Insights in the Physicochemical and Mechanical Properties and Characterization Methodology of Perlites

Abstract

Perlite is a volcanic glass that, under thermal treatment, expands, producing a highly porous and lightweight granular material which finds application in the construction, horticulture, insulation and other industrial sectors. Proper control of the feed properties and the expansion conditions allows the production of purpose-oriented grades, while the primary evaluation of its appropriateness for use in each sector is performed by the proper characterization of relevant physical, thermal or/and mechanical properties. However, due to its extreme fineness, low density, and friability, most of the available characterization methods either fail in testing or provide erroneous results, while for certain properties of interest, a method is still missing. As a consequence, the way towards the evaluation of the material is rife with uncertainties, while a well-defined methodology for the characterization of the critical properties is of practical importance towards the establishment of a pathway for its proper analysis and assessment. This article presents the available methodology for determining the main properties of interest, i.e., the size and density, water repellency/absorption and oil absorption, the microstructural composition, crushing and abrasion resistance and isostatic crushing strength, and also sampling and size reduction processes. The issues raised by the application of existing methods are analyzed and discussed, ending up to a proper methodology for the characterization of each property, based on the long-term experience of the Perlite Institute. The study is supplemented by updated insights on ore genesis, physicochemical properties, mineralogical composition and the expansion mechanism, as background information for the sufficient comprehension of the nature and properties of perlite.

1. Introduction

The petrologic term “perlite” refers to hydrated silicious volcanic glass with a characteristic structure of concentric or onion-like fractures, visible to the naked eye or under a microscope [1]. But in commercial usage, the term extends to any naturally occurring glass of igneous origin that holds water in its structure and expands when rapidly heated to yield a frothy and lightweight product. Perlite is included in industrial minerals, however, as a rock, its elemental and mineralogical composition may vary (within certain ranges), unlike that of actual minerals [2]. Expansion is a multiparametric phenomenon, and its evolution is affected by various material, as well as process parameters related to the thermal treatment of the ore. As expected, a change in one of the aforementioned parameters is capable of altering the properties of the product and potentially affecting its performance in various applications. This is because its use is based primarily on its physical properties acquired by the characteristic cellular morphology that is created during the expansion. Figure 1 depicts the key raw material properties and the thermal treatment parameters, together with the main physical characteristics of the expanded perlite and the major properties that are measured to control its quality and evaluate its suitability for different applications.

The evaluation of expanded perlite usually begins with the determination of its bulk density, which informs about the achieved expansion degree and the production yield. Furthermore, the material is subjected to analyses targeted to gain information about its physical properties, morphology, strength, and composition. A well-planned characterization campaign is necessary for deposit detection and mineral resource assessment purposes, as well as for the evaluation of the expansion process performance, quality control, and to explore the material suitability for certain applications. For most materials, the choice of characterization analyses is obvious, however, this is not the case for expanded perlite because of the fact that it uniquely combines certain extreme properties related to its size, density and mechanical strength. This is made obvious in Figure 2, which depicts the size and density ranges for the expanded perlite grades per applications.

Indeed, the desirable expanded perlite grades for most applications present granulometry below 2 mm, and in all cases do not exceed 5 mm. Also, regarding the bulk density, it is below 200 kg/m³ for the most common applications, rendering it a lightweight aggregate. The tiny size and the low density, combined with the limited resistance to mechanical stresses, as a consequence of the cellular structure and the thin glassy bubbles walls, constitute a potential source of error in many characterization analyses, and should be seriously considered before choosing the proper methodology. In this regard, the following stands for expanded perlite;

• As frothy material consisting of tiny, thin-walled glassy bubbles, expanded perlite is friable and vulnerable to fragmentation that will alter its composition if not treated accordingly.
• It presents a very low bulk density that may reach even 30 kg/m³ and in mass-based dosages it might exceed capacity of equipment.
• Expanded particles float in liquids solutions, leading inevitably to physical separation in such environments [4,5].
• Samples present non-uniform composition, since further to expanded grains of different bulk and apparent density, consist of non-expandable and unexpanded particles, and grains fragments of different size, favoring segregation [6]. Thus, further to good homogenization, a proper sampling technique should be applied to obtain representative samples.
• Each application sets a different specification, thus expanded perlite grades that are intended for different uses present radically different characteristics, mainly in terms of size, morphology, strength, density and porosity, thus the characterization strategy should be focused accordingly.

A few characteristic paradigms of typical characterization methods are used for similar granular materials but fail in their measurement of expanded perlite properties, due to one of the aforementioned reasons which are worth mentioning here. Typically, the apparent density determination for granular materials is performed by measuring the volume of water that is displaced by the sample in a flask (or pycnometer), as described in EN 1097-6 [7]. This common method is not applicable to expanded perlite because expanded particles tend to float and concentrate at the water surface, making impossible the identification of the water level, while no alternative is described for materials that, further than lightweight, are also ultrafine. As far as the determination of the particle size distribution using the classical sieving technique referred to in ASTM C136, the practice has shown that, contrary to the suggested sieving time (10 min), for expanded perlite the time should be eliminated to avoid the degradation of particles [8]. Also, the sample size should be given per volume, because with the existing mass-based specification, the volume of the suggested sample of lightweight perlite exceeds the capacity of the sieve. Another characteristic example of confusion in characterization arises in the measurement of Loss on Ignition (LOI), which is used for perlite on the determination of chemically bound water content. Depending on material type, various methods for the determination of LOI are available, suggesting sample heating at temperature range between 750 to 1000 °C for 20 min to 2 h. Since there is no clear suggestion for perlite, researchers and engineers choose arbitrarily the analysis conditions, rendering the results incomparable. Moreover, there is a risk of erroneous results due to inadequate heating that led to insufficient removal of the various water types.

