Open Access

Synthesis of n-type Mg2Si/CNT Thermoelectric Nanofibers

Nanoscale Research Letters201712:343

Received: 27 March 2017

Accepted: 1 May 2017

Published: 10 May 2017


Magnesium silicide (Mg2Si)/carbon nanotube (CNT) thermoelectric nanofibers for use as a flexible thermoelectric material were successfully synthesized through the combined processes of the sol-gel method, magnesiothermic reduction, and liquid-solid phase reaction. In the resulting product, each CNT was coated with Mg2Si which was an approximately 60-nm-thick single crystal. The synthesized Mg2Si-coated CNTs exhibited n-type thermoelectric behavior confirming that n-type thermoelectric composite nanofibers were successfully obtained.


Carbon nanotubes Magnesium silicide Thermoelectric nanofibers Flexible thermoelectric material


Thermoelectric (TE) devices have attracted much attention as energy harvesting systems because they are able to convert thermal energy directly to electrical energy [1, 2]. In recent decades, most TE research has focused on inorganic materials such as semiconductors or conducting oxides [13]. Recently, in addition to rigid inorganic materials, flexible organic materials have been receiving much attention as TE materials that may provide mechanical flexibility and low manufacturing cost [46]. Conducting polymers such as poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) exhibit attractive thermoelectric properties around room temperature [710]. However, they are unstable at temperatures over 500 K, so their operating temperature is somewhat restricted [9, 10]. For flexible TE materials that can be used over 500 K, carbon nanotubes (CNTs) are a promising candidate because of their large Seebeck coefficient as exhibited in CNT films in addition to their superior electrical properties, chemical stability, good mechanical strength with excellent flexibility, and thermo-oxidative stability up to 873 K [1113]. However, most studies on TE materials consisting of CNTs have reported on only p-type samples and there are only a few studies on that of n-type samples, because n-doped CNTs are easily oxidized in air [14, 15]. It is well known that p-n-connected TE devices are capable of efficiently generating electric power. Consequently, n-type flexible TE materials that can be used at temperatures above 500 K have been in high demanded. Inorganic materials exhibiting a stable and high TE effect even at high temperature [16] are mainly semiconductors and are inherently rigid and brittle. Thus, it is difficult to fabricate a flexible film directly from inorganic TE materials. If CNTs with good mechanical strength and high electrical conductivity are individually coated with a stable, but rigid and brittle, inorganic n-type TE material, it is expected that the obtained composite nanofibers would become high-strength, thermally stable, n-type TE nanofibers, and a flexible TE film could be fabricated from those composite TE nanofibers.

In this study, we synthesize a new composite nanofiber where each CNT is partially coated with an inorganic TE material. We further propose that an n-type TE film made from these composite nanofibers could serve as a flexible TE material that can be used at relatively high temperature, i.e., above 500 K. Magnesium silicide (Mg2Si) was selected as the TE material for coating because not only are Mg and Si both non-toxic and abundant elements but Mg2Si has superior thermo-oxidative stability up to around 720 K [17, 18].The Mg2Si coating was successfully synthesized on individual CNTs via a three-step process consisting of (1) formation of a silica coating on the CNTs by the sol-gel method [1921], (2) magnesiothermic reduction [2124] of the silica coating to a silicon (Si) coating, and (3) synthesis of a Mg2Si coating by a liquid-solid phase reaction [25] of the silicon layer with Mg powder. Subsequently, measurement of TE properties revealed that the synthesized Mg2Si-coated CNT nanofibers exhibited n-type thermoelectric behavior.


Multiwalled carbon nanotubes with diameters of 20–70 nm, which were synthesized by catalytic chemical vapor deposition (CVD) and subsequently annealed above 2473 K (Hodogaya Chemical Co., Ltd.) [13], were suspended in a 3:1 mixture of concentrated H2SO4 and HNO3 (v/v) under ultrasonication in a water bath at 323 K for 24 h [26]. The resultant suspension was diluted with deionized water and filtered to collect acid-treated CNTs on a filter paper. After washing with deionized water and ethanol, they were dried at 323 K in air. Then, 0.4 mg of the acid-treated CNTs was dispersed in 160 ml of ethanol by ultrasonication for 1 h. Subsequently 20 ml of deionized water, 2 ml of tetraethylorthosilicate (TEOS), and 35 ml of 25% NH4OH were added to the suspension. The acid-treated CNTs were chemically coated with amorphous silica during 1 h of stirring. The obtained silica-coated CNTs were washed with ethanol, filtered, and dried overnight at room temperature. The silica-coated CNTs were mixed with magnesium powder with an average particle diameter of 180 μm (Kojundo Chemical Laboratory Co., Ltd.) at a molar ratio of silica to magnesium of 2:1. The amount of silica on the silica-coated CNTs was calculated from the geometry based on the thickness of the silica layer and assuming an average CNT diameter of 40 nm. The mixed powders were sealed in a steel tube with N2 gas. Then, the tube was heated to 1023 K in a tube furnace and held for 5 h in flowing N2 gas. The silica coating on the CNTs was reduced to a Si coating during this heat-treatment; however, treatment with 20 wt% HCl was necessary to remove by-products from the reacted mixture. Finally, the Si-coated CNTs were mixed with magnesium powder at a molar ratio of Si to magnesium of 3:1. The amount of Si in the Si-coated CNTs was assumed to correspond with the amount of Si resulting from complete reduction of the original silica coating. The mixed powders were sealed in a steel tube with Ar gas. The tube was heated to 973 K at a heating rate of 5 K/min in a tube furnace, held for 2 h in flowing Ar gas, and then cooled down to room temperature at a cooling rate of 1 K/min. This process allowed the Si coating to react with Mg resulting in Mg2Si-coated CNTs.

The obtained products were evaluated by X-ray diffraction (XRD) with Cu Kα radiation (RINT, Rigaku Corporation) and field-emission transmission electron microscopy (FE-TEM) with energy dispersive X-ray spectroscopy (EDX) (JEM-2100F, JEOL Ltd.). In order to obtain a bar-shaped specimen large enough to measure thermoelectric properties, acid-treated CNTs and Mg2Si-coated CNTs were pressed in a steel die, respectively, and then a mixture of epoxy resin and curing agent (CY221 and HY956, Nagase ChemteX Corporation; 5:1 by weight) was infiltrated into the pressed compacts followed by curing at room temperature for 24 h. The obtained compacts were cut into a rectangle (2 × 2 × 8 mm), and the Seebeck coefficient and electrical resistivity were measured using a ZEM-3 instrument (ADVANCE RIKO, Inc.) at room temperature. The maximum temperature difference applied to the sample during the Seebeck coefficient measurement was 30 K.

Results and Discussion

Figure 1 shows TEM images and an XRD profile of silica-coated CNTs. A uniform, continuous, 50-nm-thick coating was synthesized on individual CNTs (Fig. 1a and b). The EDX analysis revealed that this coating consisted of Si and O. The very diffuse ring visible in the selected-area electron diffraction (SAED) pattern of the coating (inset of Fig. 1b) and the broad peak in the XRD pattern (Fig. 1c) suggest that the synthesized layer is amorphous silica.
Fig. 1

TEM images (a, b) and XRD profile (c) of silica-coated CNTs

Figure 2a shows an XRD profile of the nanofibers after magnesiothermic reduction of the silica-coated CNTs. Diffraction peaks from Si and magnesium oxide (MgO) as well as a peak from the CNTs were observed, which indicates the reduction occurred according to the following reaction [22]:
$$ 2\mathrm{Mg}\left(\mathrm{g}\right)+{\mathrm{SiO}}_2\left(\mathrm{s}\right)\to\ 2\mathrm{M}\mathrm{g}\mathrm{O}\left(\mathrm{s}\right)+\mathrm{S}\mathrm{i}\left(\mathrm{s}\right) $$
Fig. 2

XRD profile of the composite nanofibers after magnesiothermic reduction (a) and subsequent HCl treatment (b)

Figure 2b shows an XRD profile of the reduced nanofibers after HCl treatment. The major peaks originate from CNT and Si. This result indicates that MgO is a by-product of the magnesiothermic reduction of silica and was removed by HCl treatment. From these XRD profiles, it is clear that silica was reduced to Si, and Si-coated CNTs were successfully obtained.

