1 point in high energy ball milling grinding s determined by

high-energy ball milling - an overview | sciencedirect topics

high-energy ball milling - an overview | sciencedirect topics

High-energy ball milling is a ball milling process in which a powder mixture placed in a ball mill is subjected to high-energy collisions from the balls. High-energy ball milling, also called mechanical alloying, can successfully produce fine, uniform dispersions of oxide particles in nickel-base super alloys that cannot be made by conventional powder metallurgy methods. High-energy ball milling is a way of modifying the conditions in which chemical reactions usually take place, either by changing the reactivity of as-milled solids or by inducing chemical reactions during milling [20].

High-energy ball milling is a mechanical deformation process that is frequently used for producing nanocrystalline metals or alloys in powder form. This technique belongs to the comminution or attrition approach introduced in Chapter 1. In the high-energy ball milling process, coarse-grained structures undergo disassociation as the result of severe cyclic deformation induced by milling with stiff balls in a high-energy shaker mill [8,9]. This process has been successfully used to produce metals with minimum particle sizes from 4 to 26nm. The high-energy ball milling technique is simple and has high potential to scale up to produce tonnage quantities of materials [8]. However, a serious problem of this technique is the contamination from milling media (balls and vial) and/or atmosphere. Therefore, a number of improvements, including the usages of surfactants, alloy-coated milling media, and protective atmospheres, have been developed to alleviate the contamination problem [8].

The fine powder (in nano or submicron sizes) produced from ball milling can be consolidated to bulk form for large-scale applications such as hip implants and bone screws. Usually, the fine powders are compacted and sintered together via methods like hot isostatic pressing and explosive compaction under the temperatures or conditions that suppress grain growth and maintain nanocrystalline microstructure [8,10]. Bulk metallic materials produced by this approach have achieved the theoretical densities of nanocrystalline materials and greatly improved mechanical properties compared to their conventional, micron-grained counterparts.

High-energy ball milling is effective in getting well-dispersed slurry.79 The preparation procedure is summarized in Fig.24.2. First, commercially available PZT powders (APC 850) were high-energy ball milled to get the desired particle size. Secondly, a selected dispersant was added to the milled powders to get the surface-modified powders. The smaller the powder, the more important this procedure. Afterwards, PZT precursor solution was added to these surface-modified powders and mixed by further ball milling. Finally, the resultant uniform slurry was ready for further processing, such as spin coating, tape casting, screen printing and molding. The recipe for the slurry, including the concentration of xerogel solution and powder to solution mass ratio, depends on the further processing method employed. For our convenience, the recipes for the slurry were given four numbers with regard to the above two important parameters. For example, in 3025, the first two numbers represent the concentration of the xerogel solution9 in weight percent, i.e. 30wt%, and the last two numbers represent the mass ratio of the added PZT powder to xerogel solution, namely 2 to 5.

High-energy ball milling, also called mechanical attrition, can be used to reduce the grain size of materials from many micrometers to 220nm (see Mechanical Alloying). This is a result of the cold-working process creating large-angle grain boundaries. Most of the reduction in grain size occurs rapidly, but the process slows, and long times are required to reach the smallest sizes. This process has the advantage of being relatively inexpensive and can be easily scaled up to produce large quantities of material. Usually, to maximize the energy of collision, high-mass hard-steel or WC balls are used. Contamination by materials removed by the balls is a major concern. Severe mechanical deformation and plastic deformation at high strain rates (103104s1) occurs during the process. Initially, shear bands are formed consisting of a high density of dislocations. Later these dislocations annihilate and recombine as small-angle grain boundaries forming nanometer-sized grains. Finally, the orientation of these nanometer-sized grains is randomized.

The range of solubility of multicomponent systems is greatly increased by mechanical attrition. Mechanical attrition can also produce metastable materials. If the milling is done in the presence of O2 or N2, oxides or nitrides can be formed.

High-energy ball milling, a predominantly mechanical process, nevertheless results in significant structural and chemical changes in the material. Nonequilibrium synthesis of materials at low temperatures via ball milling is possible through a combination of multiple processes, which occur during milling. These processes include thermal shock, high-speed plastic deformation, mechanical grinding and fracturing, cold welding, and intimate mixing [9].

BNNTs were typically synthesized by the prolonged (approximately 150 h) high-energy milling of pure boron or h-BN powder using stainless-steel milling vessels and hardened steel balls in a pressurized (2.3 103 Torr) NH3 atmosphere. The milled material was then annealed at high temperature (>1000 C) in an N2 atmosphere for 10 h. It was found that large quantities of BNNTs can be synthesized using this method. The yield of the BNNTs depended on the duration of the milling treatment [11]. It was proposed that nanotube formation by this method was caused by two different mechanisms. The first mechanism being the nitridation of B nanoparticles in the NH3 atmosphere, which in turn served as nucleation sites for the formation of BNNTs. The second mechanism proposed was that the Fe (and other metals such as Cr and Ni) from the milling process was incorporated into the B powder during high-energy milling and that the metal particles then served as catalysts for BNNT growth [11]. In order for these two mechanisms to operate effectively, it is necessary that both the ball milling and annealing steps be carried out for long times. Other variations of this technique have been reported including the use of tungsten carbide (WC) balls, and a mixture of NiB and alumina [5,12]. Even though the yield of BNNTs can be very high using this method, the resultant nanotubes can suffer from contamination and structural defects. Figure 8.6 shows a micrograph of BNNTs synthesized using this technique.

Figure 8.6. Transmission electron microscope image of BNNTs synthesized by the ball milling process. The growth of the nanotubes from the milled material is clearly evident. The largest nanotube imaged has a bamboo-like morphology.

High energy ball milling can lead to glass formation from elemental powder mixtures as well as by amorphization of intermetallic compound powders. Solid state amorphization by high energy milling has been demonstrated in a number of Ti- and Zr-based and other alloy systems such as NiTi, CuTi, AlGeNb, SnNb, NiZr, CuZr, CoZr and FeZr. The process of ball milling is illustrated in Figure 3.56. Powder particles are severely deformed, fractured and mutually cold welded during collisions of the balls. The repeated fracturing and cold welding of powder particles result in the formation of a layered structure in which the layer thickness keeps decreasing with milling time. A part of the mechanical energy accumulates within these powder particles in the form of excess lattice defects which facilitate interdiffusion between the layers. The continuous reduction in the diffusion distance and the enhancement in the diffusivity with increasing milling time tend to bring about chemical homogeneity of the powder particles by enriching each layer with the other species being milled together. The sequence of the events that occur during milling can be followed by taking out samples from the ball mill at several intervals and by analysing these powder samples in respect of their chemical composition and structure. Let us describe one such experiment in which elemental powders of Zr and Al were milled in an attritor under an Ar atmosphere.

Elemental powders of Zr and Al of 99.5 purity, when milled in an attritor using 5 mm diameter balls of zirconia as the milling media and keeping the ball to powder weight ratio at 10:1, showed a progressive structural change as revealed in XRD patterns (Figure 3.57(a) and (b)). Diffraction peaks associated with the individual elemental species remained distinct upto 5 h of milling at a constant milling speed of 550 rpm. All particles and the balls appeared very shiny in the initial stages. With increasing milling time, the particles lost their lustre, the 111 and 200 peaks of fcc Al gradually shrunk and the three adjacent low-angle peaks of hcp -Zr, corresponding to1010, 0002 and1011, became broader. After about 15 h of milling, XRD showed only -Zr peaks which shifted towards the high angle side, implying a decrease in the lattice parameters resulting from the enrichment of the -Zr phase with Al. After 20 h of milling, all Bragg peaks except one broad peak close to the{1010} peak disappeared. Powders milled for 25 h showed an extra reflection corresponding to a lattice spacing of 5.4 nm, which matches closely to a superlattice reflection of a metastable D019 (Zr3Al) phase. On further milling, the powders transformed into an amorphous phase. The sequence of structural evolution could be described as -Zr + Al -Zr (Al) solid solution + Al nanocrystalline solid solution + localized amorphous phase Zr3Al (D019) + -Zr (Al) solid solution + amorphous phase bulk amorphous phase.

Figure 3.57. XRD patterns showing a progressive structural change for different times when elemental powders of Zr and Al of 99.5 purity were milled in an attritor using 5 mm diameter balls of zirconia with a ball to powder weight ratio of 10:1.

The mechanism of solid state amorphization during mechanical alloying has been studied on the basis of experimental observations made on several alloy systems. One of the probable mechanisms, based on local melting followed by rapid solidification, has not found acceptance as evidence of melting could not be seen in experiments. The example of ball milling of elemental Zr and Al powders has demonstrated that the amorphisation process is preceded by the enrichment of the -Zr phase to a level of approximately 15 at.% Al. The solute concentration progressively changes during milling. The various stages encountered in the course of amorphization can be explained in terms of schematic free energy versus concentration plots for the , the metastable D019, and the amorphous phases (Figure 3.58). With increasing degrees of Al enrichment, the free energy of the interface region gradually moves along the path 1-2 (Figure 3.58). Once the concentration crosses the point 2, it becomes thermodynamically feasible to nucleate the Zr3Al phase which has the metastable D019 structure. Although the equilibrium Zr3Al phase has the L12 structure, it has been shown (Mukhopadhyay et al. 1979) that the metastable D019 structure is kinetically favoured during the early stages of precipitation from the -phase. This is not unexpected as the hcp structure and the D019 structure (which is an ordered derivative of the former) follow a one-to-one lattice correspondence and exhibit perfect lattice registry.

Figure 3.58. Schematic free energy concentration plots in ZrAl system for the , the metastable D019 and the amorphous phases illustrating the various stages encountered in the course of amorphization.

With further Al enrichment, as the concentration crosses the point 3, nucleation of the amorphous phase becomes possible. It is to be emphasized that the change in composition occurs gradually from the interface to the core of the particles, with the result that the amorphous phase starts appearing at interfaces while the core remains crystalline. As the Al concentration in the powder particles crosses point 4, each particle can turn amorphous by a polymorphic process. The observed sequence of solid state amorphization in the case of ball milling of elemental Zr and Al powders suggests the occurrence of amorphization by a lattice instability mechanism which is brought about by solute enrichment of the -phase beyond a certain limit (point 4 in Figure 3.58).

The synthesis of materials by high-energy ball milling of powders was first developed by John Benjamin (1970) and his coworkers at the International Nickel Company in the late 1960s [42,43]. It was found that this method, called mechanical alloying, could successfully produce fine and uniform dispersions of oxide particles (Al2O3, Y2O3, ThO2) in nickel-base superalloys which could not be made by conventional powder metallurgy methods.

It is a ball milling process where a powder mixture placed in the ball mill is subjected to high-energy collision from the balls. Fig. I.7 shows the motions of the balls and the powder. Since the rotation directions of the bowl and balls are opposite, the centrifugal forces are alternately synchronized. Thus, friction resulted from the hardened milling balls and the powder mixture being ground alternately rolling on the inner wall of the bowl and striking the opposite wall. The impact energy of the milling balls in the normal direction attains a value of up to 40 times higher than that due to gravitational acceleration. Hence, the planetary ball mill can be used for high-speed milling [44].

During the high-energy ball milling process, the powder particles are subjected to high energetic impact. Microstructurally, the mechanical alloying process can be divided into four stages: (1) initial stage, (2) intermediate stage, (3) final stage, and (4) completion stage [44].

At the initial stage of ball milling, the powder particles are flattened by the compressive forces caused by the impact of the balls. Microforging leads to changes in the shapes of individual particles, or clusters of particles being repeatedly impacted by the balls with high kinetic energy. However, such deformation of the powders shows no net change in mass.

At the intermediate stage of the mechanical alloying process, a significant change occurs as compared to the initial stage. Cold welding becomes significant. The intimate mixture of the powder constituents decreases the diffusion distance to the micrometer range. Fracturing and cold welding are the dominant milling processes at this stage. Although some dissolution may take place, the chemical composition of the alloyed powder is still not homogeneous.

At the final stage of the mechanical alloying process, more refinement and reduction in particle size becomes evident. The microstructure of the particle also appears to be more homogeneous in microscopic scale than those at the initial and intermediate stages. True alloys may have already been formed.

At the completion stage of the mechanical alloying process, the powder particles possess an extremely deformed metastable structure. At this stage, the lamellae are no longer resolvable by optical microscopy. Further mechanical alloying beyond this stage cannot physically improve the dispersoid distribution. Real alloy with a composition similar to the starting constituents is thus formed [44].

MA in high-energy ball milling equipment is accomplished by processing an initial powder charge usually comprising a mixture of elemental, ceramic (e.g., yttria for ODS alloys), and master alloy powders, all supplied in strictly controlled size ranges. Master alloy powders are used in order to reduce in situ oxidation of highly reactive species, such as aluminum or titanium alloy additions during processing. The milling medium normally used in commercial systems is a charge of hardened steel balls, typically 2cm in diameter. The ball-to-powder weight ratio is chosen carefully for each mill and powder charge combination, but is typically around 10:1 for commercial systems. Given the enormous surface area, both of the initial powders and the fresh powder surfaces generated during MA processing, control of the milling atmosphere and its purity is essential to avoid undue alloy contamination. The principal protective atmospheres employed during commercial milling of MA powders are usually either argon or hydrogen and this protection generally extends both to pre- and post-MA powder handling. Both the purity of these gas atmospheres and the integrity of gas seals on the milling equipment are essential to control contamination, particularly when processing reactive species. For example, levels of oxide contamination in Ni3Al can double with just a few hours of milling in impure argon. Occasionally, however, deliberate doping of the milling environment has been used to facilitate alloying additions during processing.

The central event during MA is the ballpowderball collision within the milling medium during processing. It is repetition of these high-energy collisions which leads eventually to MA of the powder charge. Intimate mixing and eventual MA of the powder charge occurs in a series of identifiable, more or less discrete stages during processing (e.g., Gilman and Benjamin 1983). For ductileductile or ductilebrittle combinations of starting powders, MA initially proceeds by the flattening and work hardening of ductile powders and fragmenting of brittle constituents, which is followed by extensive cold welding between powder particles, formation of lamellar structures, and coarsening of the powder particle size distribution. Brittle powder fragments are trapped at cold weld interfaces between the evolving lamellas of the ductile constituents and thus, while continuing to comminute, become dispersed. With continued milling a balance, which is dependent on processing parameters and the composition of the constituents, is established between further cold welding and powder particle fracture, leading to relatively stable powder particle sizes.

This balance between welding and fracture is accompanied both by further decreases in lamella spacings and by folding and mixing-in of lamella fragments to produce microstructures typical of MA (Fig. 1). For ODS alloys, powder constituents are milled to the stage where light microscopy examination reveals that lamella spacings have decreased to below the resolution limit (1m). For typical levels of oxide addition (e.g., 0.5wt.% yttria) this criterion ensures average dispersoid interparticle spacings of <0.5m (Fig. 2). In other systems, milling can progress until true alloying occurs. Surprisingly, MA can also be achieved between essentially brittle powder constituents. The mechanisms by which this occurs are less well understood than in systems incorporating at least one ductile powder component. Nevertheless, granular as opposed to interlamellar mixtures of brittle powder constituents do evolve, typically with smaller, harder fragments progressively incorporated to a very fine scale within the less hard constituents, e.g., aluminanickel oxide. Moreover, MA of these brittle constituents can progress to true alloying, as has been demonstrated using lattice parameter measurements on Si28 at.% Ge progressively milled from constituent powders (Davis and Koch 1987).

Figure 1. Polished and etched metallographic section of ODM 751 FeCrAl alloy powders in the fully MA condition, showing the folded lamellar structures typical of material processed by high-energy ball milling (courtesy of D.M. Jaeger).

Figure 2. Transmission electron microscope image showing alignment of a fine-scale dispersion of oxide particles in extruded ODS alloy PM2000. The arrow shows the extrusion direction (courtesy of Y.L. Chen).