Despite the undesired consequences of the inappropriate application or incorrect choice of characterization method for the evaluation of perlite, there is a considerable lack of discussion and information on that important issue in the bibliography. Currently, the choice and implementation of a characterization plan is left to the discretion of the engineer or scientist, with the chance of erroneous but also misleading and incomparable results being considerable because of the unspecified methodology that currently exists for some of the most important physical end mechanical properties of the material. Obviously, a critical discussion on the available methodology for the evaluation of the important properties of perlite, with a focus on weaknesses, application limits and sources of uncertainties that follow each method, and would end with a suggestion of a single one for each property characterization would be valuable towards giving specific directions in the characterization procedure, and in parallel eliminating the possibility of incorrect results. It should be also highlighted that the proper implementation of methods prerequisites knowledge of material properties, mainly in terms of its structure and morphology, chemical and mineralogical composition, as well as deep understanding of the expansion phenomenon, the obeying mechanism, the parameters involved and how they affect the product properties.

The Perlite Institute was founded in 1949, a year in which it is estimated only 80.000 tons of geomaterial was mined throughout the world, while currently over 4.3 million tons per year are used in multiple industries including insulation, construction, agriculture, and filtration [3,9]. The unusual properties and advantages of this virtually new material had been known for some time, yet recognition and acceptance were impeded due to a lack of reliable technical data, uniform commercial standards and sufficient public information. Over the next few decades, the member companies worked to develop markets and product specification, and lead research in potential new applications. Special effort was spent in establishing methods for testing and evaluating the main properties of interest for raw and expanded perlite. Such characterization techniques were applied by the institute members, who exchanged information about their application to perlites of different origins as well as the obtained results, leading to strong knowledge integration and continuous streamlining of the applied methodology.

The article presents a review of properties that are of interest for raw and expanded perlite, and available methods for their characterization focused on physical, chemical and mechanical properties, as well as methods for sampling and sample reduction, that are used for similar materials. Under a more critical point of view, the suitability of each method for proper characterization of perlite is evaluated, and the sources of uncertainties and errors are determined. The necessary modifications to avoid the erroneous evaluation of the material following existing methods are presented in detail, or a new methodology is suggested, providing the necessary details that will allow the successful reproduction by the reader. Also, aiming to provide background information related to the perlite, which will facilitate better comprehension of its nature and through this understanding of the discussed characterization aspects, relevant background information about the material genesis, structure, physical properties, physicochemical composition, discussion about the expansion mechanism and involved parameters, and the role of water are reviewed.

2. Materials and Methods

2.1. Ore Genesis, Morphology and Physical Properties

Perlite is a naturally occurring volcanic rock of rhyolitic to dacitic composition and glassy structure that holds water in its structure. The composition of perlite is similar to that of obsidian and pumice with the exception of water content and the ability to expand. Its parental rock, obsidian, is formed during rapid magma raising and cooling that takes place near the earth crust. Perlite deposits are created as silicic lava domes, lava flows, welded ash-flow tuffs, glassy plugs, laccoliths and dikes. The lava flows may have thickness up to several meters and length of more than a kilometer, displaying structures and textures such as phenocrysts, stretched vesicles and flow banding [10]. Perlite domes can range in diameter from 100 m to 2 km, and height may exceed 100 m [11]. Upper parts may exhibit steep flow foliations and ramp structures with ridges on the surface. In both cases, the formations consist of a texturally zoned exterior glass unit enclosing a partially devitrified and crystallized inner glass unit, which is formed due to the decrease of the cooling rate towards their center [12].

Obsidian contains up to 2 wt.% of magmatic water that was entrapped in the material during the rapid cooling event. Further in situ and long-term hydration of obsidian by meteoric water through diffusion (also known as secondary hydration of rhyolite) causes its transformation either to perlite or pitchstone depending on the amount of the entrapped water which, for perlite and pitchstone, is typically between 2 wt.% and 5 wt.%, and over 5 wt.%, respectively [13]. Under the microscope, perlite shows concentric fracture lines, which, according to Marshall, have been created during the rapid cooling of parental rock [14]. Friedman et al. believe that the microfractures were developed after the diffusion of water molecule in material matrix and the inevitable expansion of layers [15]. The typical macroscopic morphologies of perlite are presented in Figure 3.

Perlite presents the unique property to expand when heated rapidly to temperatures that exceed its softening point, while for perlites that possess commercial interest, the expansion degree may exceed 20 times. Figure 4 depicts images of perlite rock, as well as micro- and macro-morphology of expanded perlite.

The expansion temperature depends on physicochemical properties and mineralogical composition, and usually perlites that originate from different deposits do not possess the same expansion performance, while variations are observed even among samples of the same deposit due to local inhomogeneities [18,19,20]. The expansion time is limited and typically does not exceed 3 sec in the different expansion methods applied [21,22,23,24]. After expansion, perlite grains become white and frothy, presenting low density due to the extensive network of air bubbles that are created in the glassy matrix. The unique microstructure of expanded perlite has a beneficial impact on its properties; expanded perlite is lightweight, with sound insulating properties, low thermal conductivity, fire resistant, chemically inert and nonflammable, exhibiting water retention of up to 800 wt.%. Typical values of most important physical properties of perlite are enlisted in

Table 1. Physical properties of perlite [3].

Property	Value
Color	White to off white (expanded) Gray to brownish (unexpanded)
Refractive Index	1.5 (unexpanded)
pH (water slurry)	6.5–8.0
Free Moisture (raw)	typical < 1%
Specific Gravity (raw)	2.2–2.4
Bulk Density (loose weight)	Expanded: 32–400 kg∙m−3 Crude Ore: 960–1200 kg∙m−3
Softening point	>850 °C
Specific heat (nominal)	837 J∙kg−1∙°K−1
Thermal conductivity at 20 °C	0.04–0.06 W∙m−1∙K−1 (expanded)
Solubility	Soluble in hot concentrated alkali and HF Moderately soluble (<10%) in 1 N NaOH Slightly soluble (<3%) in mineral acids (1 N) Very slightly soluble (<1%) in water or weak acids

.

