The particle size has a great effect on the recovery of gold by flotation due to its high density. The flotation is effective for gold particles in the range of 20-200 μm. For finer sizes, the selectivity of the gold decrease because of the floating of ore gangue. For thicker particle sizes, flotation must be done at high densities of the mixture (35 % solids) since this reduces the particle size of sedimentation [8].

There is a range of particle size that presents a greater metallurgical recovery, generally the decrease of this could be observed for thicker and finer sizes of the ore [17]. Recovery decreases for small sizes, which is related to the difficulty of particle/bubble adhesion, because they do not acquire enough kinetic energy to produce a stable particle/bubble aggregate. On the other hand, in the optimum size range the particles are more easily dragged into the foam, since the drainage of the pulp is favored with the increase of the sedimentation rate [18]. It is important to note that, in the primary flotation stage ("rougher" stage) the flotation is carried out with an ore granulometry in which the total release of the gold particles is not necessary, however, in the cleaning stage where the selectivity of the particles is necessary, it is essential to carry out a regrind of the concentrate of the ¨rougher¨ stage for the release of the useful species of the ore [1,8].

The function of the collector is to hydrophobize the surface of the desired mineral, the is the reason why it is the most important chemical reagent used in flotation [19]. The extensive experience in the flotation of minerals allows efficient use of certain types of collectors depending on the types of minerals and the mineralogical associations present where the collectors for the flotation of native gold usually prefer xantats and dithiophosphates, or mixtures of them [11,20,21]. The xanthates used to oscillate from ethyl xanthate to amyl xanthate, In addition, with sodium dithiophosphate, in amounts from 50 g/t to 200 g/t, respectively [2,9]. In applications in which native gold is presented along with submicrometer gold and unreleased gold included in mineral sulfides, the combination of these reagents has been shown to be efficient for the recovery of sulfides and at same way for the native gold; so later, cyanidation can be applied to the concentrates to obtain the gold.

The free gold particles can be recovered very selectively, specially against pyrite, keeping the gold surfaces clean from sludge and organic species, without the use of pH regulators, although with the use of small or zero doses of apolar collectors [20]. It should be taken into account that the flotation reagents require a certain conditioning time to be in contact with the pulp, so this way them could react efficiently on the interesting species of the ore [22]. zxThus, the conditioning stage acquires great importance, since some reagents must be added in the grinding stage to have greater contact with the ore, while others are added directly to the discharge box of the ball mills or to the conditioner [14].

For the flotation of native gold with clean surfaces it is not necessary to add activators, however activation may be required if there are coatings in the gold or if there are gold particles associated with coated pyrite. In such cases, copper sulphate, sodium carbonate (which precipitates calcium and heavy metal ions) sodium sulphide (carefully dosesed) and sulfur dioxide are implemented [7,10].The activators include salts of base metals whose metallic ions are adsorbed on the surface of the mineral particles, modifying their superficial chemical properties; as a consequence, the pH range for mineral flotation can be extended, and the rate of flotation, recovery and selectivity are increased.

The mineral used in the tests belongs to the San Nicolás mine in the municipality of Segovia-Antioquia-Colombia. We worked with a sulfide mineral that contained free gold, whose sample was taken at the discharge of the primary mill. The San Nicolás Mine corresponds to the mining district of Segovia - Remedios, an important mining district of vetiform gold and silver deposits inside of an intrusive granodiorotic monolithic (Batolito de Segovia). Wich is dated from Jurassic-Triassic [23]. The mineral moisture is 3.2 %, with an apparent density (measured with Burette) of 2.7 g/ml and a real density (measured with a Picnometer) of 2.6 g/ml. The granulometric analysis for the sampls to be floated are described in the Fig. 1.

Figure 1 Source: The authors.

The chemical analysis for the sample studied can be seen in the Table 1.

Table 1

Source: The authors.

Ignition losses are close to 41.39 % associated with sulfur, obtaining a higher percentage of pyrite, and in the case of galena, sphalerite and calcopitira as less contributors to the floated sample.

The native gold that is susceptible to flotation is determined by means of an unconventional fire proof, which consists in avoid the generation of the famous "nugget effect", using the Colombian patent generated in the research group CIMEX of the National University of Colombia [24] This method consists in the incorporation of the gravimetric concentration to the conventional tests, allowing to estimate with an accuracy between 1 % and 5%, a precision of 90%, even higher. The gold and silver levels in minerals of an alluvial or vein deposits, are discriminated if the concentration present, in the example, comes from the native element or the element associated with sulfides.

To perform the flotation tests in the laboratory, it was required to manufacture a prototype flash flotation cell (Fig. 3 and Table 3). To arrive at the design of the optimal prototype, some variables are already established in the designs suggested by the industry were we are taken into account [25].

Figure 3 Source: [25]

Table 3

Source: [25]

Based on the parameters previously established and knowing that a special condition is required in the internal geometry of the cell, an internal cone was designed inside the tank, in order to achieve the rises of the particles, therefore reducing the cells area and favoring the ascent of the heavier particles towards the mineralized foam.

(Bernoulli's Principle) The actual volume of the cell is 20 l and its design is shown in Fig. 2.

Figure 2 Source: [26,25]

The microflotation tests carried out in the Hallimond cell were previosly determinated in order the behavior in the selective gold flotation of the different preselected collectors (tionocarbamate and dithiophosphates).

Table 2

Source: The authors.

