Threatening Silicon’s Diamond Throne

Silicon Dreams

Imagine inventing a new allotrope of silicon that still functioned equally well to the diamond form of silicon as a semiconductor making it suitable for microelectronics but also provided those much desired properties the diamond form of silicon has so far been unable to deliver to enable the development of improved performance solar cells, higher efficiency light emitting diodes and lithium ion batteries with better cycle-life and safety requirements.

Envisage in the case of silicon solar cells a new allotrope of silicon able to offer a direct bandgap of around 1.3eV. This would be the ideal value that optimizes the photon-to-electron conversion efficiency for a single p-n junction to meet the theoretical Shockley-Queisser limit of 33.7%. This new allotrope of silicon with its direct bandgap would also enable the required high efficiency electron-hole radiative recombination required to drive the next generation of light emitting diodes. Conceive also of anodes for lithium ion batteries made of this new allotrope of silicon that undergo minimal volume changes and mechanical failure on electrochemical insertion and de-insertion of lithium. This property would help solve the charging-discharging cycle-life and safety problems that prevent the champion volumetric capacity of silicon from being realized in practical lithium ion batteries.

Threat to Silicon’s Diamond Throne

Materials researchers have long dreamed about this new allotrope of silicon but its discovery seemed just too much to expect. However, in my forty five years of practicing materials chemistry I have discovered that you never-say-never about so-called “non-existent” materials. Nowhere is this existential paradigm better seen than with a very recent report of a new allotrope of silicon, which is formulated as Si₂₄ and that sure enough appears to provide all of the aforementioned desirable properties demanded of silicon in one structure. This breakthrough in materials synthesis points the way to a new era of advanced silicon devices that were never thought possible before [1].

To amplify, the synthesis of Si₂₄ begins with the preparation of the known precursor Na₄Si₂₄ by a high temperature high pressure reaction of elemental sodium and silicon in an anvil cell [1]. The structure of Na₄Si₂₄ is based upon an open-framework built of sp³-bonded Si atoms forming 8-ring one-dimensional channels stuffed with linear chains of Na atoms. Of great significance is the observation that these intra-channel Na atoms can be completely removed by thermally induced diffusion and evaporation from the mouths of the channels at the rather low temperature of 400K and under dynamic vacuum conditions, as illustrated in Figure 1. This treatment causes a gradual reduction of the Na content in the channels to reach essentially zero after eight days as confirmed by XPS, PXRD and TEM-EDX analysis.

Significantly, the associated changes in the orthorhombic unit cell lattice parameters determined from Rietveld powder X-ray diffraction analysis of the transformation of Na₄Si₂₄ to Si₂₄ are only -6.8%, +1.2% and 3.0%, respectively. This corresponds to an increase in the diameter of the channels concomitant with a reduction in their length, Figure 2. This minimal volume swing of the unit cell dimensions on passing from stuffed Na₄Si₂₄ to empty Si₂₄ is unique in the context of other known allotropes of silicon and speaks well for the charging-discharging cycling stability of Si₂₄ if used as the anode in a lithium ion battery. Noteworthy in this context is that theory and experiment indicate Si₂₄ remains stable up to 750K and 10GPa [1].

Density functional theory and optical reflectivity measurements and analysis of the absorption edge concur that Si₂₄ has a quasi-direct electronic bandgap. This designation arises because in the G-Z direction of the Brillouin zone the energy of the highest valence and lowest conduction bands are essentially the same and very flat. This situation indicates that Si₂₄ is best described as a quasi-direct bandgap material with very similar indirect and direct bandgap energies of 1.29eV and 1.39eV, respectively. Of related importance is the measured temperature dependence of the electrical conductivity, which demonstrates that Si₂₄ behaves as a classical semiconductor with conductivity increasing with T according to the energy gap law, s = s_oexp(-E_g/2kT). By contrast, Na₄Si₂₄ displays metallicconductivity that as expected for a metal decreases with T. This behavior most likely originates from charge transfer of valence electrons associated with the intra-channel Na atoms to the conduction band of the Si₂₄ open-framework. Importantly, the optical absorption spectrum of Si₂₄ shows that its ability to absorb light compared to the diamond allotrope of silicon is significantly larger in the visible wavelength range where the intensity of the AM1.5 solar spectrum is at its maximum, Figure 3.

Major Challenge

Together, the extraordinary properties of Si₂₄ outlined above, which can be enhanced and enriched by n-doping and p-doping and formation of native oxide, bode well for its future utilization in high capacity lithium ion batteries, enhanced efficiency photovoltaics and next generation light emitting diodes. Realistically, however, there seems to be much stronger evidence for the promising potential in solar cells and batteries than perhaps light emitting diodes. This is based on not finding evidence of quantified and efficient luminescence – yet. As is well appreciated, a promising theoretical band structure does not always translate to an efficient luminescent structure if non-radiative defects are difficult to minimize. Within such a low dimensional and presumed high surface area structure this could be challenging but the nanocrystal passivating ligand chemistries that have been developed could help. Various silicon clathrates have been around quite a while but have not yielded efficient photoluminscence, let alone electroluminescence, to the best of my knowledge, compared to nanocrystalline silicon. For optoelectronic devices there is concern about effective doping routes and trace residual sodium within any passivating oxide structure. Finally, a major challenge that confronts the implementation of Si₂₄ in next generation advanced silicon-based devices is the urgency of discovering ways of scaling the reported high P,T synthesis of the material to industrially relevant proportions or indeed finding an entirely different pathway for its large-scale preparation at a cost that is competitive with the diamond form of silicon.

In view of what is at stake in trying to dethrone the diamond form of silicon, the Samurai of semiconductors, I predict the materials community will diligently work to discover creative ways of making tons of the new Si₂₄ allotrope in the not too distant future.

1. Synthesis of an Open-Framework Allotrope of Silicon, Kim, D.Y., Stefanoski, S., Kurakevych, O.O., T.A. Strobel, Nature Materials, 2015, 14, 169-173, DOI: 10.1038/NMAT4140.

Acknowledgments

 GAO is Government of Canada Research Chair in Materials Chemistry and Nanochemistry. This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC). The ideas and opinions expressed in this editorial have benefitted enormously from the critical reading, insightful feedback and writing skills of Leigh Canham, Markus Antonietti and Bob Davies. The aesthetics of this editorial have been greatly enhanced and enriched by the creative artistic skills of Chenxi Qian.