A commercial activated carbon (AC R003) received from Oxbow Activated Carbon was used in this study. This chemically activated carbon was further treated through plasma generated by a capacitively coupled dielectric barrier discharge. A quartz tube connected to a mechanical pump was used as a vacuum chamber. A pair of copper electrodes were attached to the outside of the tube to generate plasma. One of the electrodes was connected to a radio frequency (RF) power generator (Kurt J. Lesker, Radio Frequency Power Supply R301) while the other electrode was grounded. The operation frequency was set at 13.56 MHz. Before the plasma was generated, the quartz tube was pumped down to <1 × 10⁻² Torr and purged with the process gas to prevent cross-contamination.

For plasma activation of the carbon, ~0.05 g of carbon powder was thinly spread across a ceramic plate. The ceramic plate was placed inside the tube between the two copper electrodes. The system was sealed and pumped down to low pressure. The tube was then filled with the process gas. The plasma pressure was kept constant for each treatment at 2 Torr (measured by a vacuum meter, Kurt J. Lesker). The RF power of the plasma was fixed at 75 W and adjusted to reach a zero-watt reflection power through all experiments. This process was repeated 5x for each test in order to reach a final mass of 0.250 g of treated AC. The plasma gas used was 10% O₂ - 90% Ar mixture and treatment times were varied. The plasma treatment times for each gas were: 1 min, 2 min, 3 min, 4 min, 5 min, and 6 min.

Methylene blue (MB) was purchased from Acros Organics. A calibration curve was created by measuring the absorbance of differing concentrations of MB solution using a UV-visible spectrophotometer at a wavelength of 664 nm.

The UV-visible spectra of different methylene blue concentrations were used to construct a calibration curve of absorbance versus methylene blue concentration (mg/L). To do this, samples of methylene blue and DI water with known concentrations were measured by UV-visible spectroscopy (UV-vis). The absorbance peak height for the methylene blue identifying wavelength at 664 nm was found using Origin. The found peak heights were then plotted against concentration, and a line of best fit was generated in the software. The equation from the line of best fit was then used to calculate final methylene blue concentrations in the following experiments.

Batch adsorption tests were performed at room temperature. The experiments were used to examine the effects of plasma treatment parameters on MB adsorption by AC. The experiments were conducted on samples of AC treated with O2 plasma and different treatment times (1 - 20 min).

To measure the adsorption capacities of each AC sample, 5 mg of AC was added to a glass beaker containing 50 mL of 10 mg/L MB solution. The beaker with the AC and MB solution was then placed on a stir plate at room temperature and stirred at 180 rpm for 2 minutes. The solution with the AC was then transferred into a 50 mL centrifuge tube and was centrifuged at 6000 rpm for 10 minutes to get a clear liquid. The supernatant was then collected and placed in a glass vial. This procedure was repeated for each AC sample.

Once the supernatant from all samples was collected, the remaining concentration of MB in the solution was measured at 664 nm using a UV-visible spectrophotometer with a range of 325 nm to 700 nm. The adsorption capacities of MB on the AC samples, qt (mg/mg), were calculated using the following equation:

q t = V ( C 0 − C t ) w (1)

where C₀ (mg/L) is the liquid-phase initial concentration of methylene blue, and C_t (mg/L) is the concentration of methylene blue at each time. The mass of dry AC is represented by w (g), and V (L) is the volume of methylene blue solution.

In each experiment, 5 mg of AC was placed in a 50 mL centrifuge tube. Then 20 mL of 25 mg/L MB solution was added to the centrifuge tube. Then, the centrifuge tubes were spun in the centrifuge at 6000 rpm for 20 minutes. This allowed for the supernatant to separate from the absorbent. The supernatant was then collected and measured for the remaining MB concentration. The remaining concentration was found by measuring the absorbance of the solution by UV-visible spectroscopy at a wavelength of 664 nm. The percent adsorbed and adsorption capacities were calculated using the following equations:

R = C 0 − C t C 0 × 100 (2)

where R is the percentage of MB adsorbed, C₀ (mg/L) is the liquid-phase initial concentration of methylene blue, and C_t (mg/L) is the concentration of methylene blue at each time.