TEM images of the Si-coated CNTs are presented in Fig. 3. Each CNT was coated with a 20-nm-thick layer (Fig. 3a) that consists of nanoparticles about 5 nm in diameter (Fig. 3c). The SAED pattern (Fig. 3b) from the Si-coated CNTs also indicates that the coated fibers consist of Si and CNTs; no impurities were observed. Figure 3d shows a high-resolution image of nanoparticles. Lattice fringes with 0.31 nm spacing, which corresponds to the (111) planes of crystalline silicon, were observed. Thus, we conclude that the nanoparticles are crystalline Si. This finding regarding the magnesiothermic reduction of silica to produce Si nanoparticles is in good agreement with a previously reported result [27]. In addition, there is good contact between the Si and the CNT at the interface, and no reacted layer was observed.
Fig. 3

a TEM image of Si-coated CNTs. b SAED pattern of Si-coated CNTs. c Interface between CNT and silicon layer. d High magnification image of Si nanoparticles

Figure 4 shows an XRD profile of the nanofibers and TEM images of a single nanofiber after the reaction of Si-coated CNTs with magnesium powder. The XRD profile (Fig. 4a) indicates the presence of Mg2Si, CNTs, and MgO after the reaction; there was no evidence of unreacted Mg powder in the reacted mixture. Each CNT was coated with a layer about 60 nm thick (Fig. 4b). The SAED pattern in Fig. 4c, taken from the area indicated by the dashed circle in Fig. 4b, exhibits diffraction spots of Mg2Si. These results suggest the whole coating consists of an Mg2Si single crystal and confirm that Mg2Si-coated CNTs were successfully obtained.
Fig. 4

XRD profile and TEM images of the nanofiber after the reaction of Si with magnesium. a XRD profile of the nanofiber. b Bright field image. d The corresponding dark-field image taken with the MgO 200 diffraction spots. The area surrounded by dashed line in b indicates the observed area of SAED pattern. c SAED pattern of the nanofiber

Mg2Si is synthesized according to the following reaction:
$$ 2\mathrm{Mg}\left(\mathrm{l},\mathrm{g}\right)+\mathrm{S}\mathrm{i}\left(\mathrm{s}\right)\to\ {\mathrm{Mg}}_2\mathrm{S}\mathrm{i}\left(\mathrm{s}\right). $$

However, the mechanism of the formation of single crystal Mg2Si on CNTs is not well understood. The very slow cooling rate of 1 K/min during the liquid-solid phase reaction [25] may contribute to the growth of the single crystal. In Fig. 4c, in addition to the diffraction spots of Mg2Si, there are diffuse diffraction rings from MgO, which means nanocrystalline MgO also exists in the coating. Figure 4d shows a dark-field image of the nanofiber in Fig. 4b taken with the MgO 200 diffraction spots, and nanosized white spots are uniformly dispersed in the image of the matrix. Thus, we conclude that these white spots are MgO nanocrystals.

