Monatshefte für Chemie - Chemical Monthly

, Volume 147, Issue 2, pp 269–278 | Cite as

Self-supporting hierarchically organized silicon networks via magnesiothermic reduction

  • Michael Waitzinger
  • Michael S. Elsaesser
  • Raphael J. F. Berger
  • Johanna Akbarzadeh
  • Herwig Peterlik
  • Nicola Hüsing
Original Paper

Abstract

Using a magnesiothermic reduction process, the successful conversion of hierarchically organized meso/macroporous silica monoliths to self-supporting macroporous silicon networks comprising interparticle mesoporosity has been demonstrated. By careful variation of reaction time and temperature, the final network structure can be controlled to a large degree. Scanning and transmission electron microscopy images indicate that the cellular silicon structure is built up from aggregated 1–15 nm sized crystalline silicon particles aggregated to struts that form macropores of approximately 2 µm diameter.

Graphical abstract

Keywords

Silicon compounds Reductions Gels Material science Nanostructures 

Introduction

For many technical applications, e.g. separation science, heat insulation, adsorbents, optics, semiconductors, sensors, energy-related applications, and many more, porous materials play a vital role [1]. The pore sizes, the size distribution, tortuosity of the network, and the surface properties of such porous materials influence to a large extent its interaction with the environment, and thus the performance in the specific application. Therefore, large research efforts are devoted to the development of synthetic approaches that allow for a rational design of porous networks. Such pathways are quite well established for porous metal oxides by applying wet chemical processing routes, e.g. templating mechanisms or phase separation strategies. However, these approaches typically cannot be transferred easily to porous metals/semimetals [2].

Porous silicon is a material of high technological interest, not only due to its potential as high capacity anode material in lithium ion batteries, but also for medical applications due to its biocompatibility, e.g. as biosensor [3, 4, 5]. A typical method to produce porous silicon supported on substrates is based on anodic etching. However, in this case, control of the porous structure is limited and no self-supporting samples can be formed [6]. A very promising method that allows for a high level of structural control in the production of silicon has been published by Sandhage et al. [7]. They could show that it is possible to completely replicate the intricate and delicate structures of silica diatoms to silicon by reduction with gaseous magnesium. Typical side products in this reaction are magnesia (MgO) and magnesium silicide (Mg2Si), which both can be removed by treatment with hydrochloric acid. The only prerequisite for a successful reproduction of the silica structure is the formation of a co-continuous silicon network, which is given by the formation of aggregated silicon nanocrystals.

The already achieved high level of control in the deliberate design of porous silica provides the perfect precursors for such reduction processes and since the seminal work of Sandhage et al. several groups followed their approach to yield silicon structures. Kuryavyi et al. and Cui et al. used renewable silica sources, such as plant raw materials (e.g. rice husks and straw, scouring horsetail, larch needles) to produce porous silicon with high purity of >97 %, which can be “used as a starting material for the production of solar grade and semiconductor silicon” or as anode material in lithium ion batteries [8, 9]. Even tabasheer (an excretion from the nogal joints of certain species of bamboo) was found a suitable source for silicon of comparable quality (~96 %) [10]. In contrast to the natural silica sources, there are plenty of investigations on the application of silica synthesized from alkoxysilanes [11]. Sandhage et al. converted mesoporous silica aerogels into monolithic silicon aerogels, while Huang et al. reduced tetramethoxysilane impregnated natural cellulose to nanofibrous silicon [12, 13]. Interwoven microfilaments of silica, prepared via the application of synthetic peptides, gave interwoven microfilaments of silicon [14]. Meso- or macroporous silica prepared in the presence of structure-directing agents can be converted to mesostructured porous silicon film, silicon opal structures, hollow silicon particles or mesoporous particles [15, 16, 17, 18, 19, 20, 21, 22, 23, 24].

We report here for the first time the direct conversion of hierarchically organized silica monoliths to hierarchically organized silicon. The aim of this work was to employ the potential of the magnesiothermic reduction process for the replication of silica objects comprising a hierarchical network build-up. Besides a macroscopic, self-supporting morphology with dimensions of millimetres, a unique macroporous cellular structure consisting of amorphous silica struts that comprise mesoscopically organized pores is subjected to magnesium at high temperatures. The magnesiothermic reduction of these unique silica structures is investigated in detail with special emphasis on the degree of control of the replication at the different length scales. The influence of reaction variables, such as temperature and time, on the final structure is discussed.

Results and discussion

Figure 1 shows a mechanically stable, calcined silica monolith (c-SiO2, photograph) that is built up from an amorphous cellular silica network (SEM image) with macropores of approx. 2 µm, with every single silica strut comprising periodically arranged (2D hexagonally arranged) mesopores of about 8 nm in diameter (TEM image). The structure is also schematically depicted.
Fig. 1

The monolithic structure of c-SiO2 (photograph with slices as used for reduction, left) exhibits an interconnected, cellular macroporous network (SEM image, middle) built from struts with a highly organized 2D hexagonally arranged mesopore system (TEM, right)

In the first part of the manuscript, a representative sample is discussed to give a detailed insight into the structural evolution of the final silicon network during all synthesis steps for given reaction parameters (reaction temperature 700 °C, reaction time 1 h and Mg/SiO2 ratio of 2.1; sample 700–01), while in the following second section the focus is on the influence of reaction parameters, such as reaction time and temperature on the resulting silicon structure. For a complete conversion of silica to silicon, the Mg/SiO2 ratio needs to be 2.0 (from the chemical reaction). Higher amounts of magnesium favour the formation of Mg2Si, lower amounts the presence of residual SiO2. To ensure a complete conversion, the molar ratio Mg/SiO2 was set to 2.1 in this work.

After exposure to Mg vapour at 700 °C for 1 h, the reaction product appears deep brown with a few white spots, indicating regions of non-reacted silica. This can probably be related to the short reaction time of only 1 h. A bluish cast on the sample can be attributed to the formation of Mg2Si. The colour of the samples changes from deep to light brown upon further treatment with both hydrochloric and acetic acid in a first step and hydrofluoric acid in a second step to remove MgO/Mg2Si as well as non-reacted SiO2, respectively.

The apparent physical resilience of the monolithic silica samples increases after the magnesiothermic reduction. This is due to the formation of MgO filling up the mesopores leading to two continuous interpenetrating networks. After dissolution of MgO and SiO2, the monolithic silicon pieces can easily be broken up again. XRD analyses of samples after magnesiothermic reduction from different washing procedures are shown in Fig. 2.
Fig. 2

XRD patterns of magnesiothermically reduced silica 700-01 (a), HCl-leached sample (b), and HF-leached sample (c)

The XRD pattern of the silica starting material (not shown) only exhibits a very broad signal in the range of 2θ = 20°–30°  associated with amorphous SiO2. For the sample taken directly after the reduction process, the predominance of cubic Si, cubic MgO (periclase) and cubic Mg2Si is obvious; however, a broad background signal for amorphous SiO2 can also be detected (Fig. 2a). Mg2SiO4 as reported by Larbi and others is not found [25, 26, 27, 28]. When MgO is removed by treatment with (1) hydrochloric acid the corresponding pattern indicates the presence of silicon in an amorphous silica matrix (Fig. 2b), while after treatment with (2) hydrofluoric acid the sole presence of crystalline silicon is observed (Fig. 2c). Using the Scherrer equation, an average crystallite size of about 7 nm for silicon after reduction and washing with HF (see also Table 1) can be calculated. Crystallites of MgO and Mg2Si are larger in size with 15 and 37 nm, respectively.
Table 1

Structural parameters of the reduced and acid-leached samples

Sample

SSABET/(m2 g−1)

DBJH(ad) /(nm)

Vmax/(cm3 g−1)

Crystallite size/nm

Si (111)

MgO (200)

Mg2Si (220)

As reduced

 700-01

10

n.a.

14

7

15

37

 700-05

9

n.a.

1

11

15

n.a.

 700-10

1

n.a.

14

25

22

30

 750-01

59

n.a.

59

16

18

n.a.

 750-05

2

n.a.

15

17

21

n.a.

 750-10

3

n.a.

14

21

23

n.a.

