The plant species S. neei inhabits the seacoast of Atacama, and at the mouths of the rivers. It is a halophyte species, so it develops in populations near the coast; the sand where S. neei grows was named “native sand”, to differentiate it from other materials. The town of Bahía Salada (27º18’12 “S - 70º55’53” W, 76 km west of Copiapó) was the only source for obtaining native sand and plant material such as cuttings and seeds.

To obtain the experimental soils, two abandoned mine tailings were sampled; one of them is located on the shore of the Copiapó River (27º21’20 “S and 70º21’14” W, 360 meters above sea level); this is a deposit that contains mine tailings from copper, gold and silver mines and close to an old metal smelter. The second is a deposit that contains mine tailings from a copper mine located at 203 km NE from Copiapó; the material was obtained from a depth from 0 to 50 cm. Each sample was divided using the cone and crushing technique, an 18 mesh screen was used to disperse the material. Each one of the final mine tailings samples was used to make experimental soil mixtures for S. neei cultivation. The experimental soils were prepared from native sand and the two tailings. The prepared substrates were (1) 100% native beach sand, (2) 100% mine tailings 1, (3) 100% mine tailings 2, (4) mixture of sand and mine tailings 1 (1: 1) and ( 5) mixture of sand and mine tailings 2 (1: 1).

The seeds were collected from S. neei plants from Bahía Salada at the end of the summer (February). The seeds were ordered superficially (100 seeds) in five petri dishes (repetitions), containing a 5 mm thick layer of each experimental soil. The seeds were covered with a thin layer of the same experimental soil. The experimental plantations were watered at 50 % of their field capacity (determined for each case) every two days. The proportion (%) of germinated seeds, and emergency seedlings was recorded.

From this trial, seedlings transplanted to pots containing the different mentioned experimental soils were obtained; the growth rate of S. neei was determined in each case, measuring the plants height daily.

To prepare S. neei plants for the bioabsorption assay, cuttings of at least 20 cm of S. neei branches were obtained from Bahía Salada beach. Plant cuttings were induced to root formation by immersion in a commercial solution of IBA (20 % indole butiric acid), for a period of at least 24 hours. Rooted cuttings were transplanted into vials containing the above mentioned experimental soils; at least, 250 plants were distributed in five treatments (experimental soils and controls), five repetitions that included the double number of individual plants, to ensure the repetitions. The plants were maintained in each soil experimental and controls, under conditions similar to those mentioned above for the germination assay. The plants grew in an external environment with natural thermal variation (14 to 25 ºC). From the total of plants, at 30 days, a random sample of five individuals (complete plant) of S. neei by treatment (experimental soils) were dried under shade conditions, until the constant weight was reached.

To determinate the concentration of Fe, Cu, Mn, Mo, Hg, Cd and As in S. neei, 5 g of dry and pulverized plant material of each treatment was processed, using the same protocol for the inert materials (native sand, mine tailings and experimental soils); the procedure will be described in the following section.

The diverse materials were characterized for Fe, Cu, Mn, Mo, Hg, Cd and As in the Scientific and Technological Research Institute of the University of Atacama (Instituto de Investigación Científica y Tecnológica, Universidad de Atacama; IDICTEC-UDA) using standard methods referenced for the analysis of sludge, tailings, soils and plants (Sadzawka et al. 2007, Sandoval et al. 2011).

To analyze the samples of mine tailings (geochemical background or control) and experimental soils, 5 g of each material was dried in an oven at 105 ºC during 6 hours; the analyses were made in triplicate, and the relative humidity was used as a correction factor to calculate concentrations.

To determine the concentration of Fe, Cu, Mn, Mo, Cd and As, 1 to 2 g of each material were used; 25 mL of HNO₃ was added to each sample, the mix was digested on a heating plate at 80 ºC. Once the volume had been reduced at approximately 5 mL, the samples were cooled and 5 mL of perchloric acid was added; this is a common step to determine any element, including Hg. The mixtures were heated until reaching a high viscosity (syrupy) liquid state, at 200ºC; 2 mL of concentrated HCl was then added, and the volume was filled up to 100 mL with distilled water. The Cu and Pb concentrations were determined by flame atomic absorption spectrophotometry using a GBC Avanta atomic absorption spectrophotometer, and a HG-3000 hydride generator was coupled to the spectrophotometer to determinate arsenic concentration.

