**1. Introduction**

Alkaline granite and syenite magmatic rocks are characterized by a remarkable enrichment of high field-strength elements (HFSEs), including Zr, Nb, Y, U as well as the rare earth elements (REEs). The unusual trace element composition of alkaline rocks is in line with elevated abundances of complex and rare HFSE minerals, which have attracted both petrologists and economic geologists for a long time [1–3]. Evolving technical capabilities (e.g., high performance magnets, super-alloys, phosphors, superconductors, ceramics) and a growing high-tech based economy have enhanced the global demand for HFSEs, such as Zr, Nb, Y, or, particularly, Nd, Eu, Tb and Dy among the REE. Notwithstanding, a substantial part of these commodities (i.e., REE, Y, Nb) has a critical character indicative of supply risk for political or environmental reasons [4]. As a response to this situation, alkaline granite and syenite complexes have also come into the focus of exploration campaigns and mining feasibility studies in the last decade (e.g., Khalzan Buregtei, Strange Lake, Motzfeld, Norra Kärr [1,5–7]).

Enrichment of HFSE to enhanced grades in alkaline intrusions is generally thought to be caused by magmatic fractionation, but mineralogical and petrologic evidence often points to post-magmatic pneumatolytic or hydrothermal processes. Post-magmatic mineralization events are commonly associated with intensive and highly variable alteration types of both intrusive and wall rocks, thereby superimposing primary petrographic features [1,2,6,8–10].

As a consequence, alkaline granites and syenites show a broad range of mineralogical and textural characteristics. Most apparent, these rocks display outstanding associations of rare minerals—among these are also ore phases incorporating the HFSEs as major elements (e.g., elpidite, eudialyte, gittinsite, zircon; Nb pyrochlore; REE + Y monazite, bastnaesite, synchysite, and several REE-silicates). Furthermore, formation in multistage and partially superimposing magmatic to hydrothermal processes frequently leads to narrow and complex intergrowths of fine grained and irregularly sized and shaped crystals. Thus, such rocks display a remarkable spatial variability and heterogeneity in terms of mineral composition, grain size, mineral association, and texture [1,9,10].

In general, these mineralogical and textural characteristics of alkaline rocks represent obstacles for e fficient exploration and ore beneficiation, as will be outlined in this study. Owing to the marked mineralogical complexity, alkaline rock classification comprises a confusing multiplicity of partly unfamiliar names, many of these derived with reference to single locations only [11] (p. 558). The multitude of classifying terms is a source of incongruence in field data or core logs in exploration campaigns if work is carried out by di fferent geologists or companies. Furthermore, classification of alkaline rocks implies magmatic genesis and does not reflect intensity of post-magmatic hydrothermal mineral formation. This leads to terminological indistinctness with potential negative e ffects for data handling, e.g., in 3D numeric deposit modelling.

As a consequence of small grain sizes and intricate intergrowth, liberation sizes of ore minerals can be low. Mining operators have to cope with intensive and energy consuming grinding as well as processing of fine fractions to produce marketable concentrates with acceptable recoveries. In addition to causing high energy costs, the processing of fine fractions of polymineralic ores, comprising phases with di fferent physical properties (e.g., density, magnetic susceptibility), is likely to be ine fficient. Such handicaps for e fficient, economic, and sustained mining represent bars to the development of prospects associated with alkaline granites and syenites.

In this study the alkaline granite hosted Zr-REE-Nb deposit of Khalzan Buregtei (Mongolia), which was a target of exploration work by di fferent companies in recent years, is used for a case study to demonstrate characteristic problems associated with these types of deposits. Based on results of published studies of the magmatic and post-magmatic (i.e., hydrothermal) evolution of the complex [5,12–15], detailed sampling work was carried out in the central part of the complex, which is most a ffected by Zr-REE-Nb mineralization. In contrast to a number of detailed petrological studies addressing genesis and potential processes of either magmatic or metasomatic enrichment of rare metals in the alkaline granite of Khalzan Buregtei [5,13–15], this study particularly focusses on the identification of ore properties, which are crucial for the economic exploitation of the mineralized rocks of this deposit.