The possible reason for the presence of MgO within the Mg2Si single crystal is discussed as follows. In this study, in order to synthesize Mg2Si-coated CNTs, magnesium powder was reacted with Si-coated CNTs at a molar ratio of Si to magnesium of 3:1. In spite of an excess amount of magnesium relative to Si used for the reaction, unreacted magnesium was not observed after the synthesis process. Thus, some unknown reaction may occur during the heating process in addition to the synthesis of Mg2Si. One possibility for this reaction involves the presence of a small amount of silica, even after the magnesiothermic reduction, which was probably produced by natural oxidation on the surface of the Si nanoparticles visible in Fig. 3 [28, 29]. During the synthesis process of Mg2Si, this small amount of silica on the surface of Si nanoparticles is reduced by magnesiothermic reduction according to reaction (1), resulting in the formation of Si and MgO nanoparticles. This additional Si would also react with magnesium. Therefore, as the Mg2Si single crystal grows on the CNT, it incorporates MgO nanoparticles. This may be the reason why nanocrystalline MgO is dispersed inside the Mg2Si single crystal.

Regarding the thermoelectric properties of Mg2Si-coated CNTs, the electrical conductivity was 0.98 S/m, the Seebeck coefficient was −30 × 10−6 V/K, and the calculated power factor was 8.94 × 10−10 W/mK2 at room temperature. It is clear that Mg2Si-coated CNTs are n-type thermoelectric nanofibers, although the measured thermoelectric properties are somewhat lacking in accuracy because the measured specimen was actually an Mg2Si-coated CNT/epoxy composite. The measured thermoelectric properties are inferior to those of other CNT-based n-type thermoelectric materials [14, 15] and to those of bulk Mg2Si [18, 25]; in particular, electrical conductivity was very low. Compared to the measured electrical conductivity of acid-treated CNT/epoxy composite, which was 599 S/m and comparable with reported value of MWCNT compact [30], the electrical conductivity of Mg2Si-coated CNT was still extremely low. We believe this poor electrical conductivity is mainly due to the presence of MgO nanoparticles in the Mg2Si coating layer [31]. In addition, the contact resistance between Mg2Si coatings may be high because Mg2Si is a rigid material and the surface of the Mg2Si coating is not flat as shown in Fig. 4b. In order to improve the thermoelectric properties of Mg2Si-coated CNTs and to develop an n-type flexible TE film that can be used at relatively high temperature, it is necessary to reduce the amount of MgO nanoparticles in the Mg2Si coatings by optimizing the processing parameters. Also the volume and morphology of the Mg2Si coating should be controlled, not only to decrease the resistance between nanofibers by reducing the amount of contact between Mg2Si coatings and by increasing the amount of Mg2Si-CNT and CNT-CNT contact, but also to improve the flexibility of TE films by improving the deformability of the composite nanofibers. A study to address these problems is presently underway.


We have successfully synthesized Mg2Si-coated CNTs through the combined processes of the sol-gel method, magnesiothermic reduction, and liquid-solid phase reaction for use as a flexible TE material at temperatures above 500 K. The synthesized Mg2Si coating was an approximately 60-nm-thick single crystal, although it was not single phase as it contained MgO nanoparticles. The synthesized Mg2Si-coated CNTs exhibited n-type TE behavior confirming that n-type thermoelectric composite nanofibers were successfully obtained. Future work would involve improving the electrical conductivity and thermoelectric properties by optimizing the process parameters.



Carbon nanotubes


Chemical vapor deposition


Energy dispersive X-ray spectroscopy


Field-emission transmission electron microscopy


Magnesium silicide


Magnesium oxide




Selected-area electron diffraction






Tetraethyl orthosilicate


X-ray diffraction



The authors would like to thank Dr. T. Miyazaki from Tohoku University for performing the TEM observation. This work was supported by Grants-in Aid for Scientific Research from the Japan Society for the promotion of Science; JSPS (KAKENHI Grant Numbers JP25820363 and JP16K06790).

Authors’ Contributions

KK designed the experiment and drafted the manuscript. NN and AK participated in the experimental design and discussion of this research. KK and KY performed the experiments and the measurements. All authors read and approved the final manuscript.

Competing Interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (, which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.