To determine Hg concentration, the same initial process for non-mercury elements earlier described was used; after the acidic digestion and volume reduction, 10 mL of royal water was added, and the mixture was heated on a heating plate to 200 ºC until the brown vapor dissipated. The mixture was cooled and 15 mL of KMnO₄ was added, then the mixture was stirred and allowed the reaction overnight. If the reaction mixture turned pink, 5 mL of KMnO₄ was added, and it was analyzed after one hour. Finally, 2 mL of NaCl solution and hydroxylamine was added to neutralize the excess of oxidant; the solution was then flush to 50 mL with distilled water. The cold vapor technique to determine the presence of Hg was used, using a Hg analyzer Hg-254N. Concentrations of Hg greater than 25 g/per ton were analyzed by flame atomic absorption. The result was expressed as mg/kg of solid on dry sample basis.

To characterize the granulometry of the materials (sand, mine tailings and experimental soils) a Malvern particle size analyzer (Mastersize Hydro 2000) was used. Samples of 10 g of each material were suspended in 20 mL of distilled water and were scattered by ultrasound during one minute. Calibration curves of the experimental soils were obtained; sand, silt, and clay proportion (%) of each material was determined by interpolation. Each one of the pure and mixed inert materials to S. neei growing was assimilated to a soil type, according to the soil triangle (Sandoval 2011).

To prepare the assay of experimental soils and field capacity, experimental soils were prepared from mine tailing/native sand ratios of 100/0, 80/20, 60/40, 40/60 and 0/100 %. Field capacity was determined according to a common practical protocol: Each experimental soil (mixture) was placed in a PVC cylinder, with a fine synthetic mesh covering the lower end, the initial dry weight of the total system was recorded. Water was added to each cylinder, until it dripped by gravity. When the cylinders stopped dripping, the water content was determined by weight difference. This provided a measure of field capacity, and was used as irrigation reference (Cairo and Reyes 2017).

For the agglomeration of experimental soils by plant trial, experimental soils were placed in transparent (250 mL) polyethylene vials, with 30 replicates for each experimental soil mixture. Each pot was saturated with water at field capacity, and rooted plants were transplanted. The vials were covered with black polyethylene (15 x 20 cm) bags. The transparency of the pots allowed observing the development of the plant roots in each time of sampling. Plants were watered to at least 50 % of field capacity every three days. Assays were evaluated at 80, 110 and 140 days after transplanting. At each sampling period, 10 vials were labeled for each experimental soil (by triplicate) and irrigation was stopped one week before the analysis. In each sampling time the canopy height was measured; the vials were removed, and the weight of fresh plant plus experimental soil were determined. Then, the plants were gently shaken until the experimental soil, which was not retained, stopped pulling away from the roots; the wet plant weight plus the retained experimental soil by roots were recorded.

Each plant was separated from the retained up experimental soil; the retained experimental soil by plant roots was calculated by weight difference; the canopy height (cm) was determined and the canopy and root were separated. The weight (g) of each fresh plant fraction and wet agglomerated and free experimental soil were recorded. All of the previous fractions of plant and experimental soils were dried at 105 ºC until constant weight, and the dry weight of each one was recorded.

The data were analyzed using Student’s t-test for paired samples, that matches samples with the same number of data and different variance. The significance of the differences between treatment results is represented by the symbol “p”, a statistical error α between 5 (p = 0.05) and 10 % (p = 0.1), meaning 90 % and 95 % of confidence; if the value of p is lower than α, the result (difference) is statistically significant.

The germination of seeds (96.7 ± 1.5 %) was statistically similar in the different experimental soils. This result confirms that S. neei can germinate and grow well in mine tailings and mixes. The plants emerged were the same germinated (96.7 %), with 100 % of survival in the trial.

The concentration of elements (mg/kg of dry weight) in the different soils is shown in Table I. In the mine tailings the concentration of Fe (249 000 mg/kg dw), was 75 times higher than in S. neei related to beach sand (field control, 3333 mg/kg dw), and 55 times higher than in S. neei before the trial (4557.9 mg/kg dw). In S. neei post-assay, Fe concentration was 1.4 times higher than in the Bahía Salada wild plants. The concentration of Cu in mine tailings was 946 mg/kg dw, 95 times higher than in wild S. neei from Bahía Salada (10 mg/kg dw), and five times higher than S. neei before the trial (182.6 mg/kg dw). In S. neei post-assay, Cu concentration was 18 times higher than in control plants. The concentration of Mn in mine tailings (185 mg/kg of dw) was statistically similar to the concentration in S. neei before the trial (135.6 mg/kg of dw). No other elements were important. Sarcocornia neei accumulates mainly (99.8 %) Fe, Cu, Mn, and Hg, although the presence of this last element was low.