Whole rock chemical major and trace element composition, chemical composition of HFSE minerals dominating the ore assemblage, and comprehensive textural datasets were determined in characteristic rock specimens. In particular, this study aims to investigate the influence of post-magmatic alteration on classification and to develop solutions to handle terminological indistinctness in such highly altered alkaline rocks. A second target of the study is to characterize mineralogical and textural e ffects of ore mineralization, to indicate handicaps for mineral processing, but also to discuss the potential of mineralogical and textural characteristics of Zr-REE-Nb mineralization in alkaline rocks to yield enhanced recoveries.

## **2. Geological Setting**

The Khalzan Buregtei massif is situated in west Mongolia Altai mountains approximately 50 km north of the city of Hovd. It was discovered in 1984 by geologists of the former Soviet Union [13]. The complex covers an area of 30 × 8 km and is mainly built up by alkaline syenite to granitic rocks considered to be associated with early Devonian extensional tectonic activity [5]. The Khalzan Buregtei Zr-REE-Nb deposit itself is located in the centre of a roughly oval shaped intrusive body in the south of the massif (Figure 1). The part of the intrusion hosting the deposit is dominated by alkali granites, which are distinct to the surrounding syenitic rocks, classified as nordmarkites by the first investigators [13,14] and in the recent work of [15]. The latter authors assign the mineralized alkali granites to a fifth and seventh intrusive stage, whereas [5] classify the mineralized rocks as ore metasomatites, attesting to the strong imprint of post-magmatic hydrothermal processes. Syenites contacting the central granites were a ffected by hydrothermal activity as well. Major faulting of the working area postdates the ore formation indicated by the displacement of the rock suite as illustrated in Figure 1. According to [5] the major rare metal deposit of Khalzan Buregtei has an ore content of 2.4 × 10<sup>6</sup> t ZrO2, 3.5 × 10<sup>5</sup> t Nb2O5, and 4.9 × 10<sup>5</sup> t REE2O3 + 1.3 × 10<sup>5</sup> t Y2O3. The mechanism of rare metal enrichment to final grades has been discussed controversially in recent years by several authors [5,14,15]. The work of [5] explains the ore formation by multistage post-magmatic alteration on the basis of a previous enrichment of alkaline magma with HFS and other rare metals. The post-magmatic alteration is triggered by fluids from further unspecified carbonatite plutons, which explains the abundance of Ca bearing Zr-silicates and carbonates. However, the recent work of [15] discussed the formation of the rare metal granite (V-Phase [14,15]) and the aforementioned ore minerals by di fferentiated crystallization of alkaline magma strongly enriched in HFSE and other rare metals. The formation of Ca bearing phases is assumed to be caused by enrichment of Ca following assimilation of adjacent limestones in the melt [15].

#### **3. Materials and Methods**

#### *3.1. Sampling Method and Bulk Rock Chemical Analysis*

The ore bearing rocks of the Khalzan Buregtei deposit were investigated in two particularly well exposed areas characteristic of the central domain (Figure 1A). The deposit covers an area of approximately 1 km<sup>2</sup> of mountainous arid landscape. The central part of the deposit is built of one central ridge with an elevation up to 1940 m bordering, to the west and east, two valleys filled with sediments and debris.

A sample suite taken from the central ridge was indicated by the addition of KB G to the sample numbers. Another sample suite was taken from one more shallow ridge beyond the western, valley having good accessibility and outcrop condition. Latter samples were named KB S. Samples were taken along outcrop sections, where a macroscopic change in mineralogy and/or texture was recognizable. Hand specimens were later investigated for homogeneity prior to further treatment. In total, 18 samples, each approximately 1 kg in weight, were taken in both outcrop areas. After splitting into two sub-samples one set of sub-samples was crushed by manual crushing and further comminuted using a rotating disc mill (Siebtechnik GmbH, Mühlheim an der Ruhr, Germany). The sample powder was analysed for bulk rock chemical composition at ALS Analytical Service, a ffiliation Loughrea, Ireland, applying a combined ICP-AES and ICP-MS method after Li-borate fusion (ALS Method Codes: ME-MS81d). Representative blocks were dissected from the other set of the sub-samples for the preparation of polished thin-sections for optical petrological microscopy, electron optical investigation by QEMSCAN © (FEI/Thermo Fisher, Hillsboro, OR, USA), and electron probe micro analyser (EPMA).