Milling of very ductile metals such as aluminum and tin has to be carefully controlled to avoid complete agglomeration of the ductile phase rather than the balance between cold welding and fracture that leads to MA. This is normally achieved by adding precise amounts of organic compounds termed process control agents (PCAs) to the milling environment. Typically waxes or solvents, these compounds that interfere with cold welding progressively break down during milling to become incorporated within the final MA powders (e.g., in aluminum alloys) as fine-scale distributions of carbides or oxides. Similar restrictions to the proclivity for cold welding in ductile powders can be achieved without use of PCAs by milling at low temperatures, e.g., below 100C for aluminum.

The processing equipment used to effect MA by high-energy ball milling of powders originated in mining and conventional powder metallurgy industries. The range of high-energy ball milling equipment divides, approximately, into two categories: small, high-energy laboratory devices, and larger facilities capable of milling commercial quantities of powder. The former category includes SPEX shaker mills and planetary ball mills. Both devices are capable of rapidly effecting MA, but in quantities of powders up to no more than a few tens of grams. SPEX mills vibrate at up to 1200rpm in three orthogonal directions to achieve ball velocities approaching 5ms1. Planetary mills incorporate a rotating base plate upon which are mounted counter-rotating, smaller-radius vials containing the ball/powder charge. The kinetic energy imparted to the ball charge in the planetary mill depends on the base plate and vial radii and angular velocities. Attritor or Szigvari ball mills, depending on their size, can be used either for laboratory or commercial ball milling applications and incorporate a rotating vertical shaft with attached horizontal impellors which stirs a container housing the ball and powder charge. These devices can process batches of up to several kilograms or more of powder through the significant differential movement the impellors generate between the ball and powder charge. Balls can either cascade or tumble when leaving the mill wall during attritor processing, depending on the ball charge and impellor velocity.

The largest commercial devices applied to MA are horizontal ball mills. When these devices exceed several meters in diameter they impart sufficient kinetic energy through ball impacts to effect MA and can process over 1000kg of powder per batch in larger units. Balls either cascade or tumble during processing in these mills depending on rotational speed (see Fig. 3). The time taken to achieve MA scales approximately inversely to the size of the milling equipment used. Hence, milling which might take minutes to accomplish in a SPEX mill could take hours in an attritor or days in a horizontal ball mill. All of these processing routes, however, have very low energy conversion efficiency, in that only a small fraction of the milling energy expended effects microstructural change contributing to the MA process.

Figure 3. Configuration of a horizontal ball mill, showing the release of the powder and ball charge (at angular position ) from the inner wall of the mill rotating with angular velocity (after Lu et al. 1995).

It is worth noting that during MA, powder particles also coat (condition) the ball milling medium, which means that, to avoid cross-contamination of commercial alloys, the repeat use of ball charges is restricted to compositionally similar batches of raw materials.

Mechanical means, such as high-energy ball milling, ultrasonic or jet milling, and others, can have powder prepared into nanoparticles. This is an example of a top-down approach, which is suitable for refractory metals or materials beyond the use of chemical reactions. The disadvantages include the difficulties in classification according to the particle size and serious surface contamination.

Bombarding a metal surface with high-energy balls makes it possible to turn the surface structure into nanoscale; this can improve the abrasion and corrosion resistance of the processed material. Meanwhile, the surface is identical to the bulk material, and thus it does not peel off like nanocoating material. The main mechanism of this method is to produce a large number of defects and dislocations, which further develop into dislocation walls, and thus cut the large crystals into nanocrystalline grains (Figure 5.11).

MCP is normally a dry, high-energy ball milling technique and has been employed to produce a variety of commercially useful and scientifically interesting materials. The formation of an amorphous phase by mechanical grinding of a Y-Co intermetallic compound in 1981 (Ermakov et al., 1981) and its formation in the Ni-Nb system by ball milling of blended elemental powder mixtures (Koch et al., 1983) brought about the recognition that this technique is a potential non-equilibrium processing technique. Beginning in the mid-1980s, a number of investigations have been carried out to synthesize a variety of equilibrium and non-equilibrium phases including supersaturated solid solutions, crystalline and quasicrystalline intermediate phases, and amorphous alloys. Additionally, it has been recognized that powder mixtures can be mechanically activated to induce chemical reactions, at room temperature or at least at much lower temperatures than normally required, to produce pure metals, nanocomposites and a variety of commercially useful materials. Efforts have also been under way since the early 1990s to understand the process fundamentals of MA through modeling studies. Because of all these special attributes, this simple but effective processing technique has been applied to metals, ceramics, polymers and composite materials. The attributes of mechanochemical processing are listed below. However, in the present chapter, the focus will be on the synthesis of nanocrystalline metal particles.

Inducement of chemical (displacement) reactions at low temperatures for (a) Mineral and Waste processing, (b) Metals refining, (c) Combustion reactions, and (d) Production of discrete ultrafine particles

Nanocrystalline materials are single- or multi-phase polycrystalline solids with a grain size of the order of a few nanometers (1nm=109m=10), typically 1100nm in at least one dimension. Since the grain sizes are so small, a significant volume of the microstructure in nanocrystalline materials is composed of interfaces, mainly grain boundaries. That is, a large volume fraction of the atoms resides in the grain boundaries. Consequently, nanocrystalline materials exhibit properties that are significantly different from, and often an improvement on, their conventional coarse-grained polycrystalline counterparts. Compared to the material with a more conventional grain size, that is, larger than a few micrometers, nanocrystalline materials show increased strength, high hardness, extremely high diffusion rates and consequently reduced sintering times for powder compaction, and improved deformation characteristics. Several excellent reviews are available giving details on different aspects of processings, properties, and applications of these materials (Gleiter, 1989; Suryanarayana, 1995a, 2005).

milling energy - an overview | sciencedirect topics

milling energy - an overview | sciencedirect topics

Milling media possess significant influencing performance on the ball-milling process in terms of the milling energy and size of the end-product. The sort of materials used for manufacturing of the milling media (balls and vials) is important and crucial due to the impact of the milling balls on the inner walls of the milling vials. Selection of the milling media materials including shapes and sizes depends upon several factors, some of them are interrelated. In general, when a material is to be ball-milled there are certain critical factors, which have to be taken into account. In general, the milling media should fulfill two major requirements: (i) they should have large surface area to provide suitable contact with the material being milled; and (ii) they should be as heavy as possible to have sufficient energy required for size reduction of the powder particles. Some of these principal factors that govern the selections of the milling media are discussed below.

Hardnessthe hardness of a powder material is considered to be the most important characteristic to realize when deciding on what type of milling media to choose. The harder the milling media the better the milling efficiency. Using milling media made of hard materials such as hardened steel, tungsten carbide, agate, and zirconia leads to maximize the milling efficiency and thus minimize the milling time required to get fine and homogenous powder particles. However, using such hard milling media may lead also to contaminate the milled powders with foreign materials that come from the balls wear during the milling process. Table 3.2 presents the Vickers hardness values of selected materials used for manufacturing of the milling media.

Specific gravitythe specific gravity of the milling media also plays a very important role in the milling process. It is well known that balls with high density and large diameters would give excellent results due to the expected generation of high impact forces applied on the milled powders. In general, the balls should enjoy more dense values when compared with the ball-milled powders. The densities values of some selected materials used for producing the milling media are listed in Table 3.2.

Brittlenessanother important factor that governs the selection of the milling media is the brittleness of the milled powders. Brittleness is the degree to which a material will easily break. Almost all of ceramics, including metal oxides and metal carbides are brittle. Breaking down the brittle materials can be performed successfully by selection of proper type of the milling media based on their properties that are presented in Table 3.2. In general, materials that are enjoying ductility such as metals and some metal alloys cannot easily be ball-milled.

Balls sizesthe efficiency of ball milling depends on the surface area of the milling media that are used in the process. The milling process is affected by number of contact points between the balls and the powder particles. Hence, the angle of nip presents a very important factor so that the ball sizes must be carefully chosen in relation to the largest and hardest particles of the feed powders. In general, the capacity of a ball mill increases by decreasing the ball diameters. Moreover, the rate of ball-to-ball contacts per unit time increases with decreasing the ball diameters because the number of balls in the mill increases. It is predicted that using different ball sizes can lead to higher collision energy that will help on improving of the milling process [9]. Thus, balls should be as small as possible and the charge the balls should be graded such that the largest balls are just heavy enough to grind the largest and hardest particles in the feed, whereas the small balls are responsible for powder refining.

The dynamics of the balls of different sizes during attrition of Cu-15 vol.% Nb powders have been studied by Cook and Courtney, using cinematographic techniques [10]. In their experiments, an attritor with a clear tank was specially designed to study media dynamics. A high-speed video camera was used to record the motion of the balls with different diameters inside the attrition, while its impeller rotated at different speeds. The metallographic examinations of the samples obtained after ball milling for 216h using balls with 7.64-mm-diameter and a mixture of 7.64-mm-diameter and 4.76-mm balls are shown in Figures 3.6 and 3.7, respectively [10]. Obviously, after 16h of ball milling using a mixture of 7.64-mm-diameter and 4.76-mm balls (Figure 3.7(c)), the powder has finer particle size in comparison to that of Figure 3.6(c) indicates that a greater particle fracture rate can be obtained when milling with differently sized balls. Thus, relative particle fracture rates are increased during the fracture-dominant stage when differently sized balls were employed [10].

Recently, Vaezi et al. [11] have reached to the same conclusion but with a different binary system (Cu-50% Fe) when they ball-milled the starting material powders with high-energy planetary type, using different ball sizes. They conclude that using a mixture of 5- and 10-mm balls leads to enhance the milling kinetics.

Shape of the milling mediaas discussed in Chapter 2, the shape of the milling media can be balls, rods, or barrels. In general, impact, shear stresses, and their combination (impact+shear stresses) leads to the extreme mechanisms in any ball-milling or rod-milling systems. In the tumbling and ball mills and roller mills, use the rotation and friction of the vial shell to transfer energy to the milling media. The energy that is required to break the particles in the mill comes from the rotational energy that is supplied by the rotator drums, which are directly connected to the drive motor. Thus, the internal frictional forces and ball-powder-ball collisions lead the media to rise before gravity forces make them fall. Based on the motion of the milling media (balls), the ball mill can be classified into three zones: (i) shearing zone, where the balls are lifted due to the rotated vial action and frictional forces, (ii) cataracting zone, where the balls are falling due to the gravity, and (iii) inactive zone, where the balls are not moving (Figure 3.8). Each impact event is considered to deliver a finite amount of energy to the charge which in turn is distributed unequally to each particle that is in the neighborhood of the impacting media particles and which can therefore receive a fraction of the energy that is dissipated in the impact event.

However, balls have a greater surface area per unit weight than rods that make them better suited for fine particle-size reduction; rods have several advantages, which can be summarized as follows:Milling rods do not require cascading as do ball charges, thus enabling rod mills to be operated at lower peripheral speeds than ball mills;The existing void spaces in rod charge is less when compared with the ball charge;Rods offer great milling contact between the powder charges per surface;The action within the rods causes the energy of rods to be directed to the largest-sized particles of the powders;Since there is no collision between the rods during the rod-milling process, the volume fractions of contamination that are usually introduced by the milling tools to the powders are less.

In the recent years, a different type of milling medium called Cylpebs, which was developed for different shapes, has appeared in the market (developed by Doering International, Powerpebs by the Donhad and Millpebs by the Wheelabrator Allevard Enterprise are available in the market). These milling media have barrel- or cylindrical-like shapes of length equaling diameter, and all the edges being radiuses (Figure 3.9). It is claimed by the manufacturer (Doering International; www.donhad.com.au) that for a given charge volume, the bulk density of Cylpebs is 9% greater than steel balls, 12% greater than cast-balls. Moreover, the surface area of Cylpebs is 14.5 % greater than balls of equal weight. The density and surface area combination, deliver 25% greater grinding capacity in the mill charge. The grinding performance of the Cylpebs should then be correspondingly higher compared with the steel balls [12]. Because of their cylindrical geometry, Cylpebs have greater surface area and higher bulk density compared with balls of similar mass and size, as presented in Table 3.3. The milling mechanism using Cylpebs is considered to be a combination between ball milling and rod milling. Beside the point contact action similar to the balls, Cylpebs have other milling actions resulting from line contact along the cylindrical section and area contact between the end faces on the powder particles, being similar to the mechanism of rods. The line contact and area contact increase the tendency for milling to take place preferentially on the larger powder particles, as schematically shown in Figure 3.10. Once the large particles are caught on the line or between the face areas, this prevents the smaller particles from being broken further, which is similar to the rod mill practice [12].

Figure 3.9. Doering Cylpebs are slightly tapered cylindrical grinding media with length equaling diameter, and all the edges being radiuses. The largest Cylpebs available in the market are of the size 85mm85mm, and the smallest ones are 8mm8mm. Because of their geometry, Cylpebs have greater surface area than balls of the same mass.

It is worth mentioning here that the milling vial should not be completely filled with the powder and milling media charges. There should be enough space that allows the milling media to move around freely in the vial that provides alloying and particle size reduction. Suryanarayana [13] has suggested that 50% of the vial volume should be left free, as schematically presented in Figure 3.11.

Figure 3.11. Schematic representation of the cross-sectional view for a milling vial filled with powders charge and balls. The volume of the total charge (powders+balls) should be in the range between 40% and 50% of the total volume of the vial.

Since the dmin attained in a metal during milling is expected to depend on its mechanical properties, it is suspected that neither the nature of the mill nor the milling energy will have any effect on the minimum grain size achieved. It was reported (Galdeano et al., 2001) that there was no significant effect of milling intensity on nanostructure formation in a Cu-Fe-Co powder blend. But, in another investigation, it was shown that dmin was about 5nm when the TiNi intermetallic powder was milled in a high-energy SPEX shaker mill, but only about 15nm in the less-energetic Invicta vibratory ball mill (Yamada and Koch, 1993). Similar trends were noted for variations in the ball-to-powder weight ratios (BPR). The grain size of the niobium metal milled in the Invicta vibratory mill was about 262nm at a BPR of 5:1, but was only 181 at a BPR of 10:1 (Koch, 1993). Similar results were also reported for the pure metal Cu: 255nm for a BPR of 5:1 and 201 for a BPR of 10:1. Further, the kinetics of achieving this dmin value could also depend on the milling energy, although no such studies have been reported so far.

It was also reported that during nanocrystal formation, the average crystal size increased and the internal lattice strain decreased at higher milling intensities owing to the enhanced thermal effects (Kuhrt et al., 1993). In accordance with this argument, the grain size of Si milled at a high energy of 500kJg1 was 25nm, while that milled at a low energy of 20kJ g1 was only 4nm (Streleski et al., 2002).

The aim of this project is to affect white root fungi by genetic or other means in order that they may eliminate part of the lignin out of wood chips while the chips are still stored. The biological pulp obtained in this manner can then be submitted to the normal mechanical processing or boiling. The consumption of milling energy as well as chemicals is hereby expected to be less than usual.

New fungi suitable for biological pulp production will be developed and tested with regard to int. al. optimum growth conditions and ability to delignify and defiber various types of wood. Studies will include tests concerning controlled fiber separation as well as microscopic examinations.

