Microanalytical Analysis of Nano-Particulate Pressed Pellets made from Manganese Nodules

Introduction

Manganese nodules have been of both scientific and industrial interest due to their high content of critical metals (i.e. Co, Ni, Zn, Cr, Cu, Pb). The dawn of new microanalytical techniques such as µXRF, LIBS and LA-ICP-MS has opened the possibilities of analysing these samples quantitatively and also visualising the elemental distribution in situ without tedious sample preparation. However, the development of matrix-matched reference materials is still trailing behind. There are many powdered geological materials, whose particle size is too coarse (20-150 µm). This causes the sampling of individual grains with the analytical beam, typically 10-100 µm, instead of stoichiometric sampling of the entire sample within the beam diameter. By milling the available powders to a particle size of D90 400 nm, D50 200 nm, we are able to reduce the particle size sufficiently in order to enable the production of binder-free pressed pellets, which are suitable as reference materials for microanalytical techniques.

For this application note NOD-A-1 and NOD-P-1, purchased from the Unites States Geological Survey (USGS), were used as a starting material and prototype-pellets for µXRF with a diameter of 13 mm were produced. Additionally, one pellet from the original powder was pressed as well. NOD-A-1 was collected in the Atlantic Ocean, at the Blake Plateau (31° 02’ N, 78° 22’ W), at a water depth of 788 m. NOD-P-1 was collected from the Pacific Ocean (14° 50’ N, 124° 28’ W) at a water depth of 4300 m. An image (Google Earth) showing their respective geographical locations is shown in Fig.1.

Fig.1: Google Earth image showing the respective geographic locations of NOD-A-1 and NOD-P-1. NOD-A-1 was retrieved east of the state of Florida, USA. NOD-P-1 was retrieved west of Baja California, USA.

Analyses – Spatially resolved X-ray-fluorescence

The µXRF-analyses were performed with the M4 Tornado at Bruker Nano GmbH, Berlin, Germany. The acquisition parameters are shown in Table 1. A set of 10 nano-pellets (suffix …-NP) were inserted into a custom-made CNC-milled aluminium mount and the entire section of the mount was mapped.

Tab.1 Acquisition parameters during µXRF-mapping.

Resulting mappings, showing the colour-coded elemental distribution, highlight the improved homogeneity of nano-pellets compared to pellets pressed from the original powder (Fig.2).

Fig.2 Exemplary colour-coded elemental distribution maps (NOD-A1). The top-left pellet represents the original powder, whereas the remaining three are nano-pellets. Potassium was not lost during processing, the mineral grains containing the element were merely reduced in their particle-size and homogenously redistributed throughout the sample. The pellets’ diameter is approx. 9 mm.

This visualisation of the contrasting homogeneities can also be expressed quantitatively. Therefore, 50 analysis-spots were randomly distributed across the pellet surface of an original pellet and a nano-pellet. The statistical evaluation of this comparison can be found in Table 2. It shows the excellent within-pellet homogeneity of the nano-pellets, for the selected elements, compared to the large variations within the original powder pellet.

Tab. 2: Comparison of relative standard deviation [RSD-%] of net signal intensities acquired on 50 random spots on each respective pellet surface. One pellet made from the original powder and one nano-pellet.

The RSD-% for the original powder ranges from 5.77 to 37.1 %, whereas the RSD-% in the nano-pellet ranges from 0.28 to 4.28 %. Within-unit, or in this case, within-pellet homogeneity is an important aspect of any reference material. The same is true for between-pellet homogeneity. In order to demonstrate it, the M4 Tornado’s software is able to extract averaged data from so called objects (Fig.3), which can be created in a variety of shapes. Here, round objects were created, averaging the signal for each of the 10 (1 = orginal, 2-10 = nano) pellets.

Fig. 3: Objects created in the M4 Tornado’s software resulting in an average signal from within the objects. In this case silicon is shown as an arbitrary example of object creation. Pellet number 1 is made from original powder, and therefore shows significantly larger silicon particles.

By calculating the RSD-% between pellets 2-10 it can be demonstrated, that the variation in average signal intensity between the pellets does, with one exception (Cl), not exceed 2 % (Tab.3). The Cl signal is overlapped by the Rh-L (anode material) scattering signal, so the apparent instability for this element is not due to compositional variation in the sample.

Tab. 3.: Relative standard deviations [RSD-%] of average net signal intensities within designated objects between 9 Pellets of NOD-A1-NP, measured in a random order. Data is not drift corrected.

An example for the visualisation of the data from Tab.3 can be seen in Fig. 4. It exhibits color-coded elemental distribution maps for a selection of elements in 10 pellets NOD-P1-NP.

Fig. 4.: Elemental distribution maps of 10 nano-pellets visualising both within and between-pellet homogeneity of NOD-P1.

Conclusions & Outlook

The particle size reduction greatly improves both within and between pellet homogeneity, compared to its coarser precursor, which makes them suitable as a reference material for µXRF. It also facilitates the rapid spatially resolved in situ analysis and quantification of manganese nodules using non-destructive methods, which could be beneficial, since the logistics of retrieving them are quite elaborate and expensive.

The next step has already begun. The nano-powder is on its way for chemical characterisation.

Acknowledgment & Disclaimer

We kindly thank Dr. Roald Tagle and Falk Reinhardt from the Bruker Nano GmbH for their time and cooperation.

It is not our intention to discredit NOD-A-1 and NOD-P-1 as reference materials. They are widely used and their original intent was not in situ microanalysis. This application note intends to demonstrate the benefit of particle size reduction, thus enabling application to microanalysis.