2.2. Chemical and Mineralogical Composition

As a rock, the elemental composition of perlite may vary within certain ranges.

Table 2. Chemical composition of perlite according to Perlite Institute [3], Chatterjee [25], and typical composition of perlites from different locations.

Oxide	Perlite Institute [3]	Chatterjee [25]	Turkey [26]	Italy [26]	China [26]	Greece [20]		Pálháza, Hungary [27]	Morocco [28]
						Tsigrado	Trachilas
SiO2	72.3	70–75	71.0	69.7	71.4	72.7	73.9	68–75	76.5
Al2O3	13.6	12–15	12.8	14.1	12.4	14.9	10.3	10–15	11.8
Na2O	4.6	3–4	3.2	3.3	4.0	4.1	3.7	2.8–4.5	3.3
K2O	4.2	3–5	4.4	4.7	4.4	4.0	5.7	3.2–4.5	4.1
Fe2O3	0.9	0.5–2.0	1.6	2.1	1.1	1.2	0.9	1.0–2.5	0.7
MgO	0.3	0.2–0.7	0.4	0.4	0.3	0.3	0.1	0.2–1.5	0.3
CaO	0.9	0.5–1.5	0.9	0.8	1.0	1.5	0.8	1.5–2.0	0.8
H2O as LOI	3.0	3–5	4.7	3.8	5.0	2.9	3.3	2.0–5.0	3.4

presents the typical chemical composition of perlite from different sources and origins [25].

Usually, the silicon content exceeds 70 wt.% in oxide form, and the content in aluminum oxide exceeds 10 wt.%. The total alkaline content of the samples, namely potassium and sodium oxide do not exceed 9 wt.%, while water content ranges between 2 and 5 wt.%. Figure 5 depicts the classification of perlite using total alkali-silica diagram (TAS diagram), according to the typical composition of perlite proposed by the Perlite Institute (green circle), and the compositional range for perlite as proposed by Chatterjee (blue rectangle) [25]. Perlite lies in the region of rhyolite, except to the case where minimum silica and alkaline contents are combined. In that case it may be of dacitic type.

Regarding the mineralogical composition of perlite, further to the prevailing amorphous phase that exceeds 90% in deposits of economic interest, small portions of other glasses, mainly obsidian, and minerals are present in the amorphous aluminosilicate structure of perlite, either in the form of inclusions or phenocrystals. Common crystalline phases present in perlite are biotite, plagioclase, and feldspar [20,26,29]. Other inclusions that have been reported in bibliography are mica, hematite, magnetite, ilmenite, zircon and spinel [30,31]. Such phases are non-expandable and their presence is unwanted, while it has been shown that the co-occurrence of such minerals on grain level causes fragmentation during the expansion and the production of unwanted ultrafines [20,26]. The identification of mineral phases in perlites is carried out through the X-ray powder diffraction analysis, and a typical diffractogram is depicted in Figure 6. Further to the well-shaped mineral peaks, XRD pattern of perlite present a characteristically wide hump or bulge which extend from 15 to 35 2 theta and is attributed to the presence of non-crystalline silicates [32]. The main glassy and crystalline formulations are readily observable during the analysis of thin or polished cuts of the sample in optical microscope (Figure 7).

2.3. Perlite Expansion Mechanism

Perlite expansion occurs under the synergistic effect of two parallel acting phenomena that evolve during perlite grain heating; the decomposition of water species and release as superheated steam, and grain excess of glass transition temperature entering the softening region. When both conditions are satisfied, the plastic deformation of the material is possible, thus enabling the development of the characteristic foamy structure under the effect of the releasing superheated steam. Undoubtedly, water has a key role in the expansion phenomenon, therefore it has been at the foreground of many past research projects. Interestingly, they presented its discretization into species by following different criteria, the most common of which are the origin (magmatic or meteoric), the form and bond type with aluminosilicate matrix (moisture, molecular, hydroxyl groups), even their role in the expansion (non-effective, effective, residual after the expansion). While water in obsidian is magmatic, pre-existing the solidification procedure, the water that diffuses inside the material forming perlite and pitchstone during secondary hydration is of meteoric origin. The water form is indeed related to its origin; the magmatic water is found in perlite as hydroxyl groups that links to silicate matrix forming silanols, while during the secondary hydration of perlite, which occurs under hydrothermal conditions, mostly molecular water is incorporated in its matrix [33,34,35]. Moreover, it has been shown that OH groups are the dominant species of dissolved water in obsidian (water content below 2%), which increases very slowly beyond 3 wt.% of total water content, while excessive water is attached mostly in molecular form [26,34,36].

Upon gradual heating of perlite samples to a rate that does not exceed 20 °C∙min⁻¹, water removal occurs over a broad temperature interval, ranging from 100 °C to 900 °C, since different amount of thermal energy is required to release superficially adsorbed water, molecular water trapped in large pores, H₂O bounded by capillary forces in micropores, nanopores and microcracks and for structurally bounded OH groups in the glass itself [26,34,37,38,39]. It should be noted that, these heating times are orders of magnitude greater than the expansion times, however, prolongation of heating allows easier sample weighing, and observation of physicochemical and structural alterations that the grain experiences during the thermal treatment. Thus, thermogravimetry constitutes a relatively simple and easy way to investigate the different water types in perlite and categorize them in terms of the releasing temperature. Figure 8 presents temperature ranges for the release of various water species identified in perlites, as determined in different research studies [26,35,40].