As the name suggests, the ball milling method consists of balls and a mill chamber. Therefore, a ball mill contains a stainless steel container and many small iron, hardened steel, silicon carbide, or tungsten carbide balls are made to rotate inside a mill. The powder of a material is taken inside the steel container. This powder will be made into nanosize using the ball-milling technique. A magnet is placed outside the container to provide the pulling force to the material and this magnetic force increases the milling energy when milling container or chamber rotates the metal balls. Ball milling is a mechanical process and thus all the structural and chemical changes are produced by mechanical energy.100 Baek etal.101 recently proposed that edge-selectively functionalized graphene nanoplatelets (EFGnPs) as metal-free electrocatalysts for ORR can be large-scaled prepared by ball-milling method. The EFGnPs were obtained simply by dry ball-milling graphite in the presence of hydrogen, carbon dioxide, sulfur trioxide, or carbon dioxide/sulfur trioxide mixture. The resultant sulfonic acid- (SGnP) and carboxylic acid/sulfonic acid- (CSGnP) functionalized GnPs were found to show a superior ORR performance to commercially available platinum-based electrocatalyst in an alkaline electrolyte. It was also found that the edge polar nature of the newly prepared EFGnPs without heteroatom doping into their basal plane played an important role in regulating the ORR efficiency.

Mechanochemical synthesis involves high-energy milling techniques and is generally carried out under controlled atmospheres. Nanocomposite powders of oxide, nonoxide, and mixed oxide/nonoxide materials can be prepared using this method. The major drawbacks of this synthesis method are: (1) discrete nanoparticles in the finest size range cannot be prepared; and (2) contamination of the product by the milling media.

More or less any ceramic composite powder can be synthesized by mechanical mixing of the constituent phases. The main factors that determine the properties of the resultant nanocomposite products are the type of raw materials, purity, the particle size, size distribution, and degree of agglomeration. Maintaining purity of the powders is essential for avoiding the formation of a secondary phase during sintering. Wet ball or attrition milling techniques can be used for the synthesis of homogeneous powder mixture. Al2O3/SiC composites are widely prepared by this conventional powder mixing route by using ball milling [70]. However, the disadvantage in the milling step is that it may induce certain pollution derived from the milling media.

In this mechanical method of production of nanomaterials, which works on the principle of impact, the size reduction is achieved through the impact caused when the balls drop from the top of the chamber containing the source material.

A ball mill consists of a hollow cylindrical chamber (Fig. 6.2) which rotates about a horizontal axis, and the chamber is partially filled with small balls made of steel, tungsten carbide, zirconia, agate, alumina, or silicon nitride having diameter generally 10mm. The inner surface area of the chamber is lined with an abrasion-resistant material like manganese, steel, or rubber. The magnet, placed outside the chamber, provides the pulling force to the grinding material, and by changing the magnetic force, the milling energy can be varied as desired. The ball milling process is carried out for approximately 100150h to obtain uniform-sized fine powder. In high-energy ball milling, vacuum or a specific gaseous atmosphere is maintained inside the chamber. High-energy mills are classified into attrition ball mills, planetary ball mills, vibrating ball mills, and low-energy tumbling mills. In high-energy ball milling, formation of ceramic nano-reinforcement by in situ reaction is possible.

It is an inexpensive and easy process which enables industrial scale productivity. As grinding is done in a closed chamber, dust, or contamination from the surroundings is avoided. This technique can be used to prepare dry as well as wet nanopowders. Composition of the grinding material can be varied as desired. Even though this method has several advantages, there are some disadvantages. The major disadvantage is that the shape of the produced nanoparticles is not regular. Moreover, energy consumption is relatively high, which reduces the production efficiency. This technique is suitable for the fabrication of several nanocomposites, which include Co- and Cu-based nanomaterials, Ni-NiO nanocomposites, and nanocomposites of Ti,C [71].

Mechanical pretreatment refers to chipping, milling, and grinding of lignocellulose and is almost always the first step in all lignocellulose processes. Size reduction of biomass for easier processing and transport is the simplest form of mechanical pretreatment. Size reduction also improves the available surface area of lignocellulose that comes in contact with reactants in the process. Knife mills, hammer mill, and disc mill are some common types of size reduction instruments. The size of feedstock after chipping is in the range of 1030mm, and it can be reduced to 26mm by milling and grinding. The relationship between final size and milling energy is not linear, and further reduction of size is energetically demanding (51). Type of feedstock, moisture content, and starting milling size also influence the energy demand. Knife and hammer mills are more economical with energy consumption between 1 and 130kWht1 in comparison to disc and ball mills (52).

Extensive milling of lignocellulose also alters its physical structure and enhances its reactivity (53,54). Cellulose undergoes the most extensive change as milling disrupts the long-range ordered structure and reduces its crystallinity, making it more susceptible to chemical attacks (55). Amorphous cellulose produced by milling treatment undergoes hydrolysis at a lower temperature and exhibits higher rate of hydrolysis (56). Milling also makes lignin and hemicellulose susceptible to dissolution to facilitate their separation. The simplicity of mechanical pretreatment is attractive as it does not use solvents or corrosive chemicals. However, the high-energy requirement for milling is a major limiting factor for large-scale applications.

Mechanocatalysis, involving combined milling of lignocellulose and catalysts, is an attractive method to increase the solubility and reactivity. Combining cellulose with solid acid or small amount of mineral acid catalysts and milling it in a ball mill produced water-soluble oligomers that can be easily hydrolyzed to glucose under mild condition (5759). Acid catalyst directly depolymerized cellulose to yield soluble oligomers. NMR analysis revealed that repolymerization of fragments occurred to form -1,6 branched oligomers (59) (Fig. 7). These branched oligomers showed high reactivity for hydrolysis and hydrolytic hydrogenation reactions.

Fig. 7. Structure of branched oligomers formed by mechanocatalytic depolymerization of cellulose (A) and (B) 1H NMR of the anomeric region of branched oligomers showing cellulosic -1,4 linkages and newly formed -1,6 linkages along with and reducing ends.

Mechanocatalysis of lignocellulose produces a composite of depolymerized material that can be reacted directly to produce 5-hydroxymethylfurfural and furfural (60). Deep depolymerization of lignocellulose produces a water-soluble composite that yields sugars after hydrolysis at mild reaction conditions. The residual lignin is obtained as solid material after hydrolysis and can be recovered by simple filtration (61). Depolymerization of lignocellulose is catalyzed by the presence of acids, and influence of radicals was negligible as the presence of lignin, a radical scavenger, did not reduce the rate of depolymerization (62). It can be argued that mechanocatalysis is not merely pretreatment method as the original lignocellulose structure is chemically transformed to a large extent. Energy requirement for mechanocatalysis is analogous to mechanical pretreatment. However, the advantage of deep depolymerization reduces the cost for subsequent processing of lignocellulose. Energy requirement at gram-scale operation is 200MWht1 that can be reduced to 9.6MWht1 at kilogram scale (63). Therefore, mechanocatalysis can be feasible at large-scale operation for lignocellulose depolymerization.

The temperature rise during milling is mainly due to ball-to-ball, ball-to-powder, and ball-to-wall collisions as well as frictional effects. The overall temperature rise of the powders during milling can be due to more than one cause. First, the intense mechanical deformation due to kinetic energy of the grinding media (balls) raises the temperature of the powder. Thus, the higher the energy (milling speed, relative velocity of the balls, time of milling, size of the grinding balls, ball-to-powder weight ratio, etc.), the higher the temperature rise. Second, it is possible that exothermic processes occurring during the milling process cause the powder particles to ignite and generate additional heat [13].

It has been reported that >90% (different values have been quoted by different investigators) of the mechanical energy imparted to the powders during milling is transformed into heat [14], raising the temperature of the powder.

Bhattacharya and Arzt [15] calculated the contact temperature of the powder compact surfaces, considering that head-on collisions occur during collisions between two grinding media. Therefore, the calculated value will be the upper bound of the temperature rise. In arriving at these values, the authors have assumed that the time of impact, , could be approximated using Hertz's theory of elastic impact [16,17], and the energy flux at the compact surface is uniform over the entire contact area and is constant for times less than . They have also assumed in their calculations only a small fraction, , of the kinetic energy

where m is the mass and v is the relative velocity of the balls, is utilized in plastically deforming the powder, that is, the expended plastic energy, Up=E. By taking the appropriate heat transfer conditions into consideration, the authors [15] calculated the temperature rise, T, as

and Q represents the average quantity of heat arising from the total deformation process over a time interval , is the heat yield transfer into the powder compact, is the thermal conductivity of the powder, t0 is the thickness of the powder sheet between two colliding balls at a moment of maximum impacting force (Fig. 2.7), r0 is the radius of the powder compact, and is the thermal diffusivity.

Such as, assuming that niobium powder is milled using stainless steel balls and that the powder compact has a thickness of t0=100m and a radius r0=263m, the maximal temperature rise T reached at the time of contact surface at a moment of maximum impacting force calculated according to [15] at =0.03 by v=6m/s and 8m/s is 496 and 669K, respectively, and at =0.09 by v=6m/s and 8m/s is 941 and 1460K, respectively.

The temperature rises calculated and estimated above using theoretical models on the basis of heat build up owing to kinetic energy of the collisions of balls [15] or microstructural changes [18] are not accurate or reliable due to the assumptions made, some of which may not be substantiated. Therefore, only experimental investigations may give reasonable and accurate values for the rise in temperature. However, experimental measurements are not easy and they measure only the bulk temperature of the wall of a mill, not the instantaneous temperature at the time of impact.

Some investigators have reported very large temperature rises. The experimentally measured bulk temperature rise in different alloy systems and milling conditions is in the range 323489K [17], and more commonly it is about 373390K. A maximal temperature of 373488K was recorded [19] by milling Ni-base superalloy in an attritor. It should be realized that this is the macroscopic temperature rise, even though it is recognized that local (microscopic) temperatures can be very high, often exceeding the melting points of some component metals [20].

Example equilibrium phase diagrams and calculated Gibbs energy curves for the systems AlMn, CrCo, CuFe, FeNi, TaAl and AlTi are presented in Fig. II.20.3, Fig. II.20.4, Fig. II.20.5, Fig. II.20.6, Fig. II.20.7 and Fig. II.20.8 respectively. All Gibbs energy curves have been calculated, using assessed thermodynamic data from the SGTE Solution Database [002SGT], for a temperature of 200 C a typical ball-milling temperature. For each system, the point of intersection of the solvent-phase Gibbs energy curve with the curve for the primary precipitating phase defines the calculated metastable solubility limit (denoted in the diagrams by a dashed line to the composition axis).

II.20.3. The calculated equilibrium phase diagram for the AlMn system. (b) Calculated Gibbs energy curves for the fcc, cubic A13 and complex body-centred cubic (cbcc) phases of the AlMn system at 200 C with component Gibbs energies changed by 5000 J mol1 (stoichiometric compounds and the Al8 Mn5 D810 phase, with complex crystallography, are omitted).

II.20.4. (a) The calculated equilibrium phase diagram for the CrCo system. (b) Calculated Gibbs energy curves for the body-centred cubic (bcc), fcc and hexagonal close-packed (hcp) phases of the CrCo system at 200 C with component Gibbs energies changed by 5000 J mol1 (the phase, with complex crystallography, is omitted).

II.20.5. (a) The calculated equilibrium phase diagram for the CuFe system. (b) Calculated Gibbs energy curves for the fcc and bcc phases for the CuFe system at 200 C with component Gibbs energies changed by 5000 J mol1.

II.20.6. (a) The calculated equilibrium phase diagram for the FeNi system. (b) Calculated Gibbs energy curves for the fcc and bcc phases of the FeNi system at 200 C with component Gibbs energies changed by 5000 J mol1.

II.20.7. (a) The calculated equilibrium phase diagram for the TaAl system. (b) Calculated Gibbs energy curves for the bcc and fcc phases of the TaAl system at 200 C with component Gibbs energies changed by 5000 J mol1 (stoichiometric compounds and the phase, with complex crystallography, are omitted).

II.20.8. (a) The calculated equilibrium phase diagram for the AlTi system. (b) Calculated Gibbs energy curves for the fcc, bcc and hcp phases of the AlTi system at 200 C with component Gibbs energies changed by 5000 J mol1 (stoichiometric compounds and phases with complex crystallography are omitted).

A summary of experimental and calculated results for the above systems is presented in Table II.20.1. It can be seen from this table that results from experimental studies vary widely, although observed solubilities are in all cases significantly greater than the equilibrium values. It is likely that the various parameters associated with the different experimental ball-milling processes are in part responsible for the differing results. The selected constant milling energy (5000 J mol1) used to amend the Gibbs energies of the different powder components may, therefore, also not be a suitable value in all cases. Nevertheless, the general agreement between experimental and calculated metastable solubility boundaries, using the calculation principles listed above, is surprisingly good. An example is the CrCo system, in which the equilibrium solubility of Co in Cr is 4.9 at.% at 600 C. After mechanical alloying, the solubility from experimental measurements [001Sur] increases very significantly to 30 or 40 at.% according to two different studies. The metastable solubility limit from thermodynamic calculation, using assessed data for the system [002SGT] is 35 at.%.

Table II.20.1. Comparison of predicted and observed solid solubilities (RT, room temperature). Equilibrium solubility values and precipitated phases are taken from the Pauling File [002Vil] and solubilities after mechanical alloying from the work by Suryanarayana [001Sur]

A difficulty in making a complete comparison of calculated solubilities with the experimental results is that workers have tended to place emphasis on the observed extended solubilities and in nearly all cases provide very little information on the phases precipitating from the solvent solution. Nevertheless, in many systems and experiments, the extent of the experimentally observed solubilities is such that the compositions of intermetallic compound phases observed in the equilibrium diagram are exceeded, which supports the proposition that phases with more complex crystallographic structure are difficult to produce when starting from the pure powder components in mechanical alloying. Examples are the AlMn system for which Mn solubilities greater than the Mn concentrations of the phases Al12Mn and Al6Mn have been measured, the AlTi system, for which Ti solubilities beyond the 25 and 33 at.% Ti compositions of the phases Al3Ti and Al2Ti have been reported, and the TaAl system, with Al solubilities greater than the 2739 at.% Al range of the phase. There are some systems which, with careful systematic experimentation, could provide a sensitive test of the calculation principles used here. For example, in the CuSi system, in the Cu-rich range, not only are a number of stoichiometric compound phases found, but also a bcc and an hcp phase in addition to the fcc Cu-rich solid solution (Fig. II.20.9).

The Gibbs energies of the single-phase fcc, hcp and bcc structures have very similar values as shown by the calculated curves for a temperature of 200 C, presented in Fig. II.20.10. In this plot, no contribution to the Gibbs energies of the pure components has been made. It can be seen that the Gibbs energies of the fcc and hcp phases have very similar values and that, if no compound phase precipitates to form a two-phase structure in the alloyed material, then the hcp phase can be expected to form at compositions with xSi 0.085 up to a composition with xSi 0.175, when the bcc phase becomes stable.

At the still higher temperature of 800 C, the points of intersection of the hcp and bcc curves with the fcc curve are close to being superimposed (Fig. II.20.12). This is consistent with the temperature of the fcchcpbcc three-phase equilibrium in Fig. II.20.9. The phase formed from the fcc phase at 800 C could be hcp or bcc, depending on just very small energy changes.

All the above figures, and in particular Fig. II.20.10, Fig. II.20.11 and II.20.12, clearly demonstrate that the amount of energy imparted by the alloying process, as well as the temperature achieved during milling of the powdered components, can have a significant influence on the Gibbs energies of the potentially forming phases. The relative values of these energies will, in turn, determine which of the phases form under given conditions.

The milling process is a relatively simple route to produce ensembles of MNPs. In some cases it has been used to produce new materials from different starting compounds (mechanical alloying) and has been frequently used to produce metallic granular alloys, in which the main aim is to create ensembles of MNPs embedded in a diamagnetic matrix (M-m; M=magnetic metal, and m=nonmagnetic metal) which can give rise to interesting properties such as giant magnetoresistance, but we will not elaborate on this. The main issues can be consulted in Refs. [13, 14].

We are also not presenting here the metallurgical origin of the production of alloys by milling. The main factors for mechanical alloying have been reviewed in detail in Refs. [15, 16] and are discussed as a combination of flattening, cold welding, fracture and rewelding. These processes are influenced by the ductile/fragile nature of the components [17].