HCl leached

 700-01

323

7.0; 26.2

530

 700-05

201

6.5; 26.5

477

 700-10

193

25.9

649

 750-01

181

6.5; 37.4

484

 750-05

112

n.a.

467

 750-10

101

n.a.

434

HF leached

 700-01

189

6.5; 37.4

1512

7

 700-05

125

n.a.

794

15

 700-10

84

n.a.

734

13

 750-01

89

n.a.

485

17

 750-05

96

n.a.

266

22

 750-10

57

n.a.

148

24

Besides the crystalline lattice, the evolution of the ordered mesopore structure is followed by SAXS measurements during the reduction process (Fig. 3). For the calcined silica sample (c-SiO2), a diffraction pattern which clearly can be assigned to a 2D hexagonal arrangement of the mesopores with (10), (11), (20), and (21) reflections is obtained. The SAXS pattern of the samples after the reduction reaction shows no reflections supporting a filling of the mesopores with MgO, which leads to a smaller electron density contrast between the different phases. After removal of MgO by washing with hydrochloric acid, a small—less pronounced—reflection becomes visible, which is slightly shifted with a decreased lattice constant of 11.2 nm (from 11.6 nm) compared to the original silica sample. This is a very strong indication of the presence of remaining silica (in agreement with the white spots that have been observed macroscopically), in addition to the loss of all reflections after leaching with HF. We note, that in contrast to our findings it has been reported previously that the mesoscopic long-range ordering of a silica sample was retained under similar reduction procedures [15, 16].
Fig. 3

SAXS patterns of non-reacted c-SiO2, the magnesiothermically reduced silica 700-01 (a), HCl-leached sample (b), HF-leached sample (c)

Nitrogen adsorption/desorption isotherms obtained for silica and silicon samples are shown in Fig. 4. The original silica monolith exhibits a type IV isotherm with H1-type hysteresis loop supporting the mesoporous character of the sample with a monomodal pore size distribution with pores of about 7.9 nm. Reduction of silica to Si, MgO, and Mg2Si results in a significant decrease of the amount of adsorbed nitrogen. The low value of 14 cm3 g−1 for the maximal adsorbed volume (Vmax) for the sample after reduction shows that almost all pores are blocked by MgO or Mg2Si. Upon removal of MgO, mixed isotherms of type II and IV with a H1 hysteresis loop in the mesoporous region are obtained, indicating the presence of meso- and macroporous silica. With dissolution of MgO or Mg2Si, Vmax dramatically increases to a value of 530 cm3 g−1. Washing with hydrofluoric acid to remove non-reacted silica leads to a type II isotherm with a H1 hysteresis loop and large Vmax of 1512 cm3 g−1 (see also Table 1). The reduction approach as chosen here does not allow to retain the mesoscopic ordering of the pores in the silica monolith in the resulting silicon structure. As a possible explanation, the high temperatures which are generated locally in this exothermic reaction can be named. These high temperatures might induce sintering processes and diffusion-based rearrangements, thus resulting in materials not showing any long-range ordering of the pores. In addition, the relation of pore size and wall thickness of the silica precursor structure seems to play an important role.
Fig. 4

Adsorption–desorption isotherms from nitrogen sorption of magnesiothermically reduced silica 700-01 (a), HCl-leached sample (b), HF-leached sample (c), and c-SiO2

BET analyses yielded specific surface area (SSA) values of 583 m2 g−1 for the hierarchically organized SiO2 and values of only 10 m2 g−1 for the sample directly after reduction again confirming the filling of the mesopores (Table 1). Removal of MgO by hydrochloric acid increases the average SSA to values of about 323 m2 g−1, but post-treatment with hydrofluoric acid to remove residual silica again results in a decrease of the SSA to 189 m2 g−1.

