Getting boron concentrations in silicon right

With a major breakthrough in boron quantification in silicon, life for silicon analysts has just gotten a lot easier.

Researchers from Elkem and the University of Aberdeen have released an article revealing a major breakthrough in boron quantification in silicon. Here you see the two Elkem authors Patrick Galler (left) and Kjell Blandhol (right). Kjell Blandhol is the head of Elkem Technology lab in Kristiansand and had the initial idea of boron speciation, which Patrick Galler spun into an article.

Researchers from Elkem and the University of Aberdeen have released an article titled Boron speciation in acid digests of metallurgical grade silicon reveals problem for accurate boron quantification by inductively coupled plasma – optical emission spectroscopy.

In brief, the article describes why it is difficult and how it is possible to measure boron (B) in silicon (Si) accurately using inductively coupled plasma – optical emission spectroscopy (ICP-OES). Based on the published findings the National Institute of Standards and Technology will soon release an updated certificate for the NIST 57b Si reference material, making life for Si analysts and the industry easier.

Elkem’s research department has worked on this for the past three years. Halfway through the project, speciation analysis researchers from the University of Aberdeen got involved in the research. R&D Engineer in Elkem Patrick Galler was the lead author of the study. We spoke to him about boron speciation and the importance and implications of the results from the research effort.

What is speciation?

Speciation analysis is a collective expression for methods studying chemical differences for any element occurring in different “chemical forms” or “species” in a given sample material. One example is Chromium (Cr), which can occur as Cr3+ and Cr6+ in one and the same sample. Cr3+ is an essential nutrient, Cr6+ is carcinogenic. Hence speciation analysis is required to quantify both. The total Cr content will not say equally much about the health benefits or risks related to, for example, food materials analysed.

What is ICP-OES?

ICP-OES stands for inductively coupled plasma – optical emission spectroscopy. It is an analytical method where a sample aerosol is injected into a localized Ar plasma roughly the size of a hazelnut. The sample is excited in the plasma; i.e. ICP; and emits light. Excited sample atoms will emit light of clearly defined wavelengths. This light emission is detected and its intensity used for quantifying the concentration of different analytes. In Elkem, this method is extensively used for quantification of B, P and different metals in Si, FeSi, slag, ash and quartz. In total Elkem is operating six ICP-OES at different locations for quality and process control purposes. Roughly, one can say, that the lower the expected concentration in the sample, the higher the tendency to use a powerful method like this.

Why is this relevant for boron in silicon?

In order to make a sample accessible to analysis by ICP-OES it has to be dissolved, so it can be injected into the plasma. Sample dissolution is what analysts refer to as “digestion”. Si is digested with nitric and hydrofluoric acid. When the dissolved sample is injected into the ICP it has to pass through a tube and spray chamber of some sort, usually made of PTFE and called “sample introduction system”. We discovered that part of the B in any dissolved metallurgical grade Si sample sticks to the walls of the sample introduction system. First we thought that this was just what B usually does. Later and somewhat accidentally, we found that things were a bit more complex. In collaboration with the University of Aberdeen, we found that more than 50% of all B in digested samples occurs as some unknown B compound, very likely in conjunction with Si. And this was very confusing because this is neither intuitive nor has it been described in literature, ever. We could however show that it was exclusively this unknown B compound sticking to the walls of the sample introduction system. As a result, we reckon that globally most of the B concentration data for Si, obtained by ICP techniques, is to some extent wrong. Wrong, because the B compound sticking to the walls of the sample introduction system will not be fully accessible to analysis.

How was the problem solved?

Again, in collaboration with the University of Aberdeen we could show that the unknown B compound can be oxidised to boric acid in solution by adding hydrogen peroxide to sample digests. Boric acid in solution is a piece of cake to analyse by ICP-OES! These findings too were counter intuitive, since we tried oxidation media such as aqua regia and permanganate before without success. That hydrogen peroxide should solve our problems so effectively was unexpected. In this sense, it was very important to have access to more specialised techniques at the University of Aberdeen. Firstly, because they are top-notch scientists. Secondly, because seeing is believing!

What are the implications?

For Elkem the new method means better accuracy as well as precision with respect to B quantification in Si. This has of course ramifications for process control. On a global scale one has to mention that there is not a single certified reference material for B in Si currently on sale. Having used a reference material from the National Institute of Standards and Technology in our work we had recently confirmed, that NIST is now upgrading the B concentration for this reference material to a certified B concentration. This is huge! Apparently a lot of people have spent a lot of time and money on B analysis in Si before us without comparable success. Now Elkem at this time is per definition the key to accurate B quantification in Si. In the future we expect also a better agreement between our and customer analyses.

What surprises does the future hold?

Very uncertain! Maybe we will soon be proven wrong? Or maybe not. Nevertheless, we have a suspicion that this unknown B compound is a B-Si molecule, maybe even a blend of different B-Si compounds. We also suspect that it is formed at some point during the Si process. We could not definitely proof that, but it would be really nice. Ideally one should try extracting the unknown B compound in larger quantities from sample digests, make it accessible to methods for molecular characterization and learn more about its chemistry. More effort is also required to study B in Si in-situ, meaning directly in the sample without compromising the samples chemical and structural integrity.

A question that logically follows from our suspicion is how the unknown B compound affects electronic material properties of Elkem Solar Silicon, if it does at all. There is much talk about high temperature efficiency. Maybe B plays a role? Who knows!?



Last updated: