Divalent transition metal complexes of nitrogen, oxygen and sulfur containing ligand: design, structural, spectral, pH-metric, theoretical molecular modeling, analytical and mechanism studies

In the present study, the 1-Nicotinoyl-4-phenyl thiosemicarbazide (H2NPT) multidentate nitrogen, oxygen and sulfur containing ligand is being synthesized in a simple single step reaction by condensing 1:1 molar ratio of ethanolic solutions of both phenyl isothiocyanate and nicotinoyl hydrazine. The prepared H2NPT and its complexes with Co2+, Ni2+ and Zn2+ and Cu2+ were characterized by pH-metric titrations, elemental analysis, FTIR, electronic spectra and thermogravimetric (TGA) analyses. The association constant of the ligand and the stability constants of its complexes were calculated by pH-metric measurements in 50% ethanol–water mixture. Also, DFT method was used to draw the geometry of all compounds and the parameters such as bond lengths, bond angles, dipole moment, Frontier orbitals (HOMO, LUMO), MEP and other energetic parameters (optical energy gap, softness, hardness, electronegativity) were evaluated. The prepared H2NPT reacted with Co2+, Ni2+ and Zn2+ and Cu2+ metal ions to form colored precipitates that were readily floated to the surface solution using oleic acid (HOL) surfactant with vigorous shaking of the flotation cell. The heavy metals' concentrations in the dissolved precipitates are determined by flame atomic absorption spectrometry (FAAS). The different parameters affecting the flotation process were thoroughly investigated, viz. pH of sample solution, concentration of ligand, metal ions and HOL, temperature and interfering ions. Nano-gram quantities of Cu2+, Co2+, Ni2+ and Zn2+in 1500-ml samples are quantitatively determined with 1-Nicotinoyl-4-phenyl thiosemicarbazide (H2NPT) at pH 3.0–6.5 for Cu2+ and at 7.0–7.5 for Co2+, Ni2+ and Zn2+. The concentrations of these heavy metals are increased 300-fold. Interferences, on the flotation process, from various foreign ions were avoided by adding excess H2NPT. The mode of chelation between H2NPT and the investigated metal ions is proposed to be through the S, N and/or O atoms of the ligand. The mechanism of flotation is proposed to be due to an electrostatic interaction between the HOL surfactant and the formed metal–ligand complexes through H-bond formation. The proposed multi-element flotation methodology has been utilized for the determination of the analytes in certified reference materials, alloys, synthetic mixtures and in water samples with a recovery % more than 90% and a relative standard deviation (RSD), as a precision < 2%.


Introduction
Environmental pollution of heavy metals is increasingly becoming a problem and has become of great concern due to the adverse effects it is causing around the world. These inorganic pollutants are being discarded in our waters, soils and into the atmosphere due to the rapidly growing agriculture and metal industries, improper waste disposal, fertilizers and pesticides (Soylak et al. 2007).
Over the last few decades, water contamination due to population growth and high wastewater generation levels has become a serious concern. In order for heavy metals' ability to persist in the ecosystem, they are among the most common wastewater pollutants. The used separation and preconcentration techniques for wastewater treatment and recovery from heavy metals are being researched Tuzen et al. 2008).
Industrial operations are the main resources of heavy metal pollution, such as metal processing, fertilizers, paints, electroplating, batteries, and printing systems ; and natural operations, such as weathering, volcanoes, and the degradation of bedrocks containing some of the high-level toxic metal ions (Akl et al. 2022a;Serage et al. 2022).
Finding analyte traces in a complex sample matrix where their concentrations are close to the analytical system's detection limit and background interferences are negatively affecting the analytical signals, is one of the challenges of environmental analysis. Numerous sample cleaning, enrichment, and preconcentration techniques have been developed to address this issue. There are diverse techniques for heavy metals separation and removal from various samples, as solvent extraction, solid-phase extraction, flotation, ion exchange, and precipitation Akl et al. 2016Akl et al. , 2021aAkl et al. , 2021bAkl et al. , 2022bAkl et al. , 2023aAkl et al. , 2023bMostafa and Akl 2023;Aly et al. 2020;Ibrahim et al. 2019;Mostafa et al. 2021;Nayl et al. 2020;Saleh et al. 2021;Shoueir et al. 2017a, b). Some of these methods have drawbacks like lengthy separations, volume restrictions on the sample solutions that can be studied, time-consuming, multi-stage processes, lower enrichment factors, and the use of organic solvents that are harmful to the environment.
Due to a number of benefits, such as higher enrichment factors, the absence of emulsions, safety with regard to hazardous samples, minimal costs due to low reagent consumption, affordable instruments, flexibility, and easier incorporation into automated analytical techniques, flotation has gained popularity in comparison to other concentration techniques. Several methods have been developed for the flotation of heavy metal ions (Abdallah et al. 2004;Akl et al. 2002Akl et al. , 2003Akl et al. , 2005Akl et al. , 2006Akl 2006;Ghazy and Kabil 1994;Khalifa et al. 2001).
During the last two decades, interest in the synthesis of metal complexes of polydentate N and S ligands (El-Saied et al. 2003;El-Shazly et al. 2005)was related to their application as enzyme modulization and in medical imaging. Recently, the use of these ligands was extended to environmental problems because of their high complexing properties toward heavy metals which allowed metal de-pollution.
The current study was performed with these objectives: 1. Design and synthesis of N, O and S donor 1-nicotinoyl-4-phenylthiosemicarbazide (H 2 NPT) for the preconcentration of Cu 2+ , Co 2+ , Ni 2+ and Zn 2+ metal ions and separation by flotation followed by flame atomic absorption (FAAS) determination. 2. Characterization of H 2 NPT and its metal complexes using pH-metric titration, elemental analysis and spectroscopic (FTIR, electronic) and TGA instrumental performances.
3. Flotation separation experiments utilizing Co 2+ , Ni 2+ and Zn 2+ and Cu 2+ metal ions as pollutants. 4. Investigating the ideal parameters required for optimal separation of the studied metal ions, e.g., pH, temperature, the initial concentration of the three investigated metal ions, the concentration of H 2 NPT, and the oscillation time, as well as the interfering ions. 5. To study, also, the theoretical modeling using DFT method. 6. Studying the applicability of the proposed NOS ligand for selective separation of Cu 2+ from aqueous solutions containing Co 2+ , Ni 2+ and Zn 2+ . 7. Elucidation of the mechanisms involved in the processes of chelation between H 2 NPT and Co 2+ , Ni 2+ and Zn 2+ and Cu 2+ metal ions 8. Elucidation of the mechanisms involved in the process of flotation Co 2+ , Ni 2+ and Zn 2+ and Cu 2+ metal ions using H 2 NPT and HOL surfactant.

Reagents
Oleic acid (HOL) stock solution, 6.36 × 10 −2 moll −1 , was prepared from the food grade with sp. gr. 0.895 by dispersing 20 ml of HOL in one liter of kerosene. Two certified reference samples, lead-zinc sulfide ore-OCrO (COD 161-96) provided by MBH analytical Ltd. (Barnet, UK) and stream sediment SARM 52 prepared by MINTEX (Ranburg, Republic of South Africa), were analyzed to verify the precision and accuracy of the proposed flotation technique. The Chemistry Administration, Cairo, Egypt, provided the samples of nickel alloy and serpentine.

Apparatus
All glassware were soaked overnight in chromic mixture (K 2 Cr 2 O 7 + concentrated H 2 SO 4 ), rinsed thoroughly with double distilled water (DDW) and dried in a vacuum oven. Two types of flotation cells were used as previously reported (Ghazy and Kabil 1994). Analysis of Co 2+ , Ni 2+ and Zn 2+ and Cu 2+ was performed by Perkin Elmer 2380 flame atomic absorption spectrometer. C, H and N contents of the investigated complexes were determined at the micro analytical Unite, Cairo University, Egypt. The FTIR spectra of the ligand and its solid-state isolated complexes were recorded on. A digital pH meter (HANNA Instrument 8519) was used to measure the pH values of sample solutions.

Preparation of 1-nicotinoyl-4-phenylthiosemicarbazide (H 2 NPT)
The ligand, Scheme 1, was prepared by condensation of 1:1 molar ratio of ethanolic solutions of both phenyl isothiocyanate and nicotinoyl hydrazine with 5 drops of glacial acetic acid. The reaction mixture was heated and refluxed on a water bath for 3 h. The precipitate formed was filtrated, recrystallized from EtOH and dried.