Next, the amount of methylene blue adsorbed by 1.6 g of AC at different time increments was investigated. The AC used was not plasma treated and was left in the 36 mg/L methylene blue solution for the time increments: 0, 5, 10, 20, and 30 minutes. After the appropriate amount of time, the AC was vacuum filtered from the solution, and this solution without AC was measured to find the concentration of methylene blue remaining. The final concentration of methylene blue in the solution was measured by UV-vis.

Following the same conditions, 50 mL of MB solution with the initial concentration of 25 mg/L was placed in a glass beaker. The solution was left at room temperature and was stirred at 180 rpm. Samples were collected at different time intervals and analyzed by UV-visible spectroscopy at the wavelength of 664 nm. The amount of MB adsorbed onto the AC was calculated and different kinetic models were applied to fit the process.

Following the previous conditions, 20 mL of MB solution with different initial concentrations (5 mg/L, 10 mg/L, 15 mg/L, and 25 mg/L) was placed in 50 mL centrifuge tubes. A specific mass of AC of 5 mg was added to the solution in the centrifuge tube and the mixture was spun for an hour at 6000 rpm and then left to settle for 24 hrs. The final concentration of MB was calculated using UV-visible spectroscopy and different isotherm models were applied to fit the process.

Plasma treatment was expected to open up the original micropores and create mesopores, which had pore sizes ranging from a few nanometers to about 50 nm. High-magnification microscopy was required to examine porous structures of such scale. Therefore, transmissions electron microscopy (TEM) images were taken of the selected samples using a JEM-2200FS (JEOL, Japan).

Samples prepared with a ratio of 1 mg AC to 1 ml DI water were used for the zeta potential measurements. Before the measurement, the samples were ultrasonicated for 20 minutes. The zeta potential measurements were performed on activated carbon using a Zeta potential meter (Malvern Zetasizer Nano-ZS). Each sample was measured three times to provide an average reading.

Surface analysis by x-ray photoelectron spectroscopy (XPS) was carried out using an SSX-100 system (Surface Science Laboratories, Inc.). XPS was used to analyze surface element composition. High-resolution analysis was performed for C, O, and N.

The contact time between MB and AC was evaluated as a factor of adsorption efficiency. The UV-vis spectra for MB adsorption by AC from 5 to 30 minutes were collected. The results showed that the concentration of MB left in the solution reached a minimum between 10 and 20 minutes (shown in Figure 1), suggesting that the AC became saturated around this time. The concentration decreased 89% after incubating MB with AC for 20 and 30 minutes. The adsorption capacity at equilibrium of the AC before plasma activation was 4.13 mg∙g⁻¹.

Next, the effect of different O₂ plasma treatment times was investigated. The adsorption capacity was found similar for a plasma treatment time in the range of 3 - 6 minutes but dramatically decreased for treatments longer than 10 minutes, as shown in Figure 2. This could be due to over etching of the AC causing the internal structure to collapse. All further studies were limited to a maximum plasma treatment of 6 minutes in order to avoid over etching.

The effects on adsorption capacity with different plasma treatment times were investigated and compared to non-plasma treated AC. The adsorption capacities of the plasma treated samples significantly increased as shown in Figure 2. The non-plasma treated sample had an adsorption capacity of 4.00 × 10⁻³ mg∙mg⁻¹. As the O₂ plasma treatment was applied, the adsorption capacity of the samples

increased up to 4 minutes. The adsorption capacity for the O₂ plasma treated samples reached its maximum adsorption capacity of 0.467 mg∙mg⁻¹ at 4 minutes of plasma treatment. After 4 minutes of treatment, the adsorption capacity decreased. After 10 minutes of O₂ plasma treatment, the adsorption capacity had decreased ~88% to 0.057 mg/mg, and after 20 minutes it had decreased ~93% to 0.032 mg/mg. Overall, the O₂ plasma treated samples had adsorption capacities higher than the non-plasma treated sample. The best adsorption capacity for the O₂ plasma treatment was found to be 0.467 mg∙mg⁻¹ at 4 minutes of plasma treatment which was over a one hundred times increase. The significant increase showed that plasma treatment of AC increases adsorption capacity.

A simple experiment was conducted to ensure that the increase in adsorption capacity was not due to the removal of residual water from vacuum and localized heating. The adsorption capacity of AC left in vacuum for 10 minutes and the adsorption capacity of AC left in 110˚C oven for 10 minutes were compared to untreated AC (Figure 3). Figure 4 shows that the influence of heat and vacuum

on the adsorption capacity of MB by the AC is insignificant.

To better explain how all the plasma treatments led to similar adsorption capacities, the original AC R003 was further examined. The AC was heated to 110˚C - 300˚C for 30 minutes with and without being kept under vacuum. The treatments included in Figure 4 show that an increased temperature in concurrence with treatment under vacuum improved MB adsorption for the non-plasma treated samples. Then, the effect of treating AC with O₂ plasma at room temperature was compared with the non-plasma treated samples. The graph clearly shows that the 4-minute O₂ plasma treated sample had the highest adsorption capacity. The overall results suggest that the plasma treatment provided significant localized heating at atomic scale in addition to modifying the porous structure.

Transmission electron microscopy (TEM) is used to obtain information about the morphology and microstructure of materials [11] [12]. TEM transmits a high energy electron beam through a sample to form an image. TEM images were taken of the best performing samples from the plasma treatment, along with the untreated AC.

Figure 5 shows TEM images of non-plasma treated AC (left) and O₂ plasma treated AC (right). Brighter areas in the images indicate that more electrons were able to penetrate the sample. Hence, the brighter areas are pores. The scale bars for each image differ due to the need for higher magnification to see the porous structure in the non-plasma treated sample. The image shows that the non-plasma treated AC contains mainly micropores. With plasma treatment, meso-pores with irregular shapes appear because the AC is etched by the plasma. The larger mesopores are easily visible under lower magnification. This combined meso-pores and micro-pores facilitate the transport of MB and promote the adsorption.

The XPS analysis of AC is shown in Figure 6. XPS provides information about the surface functional species and composition. Carbon, oxygen, and nitrogen content were determined. The XPS analysis of AC that was not plasma treated gave the atomic concentrations 89.01% C 1s and 10.99% O 1s. No N 1s was detected. The high-resolution spectrum of C 1s of untreated AC could be deconvoluted into 3 peaks: 284.74, 286.35, and 287.90 eV. The O 1s region was made up of three peaks: 529.5, 531.1, and 532.5 eV.

The XPS analysis of AC that was treated with O₂ plasma for 4 minutes gave the atomic concentrations 80.06% C 1s and 19.94% O 1s with no apparent N. The high-resolution spectrum of C 1s of AC that underwent O₂ plasma treatment for 4 minutes can be broken into 5 peaks: 284.82, 285.42, 286.37, 287.92,

and 289.37 eV. The O 1s region was split into 2 peaks: 533.88 and 532.63 eV.

Compared to the untreated AC, XPS analysis showed a decrease in carbon concentration in the O₂ plasma treated sample. There was no discernable nitrogen found in the O₂ plasma treated sample, but there was an increase in oxygen. The increase in oxygen with a decrease in carbon indicates a change in polarity [4]. The samples with higher amounts of oxygen would be expected to be more polar and have less aromaticity leading to less hydrophobicity [4]. There was no discernable nitrogen in the O₂ sample or the non-plasma treated sample. XPS analysis further revealed a change in surface functional groups and atomic composition after plasma treatment.

The peak numbers and their interpretation are listed in Table 1. With plasma treatment, the C1s XPS spectra could be broken down into more peaks. Plasma treatment of AC resulted in an increase in C-O-C and C = O functional groups. In the O₂ plasma treated sample, a peak around 290 eV appeared. This peak corresponds to C = O/C = C (carbonate, occluded CO, π-electrons in aromatic rings) [9] [13] [14] [15].

With plasma treatment the O1s XPS spectra of the O₂ sample only had two peaks. The O₂ treated sample loses its O (oxide-like oxygen) and its C = O (quinone type carbonyl) peaks, and it gains a peak at 533.9 eV that corresponds with C-O (ester/amide/carboxylic anhydride) [13] [16] [17] [18] [19].

Zeta potential measurements were performed on the O₂ and the non-plasma treated AC sample. Figure 7 shows the comparison between each sample when the adsorption capacity was the highest. The O₂ plasma-treated AC has the most negative zeta potential compared to the non-plasma treated AC sample. The zeta potential curve for O₂ plasma-treated AC shifts to the left of the non-plasma treated AC. The zeta potential of the O₂ plasma-treated AC is measured at −63 mV.

Zeta potential is a measure of surface charge. It is the potential between the dispersed particle surface and the mobile dispersion medium [20]. It can be observed in Figure 7 that all the adsorbents had large negative zeta potentials. This implies that the particles in suspension resist aggregation and are inclined to disperse homogeneously throughout the solution [20]. This opposition to aggregation and the ability to spread evenly throughout a solution is beneficial to adsorption. The magnitude of the zeta potential is indicative of colloidal stability [21]. Figure 7 clearly shows that the O₂ plasma treated sample has a zeta potential between −35 mV and −65 mV. Nanoparticles with zeta potentials less than −30 mV are considered to have high degrees of stability [21].

Adsorption of dyes on AC may include chemical adsorption. To investigate the adsorption mechanism, the adsorption kinetics of MB onto AC that was not plasma treated was analyzed by using pseudo-first and pseudo-second order kinetic models. The pseudo-first-order equation can be written in the form:

log ( q e − q t ) = log ( q e ) − k 1 2.303 t (3)

where q_e represents the amount of adsorption (mg/mg) at equilibrium, k₁ is the rate constant of the pseudo-first-order equation, and q_t is the amount of adsorption (mg/mg) at t, time (min) [22]. The rate constant, k₁, can be calculated by plotting log(q_e-q_t) versus t [22].

The pseudo-second-order kinetic equation is shown below:

t q t = 1 k 2 q e 2 + 1 q e t (4)

where k₂ is the pseudo-second-order rate constant. Plotting t/q_t versus t, a line of best fit can be found (Figure 8). This straight line can then be used to calculate k₂ [22]. The R² value for the pseud-second-order model was 0.99, where the R² value for the pseudo-first-order model was only 0.80. Therefore, the adsorption kinetics are adequately described by the pseudo-second-order model. Similar results have been found for the adsorption of MB onto rejected tea leaves, AC, and different carbonaceous materials [20] [22] [23].

The pseudo-second order model fits the kinetics data of MB adsorption onto AC. This suggests that chemisorption has a dominant role in the adsorption process [16] [23].

The relationship between the initial concentration of MB and the adsorbed amount of MB onto AC are investigated and shown in Figure 9. The graphs show that the adsorption of MB increases with the increase of the equilibrium concentration of MB.

The two isotherms tested for their ability to describe the experimental results were the Langmuir and Freundlich adsorption isotherms [24]. These models provide insight into the mechanism of adsorption and the surface properties [24]. The Langmuir isotherm is based on the assumption of monolayer adsorption onto a homogeneous surface [22] [24]. The Langmuir isotherm is expressed by the following equation:

C e q e = 1 q max b + 1 q max C e (5)

where C_e is the equilibrium concentration (mg/L), q_e is the adsorption capacity at equilibrium (mg/g), b is the Langmuir constant (L/g), q_max is the maximum

adsorption capacity (mg/g).

The Freundlich isotherm is related to adsorption capacity and intensity [22]. The Freundlich isotherm assumes a heterogeneous surface with a non-uniform distribution of adsorption [22]. It is expressed by the following equation:

ln q e = ln K f + 1 n ln C e (6)

where C_e is the equilibrium concentration (mg/L), q_e is the adsorption capacity at equilibrium (mg/g), K_f is the Freundlich constant (L/mg), 1/n is the heterogeneity of the sorption sites and an indicator of isotherm nonlinearity [23].

The data was fit to the Langmuir and Freundlich isotherm adsorption models. The data fitting gave the regression coefficients Freundlich (R² > 0.996 and the Langmuir (R² > 0.977) models indicating that the models described the adsorption process adequately. This indicates that the adsorption of MB onto activated carbon occurred primarily by single molecular adsorption [23].