A direct comparison of the nitrogen sorption data with values given in other reports is difficult, because in most reported examples, no HF treatment has been applied to remove residual silica. The values obtained for the specific surface area of the pure silicon sample in our case are rather low. For instance, Bao et al. observed an increase from 1.5 m2 g−1 (starting with non-mesoporous diatoms) to 541 m2 g−1 at a reaction temperature of 650 °C and a holding time of 2.5 h [29]. This increase can be associated with the formation of new micro- and mesopores. A contrary result is given by Guo et al. [22] for the magnesiothermic reduction of mesoporous SBA 15. In this case, the SSA increased from 543 to 606 m2 g−1 still showing ordered mesopores, but no HF treatment is reported. For SBA 16, the ordered mesopores collapsed, resulting in a decrease in the SSA values from 680 m2 g−1 to 231 m2 g−1. In both cases, the authors claim that “the broader pore size distribution implies either the existence of hierarchical pores within the wall or a non-ordered mesoporous structure” [22]. As mentioned above, for the here presented samples, only non-organized mesopores from interparticle porosity can be found.

Scanning electron microscopy images and energy-dispersive X-ray analyses (EDX) of the 3D porous structure directly after the reduction process and after all washing procedures are shown in Fig. 5. For the magnesiothermically reduced sample, particles forming struts very similar to the cellular structure of the SiO2 template are observed. Additionally, crystals of Mg2Si are found on the rods indicated by the wrinkled surfaces. After acid leaching, equivalent to the removal of the MgO and Mg2Si crystals, a 3D network structure very similar to the starting silica gel remains. Due to a gradient of diffusion of Mg vapour, the size of interconnected “particles” differs (Fig. 5b, c) [30]. Lowering magnification reveals a macroporous network, which is built up of interconnected rods (Fig. 5d) as the original silica matrix. In the EDX spectrum (of a sample after all acid treatments), no signal for Mg is detected, confirming the complete removal of MgO and Mg2Si.
Fig. 5

SEM images of 700-01 (as reduced) (a), and (after leaching with HCl and HF) (be), with EDX inlay of e)

Transmission electron microscopy (TEM) images are presented in Fig. 6. The fully leached silicon product shows fine silicon struts comprising meso- and macropores. Mesoporosity is formed by agglomeration of dense silicon particles as interparticle porosity.
Fig. 6

TEM images of silicon samples after complete leaching with HCl and HF, a 700-01 and b 750-10

One major challenge in the magnesiothermic transformation of silica to silicon with preservation of the network structure is the highly exothermic character of the reaction. In the following, the different reaction parameters and their influence on the chemical composition of the product as well as the final network will be discussed.

Temperature

In the following discussion of the influence of the temperature, we focus at the reactions carried out at 700 and 750 °C. Both temperatures are above the melting point of magnesium (650 °C). In addition, it was found experimentally that with a reaction temperature of only 650 °C the yield for the given time was very low.

An increase of the reaction temperature from 700 to 750 °C (samples 700–01 and 750–01) leads to a significant increase of the crystallite sizes of Si (from 7 to 16 nm) and to a smaller extent for MgO (from 15 to 18 nm) (see also Table 1). This trend is still observable after full acid leaching (c.f. Fig. 6). Concomitantly, the specific surface area (SSABET) decreases from 189 m2 g−1 (700 °C) to 89 m2 g−1 (750 °C) and Vmax of adsorbed N2 is reduced from 1512 cm3 g−1 (700 °C) to 485 cm3 g−1 (750 °C) (Table 1). This trend can also been seen prior to the acid leaching processes, but the presence of silica renders a further discussion impossible.

Reduction time

The reaction time at 700 or 750 °C was varied from 1 h to 5 h to 10 h. For very short reaction times of only 1 h at 700 °C (700–01), a small amount of non-reacted Mg can still be observed in the stainless steel boat. For the monolithic fragments, the white spots (non-reacted silica) disappear with increasing reaction time. It can be concluded that a short reaction time and low temperatures (700 °C) cause an incomplete conversion of silica to silicon probably due to diffusion problems and thus, a slow transport of Mg vapour. Longer reaction times also result in a decrease of the amount of Mg2Si. For samples that have been reacted for 10 h, the signal of Mg2Si could not be detected anymore in the XRD patterns, consistent with the reaction of Mg2Si to MgO and Si according to Eq. (1) and a loss of the bluish colour of the samples.
$$ {\text{Mg}}_{ 2} {\text{Si}}\; ( {\text{s) }} + {\text{ SiO}}_{ 2} \; ( {\text{s) }} \to {\text{ 2 Si (s) }} + 2 {\text{ MgO}}\; ( {\text{s)}} $$
(1)

In addition, Mg2Si can decompose at a temperature of 850 °C to yield Mg vapour and elemental silicon [31]. An increase of the reaction time also facilitates crystallization through Ostwald ripening of both Si and MgO.