In aqueous solutions (a) [M 2 (NPT) 2 ]
The acetate salts of Co 2+ , Ni 2+ and/or Zn 2+ in 30 ml water (0.124g, 0.5mmol) were added to 30 ml H 2 NPT (0.2g, 1mmol) in ethanol. The solution mixtures were refluxed for 3 h. The solutions' volumes were reduced by evaporation till the formation of green, yellowish-green and/or white precipitate for the Co 2+ , Ni 2+ and Zn 2+ ions, respectively.
In oleic acid The complexes isolated from the scum layer were obtained by mixing molar amounts of both ligand and metal ion at the recommended ratio (M:L), pH and oleic acid concentration. The float was gathered by filtration in a sintered glass Gooch (G 4 ), washed several times with DDW, ethanol and finally with diethyl ether. The precipitate was dried in an oven at 80 °C and preserved in a desiccator.

pH-metric titration measurements
The pH metric measurements were performed using a Met Rohm E 536 Potentiograph equipped with a 665 DOSIMAT (Met Rohm, Herisau, Switzerland). The combined glass electrode was standardized before and checked after each titration with buffer solutions produced by Fishers, New Jersey, USA. All titrations were carried out at room temperature 25 ± 0.05 °C. This was adjusted by circulating water from an ultra-thermostat (Kottermann 3140, Germany) through the annular space of a double walled Pyrex titration cell of 50 ml capacity. The titrant (alkali) was added from the Dosimeter and the content of the titration vessel was stirred magnetically. The following solutions were prepared and titrated with standardized sodium hydroxide solution (0.99 × 10 −2 moll −1 ) at constant ionic strength (0.01 moll −1 KCl).
All the above solutions were completed to 25 ml with DDW and titrated against 0.99×10 −2 moll −1 NaOH.

Theoretical computation for geometry and binding affinity
The structures of the H 2 NPT ligand and its Co 2+ , Ni 2+ , Zn 2+ , and Cu 2+ complexes were obtained using generalized gradient approximation (GGA) with DFT calculations using the DMOL 3 tool in the Material Studio (MS) package, with optimized BLYP correlation function method and Double Scheme 1 Synthesis of 1-nicotinoyl-4-phenylthiosemicarbazide (H 2 NPT) Numerical polarization basis-set (DNP). Vibrational frequency calculations were performed to ensure that each structure had the lowest potential energy surface. The highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) are very important factors in theoretical molecular design. HOMO is the electron donor and the LUMO is the electron acceptor site. The molecular hardness and softness of the compound can be predicted from its HOMO-LUMO gap. HOMO/LUMO energies were estimated at the DFT/BLYP level and optimized geometries were obtained (Abdallah et al. 2022).

Separation procedure
For the separation and FAAS determination of Co 2+ , Ni 2+ and Zn 2+ and Cu 2+ , 2 × 10 −5 moll −1 of each analyte and 2 × 10 −3 moll −1 of H 2 NPT were introduced into an Erlenmeyer's flask. The mixture was shaken well for 2 min to allow complete complexation. The pH 3 was adjusted by HCl. The content was quantitatively transferred to flotation cell and completed to 10 ml with DDW. 2ml of 2 × 10 −4 moll −1 HOL were added and the cell was shaken well by hand. Vigorous shaking of the flotation cell creates air bubbles that help the flotation of the formed complexes. The scum layer [containing Cu 2+ complex] was separated from the mother liquor [containing Co 2+ , Ni 2+ and Zn 2+ complexes], eluted with 2 ml of concentrated HNO 3 and determined by FAAS. The pH of the mother liquor was adjusted to 7.0 by NaOH. The same flotation steps were followed in presence of 2 ml HOL (2 × 10 −4 moll −1 ). The complexes in the scum layer were eluted using 2 ml conc. HNO 3 . The eluted solution was directly aspirated to the flame for AAS determination of Co 2+ , Ni 2+ and Zn 2+ at the specified wavelength of each element. The flotation efficiency is calculated from the relation: where C i and C s are the concentrations of the analyte in the initial aqueous and scum layers, respectively.

Analysis of water samples
To 10 ml of clear filtered water samples, different concentrations of Co 2+ , Ni 2+ and Zn 2+ and Cu 2+ were introduced into the flotation cell followed by 2ml of 0.01 moll −1 of H 2 NPT and the same previous steps were carried out.

Analysis of certified reference materials
Accurately 0.5 g of the solid sample was completely dissolved in Teflon beaker with a mixture of HF (45 ml) + H 2 SO 4 (15 ml) + HNO 3 (15 ml). After complete dissolution, the solution was evaporated till dryness. The residue was dissolved in 20 ml HCl (1:1), and completed to 100 ml with DDW. 0.01 ml of this solution was taken with 2ml of 0.01moll −1 H 2 NPT. The same previous steps of flotation, metal elution and FAAS determination were carried out.

Analysis of alloys
The solid sample (0.5 g) dissolved in aqua regia was completed to 100 ml using DDW. To 0.5 ml of the dissolved sample, 2 ml (0.01 moll −1 ) H 2 NPT was introduced into a flotation cell. The same previous steps for flotation, separation, elution and FAAS determination were carried out.

pH-metric studies
The ionization constant is the most important property for the organic molecules science it measures their degree of acidity and indicates the suitable conditions for their isolation in a maximum yield. The pH-metric titration found

GCRD Equation
The energy gap (ΔE gap ) ΔE gap = (E LUMO -E HOMO ) Ionization potential (I P ) The back donation (Gomez et al., 2006) itself the more easily technique that is used to determine the deprotonation constants for the investigated ligand and abstract a suitable meaning for their values. Also, it gives an idea about the molar ratios of the complexes and their formation constants. All of such important findings can be obtained using different empirical equations. Under such investigation 1-nicotinoyl-4-phenylthiosemicarbazide (H 2 NPT) was prepared and the ionization processes were studied. The ionization constants have been calculated from the titration of the acid (HCl) in the presence and absence of the studied ligand. The titrations were carried out at constant ionic strength (0.01moll −1 KCl). The obtained curve, Fig. 1, for the investigated ligand shows that the ionization takes place in two separate steps. It is shown from the Figure that the ligand curve lies below that of HCl. Such behavior is normal due to the liberation of hydrogen ions during the ionization of the ligand. The average number of protons associated per ligand molecule (n − A ) at different pH values can be easily calculated from the ligand titration curve using Irving and Rossotti equation (Irving & Rossotti 1954): where Y is the ionizable protons of the ligand, V 1 and V 2 are the volumes of alkali required to reach the same pH value in free HCl and H 2 NPT mixture, respectively. V o is the initial volume of the titrated mixture, T L is the H 2 NPT concentration in the initial volume, A and B are the concentrations of free HCl and NaOH, respectively. The n − A values for each investigated H 2 NPT are graphically plotted against the pH (Fig. 2).
The ionization constants of H 2 NPT are determined through a plot between the pH and the average number of protons associated per ligand molecule ( n A ). Its pK 1 H and pK 2 H were determined by interpolation at n A = 0.5 and 1.5, respectively. The values of pK 1 H (10.45) and pK 2 H (7.35) refer to the stepwise dissociation of the amide and thioamide protons. The results clearly refer to H 2 NPT ionization facility. The values agree with the ability of H 2 NPT to complex through its deprotonated form.
When the same titrations were done in the presence of metal ions, their titration curves were noticeably below that of the H 2 NPT indicating chelate formation between the H 2 NPT and the studied metal ions with the liberation of amide and/or thioamide protons of H 2 NPT (Scheme 1).
The successive stability constants of metal-H 2 NPT complexes are determined by plotting graphs between the values of n − (average number of bound ligand molecules per metal ion) and pL (free ligand exponent). The obtained β 1 and β 2 values are shown in Table 2. From the calculated values we can see that: (1) The value of β 1 is larger than that of β 2 , indicating that the vacant sites of the metal ions are more freely available for binding the first ligand molecule than the second ligand molecule.
(2) The overall stability constant increases in the order: Cu 2+ > Ni 2+ > Co 2+ > Zn 2+ . This variation may be due to the difference in the electric field strength (f* = z*/r 2 ) where z* and r are the effective charge and radius of the metal ion, respectively.   (Table 3).

FTIR spectra
The FTIR spectrum of H 2 NPT, Fig. 3 Fig. 3(ib), the N 1 H and C=O bands disappeared with the appearance of new bands at 1572 and 1200 cm −1 due to ν(C=N) and ν(C-O) vibrations, respectively, indicating that H 2 NPT reacts through its enolic carbonyl. The intensity of ν(N 2 H) is lowered without any shift indicating its involvement in coordination. For the other complexes, Fig. 3(ii-iv) the ligand behaves as binegative tetradentate with the two metal ions. It chelates with one of the metal ion through C=N 2 (after deprotonation through thioenolization) and C-O where the second metal ion is bonded through C=N 1 (after deprotonation through enolization) and C-S. This conclusion is supported by the disappearance of ν(N 1 H), ν(N 2 H), ν(C=O) and ν(C=S) with the appearance of new bands at 1596, 1550, 1198 and 630 cm −1 due to ν(C=N 1 ), ν(C=N 2 ), ν(C-O) and ν(C-S), respectively. Also the appearance of new bands at 375, 450 and 495 cm −1 due to ν(M-N), ν(M-S) and ν(M-O) is a further support for the bonding through nitrogen, sulfur and nitrogen atoms (Ghali and Hyzwany 2017;Lever 1968;Monier et al. 2014).