Authors’ Affiliations

Department of Materials Processing, Graduate School of Engineering, Tohoku University


  1. Minnich AJ, Dresselhaus MS, Ren ZF, Chen G (2009) Bulk nanostructured thermoelectric materials: current research and future prospects. Energy Environ Sci 2:466–79View ArticleGoogle Scholar
  2. Zebarjadi M, Esfarjani K, Dresselhaus MS, Ren ZF, Chen G (2012) Perspectives on thermoelectrics: from fundamentals to device applications. Energy Environ Sci 5:5147–62View ArticleGoogle Scholar
  3. LeBlanc S, Yee SK, Scullin ML, Dames C, Goodson KE (2014) Material and manufacturing cost considerations for thermoelectrics. Renew Sustain Energy Rev 32:313–27View ArticleGoogle Scholar
  4. Poehler TO, Katz HE (2012) Prospects for polymer-based thermoelectrics: state of the art and theoretical analysis. Energy Environ Sci 5:8110–5View ArticleGoogle Scholar
  5. Dubey N, Leclerc M (2011) Conducting polymers: efficient thermoelectric materials. J Polym Sci Pt B-Polym Phys 49:467–75View ArticleGoogle Scholar
  6. Zhang Q, Sun YM, Xu W, Zhu DB (2014) Organic thermoelectric materials: emerging green energy materials converting heat to electricity directly and efficiently. Adv Mater 26:6829–51View ArticleGoogle Scholar
  7. Kim GH, Shao L, Zhang K, Pipe KP (2013) Engineered doping of organic semiconductors for enhanced thermoelectric efficiency. Nat Mater 12:719–23View ArticleGoogle Scholar
  8. Bubnova O, Khan ZU, Malti A, Braun S, Fahlman M, Berggren M, Crispin X (2011) Optimization of the thermoelectric figure of merit in the conducting polymer poly(3,4-ethylenedioxythiophene). Nat Mater 10:429–33View ArticleGoogle Scholar
  9. Zhang B, Sun J, Katz HE, Fang F, Opila RL (2010) Promising thermoelectric properties of commercial PEDOT:PSS materials and their Bi2Te3 powder composites. ACS Appl Mater Interfaces 2:3170–8View ArticleGoogle Scholar
  10. Sun YM, Sheng P, Di CA, Jiao F, Xu W, Qiu D, Zhu DB (2012) Organic thermoelectric materials and devices based on p- and n-type poly(metal 1,1,2,2-ethenetetrathiolate)s. Adv Mater 24:932–7View ArticleGoogle Scholar
  11. Zhao WY, Fan SF, Xiao N, Liu DY, Tay YY, Yu C, Sim DH, Hng HH, Zhang QC, Boey F, Ma J, Zhao XB, Zhang H, Yan QY (2012) Flexible carbon nanotube papers with improved thermoelectric properties. Energy Environ Sci 5:5364–9View ArticleGoogle Scholar
  12. Nakai Y, Honda K, Yanagi K, Kataura H, Kato T, Yamamoto T, Maniwa Y (2014) Giant Seebeck coefficient in semiconducting single-wall carbon nanotube film. Appl Phys Express 7:025103View ArticleGoogle Scholar
  13. Zhou WW, Sasaki S, Kawasaki A (2014) Effective control of nanodefects in multiwalled carbon nanotubes by acid treatment. Carbon 78:121–9View ArticleGoogle Scholar
  14. Yu CH, Murali A, Choi KW, Ryu Y (2012) Air-stable fabric thermoelectric modules made of N- and P-type carbon nanotubes. Energy Environ Sci 5:9481–6View ArticleGoogle Scholar
  15. Fukumaru T, Fujigaya T, Nakashima N (2015) Development of n-type cobaltocene-encapsulated carbon nanotubes with remarkable thermoelectric property. Sci Rep 5:7951View ArticleGoogle Scholar
  16. Sootsman JR, Chung DY, Kanatzidis MG (2009) New and old concepts in thermoelectric materials. Angew Chem Int Ed 48:8616–39View ArticleGoogle Scholar