As a consequence of the large macropores in our system, the diffusion of magnesium vapour is facilitated, but the thickness of the used monolith is a limiting factor. When thick monoliths (around 5 mm in diameter) are reacted, a gradient of conversion is observed; presumably caused by a thermal gradient, which leaves a white unreacted core. This can only be overcome using smaller pieces. Longer furnace residence times change both SSA and Vmax to lower values.

An increase in reaction temperature and holding time from 700 to 750 °C and 1 to 10 h, respectively, leads to an increase in the crystallite size as calculated by the Scherrer equation to >20 nm and a growth of the diameter of the silicon struts (Fig. 7). This growth is caused by sintering of both MgO and Si particles, which can easily be explained by the fact that the reaction temperature lies between the Hüttig temperature (THüttig \( \approx \) ½ TMelting temperature \( \approx \) 750 °C for MgO, T in Kelvin) and the Tammann temperature (TTammann \( \approx \) ½ TMelting temperature \( \approx \) 570 °C for Si, T in Kelvin). These temperatures predict the start of surface diffusion of atoms (Hüttig) or the formation of agglomerates (Tammann) [32].
Fig. 7

SEM images of 750-10 (after all leaching processes) at different magnifications and EDX inlay of b. The circles present sintered areas of silicon

Proposed mechanism

The structural features of the reduced silicon strongly depend on the silica source, reaction time, temperature, reaction setup and post-reduction treatments [5]. With respect to the silica source, characteristics such as sample size, available amount of silica, pore arrangement and tortuosity, pore diameter and ratio pore diameter to wall thickness have to be carefully considered when comparing data from the literature.

Based on our results and literature studies, the following mechanistic sequence for the magnesiothermic reduction of hierarchically organized silica is proposed [22, 33, 34]. Upon exposure of Mg vapour to meso/macroporous silica, magnesium gets in contact in either the gaseous or the condensed phase due to continuous cooling with diffusion into the porous silica body. STA studies have shown that in our case the onset temperature for the magnesiothermic reduction is found to be at 486 or 520 °C for heating rates of 2 or 10 K min−1, respectively. After reaching this temperature, magnesium reacts with the amorphous silica, and particles of MgO and Si are formed. We propose that the MgO particles are preferentially formed in the mesopores, which is strongly supported by the loss in mesoporosity for MgO/Si composite samples. Despite this segregation into MgO- and Si-rich domains, two interconnected networks are formed, which is a prerequisite for a 3D interconnected, self-supporting silicon network. Upon removal of MgO (and Mg2Si) as well as residual silica, a network built-up of small silicon crystallites of silicon particles comprising mesoporosity is formed. The silicon crystallites are arranged in struts comprising a macroporous, cellular structure. This indicates that the reduction reaction is spatially confined to the mesoporous silica struts by the macropores of approx. 2 µm in diameter resulting in a very good replication of the cellular structure. As a general trend, it has been found that with increasing reduction temperature and longer holding times, the silicon particles are less filigree, resulting in more stable 3D structures.

Conclusions

Hierarchically organized meso-/macroporous silica monoliths have been converted in a shape-preserving displacement reaction by magnesium vapour to 3D-interconnected meso-/macroporous, hierarchically organized silicon. After reduction, a non-porous Si/SiO2/MgO composite (with small amounts of Mg2Si) is obtained, which can be transferred by selective acid dissolution to the respective porous silicon structures. These free-standing 3D silicon networks are build up from silicon nanocrystallites (7 nm average size) agglomerated in struts comprising intercrystallite mesoporosity to yield a macroporous, cellular network. These highly porous networks possess specific surface areas of up to 190 m2 g−1. With longer reaction time (up to 10 h) and higher reaction temperatures, sintering of the silicon particles results in stabilization of the filigree network.