Thomas et al. investigated dehydration kinetics of perlite by temperature controlled high-vacuum degassing experiments and under 10 °C/min applied heating rate [35]. The water categories coincide with those proposed by Lehmann and Rössler who made the study almost 40 years ago, However, slightly different ranges of their releasing temperature are proposed [40]. The same water types and similar releasing temperatures were found by Angelopoulos et al. who investigated the kinetics of rhyolite dehydration by applying heating rate of 2.5 to 20°/min to powdery samples and proposed categorization of water according to the change in activation energy of the dehydration procedure, combined with data from IR and Raman spectroscopy [34]. As for Roulia et al. they carried out isothermal heating experiments of perlite samples of different origins and granulometry to various temperatures for 3.5 and 15 h, and categorized water in 3 different types that release in different temperature levels [26].

The same water categories were identified in all studies, however, minor differences in the temperature ranges where each water type releases occur. This is attributed to the fact that the samples that were used for the analysis originated from different mines, possessing different composition, structure and water content. Moreover, the comparison of thermogravimetric data of different samples for the identification of different water species makes sense when their thermal treatment occurred under identical conditions, and also for samples that have similar size distribution, which is not the case with the studies presented above.

It is of particular interest to consider the phenomena that unfold during perlite heating from room temperature over glass transition temperature, causing the release of various water forms, morphological alterations and finally expansion of the grain. The considered conditions do not refer to the extreme conditions that prevail during the expansion in industrial equipment, where perlite is heated to roughly 1000 °C in a matter of seconds and the phenomena unfold almost instantaneously, but to isothermal heating at various temperatures, or slow heating on rate up to 10 °C/min.

At the early heating stage, superficially bound molecular water and readily available molecular water from pores is released, provided that it is adjacent to free surfaces, i.e., micro fractures, internal and external pores or external grain surface (Figure 9, Phase 1). This type of water does not contribute to the expansion, since its removal occurs below 250 °C which is a very low temperature compared to the glass transition temperature (T_g). As the temperature increases, but before grain reaches glass transition temperature, mainly bound molecular water and water trapped in pores of the material is gasified, and free hydroxyl groups are released (Figure 9, Phase 2). Normally, the water that is located at the external layers of the grain releases unhindered. On the other hand, the release of the internal water is hindered by the still dense structure of the material, causing pressure increase and tension accumulation that leads to the formation of a network of micro-fractures which act as steam escaping routes, triggering the expansion process [31,35]. Further increase of the grain temperature over glass transition temperature, which normally occurs at around 750–850 °C, causes grain softening and the release of more water from the fractures (Figure 9, Phase 3) [38,41]. The pressure inside the grain increases significantly and the network of micro-fractures do not provide a sufficient route for gases release anymore. Finally, the relief of accumulated pressure occurs after the expansion of the softened grain by the enlargement of the fractures and the establishment of the bubbles network. After the expansion, a portion of the water is still present in the sample, mainly in the form of tightly connected hydroxyl groups [26,34]. This water type is also called “residual water” because it survives the expansion procedure. However, it may cause minor expansion of the grain once heated at an even higher temperature.

In addition to its role in the expansion as a “foaming agent”, water affects the viscosity of the grain and the glass transition temperature, thus the temperature where softening occurs and expansion initiates [42]. As a rule of thumb, the higher the water content of the sample, the lower the expansion temperature. The implementation of a theoretical model for the calculation of dynamic viscosity versus temperature for perlite with different water content shows that the reduction of water content of perlite by 1 wt.% (from 3 to 2 wt.%) causes viscosity-temperature curve shifting by 100 °C towards higher temperature [22]. Similar to water, alkali metal oxides of potassium and sodium which are present in considerable amounts in perlite reduce the viscosity and the melting point of the grain [22,24].

3. Methods for Characterization of Raw and Expanded Perlite

Figure 10 depicts the analyses applied for the characterization of both raw and expanded material. As a rule of thumb, the identification of a perlite sample is performed by the determination of its chemical and mineralogical composition, together with chemically bound water content through a simple LOI test. The non-crystalline structure is identified either with macroscopic observation or throw X-ray diffraction analysis. After the collection of all the aforementioned information, exploratory expansion trials are performed to granulometric grades of interest. The aim here is the investigation of the expansion performance, thus the expansion campaign is implemented at various conditions scoping to maximize the expansion yield, and it is repeated for the granulometries of interest. As for the primary evaluation of the expanded products, further to the expansion yield that is determined through the bulk density measurement, the composition of the expanded perlite is important, mainly in terms of its content in adequately expanded grains, as well as in shattered particles and unexpanded ones. The high portion of unexpanded particles may denote high content in non-expandable phases (crystalline phases), or inadequate heating that led to poor expansion. Also, high proportion in unwanted shattered particles can be attributed to mineralogical composition and texture. Granulometry measurement is another fundamental analysis which is applied aiming to determine the particle size distribution of the products, but also to determine the amount of ultrafines in particular. There is a number of other characterization analyses that are applied to evaluate a specific property of expanded perlite that is of interest for a certain application, as seen in Figure 10. The majority of these targets provide information on the physical properties of the product, as well as the strength that it exhibits under different stress modes.

3.1. Sampling and Sample Preparation (Size Reduction)

Acquiring a homogenous sample of perlite, either raw or expanded, prior to analyses is important to minimize analytical errors [43]. Especially for expanded perlite, the application of sampling procedure is important because of the inevitable segregation that is driven by particles’ differences in size and density, while the friability of the material should be considered too, choosing a method that applies mild conditions and avoids stressing [44]. For samples stored in bags or silos, a portion of about 2 L should be secured by means of a suitable sampling spear (also called thief tube) [43]. The sampling tube shall be inserted the full distance between diagonally opposite corners of the bag, which should be lying in a horizontal position. In the production environments, it is suggested that at least one sample shall be taken per 10 m³ of product to check production consistency and uniformity. In that case, the portions so obtained shall be combined to produce a composite sample having a total volume of at least 30 L. A “grab sample” can be taken by holding a container with a capacity of 3–5 L (or “gallon can”, or other container of known volume) in the production stream. Care should be exercised that the container is passed across the entire stream or flow so that material is selected from the entire stream or flow.

Large samples cannot be tested all at once, and their amount must be reduced to reflect the accuracy and requirements of the selected test methods. Splitter or riffle can be used for this purpose, where errors in the sample extraction are prevented by applying symmetry. The splitters usually consist of 8–16 chutes that discharge on both sides of the splitter, with one side being retained and the other rejected. The procedure is repeated until a sample of the approximate required amount is obtained. Devices of different capacities are available, depending on the application level and the volume of the final sample (Figure 11). It is suggested the implementation in a closed-bin riffle splitter because it presents minimum fine material losses [45]. Different studies have confirmed that the spinning riffle is among the best sampling methods [46,47,48]. However, it should be avoided for expanded perlite since particles experience intensive vibratory and circular motion which can potentially cause fragmentation.

When riffle is not available, cone and quartering method is an alternative. This procedure is applicable to a sample of perlite ore or expanded perlite, where the material is poured into a heap, flattened and divided by a cross-shaped cutter giving four identical samples [43]. Stepwise presentation of the method implementation on expanded perlite is depicted in Figure 12.

The relative standard deviation for sample obtained through chute splitting and cone and quartering method of sample containing mixture of particles of weight 0.05 g and 0.10 g has been investigated [43,48]. Standard deviation, and advantages and disadvantages of both discussed sampling methods are presented in

Table 3. Pros, cons, and relative standard deviation for two common sampling methods.

Sampling Method	Advantages	Disadvantages	Relative Standard Deviation (%)
Cone & Quartering	Good for powders with poor flow characteristics	Operator-dependent	6.81
Chute Riffling	Ability to reduce powder samples in half after one pass	Operator bias	1.01

[49].

3.2. Density and Size Distribution

Undoubtedly, particle size is the one property of a filler material that influences every aspect of its use and the success of many applications [50]. Indeed, for expanded perlite, together with bulk density and morphology, particle size is deterministic to the appropriateness of a grade for certain applications. Particle size and bulk density range for the most common expanded perlite applications, together with macroscopic morphology of grains are tabulated in

Table 4. Common expanded perlite grades and acceptable range of the particle size distribution, density and appearance [3].

Application	Size Range (μm)	Loose Bulk Density (LBD) (kg∙m−3)	Appearance	Use	Key Properties
Horticulture	500–5000	60–130	Coarse Granules	Hold water and air in the soil that is rendered available in mid term	Surface porosity Water absorption
Construction	63–2800	60–150	Fine and Middle Granules	Reduction of static load Provide with thermal and sound insulation	Surface porosity Density Strength
Cryogenic	150–1200	30–120	Small Granules	Minimal thermal conductivity, used as loose fill	Density Particle size
Filter Aid	<200	50–200	Powder	Crushed to produce tiny fragment. Used as precoat on filter septum to increase capacity, and as body feed	Size Thermal conductivity
Microspheres	<300	60–400	Rolling Powder	Tiny expanded particles of closed porosity used as filler	Size Strength Sphericity
Water retentive fines for agriculture	<800	60–160	Small Granules	Ultrafine highly porous particles, with great water holding capacity	Size Surface porosity

.

In most expanded perlite products, like horticulture and construction grades, a lightweight material of high porosity is desirable thus the bulk density is between 50 and 130 or 150 kg/m³. Further reduction of the density is preferable for low temperature and cryogenic storage vessels insulation only, which target the minimum thermal conductivity. Moreover, such highly porous materials are brittle and vulnerable to fragmentation under minimal strain, thus its use in applications involving stress of any kind should be avoided. This is not the case for cryogenic perlite which fills annular space in double-walled vessels. A considerably higher density is desirable for microspheres, which are mainly used as filler in various composites [51]. This is because filler should survive high shear forces and high pressures involved in manufacturing processes such as compounding, molding and extrusion. Moreover, small size and high sphericity allow better dispersion in the mixture [52]. While in all reported applications, the final product is in expanded form consisting of lightweight granules, this is not the case for the Filter Aid grade, where the expanded perlite is subjected to milling to obtain fine powder, which microscopically consists of tiny sized expanded perlite wall fragments.

The bulk and the absolute density are of interest for expanded perlite. The difference between those two terms lies in the considered sample volume. As far as the bulk density is concerned, its determination is done through dividing the sample weight by the volume occupied in a volumetric cylinder. Thus, the overall volume consists of the volume of the solid particles, including closed and open pores (intra-particle porosity) and the interparticle space. The bulk density is a robust way to determine the expansion degree, it is used to estimate volumes in loose fill applications, as well for estimating the volumes in packaging and transportation. As far as the absolute density (also called true, real, apparent and skeletal density), the volume considered in the case of sample in a volumetric cylinder is that of the solid particles including the closed porosity and excluding open pores and the intra-particle porosity. Its determination is performed by gas pycnometry [53,54]. For the analysis, the sample is placed in a cell of known volume and filled with gas and the pressure is measured, and subsequently a valve is opened connection the cell to another reference cell of known volume and the pressure is measured again. Knowing the two measured pressures and the volumes of both cells, the volume filled by the sample is calculated using the Ideal Gas Law. Helium gas is commonly used because it is inert, safe, it obeys the ideal gas law and penetrates to small pores. Mercury porosimetry has the same operating principles, but its use is limited because it is dangerous for the user and samples produced demand implementation of special disposal procedure. Another potential method for determination of absolute density is described in EN1097 using water [7]; the density is determined by measuring the mass of water that is displaced by the sample of known mass. The method fails for expanded perlite because the sample overfloats.

Sieving is traditionally applied for the determination of size distribution both for raw and expanded perlite. Despite the fact that sieving is characteristically simple, the process is more complex and governed by multiple principles, with the factors affecting process performance being the size and shape of particles, mesh size and aperture shape, amount of material on the sieve, the direction and rate of the sieve movement, presence of moisture on sieving material [55,56]. Sieve blinding is another factor with direct affection and considerable impact in the process [57,58]. This is because it reduces the effective transfer area on the surface, resulting in reduction of sieve rates and sieve efficiency. An efficient way to avoid this is by adding energy on screen surface, combining gyratory and jolting movement. The Tyler Ro-tap incorporates the combination of the forces generated by the two actions. Together with tapping, duration and method has emerged as the most important sieving factors, and should be controlled when targeting proper comparison of different materials [57].

Dry sieving procedures are suggested for raw and expanded perlite; however, wet sieving should be applied for fine samples to determine the <45 μm portion. Suggested sieving conditions and parameters related to equipment are tabulated in

Table 5. Sieving parameters for determination of particle size distribution of coarse and fine perlite grades.

Sieving Parameters	Perlite Grade
	Coarse (>150 μm)	Fine (<150 μm)
Type of sieving	Dry	Wet for 45 μm
Sieve size	200/203 mm (8 in.) diameter, and 50 mm (2 in.) height
Aperture size	150 μm, 300 μm, 600 μm, 1.18 mm, 2.36 mm	45, 75, 150 μm
Sample quantity	500 mL
Sieving time	5 min
Tapping frequency	2.5 Hz (150 taps per min)

.

The sieves characteristics should follow specifications described in ASTM E11 [59]. ASTM C136 standard method describes dry sieving process, however, two modifications are required because of the brittleness and light weight of expanded perlite [8]. The first is that the sieving time should be 5 min rather than the suggested 10 min to avoid particle degradation. The second has to do with size of sample which, for dry sieving is predefined on volume basis. In the standard procedure, the sample mass is not defined, and it is referred that to ensure that all particles have the opportunity to reach sieve opening a number of times, the quantity should be limited up to 7 kg/m² on each sieve after sieving completion. The value does not make sense for expanded perlite; assuming sample of 50 kg/m³ bulk density, the 7 kg/m² load is 4.2 L of sample which is more than twice the capacity of the sieve.

Another measurement of interest that combines granulometric analysis and bulk density determination is the fractional density. The method targets the determination of fractional loose bulk density of expanded perlite, where each fraction is obtained by sieving and consequently subjected to bulk density determination. The method is commonly used as a tool to evaluate the efficiency of the expansion process, which ideally should lead to the same expansion degree regardless of the feed size. Also, high deviation in density of the ultrafine fraction from the main bulk density may denote high proportion in expanded grain fragments produced during expansion or inadequate expansion of the ultrafine fraction.

3.3. Strength-Related Testing

Expanded perlite is widely used as filler to form organic or inorganic composites, depending on the matrix material that can be ABS [60,61], polystyrene [62], polypropylene [63,64,65], high density polyethylene [66], or geopolymer [1,67,68] and concrete [69]. A fragile filler is vulnerable to disintegration during mixing and composite formation, thus failing to produce a stiff composite. Furthermore, a filler with poor mechanical properties will affect adversely the mechanical properties of the final composite. Moreover, it is known that compressive strength of composites depends on the stiffness of the material, thus, all of the parameters which affect stiffness, including the effect of fillers, influence compressive strength [50,70]. The lightweight filler breakdown will also affect the composite density that will be higher than expected, since part of its initially occupied volume will be replace by the binder after its breakdown. But even when it is intended for loose fill application, certain strength standards are required to ensure it will withstand all applied procedures of packaging, transportation, receiving, conveying and storage, with minimal affection of its quality.

Compressive strength or crushing resistance constitutes the primary property evaluated to gain information about the strength and durability of lightweight fillers, while other strength related analyses that involve tension or shear force are not applicable [71]. A standard method for measuring crushing resistance of lightweight fillers is described in EN13055 and is applicable for expanded perlite [72]. The standard describes 2 procedures; procedure 1 is applied for lightweight aggregates in size range 4 mm to 22 mm and with bulk density over 150 kg/m³, and procedure 2 that is applicable to lightweight aggregate with bulk density below 150 kg/m³ without size range determination. For expanded perlite procedure 2 is of interest, according to the bulk density ranges of common grades as presented in Figure 2. For the analysis, a portion of expanded perlite is obtained through sampling and placed the sampler that is subjected to vibration on a vibrating table (Figure 13a). The analysis is implemented on a hydraulic press (Figure 13b), where the piston immerses in the test cylinder compressing the sample on a constant rate until a compression of 50 mm is achieved. Analysis conditions are tabulated in

Table 6. Conditions applied for crushing resistance measurement.

Property	Value
Piston speed	0.5 mm/s
Vibrating table settings	Oscillations: 3000 per min
	Amplitude: 0.5 mm (without load)
	Duration: 3 s (without and with the collar)

.

The crushing resistance (C) is calculated by the following equation:

$$C=\frac{L+F }{A}$$

(1)

where L is the force exerted by the piston in N, F is the compression force in N and A is the area of the piston in mm².

It is noteworthy that, neither of the Procedures 1 and 2 are applicable for microspheres since typically their bulk density exceeds 150 kg/m³ and the particle size is below 0.4 mm. For such type of materials, the isostatic crushing strength is commonly measured to evaluate their mechanical strength. The measurement is based on an ASTM standard developed in 1978 and withdrawn in 1984 with no replacement [73]. For the test 3–6 cm³ of microspheres are sealed in a rubber balloon together with glycerine or isopropyl alcohol. Consequently, the balloon is placed in a pressure chamber filled with hydraulic oil, pressure is applied and the change in volume and pressure is recorded. The data is used to calculate the percentage of microspheres collapsed versus the applied pressure. Despite its early withdrawal, the method is still used by glass beads, microbubbles, and microspheres [50,74,75] manufacturers as the primary mechanical test, usually expressed as maximum pressure for survival of 90% of the particles.

Abrasion resistance is another important mechanical property of expanded perlite microspheres, because it has been shown that fillers with high abrasion resistance reduce the composite wear rate [76]. There is no standard process available to evaluate such property for expanded perlite microspheres, However, a simple process applied by Bublon GmbH, a subsidiary of Binder + Co AG is suggested [77]. Around 100 mL of microspheres are obtained and placed in a steel vessel, together with stainless steel balls (Figure 14). The vessel is rotated under specific frequency and time (

Table 7. Properties and conditions for abrasion resistance measurement.

Property	Value
Container dimensions	120 mm × 300 mm
Sample volume	100 mL
Balls	Weight: 1 kg of stainless-steel balls Diameter: 6 mm each
Rotation	Speed: 380 rpm Time: 15 min

).

Abrasion resistance (R) is calculated by the following equation:

$$R=1-\frac{1/{\rho }_{out}-1/{\rho }_{raw}}{1/{\rho }_{in}-1/{\rho }_{raw}}$$

(2)

where ρ_in and ρ_out are skeletal density values of expanded microspheres before and after abrasion testing, and ρ_raw is the skeletal density of raw unexpanded perlite (typical 2.4 g∙cm⁻³).

3.4. Water Repellency/Absorption and Oil Absorption

There are two standard procedures for measuring the water absorption of aggregates which prescribe material saturation in water for 24 h and surface water removal prior to measurement [7,78]. However, neither of these methods are intended for lightweight aggregates, because the pores are not necessarily filled with water after the prescribed soaking time. Also, the important condition of no loss of particles during surface wetting is not fulfilled, since ASTM C128 is performed with a warm air stream and in EN1097-6 by gentle rolling on dry cloth. Moreover, expanded perlite particles crush with rolling on the cloth. Perlite Institute members use a simple method for the determination of water repellency by measuring the amount of passing water that will be repelled of the bulk volume of a vertically oriented, cylindrical sample of mildly compacted expanded perlite. It should be noted that the sample is neither subjected to water saturation nor to surface wetting.

For the measurements, a rigid plastic tube with inner diameter of 50 mm and length of 300 mm is used, with a 150 μm (100 mesh) screen covering firmly fastened or adhered to the bottom (Figure 15). The tube shall be marked at 400 mL from the screen covered end. First, a representative sample of about 420 mL of expanded perlite is secured and spooned into the test cylinder. The sample is compacted by dropping the tube from a height of approx. 75 mm on a large rubber stopper for a total of 10 drops. As the sample compacts to a level below the 400 mL mark, additional material shall be added so that after the tenth drop the level of the sample is within 3 mm of the 400 mL mark. The tube is weighted and the sample weight is noted. With the tube supported in a vertical position and a beaker under the tube, 250 mL of cold tap water (T ≈ 15 °C) are rapidly poured onto the perlite. Care must be taken while pouring to ensure that the stream hits the middle of the surface of the bed of perlite and does not merely slide down the side of the test cylinder. The water is allowed to drain through the column of perlite, and the amount of water repelled by the sample in the beaker is measure after 3, 5, 7, 10, and 30 min of draining time, or till drain stops. Water repellency is expressed either as ml of water repelled or per sample mass on dry basis. For water absorption calculation, the water uptake of the sample is considered, and expressed per sample mass on dry basis, while intermediate measurements can be used for graphic depiction of water removal kinetics.

The determination of oil absorption is performed according to standard procedures developed for pigments [79,80]. Linseed oil is added drop by drop in the sample, and mixed using palette knife until sample saturation in oil is achieved. The only difference in methods is the endpoint, which for the ISO method is a paste of smooth consistency and for ASTM is a very stiff, putty-like paste that does not break or separate. It is suggested to note the method followed for obtaining oil absorption together with the calculated value. The sample portion depends on the expected oil absorption value, and for expanded perlite 1 g of sample is suggested. The expected expanded perlite sample is between 100–700 g oil/100 g of expanded perlite.

3.5. Microstructural Composition of the Sample

A closer look in an expanded perlite sample informs that, together with adequately expanded grains, particles with different morphologies like tiny fragments and unexpanded grains are also present. Moreover, the proportion of a sample in adequately expanded grains is an important quality criterion, which is affected both by the feed properties and the applied expansion conditions. More specifically, main grain morphologies identified in expanded perlite sample and their behavior in aqueous solutions are the following:

• Expanded grains: white grains with cellular structure and rounded edges, having open or closed surface porosity. In adequately expanded perlite, the majority of the perlite grains are of this category. Expanded grains float once inserted in the water due to the low density.
• Unexpanded grains: are the non-crystalline, dark colored impurities commonly present in perlite such as obsidian and non-expandable crystalline minerals (quartz, biotite, plagioclase, feldspar, mica, hematite, magnetite, ilmenite, zircon and spinel, etc.). Sometimes, small portions of the feed may remain unexpanded due to the applied expansion conditions. Due to fact that their specific gravity exceeds 1, both non-expandable and unexpanded perlite grains, sink rapidly in the water.
• Fragments of shattered grains: constitute the fragments of expanded perlite that are produced either during or after the expansion. The application of extreme heating conditions and violent expansion, or the coexistence of crystalline and non-crystalline matter in a single grain leads to grain fragmentation. Additionally, due to the fragility of expanded perlite, grain fragmentation occurs due to grain–grain or grain–surface collision inside the furnace or the exhaust system. Due to their tiny size and flattened shape, fragments suspend themselves in water, sinking slowly and finally settling above the sinks. These fragments are of light white color, similar to the expanded grains.

