The clay material used for adsorption experiments performed in the previous study is essentially kaolinite and was collected from Konhontsa’a, a small town in Bamboutos Division located in the West Cameroon Region (5˚41"N and 10˚13"E) ( Fouodjouo et al., 2017 ). The adsorption experiments were carried out in 250 mL Erlenmeyer flasks by mixing together 0.1 g of clay with 50 mL of aqueous solution of Cu²⁺ of concentration 3 g·L⁻¹ ( Fouodjouo et al., 2017 ). The contents in the flasks were agitated by placing them in a constant temperature water bath thermostat for 3 minutes (equilibrium time) ( Fouodjouo et al., 2017 ). The mixture was then centrifuged (4000 rpm) and residual Cu²⁺ in the supernatant liquid was determined by atomic absorption spectroscopy ( Fouodjouo et al., 2017 ). The Cu(II)-clay composite samples used are those corresponding to maximum adsorption capacity at equilibrium found in part A of this strudy by Fouodjouo et al. (2017) .

Desorption using NTP or desorbing was carried out in batch system and the temperature was kept at 298 K using a cooling bath system. 1 g of natural dried Cu(II)-clay composite (KON-00, q₀ = 195 mg·g⁻¹) ( Fouodjouo et al., 2017 ) was dispersed in 400 mL of deionised water and the suspension introduced into the Glidarc plasma batch reactor presented in Figure 1 . Humidified air was used as flowing gas at a flow rate of 800 L·h⁻¹. The homogenisation of the suspension in the reactor was assured by the airflow and a magnetic stirrer. The Glidarc device used ( Figure 1 ) operation is described by Fouodjouo et al. (2017) and the temperature and electron density, and the distribution of active species in the plasma by Wu et al. (2015) .

The effect of exposure time, initial amount of Cu(II)-clay composite and stirring rate on the efficiency of the desorption process were studied. Furthermore, for a comparative study with other desorbing agents, 1 g of Cu(II)-clay composite was placed into flasks of 200 mL containing 50 mL of analytical grade aqueous solutions of CaCl₂, HCl or NaCl respectively. The selected concentrations of desorbing agents were 0.1 mol·L⁻¹, 0.01 mol·L⁻¹ and 0.001 mol·L⁻¹. The temperature was maintained constant by using a thermostated bath. Aliquots drawn after each of time intervals chosen, were centrifuged at 3500 rpm for 5 min and the concentration of Cu(II) ions in the supernatant was determined using atomic adsorption spectroscopy (Buck scientific 200A; Atomic Adsorption/Emission spectrophotometer). The clay and the Cu(II)-clay composite were characterized in the previous works ( Fouodjouo et al., 2017 ).

The amount of Cu(II) desorbed per unit mass of clay and the percentage of Cu²⁺ recovering as function of time were estimated using Equations (1), (2) and (3) presented below ( Wankasi et al., 2005 ; Wambu at al., 2009 ):

$q_{Des}=C_t\left(\frac{V}{m}\right)$ (1)

$q_t=q_0-q_{Des}$ (2)

$\%Des=\frac{C_t}{10m{q}_e}V$ (3)

where q_Des and q_t in mg·g⁻¹, are respectively the amount of Cu(II) desorbed per unit mass of copper-loaded clay and the amount of remaining Cu(II) onto the Cu(II)-clay composite at time t (min), C_t (mg·L⁻¹) the concentration of Cu(II) in desorption solution at time t and q₀ (mg·g⁻¹) the initial solid phase Cu(II) concentration.

Four kinetic models ( Table 1 ) were applied to desorption data and their respective parameters were determined.

where q_t is the Cu(II)-clay composite concentration (mg·g⁻¹) at any time t(min) and C_i the initial Cu(II) ions concentration loaded on Cu(II)-clay composite.

Figure 2 displays the effect of plasma exposure time on the Cu(II) desorption from a Cameroonian clay. It appears that the amount of Cu(II) in the composite decreased with the increase of the exposure time, followed by the formation of a plateau at 180 min corresponding to a 99.79% Cu(II) removal. Desorption with NTP can be explained by the collision of fast active species generated with copper ions ( Shiau et al., 2017 ; Ghosh et al., 2015 ; Johnson, 1987 ). The composite Cu(II) concentration also decreased with the decrease in solution pH of extraction solution ( Wankasi at al., 2005 ) caused by the acidifying effect of Glidarc plasma in humid air ( Brisset & Hnatiuc, 2012 ). Indeed, the humid air NTP is a source of H⁺ ions ( Njoyim-Tamunganga et al., 2011 ; Brisset & Hnatiuc, 2012 ), which can easily exchange with Cu²⁺ ions fixed on the adsorption sites of the adsorbent due to their small size and high mobility. The prevailing reaction at the surface of metal-loaded adsorbent in acidic medium and responsible for Cu(II) desorption can be described as follows:

$\left({\text{Ads}}^{2-}\right)\left({\text{M}}^{2+}\right)+2{\text{H}}^{+}\to \left({\text{Ads}}^{2-}\right)\left(2{\text{H}}^{+}\right)+{\text{M}}^{2+}$ (3)

where (Ads²⁻)(M²⁺) is the metal-loaded adsorbent and, M²⁺ the metal.

This means that loaded copper onto clay could be in form of pH-susceptible speciation such as hydrolysis products. The oxidative species such as OH• and H₂O₂ might also participate to the breaking of the bonds between Cu(II) ions and the adsorbent. The equilibrium was reached after 180 min and 99.79% of Cu(II) was recovered from the metal-laden clay.