Experimental

Tetraethoxysilane (Merck) and trimethylchlorosilane (Sigma-Aldrich) were used without further purification. Ethylene glycol (Aldrich) was purified by drying with Na2SO4 (Prolabo). Pluronic® P123 [M(av.) = 5800 g mol−1, EO20PO70EO20, BASF], petroleum ether (40–60 °C, Prolabo), magnesium powder (for synthesis, Merck) and hydrofluoric acid (40 % p. a. Merck) were used without further purification.

Preparation of hierarchically organized silica monoliths

Tetrakis(2-hydroxyethyl)orthosilicate (EGMS) was synthesized according to the procedure published by Brandhuber and Hartmann et al. [35, 36]. Hierarchically organized silica gels were prepared by sol–gel processing of EGMS in an aqueous reaction mixture containing Pluronic® P123 as the structure-directing agent and hydrochloric acid as the catalyst, according to a weight ratio of SiO2/Pluronic® P123/0.1 M HCl: 18/30/70 wt %. The reaction mixture was homogenized in an ice bath for 1 min using a magnetic stirrer and allowed to gel in closed acrylic glass cylinders at 40 °C, followed by an ageing period of 10 d at this temperature. In a next step, the whole gel bodies were immersed in a trimethylchlorosilane/petroleum ether (1/9 w/w) solution for at least 10 h. After washing with petroleum ether (until no emulsion formation is observed anymore) and ethanol (five times), the wet gel bodies were dried at room temperature for 24 h. The dry monoliths were cut in quartered slices with a scalpel and calcined for 1 h at 750 °C with a heating rate of 1 K min−1 in air to remove the remaining organic content.

Reduction of SiO2 to mesoporous Si

The calcined SiO2 slices, positioned on a stainless steel grid, were placed in a stainless steel boat filled with magnesium powder. The molar ratio of Mg/SiO2 was set to 2.1 with a typical sample weight of approximately 250 mg of the calcined silica monoliths. The boat was placed in a stainless steel tube, bolted with end caps and heated to 700 or 750 °C (denoted by 700 and 750), respectively, with a heating rate of 10 K min−1 under a slight argon flow. The reaction temperature was kept for 1 h, 5 h, or 10 h (denoted by 01, 05, or 10). The samples were allowed to slowly cool down to room temperature to yield the Si/MgO composites.

Recovery of silicon from reaction products

The reaction product was wetted with 1 cm3 of degassed water, first with 5 cm3 of glacial acetic acid and second with a mixture of 20 cm3 of 2 M hydrochloric acid for 4 h at 40 °C. While MgO is converted to water-soluble MgCl2, Mg2Si is thereby dissolved to gaseous SiH4, which is highly pyrophoric upon contact with air. Afterwards, the monolithic pieces were washed 6 times for 30 min with 40 cm3 of degassed water at room temperature resulting after drying in argon atmosphere in light brown monoliths. To remove unreacted silica, the dried monoliths were immersed in a mixture of 3 cm3 of 40 % hydrofluoric acid, 3 cm3 of 96 % ethanol, and 3 cm3 of purified water for 3 h. After extensive washing with water, the monoliths were again dried in an argon-filled desiccator.

Characterization

Nitrogen sorption

Nitrogen sorption isotherms were recorded at 77 K using a sorption porosimeter (Micromeritics, ASAP 2420). All samples were degassed for 3 h at 100 °C in vacuo. The Brunauer–Emmett–Teller (BET) surface area was evaluated using adsorption data in a relative pressure range p/p0 0.05–0.25. The mesopore size distribution was calculated on the basis of the adsorption branch using the Barrett–Joyner–Halenda (BJH) model [37].

X-ray scattering/diffraction

The small angle X-ray scattering (SAXS) measurements were performed using a NanoStar small angle X-ray scattering system (Bruker AXS) with the two-dimensional detector Våntec 2000. As X-ray source, a copper rotating anode was used with a Cu–Kα radiation of λ = 0.1542 nm. XRD measurements were performed on a PANalytical MPD PRO instrument, using Cu-Kα radiation.

Scanning electron microscopy (SEM)

The sample morphology was examined using a ZEISS/ULTRA PLUS scanning electron microscope operating at an accelerating voltage of 5 or 10 kV and using an in-lens detector. The samples were not treated for conductivity.