The simple method presented here is applied to determine the content of sample in grains of each morphological category, and it is performed by exploiting the effect of shape and density on particle kinematics in water (Figure 16). The analysis is conducted in an Imhoff cone on 300 mL of expanded perlite sample of known weight (W_in). First, 650 mL of water are added in the cylinder, followed by the addition of the sample and the rest 350 mL of water. After mild stirring, the slurry is left to rest for 45 min. The volume occupied by sinks (V_sink) and shattered particles (V_shat) is noted, and the volume basis composition of the sample in different morphological categories is performed by Equations (3)–(7). After implementation of the test, sinks and shattered are collected by removing the bottom cap, which are dried and weighted as a whole (W_sink+shat), allowing weight basis determination of the composition through Equations (6) and (7).

$$Floaters (v/v·\%)=\frac{300-(V_{sink}+V_{shat})}{300}·100$$

(3)

$$Sinks (v/v·\%)=\frac{V_{sink}}{300}·100$$

(4)

$$Shattered (v/v·\%)=\frac{S{V}_{shat}}{300}·100$$

(5)

$$Floaters (w/w·\%)=\frac{{W_{in}-W}_{sink+shat}}{W_{in}}·100$$

(6)

$$Sinks+Shattered (w/w·\%)=\frac{W_{sink+shat}}{W_{in}}·100$$

(7)

The existence of expanded perlite grains in grades intended for filtration applications is unwanted because they tend to float rather than stick on the septum of the filter. Thus, measuring of the filter aid grade content in expanded perlite is an essential quality metric, However, there is no standard procedure to do so. An easy, robust and accurate way to measure floaters content is by using a 200- or 250-mL volumetric flask with long graduated neck (Cassia flask or similar). A total of 50 mL of representative sample are obtained using a graduated measuring cylinder and introduced in the empty flask, which subsequently is filled with water to the top mark in the neck. After 5 min, or once particles no longer float to the top, the volume of the floating matter is read on the flask neck, and the result is obtaining by dividing it with the initial volume of the sample (50 mL) and given as % floaters (v/v).

3.6. Moisture and Chemically Bound Water Content

The determination of both physical adsorbed (free moisture) and chemically bound water in perlite ore and expanded perlite is of interest. The moisture that is physically and temporary adsorbed at the external surface of the material is easily removed by applying mild thermal treatment, typically at 110 ± 5 °C and 105–110 °C as describe in the relevant ASTM and ISO standard procedure, respectively [81,82]. As for the drying time, it is suggested the sample heating to constant mass in both protocols, while ISO procedure suggests that, for fine sample, drying for 16 h is sufficient. Free moisture is calculated on dry basis. The analysis is important for furnace feed material and its value should be below 0.5% to avoid decrease of the furnace temperature and energy wasting.

The determination of combined water content on raw and expanded perlite is performed by measuring the LOI of a dry sample, i.e., the weight loss as percentage of initial sample mass after thermal treatment. Further to water release, the combustion causes the removal of carbon dioxide from carbonates, the conversion of metal sulfides into metal oxides, metal sulphates and sulfur oxides, and organic matter burning. As seen in

Table 8. Heating conditions for determination of LOI on different type of materials.

Type	Heating Temperature, °C	Heating Time, min	Reference
Limestone	1000 ± 20	20 or until constant weight	[85]
Calcareous sediments	1000	60	[86]
Portland and slag cement	950	15	[87]
Pigments	900–1000	20 (phase 1) and 10 (phase 2)	[88]
Soil combustion residue	750 (Method A) 950 (Method B)	120	[89]

, different conditions are proposed for LOI of materials similar to perlite. However, the conditions are specific in terms of heating temperature and time, while such information is missing for perlite. Experiences has shown that for raw and expanded perlite, LOI determination is suggested to be done by sample heating at 950 °C for 4 h. Under such conditions, it is safe to assume that the LOI content is solely attributed to the removal of various water species, due to the minimal iron, sulfide, carbonates and organic matter content [83,84].

It is noteworthy that, ideally, LOI content is determined in thermogravimetric analyzer with simultaneous recording of the sample weight, heating time and temperature under controlled atmosphere. However, the inevitable expansion of the sample during heating causes fluctuations on weight measurements as well sample overflow from the sampler rendering the method inappropriate.

4. Conclusions

Perlite is an industrial mineral with an already well-established market, primarily in the construction, horticulture and filtration sector, but also with great prospects as core material towards the implementation of sustainability principles in buildings, constructions, horticulture and other sectors. Proper characterization of raw material and expanded perlite is aimed to the evaluation of the feed as source for the production of expanded material of high quality and yield, and for the determination of expanded product performance in applications of interest, respectively.

The article reviews the available methodology for the determination of the followings, which are critical for perlite, with a focus on methods’ applicability to perlite; Sampling and size reduction, density (bulk, skeletal and bulk fractional), particle size, water absorption/repellency and oil absorption, compressive strength, abrasion resistance, composition of expanded perlite, and moisture and combined water content. Main conclusion was that, when available, almost all relevant methods that are applicable for similar materials fail for perlite, because of the unique way that it combines ultrafine granulometry, very low density and the vulnerability to disintegration if not treated properly. As for the expanded sample composition and the abrasion resistance, no method is available. After identification of the source of uncertainty and error, the lack of proper methodology is tackled either by providing with modifications to existing methods, or through presenting new characterization techniques that has been tested and adopted throughout the last 60 years by the Perlite Institute members. The provided information allows easy and precise reproduction of the methods by interested individuals, organizations, or companies. Deep knowledge of tested material is a prerequisite towards its characterization, thus updated insights on ore genesis, physical properties, chemical and mineralogical composition and the mechanism of perlite expansion due to thermal treatment are also included, composing an integrated study which will support the proper assessment of this unique material.