If we were to underline the major advantages of using mechanical milling, we can identify: (i) the use of simple and easily affordable equipment, (ii) large variety in the production of MNPs, (iii) production of enormous quantities of MNPs, (iv) controllable particle sizes and reasonable distribution of particle sizes, and (v) in metals, control of lattice strain. The first point is supported by the low cost of the miller in comparison to other methods, although the container may become expensive and worn out with time. Point (ii) is clearly another advantage, the variety of alloys, and compounds to be produced is extremely wide [16] and different compositions are easy to be explored. The third point is important because frequently other routes that is: chemical multilayers are sometimes restricted to small reaction yields and nanometric thicknesses, respectively. In consequence, the quality of the sample may be better but the production of large quantities may render a successful technological potential via the industrial rescaling of the milling process. The next advantage is related to the fact that the production of MNPs with a determined average size (D) is easy to achieve and, for example, in RE alloy MNPs, this is achieved with low milling times. The distribution of particle sizes is not very wide although it cannot reach the narrowness obtained by some chemical routes. The reproducibility is quite high and particle sizes of between 5 and 50nm are very easy to reach. The last point (v) is related to the milling process in metals. A measurable increase of lattice strain, concomitant with the particle size decrease is always observed after milling. This possibility of promoting an increase of atomic disorder may be useful to show alterations in the magnetic intraparticle and interparticle coupling [18]. Now we will describe the mechanical milling process and the steps and parameters to be taken into account.

The mechanical milling process is a high-energy impact process, which can be performed in different mills, typically in planetary and shaker mills, with the use of balls within containers. Planetary mills consist of a number of cylindrical containers sitting on a spinning platform (see Fig. 1.6). The planetary movements involve both the horizontal rotation around the center of the base and that around the container axis. The shaker provides the milling energy thanks to the oscillating movements of a container at high velocity. In general, the planetary mill gives more flexibility for the production of samples and more quantity, whereas shaker mills are more efficient.

There are several parameters affecting the production of MNPs by mechanical milling and they are mostly based on empirical results rather than established theories and calculations. The use of a particular kind of mill already introduces particular milling parameters, which need to be taken into account. However, the common variables to control are: (i) the material of the container, (ii) ball-to-powder weight ratio, (iii) milling speed/frequency, and (iv) the milling time. These parameters are intermixed so a combination of selected values is what finally guarantees success. There are other parameters that will also require a particular comment: milling atmosphere, temperature, and eventual milling medium.

The material (and volume) of the containers must be selected for the specific mill. In planetary mills, containers of 125mL are a good compromise between the cost and the quantity of the material to be prepared. Most common vials are made of stainless steel, tungsten carbide (CW), and zirconia (ZrO2). The advantage of using stainless steel containers is their low cost, but they can contaminate the samples with Fe or Cr, which change the magnetic properties. They are also weaker than the other two types of containers, so to achieve the reduction to the nanoscale a larger milling time (t) is expected. If possible, it is very convenient to select CW containers where the total absence of magnetic components and the hardness assure an appropriate compositional control of the resulting MNPs, with a reduced milling time. In very particular processes, it would be desirable to use the expensive ZrO2 containers. It is clear that the balls should be of the same material as the container. Both containers and balls wear with time and should be substituted after a few processes, although this depends greatly on the nature of the material that is being prepared. As a matter of fact, SEM-EDX results of the milled MNPs are key prompting for a container and ball replacement, if undesired impurities are found.

The milling velocity can also be tuned. In planetary mills, a maximum rotation speed of 200rpm is usually enough. To use always the same rotation speed, and to control the value of t, it is a sensible decision to reduce the number of variables in mechanical milling. Generally speaking the increase of velocity will tend to reduce the amount of time needed to achieve the nanocrystalline state of the MNPs.

The precise number for the ball-to-sample ratio is somehow diffuse but values around 12:1 are good and give flexibility if a portion of the sample is extracted for an intermediate state analysis (see below).

The milling time (t) can be as low as a few minutes, to >300h. It needs to be stated whether the milling is performed continuously or, on the contrary, there are intercalated stops where the sample is resting (530min, approximately). In modern planetary mills, it is possible to inverse the rotation sense after that period. Going back to the optimization of the synthesis process, it is true that the mass within the vial is strictly modified, but it can be readjusted to maintain constant the value of the ball-to-powder weight ratio. In most of the analysis, the milling time is the main control parameter. It is very likely that the nature of the alloy itself is more important than the milling time. For example, some alloys become nanometric only after a few hours of milling [19], whereas others require much longer times [18,20].

There are other less studied factors. For example, an inert atmosphere (Ar 99.99%, for instance) must be used when dealing with RE-based alloys or pure Fe, as they are oxygen avid and easily form oxides. This is easily achieved in O-ring sealed containers if we follow the precaution to fill in them with the starting crushed pellets in a glove box. An alternative could be to insert the mill in a glove box itself; this seems feasible realistically only if shaker mills (e.g., SPEX) are used. In connection with the milling environment, it is possible to use oleic acid or oleylamine during the milling. This will also reduce the eventual oxidation but, more importantly, it helps to add a surfactant layer around the particles. Naturally longer t-values expose the MNPs to a more probable surface oxidizing layer. Another factor is the temperature of the containers. The mechanical milling process involves effectively an increase in the vial (container) temperature. Depending on our needs, this can be convenient or not: the increasing temperature favors the change of structure at the nanoscale resulting in a common increase in D and (in metals especially) a decrease in strain. In consequence, temperature can be a control variable, but direct temperature control (either by a coolant or furnace) is discouraging. Precisely, our experience recommends the use of stop periods so that the containers are allowed to cool down to avoid undesired recrystallizations.

In short, the high-energy mechanical milling procedure is a powerful technique to produce nanoparticles of different natures in enormous quantities. The latter is not only an important point for industrial rescaling of the production, but also to study the magnetic arrangement at a microscopic scale. The process is simple, cost effective, and the quality of the samples is generally enough for magnetic studies.

Mechanical pretreatment of biomass aims to enhance the digestibility of biomass. Coarse size reduction, cutting, shredding, chipping, grinding, or milling are among the different mechanical methods that can be used to decrease particle size, increase accessible specific surface area, increase pore size of particles and the number of contact points, and reduce the DP or crystallinity of the cellulose, although different fractions (cellulose, hemicellulose, and lignin) will not be separate.

Commonly, starting materials are presized during harvesting or preconditioning, using methods such as shredding, forage cutting, or chipping to sizes of about 1050mm. This is the minimum pretreatment needed prior to biomass processing. The size of material can be further reduced to 0.22mm by milling or grinding through different machines: vibratory ball mills, hammers, knifes, balls, discs, colloids, and extruders.

Chipping is used to reduce heat and mass transfer limitations. Grinding and milling are more effective at reducing the particle size and cellulose crystallinity (Dumas etal., 2015). Vibratory ball milling is more effective than ordinary ball milling in reducing cellulose crystallinity. Disk milling, which produces fibers, is more efficient in enhancing cellulose hydrolysis than hammer milling. The type and duration of milling, as well as the kind of biomass, determine the digestibility of the biomass (Hideno etal., 2013; Zakaria etal., 2015).

This method has some disadvantages in terms of the energy demand. The final particle sized desired and the biomass characteristics determine the energy requirements for mechanical comminution, which is usually very high. There is a critical size below which further reduction will not affect the biomass treatment significantly (Dumas etal., 2015; Liu, 2015). Initial moisture content and biomass composition have shown to be important parameters impacting the specific energy requirement (Barakat etal., 2015). In this sense, mechanical comminution is not considered as an attractive option in biomass pretreatment.

The combination of biological or chemical treatment prior to mechanical diminution has confirmed the feasibility of reduction of the energy consumption of mechanical processes (Fougere etal., 2015; Motte etal., 2015). Studies show that milling after chemical pretreatment can significantly reduce milling energy consumption, cost of solid liquid separation, and liquid to solid ratio, and does not result in the production of fermentation inhibitors (Zhu etal., 2009).

Extrusion pretreatment is a physical pretreatment in which materials are exposed to mixing, heating, and shearing, suffering physical and chemical modification. The shear forces applied in the extrusion process serve to remove the softened surface regions, exposing the interior to chemical and/or thermal action and therefore improving the cellulose conversion (Mood etal., 2013).

Factors such as extruder temperature, screw speed, feedstock particle size, and moisture content have been investigated to determine their influence in energy requirements of the pretreatment (Karunanithy and Muthukumarappan, 2011a).

Main advantages of this method include short residence time, moderate temperature, no formation of inhibitors such as furfural or 5-hydroxymethylfurfural (HMF), no need of washing step, no solid loss, rapid mixing, feasibility of scale-up, and possibilities of continuous operation (Karunanithy and Muthukumarappan, 2011b,c).

Recent studies about alkali-combined extrusion pretreatment indicated that combined pretreatment increased the number of pores in the biomass structure, giving improved sugar yields (Zhang etal., 2012).

LHW processes are biomass pretreatments based on the use of pressure to keep water at high temperatures (160240C). It is also referred in the literature as autohydrolysis, hydrothermolysis, hydrothermal pretreatment, aqueous fractionation liquefaction or extraction, solvolysis, aquasolv, steam pretreatment, or water prehydrolysis (Mosier etal., 2005).

The reaction is initiated by the hydro ions [H3O+] generated from the dissociation of water molecules. This process changes the biomass structure, resulting in the hydrolysis of hemicellulose and removal of a small portion of the lignin, which makes the cellulose more accessible for further hydrolysis while avoiding the formation of fermentation inhibitors that occurs at higher temperatures (Vallejos etal., 2015).

It is important to maintain the pH between 4 and 7 during the pretreatment because at this pH, the dissolved hemicellulose exists mainly in oligomeric form, and the formation of monosaccharides and the subsequent degradation products that further catalyze hydrolysis of cellulosic material are minimized.

Pretreatment of some biomass feedstock can be carried out under mild conditions (140180C), but for most raw biomass, it needs to be performed at higher temperatures (up to 190230C). The water and the biomass (18%) are brought in contact up to 5min at most severe conditions. At this high temperature, sugar degradation may increase significantly. Between 40% and 60% of the total biomass can be dissolved in the process with the removal of 422% of the cellulose, 3560% of the lignin and the majority of the hemicellulose (Rogalinski etal., 2008).

The process itself is simplified. The use of lower temperatures with minimization of degradation products eliminates the need for a final washing step or neutralization because the pretreatment solvent here is water. Neither sludge handling nor acid recycling result.

The biomass glucan content is not modified. The physicochemical modification caused by treatment on lignin and cellulose facilitates the further separation of different fractions. Hemicelluloses can be converted into hemicellulosic sugars at good yields with low byproduct generation.

The main disadvantage of LWH pretreatment is related to the downstream processing. The amount of solubilized product is higher, while the concentration of these products is lower compared to other pretreatments. High energy is demanded due to the large volumes of water involved.

The recovery of hemicellulose from solution is impeded by high-lignin solubilization. A catalyst such as an acid can be added making the process similar to dilute acid pretreatment. However, degradation of sugars can result in undesirable inhibitory products. During pretreatment, the pH and pKa of water is affected by temperature, so KOH can be used to maintain the pH above 5 and below 7 to minimize the formation of monosaccharides that are degraded to fermentation inhibitors (Mosier etal., 2005).

Pyrolysis is a thermal pretreatment of lignocellulosic biomass where raw material is heated in an inert atmosphere at temperatures between 350 and 650C. It is usually employed to enhance the energy density of fuels produced from biomass. Nitrogen is the commonly used carrier gas to provide a nonoxidizing atmosphere. Torrefaction takes place at similar conditions to those of pyrolysis but at lower temperatures of 200300C, due to which, it is also called mild pyrolysis.

Cellulose decomposes rapidly to gaseous products and residual char when biomass is heated above 300C. At lower temperatures, decomposition is much slower, and resulting products are less volatile. Under severe conditions, hemicellulose is almost depleted completely, and cellulose is oxidized to a great extent. Lignin is the most difficult component to be degraded, and thus, its removal is very low under torrefaction conditions.

After thermal pretreatment, the properties of biomass are improved to a great extent. Main benefits from torrefaction are more uniform properties in the biomass, which include improved grindability and reactivity, higher energy density, lower atomic O/C and H/C ratios and moisture content, and higher hydrophobicity (Chen etal., 2015b).

Recent studies have demonstrated that torrefaction pretreatment causes mechanical disruption of biomass fibers, resulting in their size reduction as well as high solid product yield (Das and Sarmah, 2015).

Torrefaction and pyrolysis have been studied as pretreatment processes for biomass to fuel conversion. Pretreatment by torrefaction was found to be far more attractive than pyrolysis (Kumar etal., 2009).

Freeze/thaw pretreatment is a novel approach for physical pretreatment of biomass. In this process, biomass is frozen in a conventional freezer at temperature below 20C for a certain period of time (between 2 and 24h) and then immediately thawed in hot water or at room temperature (Chang etal., 2011).

Treatment with freezing/thawing could be an efficient alternative for pretreatment of lignocellulosic biomass due to significantly increasing the enzyme digestibility of substrates (Smichi etal., 2015).

Only a few studies have been carried out, but despite the high cost involved, its attractive characteristics, i.e. lower negative environmental impact, application of less dangerous chemicals, and high effectiveness, make freeze/thaw process a promising pretreatment in biomass processing.

Some authors have suggested treatments involving the use of gamma rays that give a larger surface area and lower crystallinity by cleaving the -1,4 glycosidic bonds. Gamma irradiation after sulfuric acid pretreatment on wheat straw showed a great influence on enzymatic hydrolysis, owing to disruption of cellulose crystallinity, removal of hemicelluloses, and structural modification of lignin polymers (Hong etal., 2014). This method would be very expensive on a large scale with huge environmental and safety concerns.

Microwave irradiation could be an alternative to the conventional heating in order to modify the structure of cellulose, degrade and partially remove lignin and hemicelluloses, and enhance the enzymatic susceptibility of reducing sugars. The advantages of this method include short process time, high uniformity and selectivity, and less energy input than conventional heating. Microwave-assisted pretreatment has demonstrated the improvement in enzymatic hydrolysis of corn straw and rice husk (Diaz etal., 2015).

Pulsed electric field (PEF) pretreatment is a physical pretreatment of lignocellulosic biomass that involves the application of a short burst of high voltage to a sample (biomass) situated between two electrodes. The sample can be either placed or transported between the electrodes, and the electric discharge is applied in the form of pulses. High intensity electric field produced structural changes in the cell membrane, resulting in an increase in mass permeability and mechanical rupture (Kumar etal., 2009). The creation of permanent pores in the cell walls facilitate the entry of acids (in the case of chemical treatments) or enzymes (biological processes) used to break down the cellulose into its constituent sugars, and thus PEF increases the hydrolysis rate.

Most important factors in PEF pretreatment include electric field strength, which is usually above 1kV/cm (520kV/cm), number of pulses, and treatment time, in the microsecond range. Major benefits of PEF pretreatment are that it can be carried out at ambient conditions, energy requirement is low, and the process itself is not very complex as it does not involve moving parts.

Ultrasonic pretreatment (USP) of lignocellulosic biomass has been studied at laboratory scale, although it is a well-established technique for industrial wastewater treatment. It promotes the pretreatment and conversion process through cavitation phenomenon. Ultrasonic energy allows destruction of the lignocellulosic structure and fractionation of biomass components, with increased yields of sugars, bioethanol, and gas products. Sonication promotes hydrolysis and leads to reduced reaction time, lower reaction temperature, and less amounts of solvents (Luo etal., 2014).

Experiments carried out on a model compound (carboxymethyl cellulose) showed that reaction time was dramatically increased (Imai etal., 2004). Other studies on lignocellulosic biomass in combination with hydrogen peroxide for bioethanol production gave higher yields of cellulose recovery and delignification (Ramadoss and Muthukumar, 2016).

high energy ball milling process for nanomaterial synthesis

high energy ball milling process for nanomaterial synthesis

It is a ball milling process where a powder mixture placed in the ball mill is subjected to high-energy collision from the balls. This process was developed by Benjamin and his coworkers at the International Nickel Company in the late of 1960. It was found that this method, termed mechanical alloying, could successfully produce fine, uniform dispersions of oxide particles (Al2O3, Y2O3, ThO2) in nickel-base superalloys that could not be made by more conventional powder metallurgy methods. Their innovation has changed the traditional method in which production of materials is carried out by high temperature synthesis. Besides materials synthesis, high-energy ball milling is a way of modifying the conditions in which chemical reactions usually take place either by changing the reactivity of as-milled solids (mechanical activation increasing reaction rates, lowering reaction temperature of the ground powders)or by inducing chemical reactions during milling (mechanochemistry). It is, furthermore, a way of inducing phase transformations in starting powders whose particles have all the same chemical composition: amorphization or polymorphic transformations of compounds, disordering of ordered alloys, etc.

The alloying process can be carried out using different apparatus, namely, attritor, planetary mill or a horizontal ball mill. However, the principles of these operations are same for all the techniques. Since the powders are cold welded and fractured during mechanical alloying, it is critical to establish a balance between the two processes in order to alloy successfully. Planetary ball mill is a most frequently used system for mechanical alloying since only a very small amount of powder is required. Therefore, the system is particularly suitable for research purpose in the laboratory. The ball mill system consists of one turn disc (turn table) and two or four bowls. The turn disc rotates in one direction while the bowls rotate in the opposite direction. The centrifugal forces, created by the rotation of the bowl around its own axis together with the rotation of the turn disc, are applied to the powder mixture and milling balls in the bowl. The powder mixture is fractured and cold welded under high energy impact.

The figure below shows the motions of the balls and the powder. Since the rotation directions of the bowl and turn disc are opposite, the centrifugal forces are alternately synchronized. Thus friction resulted from the hardened milling balls and the powder mixture being ground alternately rolling on the inner wall of the bowl and striking the opposite wall. The impact energy of the milling balls in the normal direction attains a value of up to 40 times higher than that due to gravitational acceleration. Hence, the planetary ball mill can be used for high-speed milling.

During the high-energy ball milling process, the powder particles are subjected to high energetic impact. Microstructurally, the mechanical alloying process can be divided into four stages: (a) initial stage, (b) intermediate stage, (c) final stage, and (d) completion stage.

(a) At the initial stage of ball milling, the powder particles are flattened by the compressive forces due to the collision of the balls. Micro-forging leads to changes in the shapes of individual particles, or cluster of particles being impacted repeatedly by the milling balls with high kinetic energy. However, such deformation of the powders shows no net change in mass.

(b) At the intermediate stage of the mechanical alloying process, significant changes occur in comparison with those in the initial stage. Cold welding is now significant. The intimate mixture of the powder constituents decreases the diffusion distance to the micrometer range. Fracturing and cold welding are the dominant milling processes at this stage. Although some dissolution may take place, the chemical composition of the alloyed powder is still not homogeneous.

(c) At the final stage of the mechanical alloying process, considerable refinement and reduction in particle size is evident. The microstructure of the particle also appears to be more homogenous in microscopic scale than those at the initial and intermediate stages. True alloys may have already been formed.

(d) At the completion stage of the mechanical alloying process, the powder particles possess an extremely deformed metastable structure. At this stage, the lamellae are no longer resolvable by optical microscopy. Further mechanical alloying beyond this stage cannot physically improve the dispersoid distribution. Real alloy with composition similar to the starting constituents is thus formed.

Theoretical considerations and explorations of planetary milling process have been broadly studied in order to better understand and inteprate the concept. Joisels work is the first report to study the shock kinematics of a satellite milling machine. This work focused on the determination of the milling parameters that were optimized for shock energy. The various parameters were determined geometrically and the theoretical predictions were examined experimentally using a specifically designed planetary mill. Schilz et al. reported that from a macroscopical point of view, the geometry of the mill and the ratio of angular velocities of the planetary to the system wheel played crucial roles in the milling performance. For a particular ductile-brittle MgSi system, the milling efficiency of the planetary ball was found to be heavily influenced by the ratio of the angular velocity of the planetary wheel to that of the system wheel as well as the amount of sample load. Mio et al. studied the effect of rotational direction and rotation-to-revolution speed ratio in planetary ball milling. Some more theoretical issues and kinematic modeling of the planetary ball mill were reported later in related references. Because mechanical alloying of materials are complex processes which depend on many factors, for instance on physical and chemical parameters such as the precise dynamical conditions, temperature, nature of the grinding atmosphere, chemical composition of the powder mixtures, chemical nature of the grinding tools, etc., some theoretical problems, like predicting nonequilibrium phase transitions under milling, are still in debate.

For all nanocrystalline materials prepared by high-energy ball milling synthesis route, surface and interface contamination is a major concern. In particular, mechanical attributed contamination by the milling tools (Fe or WC) as well as ambient gas (trace impurities such as O2, N2 in rare gases) can be problems for high-energy ball milling. However, using optimized milling speed and milling time may effectively reduce the contamination. Moreover, ductile materials can form a thin coating layer on the milling tools that reduces contamination tremendously. Atmospheric contamination can be minimized or eliminated by sealing the vial with a flexible O ring after the powder has been loaded in an inert gas glove box. Small experimental ball mills can also be completely enclosed in an inert gas glove box. As a consequence, the contamination with Fe-based wear debris can be reduced to less than 12 at.% and oxygen and nitrogen contamination to less than 300 ppm. Besides the contamination, long processing time, no control on particle morphology, agglomerates, and residual strain in the crystallized phase are the other disadvantages of high-energy ball milling process.

Notwithstanding the drawbacks, high-energy ball milling process has attracted much attention and inspired numerous research interests because of its promising results, various applications and potential scientific values. The synthesis of nanostructured metal oxides for gas detection is one of the most promising applications of high-energy ball milling. Some significant works have been reported in recent years. Jiang et al. prepared metastable a-Fe2O3MO2 (M: Ti and Sn) solid solutions by high-energy milling for C2H5OH detection. The 85 mol% a-Fe2O3SnO2 sample milled for 110 hours showed the highest sensitivity among all the samples studied. The best sensitivity to 1000 ppm C2H5OH in air at an operating temperature of 250 C was about 20. Zhang et al. synthesized FeSbO4 for LPG detection. They found that there were two-step solid-state reactions occurring in the raw powders during the ball milling:

The response and recovery times of their sensor were less than one second. The sensitivity to 1000 ppm C2H5OH at an operating temperature of 375 C was about 45. Diguez et al. employed precipitation method to prepare nanocrystalline SnO2 and planetary milling to grind the obtained powder for NO2 detection. They found that the grinding procedure of the precursor and/or of the oxide had critical effect on the resistance in air. As a result, the gas sensing properties to NO2 had been considerably improved. Cukrov et al. and Kersen et al. synthesized SnO2 powders by mechanochemical processing for O2 and H2S sensing applications, respectively. The O2 sensor exhibited stable, repeatable and reproducible electrical response to O2. More recently, Yamazoes group reported the sensing properties of SnO2Co3O4 composites to CO and H2. A series of SnO2Co3O4 thick films containing 0100% Co3O4 in mass were prepared from the component oxides through mixing by ball-milling for 24 h, screen-printing and sintering at 700 C for 3 h. The composite films were found to exhibit n- or p-type response to CO and H2 depending on the Co3O4 contents in the composites. The n-type response was exhibited at 200 C or above by SnO2-rich composites (Co3O4 content up to 5 mass%). The sensor response to both CO and H2 was significantly enhanced by the addition of small amounts of Co3O4 to SnO2, and the response at 250 C achieved a sharp maximum at 1 mass% Co3O4. The p-type response was obtained at 200 C or below by the composites containing 25100 mass% Co3O4. The sensitivity as well as selectivity to CO over H2 could thus be increased by the addition of SnO2 to Co3O4.

Besides the above mentioned researches, significant efforts on the synthesis of nanostructured metal oxide with high-energy ball milling method for gas sensing have been actively pursued by the authors of this chapter. In our research, we use the high-energy ball milling technique to synthesize various nanometer powders with an average particle size down to several nm, including nano-sized a-Fe2O3 based solid solutions mixed with varied mole percentages of SnO2, ZrO2 and TiO2 separately for ethanol gas sensing application, stabilized ZrO2 based and TiO2 based solid solutions mixed with different mole percentages of a-Fe2O3 and synthesized SrTiO3 for oxygen gas sensing. The synthesized powders were characterized with XRD, TEM, SEM, XPS, and DTA. Their sensing properties were systematically investigated and sensing mechanisms were explored and discussed as well.

effect of high-energy ball milling time on structural and magnetic properties of nanocrystalline cobalt ferrite powders - sciencedirect

effect of high-energy ball milling time on structural and magnetic properties of nanocrystalline cobalt ferrite powders - sciencedirect

Effect of ball milling on the structural and magnetic properties of cobalt ferrite.Prolonged milling promotes interparticle aggregation.Tunable coercivity due to crystal size, strain, and cation distribution.

Cobalt ferrite nanocrystals synthesized by conventional and size-controlled coprecipitation methods were treated by high-energy ball milling, HEBM, in order to study the effect of crystal size reduction and/or strain on the resulting magnetic properties. Processed nanocrystals were characterized by X-ray diffraction, Brunauer, Emmett, and Teller surface area analysis, transmission electron microscopy (TEM), and vibrating sample magnetometry. The cobalt ferrite nanocrystals exhibited crystal size reduction from initial values (average crystallite sizes of 121nm and 183nm, respectively) down to 10nm after HEBM for 10h. The specific surface area was decreased by milling (from 96.5 to 59.4m2/g; for the 12nm cobalt ferrite nanocrystals), due to particles aggregation. TEM analyses corroborated the aggregation of the nanoparticles at such long milling times. The same cobalt ferrite nanocrystals exhibited a rise in coercivity from 394 to 560Oe after 5h ball milling which was attributed to the introduction of strain anisotropy, namely point defects, as suggested by the systematic shift of the diffraction peaks towards higher angles. In turn, the magnetic characterization of the starting 18nm-nanocrystals reported a drop in coercivity from 4506Oe to 491Oe that was attributed predominantly to size reduction within the single domain region. A correlation between particle size, cationic distribution, and HEBM processing conditions became evident.

photocatalytic hydrogen production using fetio 3 concentrates modified by high energy ball milling and the presence of mg precursors | springerlink

photocatalytic hydrogen production using fetio 3 concentrates modified by high energy ball milling and the presence of mg precursors | springerlink

Ilmenite (FeTiO3) concentrates were modified by high energy ball milling (HEBM) at different times (1, 2, and 3h) and in the presence of MgO or metallic Mg at different contents (0.5, 1.0, and 3.0wt%), with the aim of obtaining a low-cost and highly available photocatalyst with enhanced performance towards H2 production. FeTiO3 concentrates were obtained from ilmenite-rich black sand by gravimetric concentration followed by wet electromagnetic and dry magnetic separation. HEBM was performed to insert Mg into FeTiO3 structure and to simultaneously reduce the particle size. The insertion of Mg intended to shift the conduction band of the materials to more negative potentials, in spite of the band-gap widening. XRD patterns of the milled samples showed a decrease in the intensity of ilmenite peaks without any displacement, indicating slight amorphization without significant changes in the crystalline structure. However, the increase in the intensity of hematite peaks suggests substitutional doping according to the proposed solid-state reactions. Although thermodynamic analyses showed that doping with metallic Mg should be more favorable, doping with MgO appears to be kinetically favored due to the physical characteristics of the precursor, as it was revealed by XRD, XPS, Raman spectroscopy, SEMEDS, and UVVis DRS results. Deconvolution of high-resolution XPS spectra corresponding to Mg1s exhibited the presence of an additional component peak for the sample milled with metallic Mg, in comparison to that with MgO. This result evidenced the formation of a new phase, suggesting that part of the metallic Mg was not inserted into FeTiO3, as it was confirmed by Raman spectroscopy. UVVis DRS analyses showed an increase in theband-gap of the samples milled in the presence of Mg precursors (from 2.51 to 2.552.59eV) which were attributed to slight modifications of the conduction band due to the insertion of Mg. The performance towards hydrogen production under UV irradiation was improved from 240.5molg1h1 (unmilled sample) to 255.3molg1h1 (2h milled in the absence of Mg precursor), 296.0molg1h1 (2h milled with 1.0wt% MgO), and 265.2molg1h1 (2h milled with 1.0wt% metallic Mg). This improvement was attributed mainly to the insertion of Mg, and the consequent modification of the band structure, rather than the modification in the surface area.

The increase in energy demand and the growing concern over the negative effects that fossil fuels have on the environment has led to the search for new sources of sustainable energy [1,2,3]. In this regard, hydrogen (H2) is profiled as a potential fuel due to several advantages associated to its use; to mention a few, H2 has high energy efficiency, it can be transformed directly to electric energy, its combustion does not induce to the formation of harmful emissions such as carbon and nitrogen oxides, as it occurs with hydrocarbon combustion, and more importantly, H2 can be obtained from renewable sources like water and solar radiation [2, 4,5,6]. Among the techniques that enable the use of light for hydrogen production, the photocatalytic splitting of water, depicted in Fig.1, has drawn significant attention from scientists around the world [7,8,9,10]. Titanium dioxide (TiO2) has been the most widely studied semiconductor photocatalyst [10,11,12,13]; however, it holds a major disadvantage for its use in hydrogen production related to the location of its Conduction Band (CB); this drawback is explained considering that the photocatalytic production of H2 can only occur when the CB of the semiconductor is located at more negative potentials compared to the redox potential of H2O reduction. This condition is slightly satisfied by TiO2 in the dark; however, it has been reported that once TiO2 is illuminated, a band displacement takes place, inhibiting H2 production [14, 15]. Taking these factors into consideration and looking for an appropriate semiconductor to carry out H2 production and other photoactivated processes, it is necessary to count with highly available semiconductors that enable the reactions of interest. In that regard, the use of ilmenite (FeTiO3) as an alternative semiconductor in photo-activated processes has been the center of interest for several authors [16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32] due to its content of Fe and Ti, its band-gap between 2.592.9eV [14, 33], its optic and magnetic properties, in addition to its high abundancy, especially when referring to natural ilmenite [17, 28, 34]. In previous research performed with mineral ilmenite, we reported its potential for photocatalytic reactions and highlighted its photo-reductive capabilities [35], which suggest a good performance towards H2O reduction for H2 production. Indeed, the potential of the conduction band of ilmenite is theoretically located at a more negative value, compared to TiO2 [36], which could imply an enhancement in the photocatalytic H2 production due to the increase in the thermodynamic driving force for this reaction. In addition to FeTiO3, MgTiO3 was found to possess a conduction band located at even more negative potentials [33] and this property has made MgTiO3 the object of some studies concerning its use on H2 production [37,38,39]. Nonetheless, MgTiO3 holds a wide band-gap between 3.43.7eV [37] which increases the energy requirement for its activation.

Considering the appealing properties and band structure of ilmenite, as well as the reductive ability of the photogenerated electrons in both FeTiO3 and MgTiO3, it is proposed that the insertion of Mg into FeTiO3 structure could shift the conduction band of the material to more negative potentials (Fig.1), increasing the driving force for H2 production, and therefore, enhancing the performance of the material.

Regarding the feasibility of obtaining (Mg,Fe)TiO3 materials, in 1998 Welham reported the formation of mixed FeMg titanates during the mecanochemically induced reduction of ilmenite by magnesium [40]. Furthermore, Welham [41] and Linton et al. [42, 43] have devoted efforts on studying the formation of Fe and Mg mixed titanates, using both mineral [41] and synthetic ilmenite [42, 43]. These works focused on studying the formation of the materials analyzing the reactions involved in the synthesis, as well as the structure and thermodynamics of the system, but they did not include a practical application of the obtained materials. In 2014, Canaguier investigated the synthesis of FeTiO3-MgTiO3 solid solutions in order to evaluate the effect of the inclusion of Mg impurities into synthetic FeTiO3 over the reduction of the material [44]. Additional research has been performed using MgTiO3 as base material. In 2017, Yang et al. synthesized Fe-doped and Fe/N-codoped MgTiO3 by the solgel method to obtain near-infrared reflective inorganic pigments [45]. Similarly, in 2019, Zhao et al. prepared Fe-doped MgTiO3 films supported on silicon wafers and evaluated their structure along with the optical and magnetic properties, to determine the potential of these materials in optoelectronic and magnetic applications [46]. However, to the best of our knowledge, there are no reports on the use of materials in the system (Mg,Fe)TiO3 prepared from natural ilmenite applied for photocatalytic applications nor H2 production and it has not been found the influence of Mg insertion over the physicochemical properties of FeTiO3.

In this investigation, semiconductors of the type (MgxFe1-x)TiO3 were prepared by submitting FeTiO3 mineral concentrates to low-cost High Energy Ball Milling (HEBM) process in the presence of MgO or metallic Mg as precursors. The effect of the insertion of Mg into FeTiO3 was evaluated in the photocatalytic production of H2 under UV irradiation.

FeTiO3 concentrates were obtained from an ilmenite-rich black sand alluvial deposit by gravimetric concentration performed on a Reichert spiral, followed by two consecutive stages of wet high-intensity electromagnetic and dry low-intensity magnetic separation. The protocol and variables of the concentration process will be described in a separate paper.

HEBM tests were carried out in a 900mL stainless steel attritor mill to insert Mg into FeTiO3 structure and to simultaneously reduce the particle size. Preliminary tests were performed in the absence of Mg precursor to set the conditions to reach smaller particles. The mill was filled up to 20% of its volumetric capacity, and every experiment was conducted at autogenous temperature starting from room temperature, under air atmosphere and without any process control agent. The effect of time (0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.0, 5.0, and 6.0h), milling speed (400 and 600rpm), ball to powder weight ratio-BPR (50:1 and 100:1), and size of grinding media (3.0, 4.0 and 4.8mm) was evaluated. It was found that after 2h of milling at a speed of 600rpm, with a BPR of 50:1, and using grinding media of 4.8mm, a minimum particle size of 278nm was reached [47].

Using the aforementioned conditions, HEBM tests were performed in the presence of Mg precursors: MgO (Laboratorios Len, 100%, powder) or metallic Mg (PanReac, 99.7%, ribbon). Filings of the metallic Mg ribbon were manually obtained using a stainless-steel lime. The nominal concentrations added of Mg were 0.5, 1.0, and 3.0 wt%.Those values of concentration were chosen looking for doping without inducing significant changes in the crystal structure of the host material [48, 49]. However, it is important to note that it has been reported that Mg exhibits complete solubility in FeTiO3 [42]. The mass of MgO or metallic Mg added to the mill were calculated accoording to Eqs. 1 and 2, respectively.

where \(wt_{MgO}\) and \(wt_{Mg}\) are the masses of MgO and metallic Mg added to the mill, respectively; \(wt_{Ilm}\) is the mass of ilmenite concentrate; \(NC\) is the nominal concentration added of precursor; and \(M_{Mg}\) and \(M_{MgO}\) are the molar masses of Mg and MgO (24.305 and 40.304gmol1, respectively).

The samples obtained by HEBM are named hereafter indicating the nominal concentration added of a given precursor and the milling time: e.g. FeTiO3-0.5wt%MgO-2h corresponds to the sample milled with 0.5wt% of MgO during 2h, whereas FeTiO3-1h is the sample milled for 1h without any Mg precursor. The unmilled concentrate is referred to as FeTiO3-0.

X-ray Diffraction (XRD) patterns were recorded in a Bruker D8 ADVANCE diffractometer using Cu- K radiation, operating at 40kV and 40mA. Qualitative and semi-quantitative analyses of the XRD patterns were performed using the software DIFFRAC EVA V4.2. X-ray Photoelectron Spectroscopy (XPS) was carried out in a SPECS XPS/ISS/UPS platform, using Al-K radiation operated at 100W; the binding energies (BE) were referenced respect to the C1s peak at 284.8eV (adventitious carbon). Raman spectroscopy was performed on a Raman microscopy XploRA PLUS Horiba Scientific using 785, 638, and 532nm wavelength lasers. UVVis Diffuse Reflectance Spectroscopy (UVVisDRS) data were collected by a UVVis spectrometer Varian Cary 100Scan, equipped with an integrating sphere and using BaSO4 as reference. Scanning Electron Microscopy coupled with Energy Dispersive Spectroscopy (SEMEDS) images were recorded in a high-resolution scanning electron microscope Quanta FEG 650, operating at 20kV. N2 Physisorption data were recorded by a gas adsorption analyzer Quantachrome Nova 4200e at the temperature of liquid nitrogen (77K).

Photocatalytic hydrogen production was evaluated using a closed glass photoreactor system, filled with 200mL of an aqueous solution, composed of 150mL of distilled water and 50mL of methanol as hole scavenger. To avoid the competition with oxygen for the consumption of photo-generated electrons, the solution was bubbled with nitrogen for 8min prior to the addition of the photocatalysts, to expel dissolved oxygen. 0.5mgmL1 of photocatalyst were suspended in the solution. The reaction system was magnetically stirred (~1000rpm) during the process. The solution was maintained in the dark for 30min to reach adsorptiondesorption equilibrium, and afterward, it was illuminated with a UV radiation Pen-Ray lamp for 5h. Hydrogen measurements were performed using a gas chromatograph Shimadzu G-08 equipped with a thermal conductivity detector (TCD) and a Shincarbon packed column (2m length, 1mm ID and, 25mm OD), using N2 as carrier gas. The quantification of produced H2 was carried out after the first 30min of reaction in the dark, and hourly after starting irradiation. Tests in the absence of catalyst as well as using TiO2 P25 as photocatalyst, were carried out to compare the performance of the obtained samples.

To establish the effects of including MgO or metallic Mg in the HEBM process, the XRD patterns of the samples were analyzed (Fig.2). The XRD pattern of the unmilled concentrate (FeTiO3-0) showed the presence of ilmeniteFeTiO3 (PDF No. 75-1209), hematiteFe2O3 (PDF No. 73603), and rutileTiO2 (PDF No. 87-710) as main phases.

Effect of (I) milling time, (II) nominal concentration added of Mg precursor and (III) type of Mg precursor in the crystal structure in the XRD patterns of samples: a FeTiO3-0; b FeTiO3-1.0wt%MgO-1h; c FeTiO3-1.0wt%MgO-2h; d FeTiO3-1.0wt%MgO-3h; e FeTiO3-0.5wt%Mg-2h; f FeTiO3-1.0wt%Mg-2h; g FeTiO3-3.0wt%Mg-2h. :Ilmenite

The effect of milling time (Fig.2(I)), nominal concentration added of Mg (Fig.2(II)), and the Mg precursor (Fig.2(III)) was analyzed. As it can be noted, the crystalline structure of ilmenite was still identified in the samples after milling in the absence and presence of Mg precursors; slight amorphization was evidenced by the broadening and diminishing in the intensity of the ilmenite peaks. If the insertion of Mg into FeTiO3 structure effectively took place, it is believed that the Mg could substitute Fe in the ilmenite lattice, inducing distortions in the crystals; however, no significant changes in the 2 position of the characteristic peaks of ilmenite were detected. This indicates that the insertion of Mg did not lead to a considerable modification of lattice parameters, perhaps due to the small amount of Mg added. Furthermore, the d spacing of Fe and Mg titanates overlap [41], which hinders the differentiation in the XRD patterns of these materials.

The patterns of the samples milled in the presence of MgO or metallic Mg did not exhibit signals associated with either one of the precursors (Fig.2(III)). With comparative purposes, physical mixtures of ilmenite with MgO or metallic Mg without milling were prepared. The XRD patterns of these mixtures (not shown) were also characterized by the absence of signals ascribed to MgO or metallic Mg. This can be attributed to the small amount of precursor, especially considering that the XRD patterns of MgO and metallic Mg are composed only by two low-intensity peaks, and therefore their contribution can be mistaken with the background.

The substitution of Fe for Mg can follow the reactions proposed in Eqs. 3a and 3b, which would lead to the production of FeO or Fe. The absence of these phases in the XRD pattern of the samples can be attributed to the instability and high reactivity of FeO and Fe, as they can react with atmospheric oxygen and form hematite (Fe2O3). Nonetheless, the absence of these phases also suggests the occurrence of a substitution reaction that involves atmospheric oxygen, forming Fe2O3 as the main product (Eqs.4a and 4b), as stated by Welham [41]. The thermodynamic feasibility of the substitution reaction, both neglecting and considering the involvement of oxygen, was examined by comparing the Gibbs free energy of Eqs. (34). It is worth noting that, for both MgO and metallic Mg, reactions with oxygen are thermodynamically more favorable than their equivalent without the presence of O2 (Fig.3).

Considering that the substitution reactions may lead to the formation of Fe2O3, semi-quantitative analyses were performed to follow the formation of hematite after milling the samples and the results are summarized in Table1. It is important to note that, in all the cases, when the amount of Fe2O3 increases, the amount of FeTiO3 decreases and vice versa, confirming that the formation of hematite is related to the modification of ilmenite. Additionally, the amount of TiO2 remains nearly constant in most of the samples, which rules out the decomposition of FeTiO3 to form TiO2 and Fe2O3 separately.

As it is observed, when using MgO as precursor, the percentage of Fe2O3 increased progressively with milling time until 2h. However, after 3h of milling, the amount of Fe2O3 decreased, which evidenced reversibility of the substitution reaction. Similarly, the amount of hematite increased when the nominal concentration added of Mg precursor was risen from 0.5 to 1.0wt%. However, when adding 3.0wt% of MgO to the process, the percentage of Fe2O3 decreased again, which exhibited saturation of Mg at the given conditions. This saturation is presumed to be reached due to the lack of new reactive surface. The substitution reaction occurs at the surface of the grain (since the Mg added cannot reach the bulk); therefore, the particle size reached during the HEBM might have not been small enough to promote the complete insertion of Mg into FeTiO3 structure with the highest nominal concentration added. In that sense, increasing the amount of Mg precursor results to be ineffective without further reducing the particle size (increasing surface area), and thus, creating more available sites for Mg to substitute Fe in the structure.

On the other side, when using metallic Mg as precursor, the amount of hematite slightly increased compared to the unmilled concentrate, but it remained nearly constant when changing the milling time and the nominal concentration added of precursor, which indicates that the insertion of Mg into FeTiO3 was not equally effective with both precursors.

Although it was shown that doping with metallic Mg is thermodynamically more favorable (Fig.3), the semi-quantitative analyses of XRD data indicated that the insertion of Mg into FeTiO3 was more effective when milling the samples with MgO as opposed to metallic Mg. This was attributed to the physical properties of the metallic filings which were manually obtained, and therefore, their hardness, ductility and considerably big size, inhibited the correct fracture of the material and the insertion of Mg into FeTiO3. In contrast, MgO powder was composed of smaller and more fragile particles, and thus it was easily fractured. Thus, it was concluded that the physical properties of the precursor play a crucial role in the modification of the crystalline structure by high energy ball milling.

XPS spectra provided information regarding the chemical state of the elements in the samples. As observed in the general XPS spectra (Fig.4(I)), the unmilled concentrate exhibited signals related to C, Fe, Ti, O, and small traces of Mn within FeTiO3, as well as Ca, and Si, which confirm the presence of amorphous aluminosilicates associated with the ilmenite grains, as reported previously [35]. It was noted that the general spectra of the unmilled sample, as well as sample FeTiO3-2h, did not exhibit the presence of Mg. However, the signal ascribed to Mg increases in samples FeTiO3-1.0wt%MgO-2h and FeTiO3-1.0wt%Mg-2h, evidencing the presence of this element in the modified samples.

(I) XPS general spectra, (II) HRXPS spectrum for Fe2p, (III) HR XPS spectrum for Ti2p, (IV) HR XPS spectrum for Mg1s for the sample milled with MgO, (V) HR XPS spectrum for Mg1s for the sample milled with metallic Mg, and (VI) HR XPS spectrum in the valence band region. Samples: aFeTiO3-0; cFeTiO3-1.0wt%MgO-2h; cFeTiO3-1.0wt%Mg-2h; and dFeTiO3-2h

In the High-Resolution (HR) spectra of the Fe2p region (Fig.4(II)), the components of the signals ascribed to Fe2p3/2 and Fe2p1/2 located at 710.91 and 728.23eV, respectively, are related to the presence of Fe2+. Meanwhile, the signals at 712.25eV (Fe2p3/2) and 724.86eV (Fe2p1/2) along with the satellite located at 718.49eV, are characteristic of Fe3+ [19, 30]. These results evidence the presence of hematite (Fe2O3) in the sample, as found in the XRD analyses and suggest that ilmenite holds mixed-valence cations Fe2+/Fe3+ in its structure, as it has been reported by other authors [19, 50].

HR spectra of Ti2p (Fig.4(III)) showed that the signals of Ti2p3/2 and Ti2p1/2 were composed of two contributions each; the peaks centered at 458.33eV (Ti2p3/2) and 464.12eV (Ti2p1/2), with a difference of 5.79eV, are consistent with Ti4+ involved in an O-TiO bond [37, 51, 52], while the signals located at higher binding energies(459.44eV for Ti2p3/2 and 464.25eV for Ti2p1/2) can be related to the formation of Fe-TiO bonds [52]. All the samples exhibited similar behavior in the peaks ascribed to Fe and Ti.

In the case of HR spectra for Mg1s, the sample milled with MgO exhibited a signal adjusted with only one component centered at 1304.65eV (Fig.4(IV)), which matches quite well with Mg2+. The BE of this signal, located at the same position reported for MgTiO3 [37, 38, 51], might be an indication of the replacement of Fe by Mg within the lattice of the modified samples. For the sample milled with metallic Mg, it is important to point out that the XPS HR spectra did not exhibit signal ascribed to metallic Mg (Mg0), which can be found at around 1303eV [53, 54] (lower binding energies compared to Mg2+), this suggests that the metallic Mg reacted during the HEBM process either with FeTiO3 or with elements in the milling atmosphere. For the sample FeTiO3-1.0wt%Mg-2h, the signal of Mg1s was deconvoluted into two components located at 1304.55 and 1306.78eV (Fig.4(V)). Similar as the sample milled with MgO, the peak at 1304.55eV was associated with Mg2+. On the other hand, the signal located at 1306.78eV is believed to be associated with products of the reaction of the metallic Mg not included into FeTiO3, with species at the surface of the material or in the milling atmosphere, probably forming ionic compounds [55, 56]. It is worth mentioning that no significant changes within the HR spectra of O, Fe nor Ti were observed after the insertion of Mg (using either one of the precursors), probably due to the similarities between Fe and Mg located as cations in the MTiO3 (M=Fe,Mg) system.

The appearance of an additional deconvoluted component in the HR XPS spectrum for Mg1s of the sample milled with metallic Mg (i.e., FeTiO3-1.0wt%Mg-2h), along with the general spectra of the samples (Fig.4(I)), in which the Mg1s signal of the sample milled with MgO (i.e., (b) in Fig.4(I)) exhibited greater intensity compared to that of metallic Mg (i.e., (c) in Fig.4(I)), reaffirm our previous statement that doping with MgO reached greater efficiency in the insertion of Mg into FeTiO3.