Energy-dispersive X-ray spectroscopy (EDX)

Elemental composition was acquired using an Oxford INCA Synergy 450X-Max 50 energy-dispersive X-ray microanalysis system attached to the SEM with an area analytical silicon drift detector at an acceleration voltage of 10 keV.

Transmission electron microscopy (TEM)

The images were recorded on a Philips EM 400 transmission electron microscope (TEM). For the TEM images, the materials were first embedded in epoxy resin. The resin part was cut into slices using a diatome diamond knife and individual slices were placed onto carbon-coated Cu grids.

Differential thermal analysis

The analysis was performed in argon atmosphere with a Netzsch STA 449 C Jupiter with a heating rate of 2 K min−1 or 10 K min−1. Silica was ground in an agate mortar, mixed with magnesium powder in a molar ratio of Mg/SiO2 with 2.0/1 and filled in a corundum crucible.

Notes

Acknowledgments

This work has been financially supported by the Bundesministerium für Bildung und Forschung (BMBF) (project “KoLiWIn”/03SF0343C). Dr. J. Holzbock and M. Rapp are gratefully acknowledged for helpful discussions. We thank S. Blessing for XRD analysis (Ulm University) and Mubera Suljic for the nitrogen sorption measurements. We are also grateful to the Central Facility for Electron Microscopy (Ulm University) for TEM measurements.