Among the used reagents were: monoalkyl dithioncarbamate (8474), isoamyl dialkyl dithiophosphate (Aero 3501), dithiophosphinate (AEROPHINE 3418 A), dithionocarbamate (S-9411), allyl thionocarbamates allyl thionocarbamates alkyls (5100), isoamyl dialkyldithiophosphate plus amyl xanthate. (PAX) and allyl, a type dithionocarbamate alkyls plus dithionocarbamate, emphasizing the highest recovery of the first two (88-100%). The flotation conditions were as follows: collector conditioning time 10, 20 or 30 minutes, pH of the solution varying between 4 and 11, water volume of 30 ml, and a glycols type foaming agent.

The Thiol collector-type reagents were selected to favor the chemisorption on metal-type bonds, frequently used in the flotation of metal sulfides.

In Table 4, are present the general characteristics of the used collectors, whose were selected through the microflotation tests of free gold. Both collectors were used for the quantification of gold hydrophobicity by measuring the contact angle, and at the same way to carrying out the flotation tests.

Table 4

Source: The authors.

According to the results obtained from these measurements, could seen that the reagents induce greater hydrophobicity to the gold particles and a better conditioning time for the reagent. Additionally, it was possible to identify the variables to be taken into account in the design of experiments for the flash flotation tests.

Fig. 4 shows the results obtained in the contact angle measurements with the gold particle trough the collector AERO 3501^®, selected for the flotation at different conditioning times.

Figure 4 Source: The authors.

The greater value obtained for the contact angle indicates that the collector induces greater selective hydrophobicity to the gold surface when there is a longer conditioning time and the pH is 11. On the other hand, the reagent S-8474^®, provides selective hydrophobicity to the gold surface with a natural pH and a conditioning time of 10 min.

In Fig. 5 the measurement of the zeta potential for gold, pyrite and galena at different basic pH, are shown. A region of zeta potential is observed (between -25 and -50 mV), of which can be said that the gold particle, pyrite and galena are unstable and susceptible to be collected by a surfactant in the flotation process. When it is taken in place at a pH between 8-11. At pH 11, the zeta potential for gold, pyrite and galena, is close to -50 mV, meaning that at a pH of 11, there will be no separation of gold from its accompanying minerals. This result confirms that activators and depressants should be used.

Figure 5 Source: The authors.

For the measurement of surface free energy, Atomic Force Microscopy (AFM), we made some measurements to a gold particle without a collector and to a gold particle with a collector (Fig. 6).

Figure 6 Source: The authors.

With the AFM measurements, it is sought to establish if the surfactants are decreasing the superficial tension in the surface of the gold particle.

The results obtained from the first measurements are shown in Fig. 7.

Figure 7 Source: The authors.

Comparing both graphs, it can be determined that the energy of the surface with the presence of the collector decreased substantially, which indicates a process of adsorption of the surfactant in the gold-water interface.

With the measurements in the AFM, it is not quantitatively determined which collector induces greater hydrophobicity to the gold surface, but at the same, the effect of different collectors on the surface free energy, as well can be determined.

The flash cell can be used in the discharge of primary mill, in a small mining and/or in the discharge of classifiers in the return, when grinding circuits are closed (see Fig. 8.). This process competes with the use of mercury and the use of jig and centrifugal gravimetric concentrations, whose are currently used.

Figure 8 Source: The authors.

After the results described above, the best upshots of the tests in Flash flotation are shown in Table 5.

Table 5

Source: The authors.

Analyzing the recovery of free gold from the flotation tests carried out with the AERO 3501^® and S-8474^® collectors, it is observed that the best results were achieved by implementing a 10-min conditioning time and 2-min flotation times, with recoveries of free gold particles of 99.28 % and a conditioning time of 10 min and a flotation times of 5 min., with recoveries of free gold particles of 97.22 %, respectively.

Because of the characteristics of the stream in which a Flash cell is installed, the float gold is thick (greater than 150 μm), therefore this technique is frequently used for the rapid concentration of minerals that have reached their releases in the discharges of the classifiers, allowing to increase the global recovery of the plants, whose main objectives are to avoid the overmoulding of released minerals and at the same way in the recirculation of the stream in the mill.

The main benefits with the use of flash flotation Cells are:

Lower stock market

The use of mercury in Colombian mining has generated serious environmental and technological impacts, which historically make a detrimental association between gold mining and the soil contamination, the surface water and other actors in the environments around the mines [27].

There is a good correlation between artisanal mining and the increase of mercury losses in th the enrichment processes of gold-bearing minerals. While the large-scale mining reports orders of 0.001 g of mercury per gram of gold produced, small-scale mining reports losses of up to 14.5 g of mercury per gram of produced gold on average, whose can amount to 66.5 g of mercury per gram of gold with the minimum values of 7.5 g of mercury per gram of the produced gold [28], which implies global losses of Mercury around 55.1 t/year just in 5 municipalities of Northeast Antioquia: (Segovia, remedies, Nechí, El Bagre, Zaragoza).

It is estimated from the 100 % of the lost, 34 % is emitted to the atmosphere and 66 % goes to the water streams [28]. The technology, traditionally used by miners in these regions, consists of the use of amalgamating coconuts, in which it is estimated that around 50% of the mercury is lost in the process. This project is basically aimed at the elimination of a historical problem, which has been the use of mercury in gold mining, to search the developing of the least environmental impact and to generate conditions that allow improving the relations between mining and the environment in pro of the sustainability of the exploitation of gold.