The XPS spectra in the lowest BE were analyzed to examine the Valence Band(VB) of the samples. As it is shown in Fig.4(VI), the estimation of the VB position provided a value of EVB=0.92340.0332eV for the samples milled in the absence and presence of Mg precursors. Although the insertion of Mg into FeTiO3 might influence the VB structure, the conditions of the XPS spectra did not allow an accurate differentiation of the values [57]. A more precise estimation of the VB should be carried out using XPS spectra with higher intensity (the VB spectra used in this work reached only 40CPS) or through electrochemical techniques.

Four signals located at 231, 370, 463, and 682cm1 can be noted in the Raman spectrum of the unmilled concentrate (i.e., (a) FeTiO3-0 in Fig.5). These signals were associated to FeTiO3 [50, 58, 59]. After 2h of milling without any precursor (i.e., (b) FeTiO3-2h in Fig.5(I)), the signals remained in the same position, but they got broader and more intense, which was attributed to the amorphization of the material [60], as established with the XRD results.

Raman spectrum of MgO (i.e., (d) in Fig.5(II)) showed its characteristic signals at 975, 1146, 1385, 1633, 1796, and 2202cm1. When comparing the spectra of the unmilled concentrate, the MgO and the sample milled with MgO (i.e., (c) FeTiO3-1.0wt%MgO-2h in Fig.5(II)), it can be noted that the samples FeTiO3-0 and FeTiO3-1.0wt%MgO-2h presented similar spectra with signals associated to ilmenite located at the same frequencies, with slight differences in the shape and intensity of the peaks; no track of any of the MgO signals was observed in the FeTiO3-1.0wt%MgO-2h spectrum, which indicates that the substitution reaction between ilmenite and Mg could have taken place during the milling with MgO. Similarly, the spectra of the unmilled concentrate, the metallic Mg (i.e., (f) in Fig.5(III)), and the sample milled with metallic (i.e., (e) Mg FeTiO3-1.0wt%Mg-2h in Fig.5(III)) were compared. In this case, the spectrum of FeTiO3-1.0wt%Mg-2h exhibited significant changes compared to that of FeTiO3-0. In the first place, the most intense peak of ilmenite, located initially at 682cm1 was displaced to a lower frequency (635cm1); additionally, an intense signal at 137cm1 was appreciable in the milled sample, similar to the one observed in the Raman spectrum of metallic Mg. These results imply that the sample FeTiO3-1.0wt%Mg-2h is composed of two phases. The first phase possibly corresponds to ilmenite modified with some of the metallic Mg, while the second phase may correspond to an additional new phase formed by the metallic Mg that was not fractured and did not react with ilmenite. These findings agree with the XRD and XPS results regarding the difficulties to mill the filings of metallic Mg and the formation of a new phase from the unreacted metallic Mg.

The comparison between the micrographs of the unmilled concentrate and the samples milled the absence and presence of Mg precursor showed that HEBM process did not modify the mineral morphology (Fig.6). In all the cases, the samples exhibited polyhedral grains with defined boundaries. Despite the continuous fracture of the material during HEBM, the samples still display a wide dispersion in the particle size. The formation of agglomerates was observed, in agreement with the reported in the literature [61].

EDS analyses were performed on some zones of the samples. The results of the analysis performed on a grain of the unmilled concentrate, confirm the purity of ilmenite as well as the presence of Mn within this phase. This observation is in agreement with previous findings, in which it was proposed the stoichiometric formula (MnxFe1x)TiO3 for the ilmenite contained in the black sand deposit used as raw material in this work [35]. By widening the range of EDS evaluation, elements such as Mg, Al, Si, and Ca were detected, and their presence was attributed to intergranular aluminosilicates in ilmenite [35], which were also detected by XPS.

The EDS of the sample milled with MgO (i.e., (b) FeTiO3-1.0wt%MgO-2h in Fig.6) exhibited an increased content of Mg, reaching amounts of 0.93wt%. Contrary, when adding metallic Mg (i.e., (c) FeTiO3-1.0wt%Mg-2h inf Fig.6), the amount of Mg did not rise significantly. This result reinforces the fact that metallic Mg was not effectively fractured during the milling process.

Comparisons between the micrographs registered with both secondary and backscattered electrons detectors were performed (not shown) and it was possible to rule out the deposition of MgO or metallic Mg on the surface of ilmenite; consequently, it is inferred that the Mg detected during EDS and XPS analyses was inserted into the ilmenite structure.

The diffuse reflectance spectra (not shown) of the samples evidenced that the modified concentrates exhibit photoactivity in the visible range of the electromagnetic spectrum, following the optical behavior expected for black materials. The samples showed reflectance peaks at 542, 483, and 391nm, thus indicating absorption of radiation in such wavelengths.

The band-gap energy of the samples was estimated applying the Tauc method, as depicted in Fig.7, and the values obtained are summarized in Table 1. By applying this method, the accuracy of the estimated Eg values was0.01eV [62]. As it can be noted, milling the concentrate in the absence of Mg precursors led to an increase of the band-gap from 2.38eV to 2.52eV. This was attributed to the release of granular inclusions of aluminosilicates found inside the ilmenite [35]. Without any milling, these phases were contained inside the ilmenite grains and their contribution was negligible (the measured reflectance does not correspond to the bulk, but to the surface of the material). Once the grains are milled, the insulating oxides are released, affecting the band-gap value.

Beyond the increase caused by milling, slight changes in the Eg were observed and are presumed to be associated to the content of Mg included in the FeTiO3 structure, since it is expected for Mg insertion to promote modifications of the conduction band to more negative values (possibly increasing the band-gap as consequence). As it was discussed previously, insertion of Mg appears to be more effective when using MgO as precursor; this can be also supported by the variation of the band-gap of the samples milled with MgO, which range between 2.552.59eV, meanwhile, the band-gap of the samples milled with metallic Mg remained nearly constant at about 2.56eV.

Regarding the milling time and nominal concentration added of Mg, the samples followed similar tendencies to those observed with the XRD results. Specifically, when using MgO, it can be noted that increasing the nominal concentration added of Mg precursor from 0.5 to 1.0wt% led to a slight increase in the band-gap (from 2.55eV to 2.59eV), which was kept when further increasing nominal concentration added to 3.0wt% (2.59eV). Negligible variations were observed when milling for 1 and 2h; however, increasing milling time from 2 to 3h resulted in a decrease in the band-gap, suggesting reversibility of the substitution reaction.

The adsorption isotherms of the samples FeTiO3-0, FeTiO3-2h, FeTiO3-1.0wt%MgO-2h and FeTiO3-1.0wt%Mg-2h are presented in Fig.8. The samples exhibited type IV isotherms with the presence of hysteresis, related to mesoporous materials with pores wider than~4nm [63]. This behavior agrees with the calculated average pore size, which ranges between 5.44 and 7.37nm.

The surface area of the samples was determined using the Brunauer-Emmet-Teller (BET) equation on the N2 physisorption data, and it was found that the samples exhibited low specific surface areas (Table1) compared to those of commonly used photocatalysts (e.g. commercial TiO2 such as EvonikP25 and EvonikP90 exhibit surface areas of around 60 and 100 m2g1, respectively [64, 65]). By milling the sample during 2h in the absence of Mg precursor, an increment in the specific surface area was observed (from 2.9 to 14.7 m2g1) and it was associated with the reduction of particle size. The presence of Mg precursor during HEBM process seems to have affected the fracture of the material since no further increase of surface area was reached. The samples milled for 2h with MgO and metallic Mg exhibited lower surface areas (8.2 and 4.2m2g1, respectively) compared to the sample milled for the same time without any precursor. Particularly, when using metallic Mg, the surface area was maintained nearly constant with milling time and did not increase significantly compared to the unmilled concentrate; this can be attributed to the ductility of metallic Mg, which could have led to the formation of thin films that inhibited fracturing of the mineral grains.

Considering the relevance of surface area during photocatalytic processes, additional efforts should be made to enhance the efficiency of HEBM processes on diminishing particle size and increasing surface area. As it was previously mentioned, during this work the samples were milled in a dry atmosphere without any process control agent. Nonetheless, the presence of surfactant during HEBM could be considered, given that it might help reach smaller particle sizes. Due to the electrostatic and steric forces, as well as the reduction in surface tension, surfactants can both stabilize the particles in the milling media, minimizing the uncontrolled fracture of the particles, and reduce the effect of the cold-welding phenomena that take place [61, 66, 67], thus, allowing to attain smaller particles and to control the final size distribution, shape and purity of the obtained material [61, 68]. Despite these advantages, the surfactant can interact with the powder and they may be inserted into the material structure or lead to the formation of undesired compounds [61]. Another alternative to improve HEBM processes, is the inclusion of acid solutions, as it is known that an appropriate acid treatment can lead to the enlargement of the surface area and the modification of properties such as the acidity and porosity [69,70,71], which are quite relevant for catalytic and photocatalytic materials. Similarly, as it occurs with the inclusion of surfactants, the presence of acid solutions might help mitigate the agglomeration of particles by reducing milling time. However, the selection of the acid and its concentration is fundamental if the digestion or leaching of metal (such as Fe2+ in ilmenite) is to be avoided [69, 72].

The photocatalytic performance of the samples under UV irradiation is presented and compared in Fig.9. The method and equipment used to measure the evolution of H2 provided an accuracy of0.8molh1. As it can be noted, the production of H2 in the absence of catalyst (photolysis) is negligible (i.e., (a) in Fig.9(I)). A test performed using TiO2 (EvonikP25) as photocatalyst (i.e., (b) in Fig.9(I)) showed that, under the conditions evaluated, TiO2 does not hold the capacity to produce H2, as it was previously stated. Meanwhile, the FeTiO3 concentrates prepared in this work exhibited increased hydrogen production (i.e., (c) to (i) in Fig.9).

Photocatalytic hydrogen production. Effect of (I) photocatalyst, (II) milling time and (III) concentration of Mg precursor. Samples: a absence of catalyst (photolysis); b TiO2P25; c FeTiO3-0; d FeTiO3-2h; e FeTiO3-1.0wt%MgO-2h; f FeTiO3-1.0wt%Mg-2h; g FeTiO3-1.0wt%MgO-1h; h FeTiO3-0.5wt%MgO-2h; and i FeTiO3-3.0wt%MgO-2h. Conditions: 0.5mgmL1 of photocatalyst, 150mL H2O and 50mL methanol

The amount of evolved hydrogen showed that milling the sample for 2h had a positive effect on its performance (i.e., (c) FeTiO3-0 exhibited lower H2 production compared to (d) FeTiO3-2h, Fig.9(I)). Moreover, the addition of Mg precursors enhanced the performance of the samples. The effect of milling time when using MgO as precursor is compared in Fig.9(II), and it can be observed that increasing milling time from 1 to 2h had a positive effect on H2 production (i.e., (g)FeTiO3-1.0wt%MgO-1h and (e) FeTiO3-1.0wt%MgO-2h in Fig.9(II)). The effect of the nominal concentration added of Mg when using MgO as precursor is presented in Fig.9(III). It was noted that the lowest H2 production was reached when milling with 0.5 wt% (i.e., (h) FeTiO3-0.5wt%MgO-2h in Fig.9(III)), and the performance was improved when increasing the amount of precursor to 1.0 wt% (i.e., (e) FeTiO3-1.0wt%MgO-2h in Fig.9(III)); nonetheless, further increasing the Mg concentration to 3.0 wt% led to a diminishing in the photocatalytic performance (i.e., (i) FeTiO3-3.0wt% MgO-2h in Fig.9(III)). This behavior is related to the saturation of FeTiO3, as previously stated. The exact same tendency, for both milling time and nominal concentration added of Mg, was observed for the samples milled with metallic Mg (not shown) and such behavior concur with the conclusions drawn from the XRD analyses.

The photocatalytic activity towards H2 production was the result of the combined effect of the surface area, the band-gap, and the potential of the conduction band of the samples. In that sense, sample FeTiO3-2h exhibited improved performance by having been milled in the absence of Mg precursor (255.3molg1h1); this was attributed to the increase in both surface area (from 2.9 to 14.7 m2g1) and the band-gap (from 2.38 to 2.51eV) of the sample, which implies a reduction in the recombination rate. Furthermore, samples FeTiO3-1.0wt%MgO-2h and FeTiO3-1.0wt% Mg-2h exhibited greater activity (296.0 and 265.2molg1h1, respectively) despite the diminishing in their specific surface area (8.2 and 4.2m2g1, respectively) compared to that of FeTiO3-2h. This behavior is associated with the insertion of Mg into FeTiO3 structure, which could have modified the potential of the conduction band, slightly increasing the band-gap as consequence (Eg=2.59 and 2.55eV for the samples milled with 1wt% of MgO and metallic Mg, respectively). The maximum H2 production was reached with the sample FeTiO3-1.0wt% MgO-2h, which at the same time was one of the samples that exhibited the higher band-gap value; meanwhile the H2 production of the sample milled with metallic Mg (FeTiO3-1.0wt% Mg-2h) was slightly superior to the reached with FeTiO3-2h; this tendency is also followed by the band-gap value of the samples (i.e., Eg(FeTiO3-1.0wt% Mg-2h)>Eg(FeTiO3-2h)).

To the best of our knowledge, there are no reports on the use of materials in the (Mg,Fe)TiO3 system for photocatalytic production of hydrogen. Nonetheless, the performance of the samples used in the present work was compared to the reported in the literature for related systems such as MgTiO3-based materials and Fe-doped TiO2; the evolved amount of hydrogen are summarized in Table2. It can be noted that the materials obtained in this work exhibited superior hydrogen evolution compared to all the Fe-TiO2 materials reported, which is attributed to the position of the conduction band for TiO2, located at less negative potentials compared to FeTiO3. Regarding MgTiO3-based materials, the evolution in this work was found to be superior to the reported by Zhu et al. [39] and Zhang et al. [38], and close to the reported by Wang et al. for MgTiO3 nanofibers [37]. This is an encouraging observation considering that the conduction band of MgTiO3 is theoretically located at more negative potentials than the materials evaluated in this work (i.e., H2 production is thermodynamically favored on MgTiO3 rather than FeTiO3). In 2017, Meng et al. [73] reached significant H2 production using MgTiO3-related materials. However, the complexity of the system (composed of MgTiO3/MgTi2O5/TiO2 heterogeneous belt-junction) along with the synthesis methods, which involved reagents of analytic grade and thermal processing, could represent an important drawback in the scaling-up of the process. Meanwhile, the samples used in this work were obtained by low-cost processing of a natural source. It is believed that the performance of the samples from mineral ilmenite can be further improved by increasing surface area.

Ilmenite concentrates were submitted to HEBM process in the absence and presence of Mg precursors. It was found that neither the crystalline structure of ilmenite nor the morphology of the samples was significantly varied after the milling. XRD, XPS, Raman spectroscopy, and SEMEDS analyses suggested the inclusion of Mg into FeTiO3 structure and evidenced that the substitution of Fe by Mg was more effective when using MgO as Mg precursor, compared to metallic Mg; this was attributed to the difficulty on milling and fracturing the filings of metallic Mg. Both, the HEBM process and the inclusion of Mg precursors led to an increase in the band-gap (Eg) of the samples, estimated from the UVVisDRS results. The first increase, from 2.38 to 2.51eV, was attributed to the release of insulating aluminosilicates into ilmenite grains, while the second increase, from 2.51 to 2.552.59eV, was related to the possible modifications on the conduction band caused by the insertion of Mg into FeTiO3. Greater increases in Eg were observed when milling with MgO. The surface area of the samples was increased from 2.9 to 14.7m2g1 after 2h of milling in the absence of precursors. The presence of Mg precursors affected the fracture of the mineral which hindered the increase of surface area, reaching values of 8.2 and 4.2 m2g1 when milling with MgO and metallic Mg, respectively. The results evidenced reversibility of the substitution reaction, as well as saturation of the FeTiO3, due to the lack of available sites for the substitutional doping to take place. The photocatalytic performance towards H2 production was improved by simply milling the samples (from 240.5 to 255.3molg1h1), due to the increase of surface area. The photoactivity was further enhanced when the samples were milled in the presence of Mg precursors (from 255.3 to 296.9 and 265.2molg1h1 with MgO and metallic Mg, respectively), despite the lower surface area exhibited by these samples; this was attributed to the modifications of the conduction band of the materials that might have been induced by the insertion of Mg, showing that the performance of the samples was affected more by the insertion of Mg than by modifications in surface area. The decrease in H2 production when the nominal concentration added of Mg precursor was increased from 1.0 to 3.0wt%, provided additional evidence of the saturation of the FeTiO3.