**Figure 1.** (**A**) Satellite image with the outline (magenta) of the entire alkaline massif of Khalzan Buregtei as proposed by [13,15]. The red box marks the area of the Khalzan Buregtei deposit shown as sketch map in B [16]. (**B**) Geological sketch map of the Khalzan Buregtei Zr-REE-Nb deposit with sampling areas KB S and KB G, see text for further details. Map basis according to [12,15] with nomenclature following [5].

Zirconium concentration data were obtained by X-ray fluorescence (XRF) measurement at the Unit of Mineral Processing, RWTH-Aachen University, Niton XL3t (Thermo Fisher, Waltham, MA, USA) with the handheld XRF system. The measurements were performed with three replicates on pressed powder pellets created with 8 g of air-dried sample powder. Data was validated by an external quality check using NIM-L lujavrite standard reference material [17] prepared and handled like the unknowns.

#### *3.2. SEM Based Semi-Automated Mineralogy (QEMSCAN*©*)*

Semi-automated mineralogical analyses were performed by applying a Quanta 650-F QEMSCAN © (FEI/Thermo Fischer) scanning electron microscope (SEM) at the Institute of Applied Mineralogy and Economic Geology, RWTH-Aachen University [18]. Polished thin sections and polished sections were analysed after carbon coating. The measurements were conducted with an acceleration voltage of 25 kV and a fixed sample current of 10 nA. The surface of each sample section was scanned with a spatial resolution of 5 μm. Back scatter (BSE) intensities and individual X-ray spectra were recorded for each pixel with a 4-quadrant BSE detector and two DualXFlash 5030 SDD (Bruker AXS, Karlsruhe, Germany) energy dispersive x-ray spectrometers (EDX). Phase assignment was carried out by comparison of spectral data obtained for each pixel with library information using the iDiscover (Version 5.3.2.501, FEI/Thermo Fisher, Hillsboro, OR, USA) software suite. Automated image analysis was applied to phase maps to compute quantitative mineralogical and textural parameters. The modal composition of each sample in volume (vol.) % was calculated from the volumetric abundance of each mineral phase. Particle size calculations were conducted by measurement of the diameter of the virtual sphere with equivalent perimeter length assigned to each individual target phase [19]. Particle populations were obtained by segmentation of the target minerals or mineral assemblage from QEMSCAN © phase maps.

#### *3.3. Electron Probe Micro Analyzer (EPMA)*

The JXA-8900R electron probe micro analyser (Jeol, Jeol Germany GmbH, Freising, Germany), of the Institute of Applied Mineralogy and Economic Geology, RWTH-Aachen University, was used for high-resolution element mapping and for quantitative chemical analyses of ore-forming minerals. For element mapping, the instrument was operated with a focused electron beam and an acceleration voltage of 20 kV. Each pixel of 1 μm size was measured with a dwell time of 50 ms. The setup for the five wavelength dispersive spectrometers of the EPMA is given in the Appendix A (Table A1).

Quantitative analyses of the mineral chemistry were conducted on di fferent ore minerals. REE-carbonate minerals were analysed with an acceleration voltage of 15 kV and an electron beam of 10 μm diameter to avoid intensive damaging of the target minerals. Zircon, in contrast, was analysed with an increased acceleration voltage of 20 kV and a focused electron beam. The beam current was set in all analytical sessions to 24 nA. Detailed spectrometer setup parameters, like peak and background recording times, as well as standards, are given in the Appendix A (Tables A2 and A3).