Electronic spectra
The geometry of the investigated complexes has been elucidated by the measurement of the magnetic moment and recording the electronic spectra. The spectrum recorded in Nujol for the green Co 2+ complex shows broad bands at shows two bands at 14.367 and 24.39 cm −1 attributed to the 3 A 2 g → 3 T 1 g (F) and 3 A 2 g → 3 T 1 g (P) transitions, respectively, in an octahedral geometry. Also, the magnetic moment value (3.1 B.M.) for one nickel atom is in the range (2.8-3.2 B.M.) measured for the aforementioned geometry (Mostafa et al. 2000).The spectrum of the Cu 2+ complex (Scheme 2) shows one broad band at 17.391 cm −1 assigned to the 2 E 2 g → 2 T 2 g transition in a distorted octahedral geometry. The distortion is due to Jahn-Teller effect (Beecroft et al. 1974) (Fig. 4).

Thermogravimetric analysis
The thermogravimetric (TGA) and derivative thermogravimetric (DTG) analyses in the temperature range 25-1000 °C for the Cu 2+ and Ni 2+ complexes recorded the weight loss against temperature. The Cu 2+ complex has a thermal stability till 137 °C and the first degradation step (137-240 °C) refers to the removal of the coordination water and C 6 H 5 NH as a fragment of the ligand. The Ni 2 (NPT) 2 complex shows a thermal stability till 192 °C after which two decomposition steps are observed. The first decomposition is due to the removal of C 7 H 6 N 3 OS with weight loss of 46.6 (Calcd. 49.8%). The second stage starts at 360 °C which includes the removal of the ligand rest with loss of 29.9 (Calcd. 27.9%). The difference in the values may be due to overlapping between the two steps. The residue is NiO with a percent of 21.0 (22.4%) (Fig. 5).

Geometry optimization
The optimized geometry of the H 2 NPT ligand and its Co 2+ , Ni 2+ , Zn 2+ and Cu 2+ complexes are shown in Fig. 6.

Global Chemical Reactivity Descriptors (GCRD)
The two most important components of theoretical molecular design are highest occupied molecular orbitals (HOMO) and lowest unoccupied molecular orbitals (LUMO). The HOMO-LUMO gap can be used to predict the molecular hardness and softness of a compound because the HOMO and LUMO sites are electron donor and acceptor sites, respectively, which the HOMO and LUMO of optimized geometry of the H 2 NPT ligand and its Co 2+ , Ni 2+ , Zn 2+ and Cu 2+ complexes are shown in Fig. 7. The Global Reactivity such as ionization potential (IP), electron affinity (EA), hardness (η), softness (σ), electronegativity (χ), electrophilicity index (ω), chemical potentials (µ)] can be determined from the HOMO and LUMO orbital energies through Koopman's theorem. The energy gap (ΔE), electronegativity ( ), hardness ( ), chemical potentials (µ), softness ( ), and electrophilicity index (ω) are illustrated in Table 5. The energy gap values designate the reactivity of the compound for the metal surface (as the energy required for the transition from HOMO to LUMO increase, the reactivity decrease).
The ΔE (LUMO-HOMO) for Cu-complex was found to be less reactive than of the ligand H 2 NPT. On other hand Co(II), Ni(II) and Zn(II) complexes ware found to be more reactive than of the ligand H 2 NPT.
Soft molecules (σ) have a small energy gap compared to hard molecules (η), which have a larger energy gap. A soft molecule is more reactive than a hard molecule because a soft molecule has a lower ΔE (LUMO-HOMO). From Table 1, Co 2+ , Ni 2+ and Zn 2+ complexes are softer than ligand H 2 NPT  (NPT) and this confirms that Co 2+ , Ni 2+ and Zn 2+ complexes are more reactive than ligand H 2 NPT.
The (x) is a measure of power of atom(s) to attract the electrons. A high value of electronegativity (χ) suggests strong ability to attract electrons from the ligand, which leads to greater interaction to form the complex, decreases according to the following order: Chemical potential (μ), which estimates the escaping capability of electrons from the equilibrium framework, increases according to the following order: The energy components assessed by DFT are shown in Table 5. The increased value of the calculated binding energy of complex linked to that of the ligand, signifies the higher stability of the designed complexes than the free ligand (Abdallah et al. 2022;Younis et al. 2021).