Wang C, Hu Q, Huang J, Wu L, Deng Z, Liu Z, Liu Y, Cao Y (2013) Efficient hydrogen production by photocatalytic water splitting using N-doped TiO2 film. Appl Surf Sci 283:188192. https://doi.org/10.1016/j.apsusc.2013.06.080

Ropero-Vega JL, Pedraza-Avella JA, Nio-Gmez ME (2015) Hydrogen production by photoelectrolysis of aqueous solutions of phenol using mixed oxide semiconductor films of BiNbMO (M=Al, Fe, Ga, In) as photoanodes. Catal Today 252:150156. https://doi.org/10.1016/j.cattod.2014.11.007

Malato S, Fernndez-Ibez P, Maldonado MI, Blanco J, Gernjak W (2009) Decontamination and disinfection of water by solar photocatalysis: recent overview and trends. Catal Today 147(1):159. https://doi.org/10.1016/j.cattod.2009.06.018

Malato S, Maldonado MI, Fernndez-Ibez P, Oller I, Polo I, Snchez-Moreno R (2016) Decontamination and disinfection of water by solar photocatalysis: the pilot plants of the Plataforma Solar de Almeria. Mater Sci Semicond Process 42:1523. https://doi.org/10.1016/j.mssp.2015.07.017

Wang Q, Lian J, Bai Y, Hui J, Zhong J, Li J, An N, Yu J, Wang F (2015) Photocatalytic activity of hydrogen production from water over TiO2 with different crystal structures. Mater Sci Semicond Process 40:418423. https://doi.org/10.1016/j.mssp.2015.06.089

Zaraza-Morn ME, Torres-Martnez LM, Moctezuma E, Jurez-Ramrez I, Zermeo BB (2016) Synthesis, characterization, and catalytic activity of FeTiO3/TiO2 for photodegradation of organic pollutants with visible light. Res Chem Intermed 42(2):10291043. https://doi.org/10.1007/s11164-015-2071-9

Gao B, Kim YJ, Chakraborty AK, Lee WI (2008) Efficient decomposition of organic compounds with FeTiO3/TiO2 heterojunction under visible light irradiation. Appl Catal B 83(3):202207. https://doi.org/10.1016/j.apcatb.2008.02.017

Gu D, Qin Y, Wen Y, Qin L, Seo HJ (2017) Photochemical and magnetic activities of FeTiO3 nanoparticles by electro-spinning synthesis. J Taiwan Inst Chem Eng 78:431437. https://doi.org/10.1016/j.jtice.2017.04.003

Kim YJ, Gao B, Han SY, Jung MH, Chakraborty AK, Ko T, Lee C, Lee WI (2009) Heterojunction of FeTiO3 nanodisc and TiO2 nanoparticle for a novel visible light photocatalyst. J Phys Chem C 113(44):1917919184. https://doi.org/10.1021/jp908874k

Zhang X, Li T, Gong Z, Zhao H, Wang L, Wan J, Wang D, Li X, Fu W (2015) Shape controlled FeTiO3 nanostructures: Crystal facet and photocatalytic property. J Alloy Compd 653:619623. https://doi.org/10.1016/j.jallcom.2015.09.029

Ru J, Hua Y, Xu C, Li J, Li Y, Wang D, Gong K, Wang R, Zhou Z (2014) Microwave-assisted preparation of submicron-sized FeTiO3 powders. Ceram Int 40(5):67996805. https://doi.org/10.1016/j.ceramint.2013.11.142

Dai Z, Zhu P, Yamamoto S, Miyashita A, Narum K, Naramoto H (1999) Pulsed laser deposition of ilmenite FeTiO3 epitaxial thin film onto sapphire substrate. Thin Solid Films 339(12):114116. https://doi.org/10.1016/S0040-6090(98)01161-4

Xiao W, Lu XG, Zou XL, Wei XM, Ding WZ (2013) Phase transitions, micro-morphology and its oxidation mechanism in oxidation of ilmenite (FeTiO3) powder. Trans Nonferrous Met Soc China 23:24392445. https://doi.org/10.1016/S1003-6326(13)62752-1

Guan X-F, Zheng J, Zhao M-L, Li L-P, Li G-S (2013) Synthesis of FeTiO3 nanosheets with 0001 facets exposed: enhanced electrochemical performance and catalytic activity. RSC Adv 3(33):1363513635. https://doi.org/10.1039/c3ra22125c

Acosta-Santamara P, Ibat-Soto A, Lpez-Vsquez L (2016) Evaluation of the discoloration of methyl orange using black sand as semiconductor through photocatalytic oxidation and reduction. Int J Chem Mol Nucl Mater Metall Eng 10(10):13351339

Garca-Muoz P, Pliego G, Zazo JA, Bahamonde A, Casas JA (2016) Ilmenite (FeTiO3) as low cost catalyst for advanced oxidation processes. J Environ Chem Eng 4(1):542548. https://doi.org/10.1016/j.jece.2015.11.037

Habibi MH, Habibi AH (2014) Photocatalytic degradation of Brilliant Red M5B using four different nanocomposites (ZnFe2O4, porous ZnFe2O4, ZnFe2O4-TiO2, FeTiO3) coated on glass. J Ind Eng Chem 20(5):29642968. https://doi.org/10.1016/j.jiec.2013.10.066

Tao T, Chen Y, Zhou D, Zhang H, Liu S, Amal R, Sharma N, Glushenkov AM (2013) Expanding the applications of the ilmenite mineral to the preparation of nanostructures: TiO2 nanorods and their photocatalytic properties in the degradation of oxalic acid. Chem Eur J 19(3):10911096. https://doi.org/10.1002/chem.201202451

Caas-Martnez DM, Gauthier GH, Pedraza-Avella JA (2019) Photo-oxidative and photo-reductive capabilities of ilmenite-rich black sand concentrates using methyl orange as a probe molecule. Photochem Photobiol Sci 18:912919. https://doi.org/10.1039/C8PP00315G

Wang L, Yang G, Peng S, Wang J, Ji D, Yan W, Ramakrishna S (2017) Fabrication of MgTiO3 nanofibers by electrospinning and their photocatalytic water splitting activity. Int J Hydrogen Energy 42(41):2588225890. https://doi.org/10.1016/j.ijhydene.2017.08.194

Zhu W, Han D, Niu L, Wu T, Guan H (2016) Z-scheme Si/MgTiO3 porous heterostructures: Noble metal and sacrificial agent free photocatalytic hydrogen evolution. Int J Hydrogen Energy 41(33):1471314720. https://doi.org/10.1016/j.ijhydene.2016.06.118

Linton JA, Fei Y, Navrotsky A (1997) Complete Fe-Mg solid solution in lithium niobate and perovskite structures in titanates at high pressures and temperatures. Am Miner 82(56):639642. https://doi.org/10.2138/am-1997-5-624%JAmericanMineralogist

Yang R, Han A, Ye M, Chen X, Yuan L (2017) Synthesis, characterization and thermal performance of Fe/N co-doped MgTiO3 as a novel high near-infrared reflective pigment. Sol Energy Mater Sol Cells 160:307318. https://doi.org/10.1016/j.solmat.2016.10.045

Kumaravel V, Mathew S, Bartlett J, Pillai SC (2019) Photocatalytic hydrogen production using metal doped TiO2: a review of recent advances. Appl Catal B 244:10211064. https://doi.org/10.1016/j.apcatb.2018.11.080

Raghavender AT, Hoa Hong N, Joon Lee K, Jung M-H, Skoko Z, Vasilevskiy M, Cerqueira MF, Samantilleke AP (2013) Nano-ilmenite FeTiO3: synthesis and characterization. J Magn Magn Mater 331:129132. https://doi.org/10.1016/j.jmmm.2012.11.028

Yang Z, Jiang Y, Zhang W, Ding Y, Jiang Y, Yin J, Zhang P, Luo H (2019) Solid-state, low-cost, and green synthesis and robust photochemical hydrogen evolution performance of ternary TiO2/MgTiO3/C photocatalysts. iScience 14:1526

Linnik O, Chorna N, Smirnova N (2017) Non-porous iron titanate thin films doped with nitrogen: optical, structural, and photocatalytic properties. Nanoscale Res Lett 12(1):249249. https://doi.org/10.1186/s11671-017-2027-7

Jennison DR, Weightman P, Hannah P, Davies M (1984) Calculation of Mg atom-metals XPS and Auger shifts using a SCF excited atom model. J Phys C 17(20):37013710. https://doi.org/10.1088/0022-3719/17/20/018

Fischer A, Kstler H, Schlapbach L (1991) Hydrogen in magnesium alloys and magnesium interfaces: preparation, electronic properties and interdiffusion. J Less Common Met 172174:808815. https://doi.org/10.1016/0022-5088(91)90207-K

Sharma YK, Kharkwal M, Uma S, Nagarajan R (2009) Synthesis and characterization of titanates of the formula MTiO3 (M=Mn, Fe Co, Ni and Cd) by co-precipitation of mixed metal oxalates. Polyhedron 28(3):579585. https://doi.org/10.1016/j.poly.2008.11.056

El-Eskandarany MS (2017) Mechanically induced graphite-nanodiamonds-phase transformations during high-energy ball milling. J Mater Eng Perform 26(6):29742982. https://doi.org/10.1007/s11665-017-2711-x

Lpez T, Hernandez-Ventura J, Gmez R, Tzompantzi F, Snchez E, Bokhimi X, Garca A, (2001) Photodecomposition of 2,4-dinitroaniline on Li/TiO2 and Rb/TiO2 nanocrystallite solgel derived catalysts. J Mol Catal A 167(1):101107. https://doi.org/10.1016/S1381-1169(00)00496-9

Thommes M, Kaneko K, Neimark AV, Olivier JP, Rodriguez-Reinoso F, Rouquerol J, Sing KSJP, Chemistry A (2015) Physisorption of gases, with special reference to the evaluation of surface area and pore size distribution (IUPAC Technical Report). 87 (910):10511069

Qin X, Jing L, Tian G, Qu Y, Feng Y (2009) Enhanced photocatalytic activity for degrading Rhodamine B solution of commercial Degussa P25 TiO2 and its mechanisms. J Hazard Mater 172(2):11681174. https://doi.org/10.1016/j.jhazmat.2009.07.120

Doudrick K, Monzn O, Mangonon A, Hristovski K, Westerhoff P (2012) Nitrate reduction in water using commercial titanium dioxide photocatalysts (P25, P90, and Hombikat UV100). J Environ Eng 138(8):852861. https://doi.org/10.1061/(ASCE)EE.1943-7870.0000529

Saravanan P, Deepika D, Hsu J-H, Vinod VTP, ernk M, Kamat SV (2015) A surfactant-assisted high energy ball milling technique to produce colloidal nanoparticles and nanocrystalline flakes in MnAl alloys. RSC Adv 5(112):9240692417. https://doi.org/10.1039/C5RA16550D

Myriam M, Surez M, Martn-Pozas JM (1998) Structural and textural modifications of palygorskite and sepiolite under acid treatment. Clays Clay Miner 46(3):225231. https://doi.org/10.1346/CCMN.1998.0460301

Shigapov AN, Graham GW, McCabe RW, Paputa Peck M, Kiel Plummer H (1999) The preparation of high-surface-area cordierite monolith by acid treatment. Appl Catal A 182(1):137146. https://doi.org/10.1016/S0926-860X(99)00003-4

Valkov M, Barabaszov K, Hundkov M, Ritz M, Plevov E (2011) Effects of brief milling and acid treatment on two ordered and disordered kaolinite structures. Appl Clay Sci 54(1):7076. https://doi.org/10.1016/j.clay.2011.07.014

Meng L, Ren Z, Zhou W, Qu Y, Wang G (2017) MgTiO3/MgTi2O5/TiO2 heterogeneous belt-junctions with high photocatalytic hydrogen production activity. Nano Res 10(1):295304. https://doi.org/10.1007/s12274-016-1292-6

Dholam R, Patel N, Adami M, Miotello A (2009) Hydrogen production by photocatalytic water-splitting using Cr- or Fe-doped TiO2 composite thin films photocatalyst. Int J Hydrogen Energy 34(13):53375346. https://doi.org/10.1016/j.ijhydene.2009.05.011

Khan MA, Woo SI, Yang OB (2008) Hydrothermally stabilized Fe(III) doped titania active under visible light for water splitting reaction. Int J Hydrogen Energy 33(20):53455351. https://doi.org/10.1016/j.ijhydene.2008.07.119

Sun T, Fan J, Liu E, Liu L, Wang Y, Dai H, Yang Y, Hou W, Hu X, Jiang Z (2012) Fe and Ni co-doped TiO2 nanoparticles prepared by alcohol-thermal method: application in hydrogen evolution by water splitting under visible light irradiation. Powder Technol 228:210218. https://doi.org/10.1016/j.powtec.2012.05.018

Zhao W, Li Y, Zhang M, Chen J, Xie L, Shi Q, Zhu X (2016) Direct microwavehydrothermal synthesis of Fe-doped titania with extended visible-light response and enhanced H2-production performance. Chem Eng J 283:105113. https://doi.org/10.1016/j.cej.2015.07.064

Yan X, Xue C, Yang B, Yang G (2017) Novel three-dimensionally ordered macroporous Fe3+-doped TiO2 photocatalysts for H2 production and degradation applications. Appl Surf Sci 394:248257. https://doi.org/10.1016/j.apsusc.2016.10.077

The authors acknowledge COLCIENCIAS for the partial funding provided through the investigation project No. 20 3. D.M.Caas-Martnez acknowledges the School of Chemical Engineering for the financial support through the Master Scholarship assigned with the contract No.51 of 2015 and through the Doctoral Scholarship assigned with the contract No.69 of 2018. The authors also acknowledge the support of Professor J.A.Henao-Martnez of the X-ray Laboratory (UIS, Colombia) for his collaboration with XRD data, and the Laboratory of Microscopy at UIS, Colombia for the assistance with the SEM-EDS recording.

The authors report partial funding provided by COLCIENCIAS through the investigation project No. 20 3. D.M.Caas-Martnez acknowledges the School of Chemical Engineering at UIS for the financial support through the Master Scholarship assigned with the contract No.51 of 2015 and through the Doctoral Scholarship assigned with the contract No.69 of 2018.

Grupo de Investigaciones en Minerales, Biohidrometalurgia y Ambiente - GIMBA, Universidad Industrial de Santander - UIS, Sede Guatiguar, Transv. Guatiguar, Calle 8N No. 3W-60, Barrio El Refugio, C.P. 681011, Piedecuesta, Santander, Colombia

Caas-Martnez, D.M., Cipagauta-Daz, S., Manrique, M. et al. Photocatalytic hydrogen production using FeTiO3 concentrates modified by high energy ball milling and the presence of Mg precursors. Top Catal 64, 216 (2021). https://doi.org/10.1007/s11244-020-01396-8

Related Equipments