The effect of the stirring rate of the desorption medium on Cu(II) desorption is depicted in Figure 3 . The results obtained suggests that, Cu(II) desorption efficiency increased with increase of stirring rate. This signifies that, stirring could speed up the hydrogen ions or the oxidative species diffusion and thus accelerated the displacement of Cu(II) ions from the solid phase. The increasing of the stirring rate reduced the time necessary for complete desorption. At 1250 rpm, 87.97% of copper was desorbed after only 30 min.

Reusability of the clay has been checked via six adsorption-desorption cycles with NTP plasma as desorbing agent. Figure 5 shows that after six cycles the adsorption capacity of the cycles decreased only of 38.97%. The adsorption capacity dropped from 195 mg·g⁻¹ to 119.23 mg·g⁻¹. The use of NTP as desorbing agent did not noticeably change the adsorption capacity of the adsorbent and thus, the clay could be repeatedly used in Cu(II) ions adsorption without significant loss of initial adsorption efficiency. This is due to the fact that NTP at each cycle increase the oxygen content on the adsorbent surface ( Zeng et al., 2017 ). Indeed, several previous works reported that the plasma does not significantly affect the adsorption capacity of the clay ( Pinard & Batiot-Dupeyra, 2024 ; Srour et al., 2021 ). NTP plasma is suitable to keep low the cost of Adsorption-desorption processing and therefore the overall cost of heavy metal control in water using Adsorption technique.

Kinetics study of Cu(II) desorption from Copper loaded clay using NTP as desorbing agent was important to understand the process. As shown in Figure 2 in section 3.1, the desorption rate was rapid during the first 60 minutes, then, it slowed down until the desorption process approached completion. The desorption attained equilibrium in 180 min at a stirring rate of 500 rpm. The kinetics of Cu(II) desorption by NTP and HCl as desorbing agents were evaluated using the first order, Elovich, modified Freundlich and parabolic diffusion equation models. These multiple models were applied to point out the efficiency of NTP compared to others desorbing efficient agent such as HCl. The kinetic parameters obtained were confined in Table 2 .

First order kinetics

Figure 6(a) shows the plot of lnq_t versus t at 298 K. It appeared that the experimental data fitted linear with first order model. The first order equation assumes that the rate of Cu desorption is proportional to either the number of occupied sites on the surface or the concentration of Cu in the occupied sites. The first order equation constants (k₁ and b) and the correlation factor R² were determined and presented in Table 2 . From those constants it can be seen that NTP was more efficient and faster than HCl for Cu(II). Furthermore, the half time of desorption t_1/2 was found to be 45 min for NTP and 61 min for HCl.

Elovich kinetics model

The empirical Elovich equation ( Chien & Clayton, 1980 ; Kuo & Lotse, 1973 ) was also used to describe the Cu desorption rate. Figure 6(b) depicts the plot of qt versus logarithm of exposure time at 298 K. It showed a linear relationship exists between q_t and lnt for desorption with NTP as desorbing agent. Table 2 recapitulated all the elovich parameters for Cu(II) desorption using NTP and HCl as desorbing agent. In fact, a decrease in α value, evidences a reduction in the desorption reaction rate ( Motaghian & Hosseinpur, 2013 ). The α value of NTP as desorbing agent was higher than that of HCl, showing the high capacity of this environmentally friendly technique (NTP) to be the clean and efficient desorbing agent for heavy metals.

Parabolic diffusion

Parabolic Diffusion Equation was also used to describe the kinetic data of Cu(II) release with NTP or HCl as desorbing agent. The plot of q_t against t^1/2 at 298 K ( Figure 6(c) ) was linear, showing that experimental data fitted well this kinetic model. Diffusion law parameters determined from the slope and intercept of the linear plots are shown in Table 2 . The absolute value of parabolic diffusion rate constant of desorption with NTP is once higher that of HCl.

Modified Freundlich kinetics

The Modified Freundlich Equation was also used to describe the Cu desorption data. The plot of lnq_t versus lnt at 298K ( Figure 6(d) ) gave a straight line. This kinetics models as the three others models studied above fitted well experimental data and can be used for the description of the mechanism of desorption of Cu(II) from metal loaded clay. The modified Freundlich equation parameters k_D and 1/m, calculated from the linear plots are listed in Table 2 . From Table 2 , it can be seen that the k_D of NTP is highest in comparison to that of HCl.

All the four kinetics models studied described well the desorption of Cu(II) from metal loaded clay (R² close to 1). Similar observation was done by Elkhatib et al. ( Elkhatib et al., 2007b ). However, for desorption kinetics of Cu(II) with NTP, experimental data correlated well with the different models studied in the following order: First order > Elovich > Parabolic diffusion > modified Freundlich.

The histogram of Figure 7 depicts the performances of some desorption medium (HCl, CaCl₂, NaCl) of various concentration compared to the plasma process. For comparison, desorption was also performed in deionised water. Plasma (99.79%) and HCl (97.63% at 0.1 M) showed the best performances compared to CaCl₂ (max value), NaCl (max value) and deionised water (max value). Although HCl had effectiveness close to plasma efficiency, NTP is the best desorbing agent because of its environmental friendliness, its low cost and ease to set up. The most significant difference between NTP desorption process and conventional desorption is the existence of abundant short-lived active species (i.e. electrons, excited species, radicals, and positive/negative ions) ( Shiau et al., 2017 ; Benstaali et al., 2002 , Czernichowski et al., 1996 ). Thus, Cu(II) ions adsorbed can be effectively desorbed and even converted during NTP process. The desorption efficiency followed the order: plasma > HCl > CaCl₂ > NaCl > deionised water. From Figure 5 , it was also found that for chemical desorbing agents, the efficiency of desorption increased with the increase in concentration of the desorbing agent. Earlier, Gupta & Rastogi (2008) and Kwon & Jeon (2012) found the same observation with Cr (VI) and Indium ions respectively.