TABLE I

Element	Sample
	Mine tailings (mg/kg)	Sarcocornia control (from beach) (mg/kg dw) (A)	Sarcocornia post-assay (mg/kg dw) (B)	A-B (%)
Cu	946a	10b	182,6c	94.5
Fe	249000d	3333e	4557,9f	26.9
Mn	185c	0	135,6c	100
Mo	15.6b	ND	<5	ND
As	12.5	0.43	2,84	ND
Hg	0.30	0.14	0,49	ND
Cd	<0.5	ND	<0,25	ND

mg/kg dw = mg/kg per dry weight. ND = Not detected, < x = below detectable limits of equipment. The same letter (a, b, c, d, e or f) added at right of more than one value indicates statistical similarity between them.

The difference of concentration in S. neei between treatments (native sand as control), showed that the plants absorbed 94.5 % Cu, 27 % Fe and 100 % Mn from the environmental content in natural conditions. Although the element present at highest concentrations was Fe, the main gain of S. neei was in manganese, since the native concentration (sand beach) in S. neei was close to zero, showing high affinity to this element when it was available in the experimental soil.

Changes in the growth rate of canopy of S. neei in both experimental soils with mine tailings, between sampling times were evaluated (Fig. 1). The growth response was essentially homogeneous, there was no significant difference (p = 0.15) neither between different experimental soils nor sampling times. This result allowed us to calculate an average overall growth rate of 0.102 + 0.01 cm/day.

Fig. 1

On the other hand, there was no significant difference (p = 0.23) in weight gain response of plants between the different experimental soils, in all the sampling period. However, from T80 (T = time) to T110 days, the weight change rate (Fig. 2) was 40.5 % higher in the experimental soil containing mine tailings 1, and 27.7 % higher in the experimental soil containing mine tailings 2 (p = 0.04). This result had significant implications for the experimental soil retention trial.

Fig. 2

In the evaluation of experimental soil retention (Fig. 3), the proportion of agglomerated experimental soil was statistically similar (p > 0.05) for all the experimental soils in each sampling time; allowing to obtain an average for each sampling time and mine tailings. The proportion (%) of retained experimental soil was significantly different (p < 0.005) for each sampling time and related to T0 sample, indicating that experimental soil retention is a function of plant root development. Confronting the retention of experimental soil, just the retention of experimental soil at T140 was higher than T80 (p = 0.01) using mine tailings 1.

Fig. 3

Using mine tailings 2, the retention of experimental soil was significantly different (p = 0.03) in the three sampling times, it increased over time T80 (19.2 %) < T110 (46 %) < T140 (69.4 %). The matching of retention of experimental soil at each sampling time between mine tailings 1 and 2 showed that retention was significantly higher in experimental soils containing the mine tailings 1 at T80 days (60.2 %, p = 0.02), 110 days (66.7 %, p = 0.03), and at 140 days (82.3 %, p = 0.01).

The size particle analysis of the original mine tailings (Fig. 4) showed that the two tailings are equivalent to a sandy soil (particles above 50 μm), and both of them had lower sand content than the beach sand (100 % sand) (p = 0.01 for mine tailings 1, p = 0.003 for mine tailings 2). The proportion of sand equivalent in the mine tailings was significantly higher than the proportion of silt and clay (p = 0.00013). The native sand from the beach, did not present particles under 50 μm; but, the silt content (50 < x <2 μm) was significantly higher (p = 0.00008) in mine tailings 2 than 1, while clay (x < 2 μm) was significantly higher in mine tailings 1 (12.7 %, p = 0.01) than 2. The difference in the particle size distribution between mine tailings did not mean a difference in field capacity (p = 0.23) between the experimental soils containing different proportion of each mine tailings, then the field capacity of the experimental soils was 20.8 + 0.8 % of soil weight. According to the soil triangle of Sandoval et al. (2011), the native beach sand was equivalent to sandy soil and the experimental soils containing mine tailings 1 and 2 were equivalent to sandy-loam soil, with different proportions of silt and clay. Then, adding the silt and clay fractions, the mine tailing 1 showed higher fine fraction content than 2.

Fig. 4