References

  1. 1.
    Weitkamp J, Sing KSW, Schüth F (2002) Handbook of porous solids. Wiley-VCH, WeinheimGoogle Scholar
  2. 2.
    Triantafillidis C, Elsässer MS, Hüsing N (2013) Chem Soc Rev 42:3833CrossRefGoogle Scholar
  3. 3.
    Bang BM, Lee J-I, Kim H, Cho J, Park S (2012) Adv Energy Mater 2:878CrossRefGoogle Scholar
  4. 4.
    Lin VS-Y, Motesharei K, Dancil K-PS, Sailor MJ, Ghadiri MR (1997) Science 278:840Google Scholar
  5. 5.
    Canham L (2014) Porous Silicon Formation by Porous Silica Reduction. In: Canham L (ed), Handbook of Porous Silicon. Springer International Publishing, p 1Google Scholar
  6. 6.
    Zhu J, Gladden C, Liu N, Cui Y, Zhang X (2012) Phys Chem Chem Phys 15:440CrossRefGoogle Scholar
  7. 7.
    Bao Z, Weatherspoon MR, Shian S, Cai Y, Graham PD, Allan SM, Ahmad G, Dickerson MB, Church BC, Kang Z, Abernathy HW, Summers CJ, Liu M, Sandhage KH (2007) Nature 446:172CrossRefGoogle Scholar
  8. 8.
    Zemnukhova L, Panasenko A, Fedorishcheva G, Ziatdinov A, Polyakova N, Kuryavyi V (2012) Inorg Mater 48:971CrossRefGoogle Scholar
  9. 9.
    Liu N, Huo K, McDowell MT, Zhao J, Cui Y (2013) Sci Rep 3:1919Google Scholar
  10. 10.
    Batchelor L, Loni A, Canham LT, Hasan M, Coffer JL (2012) Silicon 4:259CrossRefGoogle Scholar
  11. 11.
    Dasog M, Yang Z, Veinot JGC (2012) CrystEngComm 14:7576CrossRefGoogle Scholar
  12. 12.
    Chen K, Bao ZH, Shen J, Wu GM, Zhou B, Sandhage KH (2012) J Mater Chem 22:16196CrossRefGoogle Scholar
  13. 13.
    Zhang Y, Huang J (2011) J Mater Chem 21:7161CrossRefGoogle Scholar
  14. 14.
    Dickerson M, Naik R, Sarosi P, Agarwal G, Stone M, Sandhage K (2005) J Nanosci Nanotechnol 5:63CrossRefGoogle Scholar
  15. 15.
    Richman EK, Kang CB, Brezesinski T, Tolbert SH (2008) Nano Lett 8:3075CrossRefGoogle Scholar
  16. 16.
    Fang J, Kang CB, Huang Y, Tolbert SH, Pilon L (2012) J Phys Chem C 116:12926CrossRefGoogle Scholar
  17. 17.
    Ibisate M, Golmayo D, López C (2009) Adv Mater 21:2899CrossRefGoogle Scholar
  18. 18.
    Pallavidino L, Liscidini M, Virga A, Chiodoni A, Descrovi E, Cos J, Andreani LC, Pirri CF, Geobaldo F, Giorgis F (2011) Opt Mater 33:563CrossRefGoogle Scholar
  19. 19.
    Chen D, Mei X, Ji G, Lu M, Xie J, Lu J, Lee JY (2012) Angew Chem Int Ed 51:2409CrossRefGoogle Scholar
  20. 20.
    Liu MP, Li CH, Du HB, You XZ (2012) Chem Commun 48:4950CrossRefGoogle Scholar
  21. 21.
    Chen W, Fan Z, Dhanabalan A, Chen C, Wang C (2011) J Electrochem Soc 158:A1055CrossRefGoogle Scholar
  22. 22.
    Guo M, Zou X, Ren H, Muhammad F, Huang C, Qiu S, Zhu G (2011) Microporous Mesoporous Mater 142:194CrossRefGoogle Scholar
  23. 23.
    Jia H, Gao P, Yang J, Wang J, Nuli Y, Yang Z (2011) Adv Energy Mater 1:1036CrossRefGoogle Scholar
  24. 24.
    Tao HC, Huang M, Fan LZ, Qu X (2012) Solid State Ionics 220:1CrossRefGoogle Scholar
  25. 25.
    Larbi KK, Barati M, McLean A (2011) Can Metall Q 50:341CrossRefGoogle Scholar
  26. 26.
    Banerjee HD, Sen S, Acharya HN (1982) Mater Sci Eng 52:173CrossRefGoogle Scholar
  27. 27.
    Barati M, Sarder S, McLean A, Roy R (2011) J Non-Cryst Solids 357:18CrossRefGoogle Scholar
  28. 28.
    Swatsitang E, Krochai M (2009) J Met Mater Miner 19:91Google Scholar
  29. 29.
    Bao Z, Ernst EM, Yoo S, Sandhage KH (2009) Adv Mater 21:474CrossRefGoogle Scholar
  30. 30.
    Hai N, Grigoriants I, Gedanken A (2009) J Phys Chem C 113:10521CrossRefGoogle Scholar
  31. 31.
    Shepherd RF, Panda P, Bao Z, Sandhage KH, Hatton TA, Lewis JA, Doyle PS (2008) Adv Mater 20:4734CrossRefGoogle Scholar
  32. 32.
    Winnacker K, Küchler L, Dittmeyer R (2004) Chemische Technik: Methodische Grundlagen, vol 1, 5th edn. Wiley-VCH, WeinheimGoogle Scholar
  33. 33.
    Wong DP, Lien H-T, Chen Y-T, Chen K-H, Chen L-C (2012) Green Chem 14:896CrossRefGoogle Scholar
  34. 34.
    Won CW, Nersisyan HH, Shin CY, Lee JH (2009) Microporous Mesoporous Mater 126:166CrossRefGoogle Scholar
  35. 35.
    Brandhuber D, Torma V, Raab C, Peterlik H, Kulak A, Hüsing N (2005) Chem Mater 17:4262CrossRefGoogle Scholar
  36. 36.
    Hartmann S, Brandhuber D, Hüsing N (2007) Acc Chem Res 40:885CrossRefGoogle Scholar
  37. 37.
    Barrett EP, Joyner LG, Halenda PP (1951) J Am Chem Soc 73:373CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Wien 2015

Authors and Affiliations

  • Michael Waitzinger
    • 1
  • Michael S. Elsaesser
    • 1
  • Raphael J. F. Berger
    • 1
  • Johanna Akbarzadeh
    • 2
  • Herwig Peterlik
    • 2
  • Nicola Hüsing
    • 1
  1. 1.Materials ChemistryParis Lodron University SalzburgSalzburgAustria
  2. 2.Faculty of PhysicsUniversity of ViennaViennaAustria

Personalised recommendations