correlation between tensile modulus and hydrogen content. A peak around 3000 cm−1 was assigned to carbon–hydrogen stretching modes, and the area under this peak increased (i.e., higher hydrogen concentrations) for CF with lower tensile modulus. Additionally, Brubaker et al. identified INS features around 900–950 cm−1 and 1100 cm−1 that varied with tensile modulus. Guided by previous work on activated carbons, they assigned these features to out-of-plane and in-plane carbon–hydrogen vibrations, respectively [22], [23], [24].
Although the work of Brubaker et al. established a connection between hydrogen defects and the mechanical properties of CF, the atomic structure of the hydrogen defects in CF is not clear. The work on activated carbons investigated hydrogen atoms bound to the edges of graphene flakes, a local environment which does not include the interlayer interactions expected in CF. Previous work has shown that the extent of out-of-plane displacement of hydrogen-bound carbon atoms is reduced by approximately 0.2 Å in graphite compared to graphene because of the interlayer van der Waals interactions between the graphitic layers [25]. Additionally, the carbon–hydrogen bond of hydrogen adsorbed to graphene vacancies, which is highly favorable at standard atmospheric conditions, has been calculated to be 0.01–0.08 Å less than hydrogen adsorbed to pristine graphene [25], [26]. Consequently, it is important to extend the knowledge of the vibrational spectra of hydrogen adsorbed on graphene flakes to include interlayer interactions and hydrogen adsorption to other common CF defects, particularly monovacancies, which we investigated recently [27].
In this work, we have performed density functional theory (DFT) calculations to simulate the INS spectra of several hydrogen defects in graphite as a surrogate for CF, and we have investigated the changes to the phonons induced by these defect structures. In addition to providing support for the experimental peak assignments, we showed that the INS features previously reported for terminal hydrogen adsorption on graphene flakes can be extended to hydrogen substitution and intercalation in bulk graphite, although the designation of in-plane versus out-of-plane bending modes is complicated in the multilayer material. This is especially true when defect–defect interactions are present, which can also notably change the shape and position of the phonon density of states and INS peaks.
Section snippets
Methods
We investigated the vibrational properties of hydrogen-defective graphite using spin-polarized DFT as implemented in the electronic structure theory package Real-space Multigrid (RMG) DFT [28], [29], which was executed on the Summit supercomputer at the Oak Ridge Leadership Computing Facility. RMG uses real-space descriptions of the wave functions, charge densities, potentials, and the Kohn–Sham Hamiltonian to reduce the number of Fourier transforms between real and reciprocal space needed
Defect structures
There are two nonequivalent carbon atoms in pristine graphite. The first, which we refer to as A-type atoms, sit directly between carbon atoms in the neighboring graphene sheets, while the second (B-type carbon atoms) sit directly between hexagon centers (Fig. 1A). As such, there are two nonequivalent substitution sites in graphite, though their formation energies differ by only 0.006 eV. We find the substitutions of B-type carbon atoms are slightly lower in energy (formation energy of 6.03 eV)
Conclusions
In conclusion, we used DFT to simulate the INS spectra of hydrogen substitution and intercalation in bulk graphite to aid in the interpretation of recent experiments on commercially available CF. We provided additional support for the assignment of the peaks near 950, 1100, and 3000 cm−1 to carbon–hydrogen vibrations and showed that the interlayer interactions of graphite are not strong enough to alter the frequency of the modes as reported for activated carbons. However, the designation of
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
This research used resources of the Oak Ridge Leadership Computing Facility, which is a US DOE Office of Science User Facility supported under Contract no. DE-AC05-00OR22725.
The authors would like to thank Roger J. Kapsimalis for insightful discussions on the work, the developers of RMG for user assistance and access to their RMG-Phonopy interface, and Y. Q. Cheng for assistance with OCLIMAX.
Data availability
The data presented in this manuscript is available upon request from the corresponding author.
References (43)
Processing, structure, and properties of carbon fibers
Compos. A Appl. Sci. Manuf.
(2016)- et al.
Raman microprobe studies on carbon materials
Carbon
(1994) - et al.
Study of carbon-fibre strain in model composites by Raman spectroscopy
Compos. Sci. Technol.
(1996) - et al.
Raman microspectroscopy of soot and related carbonaceous materials: Spectral analysis and structural information
Carbon
(2005) - et al.
A reconsideration of the relationship between the crystallite size La of carbons determined by X-ray diffraction and Raman spectroscopy
Carbon
(2006) - et al.
Densification mechanism of polyacrylonitrile-based carbon fiber during heat treatment
J. Phys. Chem. Solids
(2011) - et al.
A Raman study to obtain crystallite size of carbon materials: A better alternative to the Tuinstra-Koenig law
Carbon
(2014) - et al.
Investigating nanostructures in carbon fibres using Raman spectroscopy
Carbon
(2018) - et al.
Carbonization of polyacrylonitrile-based fibers under defined tensile load: Influence on shrinkage behavior, microstructure, and mechanical properties
Polym. Degrad. Stab.
(2019) - et al.
A multi wavelength Raman scattering study of defective graphitic carbon materials: The first order Raman spectra revisited
Carbon
(2016)
Neutron scattering study of the terminating protons in the basic structural units of non-graphitising and graphitising carbons
Carbon
Structure and stability of hydrogenated carbon atom vacancies in graphene
Carbon
Pseudopotentials for high-throughput DFT calculations
Comput. Mater. Sci.
Reviewing computational studies of defect formation and behaviors in carbon fiber structural units
Comput. Mater. Sci.
Resonant Raman spectra of graphene with point defects
Carbon
The stretching vibration of hydrogen adsorbed on epitaxial graphene studied by sum-frequency generation spectroscopy
Chem. Phys. Lett.
Carbon fibers: precursor systems, processing, structure, and properties
Angew. Chem. Int. Ed. Engl.
Reinforcement systems for carbon concrete composites based on low-cost carbon fibers
Fibers
Carbon Fiber Composities
Fabrication and properties of carbon fibers
Materials
Mesophase pitch and its carbon fibers
Pure Appl. Chem.
Cited by (1)
Connecting mechanical properties to hydrogen defects in PAN-based carbon fibers
2023, Physical Review Materials
- ☆
Notice: This manuscript has been authored by UT-Battelle, LLC, under contract DE-AC05-00OR22725 with the US Department of Energy (DOE). The US government retains and the publisher, by accepting the article for publication, acknowledges that the US government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for US government purposes. DOE will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (https://energy.gov/downloads/doe-public-access-plan).