Molecular electrostatic potential (MEP) maps
Molecular electrostatic potential (MEP) maps, that are used to locate positive and negative regions as well as to predict reactive sites in molecules, are shown in Fig. 8. These maps can be used to comprehend the potential for nucleophiles or electrophiles to interact in chemical species at more advantageous locations. Different colors on the MEP surfaces, including green, orange, blue, red, and yellow, indicate different levels of electrostatic potential within the chemical structures. It was found that the ascending order of electrostatic potential is represented by the colors in the following ranges: red > orange > yellow > green > blue. The MEP map's red-highlighted area, which shows the area of negative potential and might be the most pertinent, was provided by the atoms. Contrarily, the area of positive potential and perhaps the most suitable location for nucleophilic engagement is indicated by the region that is highlighted in blue or green. In the current study, ligand oxygen atoms exhibit negative electrostatic potential, whereas nitrogen and sulfur atoms exhibit positive electrostatic potential (Noreen and Sumrra 2021). Table 6 displays the energy components that were evaluated by DFT. The calculated binding energy of the complexes increased in value when compared to that of the ligand, indicating that the designed complexes are more stable than the free ligand.

Effect of pH
A group of experiments were performed to float 2 × 10 −5 moll −1 of each Co 2+ , Ni 2+ , Cu(II) and Zn(II) individually or     in combination in presence of 2 × 10 −4 moll −1 H 2 NPT and 2 × 10 −4 moll −1 HOL that is still less than the critical micelle concentration (CMC) at different pH values. Figure 9 shows that, the floatability of Cu 2+ gradually increases from 35% to nearly 100% at the pH range 3.0-7.5. On the other hand, the floatability of Co 2+ , Ni 2+ , and Zn 2+ gradually rises from zero to nearly 100% at pH 6.6-7.5 above which it decreases. The decrease may be attributed to the formation of white emulsion and excessive foam of sodium oleate that is formed in alkaline medium. The results in Fig. 9 proved that selective and quantitative separation of Cu 2+ was carried out at pH 3.5, whereas Co 2+ , Ni 2+ , and Zn 2+ were separated at pH ~ 7. Accordingly, all subsequent experiments were carried out at the recommended pH values.

Effect of H 2 NPT concentration
The separation efficiency of Co 2+ , Ni 2+ , Cu 2+ and Zn 2+ was determined in presence of only different HOL concentrations (15-40%) in the absence of chelating ligand. Unsatisfactory results obtained entailed the use of a complexing reagent. 1-Nicotinoyl-4-phenylthiosemicarbazide (H 2 NPT) has been found suitable for reaching the maximum separation percentage of traces of Co 2+ , Ni 2+ , Cu 2+ and Zn 2+ . The data in Fig. 10 shows the role of H 2 NPT on the flotation of the analytes. Suitable concentration of HOL (2 × 10 −4 moll −1 ) was added to (2 × 10 −5 moll −1 ) of each analyte in five series of flotation cells. To each series, different concentrations of H 2 NPT were added after adjusting the pH at the recommended values. Figure 10 shows that Cu(II) reaches complete separation at Cu 2+ :4H 2 NPT ratio, where Co 2+ , Ni 2+ , and Zn 2+ at 1:5 ratio. Excess of H 2 NPT does not affect the separation efficiency facilitating the work in real samples with unknown analyte concentrations. 2 × 10 −4 moll −1 of H 2 NPT was selected as a suitable concentration for subsequent experiments.

Effect of HOL concentration
The impact of HOL on the separation of 2 × 10 −5 moll −1 of each analyte was studied in the presence of 2 × 10 −4 moll −1 of H 2 NPT at the recommended pH values. From the data (Fig. S1) it can be noticed that, all the analytes have maximum separation at a wide range of HOL concentrations. 2 × 10 −4 moll −1 of HOL was selected throughout this work.