  17. Kim HS, Kikuchi K, Itoh T, Iida T, Taya M (2014) Design of segmented thermoelectric generator based on cost-effective and light-weight thermoelectric alloys. Mater Sci Eng, B 185:45–52View ArticleGoogle Scholar
  18. Stathokostopoulos D, Chaliampalias D, Pavlidou E, Paraskevopoulos KM, Chrissafis K, Vourlias G (2015) Oxidation resistance of magnesium silicide under high-temperature air exposure. J Therm Anal Calorim 121:169–75View ArticleGoogle Scholar
  19. Satishkumar BC, Govindaraj A, Vogl EM, Basumallick L, Rao CNR (1997) Oxide nanotubes prepared using carbon nanotubes as templates. J Mater Res 12:604–6View ArticleGoogle Scholar
  20. Cui W, Du FP, Zhao JC, Zhang W, Yang YK, Xie XL, Mai YW (2011) Improving thermal conductivity while retaining high electrical resistivity of epoxy composites by incorporating silica-coated multi-walled carbon nanotubes. Carbon 49:495–500View ArticleGoogle Scholar
  21. Zhu XF, Xia BY, Guo MY, Zhang QA, Li JX (2010) Synthesis of carbon nanotube composites with size-controlled silicon nanoparticles. Carbon 48:3296–9View ArticleGoogle Scholar
  22. Cai Y, Allan SM, Sandhage KH (2005) Three-dimensional magnesia-based nanocrystal assemblies via low-temperature magnesiothermic reaction of diatom microshells. J Am Ceram Soc 88:2005–10View ArticleGoogle Scholar
  23. Richman EK, Kang CB, Brezesinski T, Tolbert SH (2008) Ordered mesoporous silicon through magnesium reduction of polymer templated silica thin films. Nano Lett 8:3075–9View ArticleGoogle Scholar
  24. Hai NH, Grigoriants I, Gedanken A (2009) Converting stober silica and Mediterranean sand to high surface area silicon by a reaction under autogenic pressure at elevated temperatures. J Phys Chem C 113:10521–6View ArticleGoogle Scholar
  25. Itoh T, Hagio K (2012) Thermoelectric properties of Mg2Si-based compounds synthesized partially using magnesium alloy. AIP Conf Proc 1449:207–10View ArticleGoogle Scholar
  26. Cho S, Kikuchi K, Kawasaki A (2012) On the role of amorphous intergranular and interfacial layers in the thermal conductivity of a multi-walled carbon nanotube-copper matrix composite. Acta Mater 60:726–36View ArticleGoogle Scholar
  27. Kim KH, Lee DJ, Cho KM, Kim SJ, Park JK, Jung HT (2015) Complete magnesiothermic reduction reaction of vertically aligned mesoporous silica channels to form pure silicon nanoparticles. Sci Rep 5:9014View ArticleGoogle Scholar
  28. Li X, Yan P, Arey BW, Luo W, Ji X, Wang C, Liu J, Zhang JG (2016) A stable nanoporous silicon anode prepared by modified magnesiothermic reactions. Nano Energy 20:68–75View ArticleGoogle Scholar
  29. Nakamine Y, Kodera T, Uchida K, Oda S (2011) Removal of surface oxide layer from silicon nanocrystals by hydrogen fluoride vapor etching. Jpn J Appl Phys 50:115002View ArticleGoogle Scholar
  30. Tang Y, Gou JH (2010) Synergistic effect on electrical conductivity of few-layer graphene/multi-walled carbon nanotube paper. Mater Lett 64:2513–6View ArticleGoogle Scholar
  31. de Boor J, Dasgupta T, Kolb H, Compere C, Kelm K, Mueller E (2014) Microstructural effects on thermoelectric efficiency: a case study on magnesium suicide. Acta Mater 77:68–75View ArticleGoogle Scholar


© The Author(s). 2017