Elimination of Boron from Soluble Silica Via Solvent Extraction with 2,2,4-Trimethyl-1,3-Pentanediol Using a Multistage Flow-Type Reactor

The sustainable production of solar-grade silicon (SOG-Si) with inexpensive refining process is highly demanded for Si photovoltaic power generation. Diatomaceous earth, an Earth-abundant silica resource, with aqueous chemical refining is a candidate for high-purity source, which could be directly reduced to SOG-Si.¹ Since boron and phosphorus act as dopant in Si semiconductor, efficient way of elimination of these light elements is strongly required. We have proposed the application of microchannel device for the solvent extraction of boron with 2-ethyl-1,3-hexanediol (EHD).^2,3 While the short diffusion distance as well as the large specific interfacial area of microchannel enabled fast and efficient elimination of boron to SOG-purity, however, improvement of throughput was required for the practical application. In order to take the advantage of microchannel device together with large productivity, further improvement of the elimination efficiency and increase of the area of liquid-liquid interface for the reaction with maintaining short diffusion distance comparable to the microchannel, is necessary. Here, we report the optimization of the extraction reaction and the channel design for increasing the throughput on the boron elimination from soluble silica via solvent extraction using the flow-type reactor.

    A flow-type reactor with 0.10 mm depth and 1.0 mm width was fabricated on a silicon and glass substrate using photolithography techniques. The number of stages was one or two, and each channel length was 10 mm or 25 mm. As model solution sample, the refined silica spiked with a trace amount of boric acid was dissolved in 2.5M NaOH aqueous solution prior to adjusting pH to 1.0 with 4.0M HCl aqueous solution. In order to improve the elimination efficiency, 2,2,4-trimethyl-1,3-pentanediol (TMPD) was used instead of EHD and added to toluene to be 0.5 M solution. The model solution sample was injected to the flow-type reactors with the flow rate of 1.5 mL·min^-1 and the flow rate of the extractant was adjusted to form parallel flow and be separated at the outlet. The residual boron content was determined with inductively coupled plasma mass spectroscopy (ICP-MS). Table 1 shows representative results. The contact duration, i.e., the reaction period was estimated to be 20 ms for 10 mm length and 50 ms for 25 mm length. In the both cases of EHD and TMPD with 25 mm length, elimination rate was almost same and the residual boron content decreased to less than 0.1 ppm. On the other hand, the elimination rate with TMPD on other cases was higher than that with EHD. In addition, the residual boron content decreased to almost 0.1 ppm in the case of TMPD with 40 ms contact duration. These results indicate that the contact duration on the extraction with TMPD could be optimized between 40 ms and 100 ms. Thus, increase of the flow rate with the optimized total duration and area of liquid–liquid interface can improve the throughput. The application of optimized channel design and extraction condition with TMPD based on these results will be discussed.

This study was financially supported in part by CREST program by Japan Science and Technology Agency (JST).

1. M. Bessho, Y. Fukunaka, H. Kusuda, T. Nishiyama, Energy & Fuels, 23, 4160 (2009).
2. N. Matsuo, Y. Matsui, Y. Fukunaka, T. Homma, ECS Trans. 50, 103 (2013).
3. N. Matsuo, Y. Matsui, T. Ishihara, Y. Fukunaka, T. Homma, 224th ECS Meeting, Abstract #2261 (2013).