Effect of temperature
A separation efficiency of 2 × 10 −4 moll −1 for each analyte at different temperatures was investigated under the previously recommended conditions of pH, H 2 NPT and HOL concentrations. The results (Fig. 12) show that the flotation efficiency of the investigated specimens is unaffected up to 40 °C and decreases with increasing temperature. The decrease in flotation efficiency at high temperatures could be attributed to increased sediment solubility and foam instability, resulting in partial dissolution of the foam and reduced foam consistency to retain it. Weak hydrogen bonds between the HOL surfactant and the complex are also broken by the temperature increase. All experiments were performed at room temperature (25 ± 2°C).

Induction time (τ)
The induction time (τ) is the time necessary to incorporate traces of the investigated elements in the collector precipitate. The experiment showed that, under the prevailing conditions, H 2 NPT could quantitatively collect Cu 2+ , Co 2+ , Ni 2+ and Zn 2+ during 1-5 min. However, in practice an induction time of 5 min was sufficient to perform the investigation. The reasonably fast kinetics of the ligand-metal ion interaction at optimum pH reflects a good accessibility of the chelating sites of the ligand and higher binding constants of metal ions with H 2 NPT.

Effect of sample volume
To determine the maximum applicable sample solution (or minimum analyte concentration), we investigated the effect of sample solution volume (100-2000 mL) on the recovery of each analyte (Fig. S2). Analyte recovery is constant up to 1500 mL, even with varying volumes of solution containing a constant amount of each analyte (100 μg). A concentration factor of 300 can be achieved using 5 ml of eluate. From this we can conclude that Co, Cu, Zn, or Ni ions can be measured at a concentration of 66 ng ml −1 .

Effect of foreign ions
The effect of foreign ions on the separation of 2 × 10 −5 moll −1 of the analytes in presence of 2 × 10 −4 moll −1 of H 2 NPT using 2 × 10 −4 moll −1 HOL was studied at the recommended pH. The data are demonstrated in Table 7. The dipositive Fig. 12 Effect of temperature on the separation of 2 × 10 −5 moll −1 of Co 2+ , Ni 2+ , Zn 2+ and Cu 2+ at the recommended conditions Table 7 Effect of different foreign ions on the separation of 2 × 10 −5 moll −1 of Co 2+ , Ni 2+ , Cu 2+ and Zn 2+ in presence of (a) 2 × 10 −4 moll −1 , (b) 2 × 10 −3 moll −1 of H 2 NPT using 2 × 10 −4 moll −1 of HOL at the recommended pH ions of Mg, Ca, Sr, Ba, Pb and Cd have no effect on the separation efficiency of the analytes; the other studied foreign ions interfere. The interferences were completely eliminated by using excess concentration of H 2 NPT (2 × 10 −3 moll −1 ) to cover the requirements for complexation of foreign ions.
To study the effect of ionic strength, sodium, potassium, calcium and magnesium as chloride, nitrate, sulfate and iodide salts were added during the separation of analytes at the recommended conditions. It was found that most of these salts, even up to 0.5 moll −1 have no effect on the separation efficiency, this facilitates the separation of analytes from natural seawater solutions.

Analytical figures of merits
The selected wavelengths of the resonance lines and other standard conditions for AAS measurements of the examined metal ions were obtained from the instrument manual. Absorbance was found to be linear over the concentration range of 0.1 to 10 mg ml −1 . Each measurement was recorded three times to ensure reproducibility (erasure) of the recorded signal. The instrumental quantification limits (DL)i for the tested metal ions were calculated using Eq. (3).
The selected wavelength of the resonance lines and other standard conditions for AAS determination of the investigated metal ions was taken from the instrumental manuals. The absorbance was found to be linear in the concentration range 0.1-10 µg ml −1 . For reproducibility of the recorded signals (absorbance), the measurements were reported three times in each case. The instrumental determination limits (DL) i of the investigated metal ions were calculated using Eq. (3).
The S.D. and the mean absorbance in Eq. (3) are for the least possible measured concentration, C − M n+ by the used instrument. The detection limits are very impressive for the studied metal ions. The analytical detection limits (DL) a were calculated by dividing the instrumental detection limit by the preconcentration factor (300 in the present study). The instrumental detection limit for the investigated metal ions is 4-10 ng ml −1 , while the analytical limit is 0.013-0.035 ng ml −1 .
In order to validate the multielement flotation procedure, the proposed method was applied to two certified reference materials, viz.: lead zinc sulfide ore-OCrO  and stream sediment SARM 52 and two alloy samples; nickel and serpentine alloy samples. The average contents of metal ions (n = 5) in the certified reference materials and alloy samples are shown in Table 8 with a precision as relative standard deviation (R.S.D.) < 1%.

Mechanism of flotation
There are two main types of interactions involved in flotation viz physical and electrostatic interactions (Beecroft et al. 1974;Ghazy 1994;Mostafa et al. 2000)The predominance of the electrostatic mechanism is mainly governed by such important factors as the size and charge of ionic species involved, the presence of electronegative atoms capable of forming hydrogen bonds in the ligand structure, and charged sites on the solid species and precipitates. Based on the existence of some experimental factors, it seems reasonable to assume the electrostatic mechanism for the investigated system.   Fig. 3(i), isolated in the scum layer is characterized by the appearance of new strong bands at 3431 and 1371 cm −1 due to the stretching and bending vibrations of OH group. The hydroxo group may be present due to the addition of NaOH during the flotation process. The appearance of new bands at 2850 and 1100 cm −1 is due to the stretching and bending vibrations of the aliphatic hydrocarbons in oleic acid. The appearance of new bands at 1715 and 463 cm −1 due to ν(CO) of carboxylic group and ν(M-O) is strongly supported the OH bonding.
The OH group may form hydrogen bond with the complex and oleate. Strong evidence comes from the dissolution of the floated precipitate on heating to 50°C.
(ii) Examining the spectra of both aqueous and floated Co 2+ , Ni 2+ and Zn 2+ complexes, Fig. 3(ii-iv), reveals the appearance of new band at ~ 2850 cm −1 due to ν(C-H) of oleic acid; its intensity increases in the order: Ni 2+ > Co 2+ > Zn 2+ . This may be explained by the high bonded of oleic acid to the Ni 2+ complex rather than the Zn 2+ . The bands at ~ 2330 and 2360 cm −1 due to ν (N-H…. O) vibration resulting from the intermolecular hydrogen bonding between PhNH proton and carboxylate oxygen of oleic acid. The band centered at ~ 3420 cm −1 may be due to the existence of hydrated water. The appearance of new band at ~ 1720 cm −1 due to ν(CO) of COO − group suggesting the participation of oleic acid in complex formation. The above data suggest that oleic acid takes part in hydrogen bonding.
(iii) The decrease in the flotation efficiency upon increasing the temperature over 40 °C reflecting a weak H-bond that is easily destroyed on heating.
(iv) Preliminary conductometric information about the 1:2 complex of H 2 NPT with Cu(II) and the 2:2 complexes of H 2 NPT with other metal ions. These suggest Scheme 3 for the proposed mechanism:

Conclusion
In the current study, we prepared novel NOS containing 1-nicotinoyl-4-phenylthiosemicarbazide (H 2 NPT) and its Co 2+ , Ni 2+ 2+ , Zn 2+ , and Cu 2+ complexes and characterized their structures by elemental, spectroscopic, and thermogravimetric methods. The association constant of H 2 NPT ligand was determined by pH-metric measurements. Theoretical modeling using the DFT method was performed to evaluate the extracted parameters such as bond lengths, bond angles, HOMO and LUMO and other energy parameters such as softness, hardness, electronegativity and MEP maps. The mode of chelation between H 2 NPT and the metal ions under investigation is proposed to occur via the ligand's S, N, and/or O atoms. The prepared H 2 NPT formed hydrophobic chelates with Co 2+ , Ni 2+ , Zn 2+ , and Cu 2+ ions. These chelates readily floated using HOL anionic surfactant. Various experimental variables affecting the flotation process have been thoroughly studied. A combination of the flotation-FAAS method enabled the nanogram determination of the metal ions of interest with a 300-fold enrichment factor and detection limit < 35 pg. It is proposed that the flotation mechanism relies on electrostatic interactions between the HOL surfactant and the formed metal-ligand complexes via H-bond formation. The subsequent steps of synthesis of H 2 NPT and its complexes, characterization and theoretical molecular modeling are graphically represented in Fig. 13.
Funding Open access funding provided by The Science, Technology & Innovation Funding Authority (STDF) in cooperation with The Egyptian Knowledge Bank (EKB). This research did not receive external funding.

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Fig. 13
Schematic representation of synthesis of H 2 NPT and its complexes, characterization and theoretical molecular modeling