Adsorption Desulfurization of Iraqi Light Naphtha Using Metals Modified Activated Carbon

of sulfur content from Iraqi light naphtha produced in Al aims to evaluate the removal he study T Dora refinery by adsorption desulfurization DS technique using modified activated carbon MAC loaded with nickel Ni and copper Cu as single binary metals. The experiments were carried in a batch unit with various operating parameters; MAC dosage, agitation speed, and a contact time of 300 min at constant initial sulfur concentration 155 ppm and temperature. The results showed higher DS% by AC/Ni-Cu (66.45)% at 500 rpm and 1 g dosage than DS (29.03)% by activated carbon AC, increasing MAC dosage, agitation speed, and contact time led to increasing DS% values. The adsorption capacity of MAC results was recorded (16, 15, and 20) mg sulfur/g MAC for AC/Ni, AC/Cu and, AC/Ni-Cu, respectively. Equilibrium isotherm study results show good Cu. The kinetic study value (0.95) for AC/Ni 2 fitting with Freundlich isotherm model with R and 0.95) by pseudo first order and (0.96, 0.916 and, 0.909) value (0.974, 0.981, 2 R ed results show ) e(cal calculated q e h T . respectively , Cu , AC/Cu, and AC/Ni by pseudo second order for AC/Ni (5.125) mg/g by the e(exp) obtained q the nearest to was the mg/g by first order model ) 4.79 4.337 ( experiments where no interparticle diffusion referring to more than one process is controlling the adsorption process of sulfur compounds by MAC.


INTRODUCTION
Full-range naphtha is the production of crude oil thermal distillation at a boiling range between 30-200 o C, and represent about 15-30% by w/w of the crude oil. It includes light naphtha with boiling range from 30-90 o C, containing the C5 -C6 hydrocarbons while heavy naphtha is the fraction boiling from 90-200 o C, naphtha of different origin contain small amounts of additional compounds containing elements such as sulfur (Goarge and Abdullah, 1995). Light naphtha is used for producing of gasoline in the petroleum refinery (Yahya and Hussein, 2019) Sulfur content in fuel oil, such as mercaptans, thiophenes (T), benzothiophenes (BT), and dibenzothiophenes (DBT), produce sulfur oxide (SOx) upon combustion, which are the main sources of acid rain and air pollution, (Muhammad, et al., 2019), (Ibrahim, et al., 2016). The US Environmental Protection Agency (EPA) has set the primary National Ambient Air Quality Standards (NAAQS) for SO2 exposure at 78 g/m 3 . The annual average ambient concentration of SO 2 in the US is 15 g/m 3. These compounds also cause corrosion in refinery equipment and deactivate ("poison") the catalysts that promote desired chemical reactions in certain refining processes, (González-García, et al., 2018). Therefore, removal of such sulfur-containing compounds is imperative for the production of green fuel oils and to meet the new standards of sulfur content (10-15 ppm) as per the recommendations of the United State Protection Agency (USEPA), given the environmental concerns surrounding sulfur, (Subhan, et al., 2019). The removal of sulfur from transportation fuel and petrochemicals is gaining more attention due to the increased awareness of the adverse effects of burning sulfurcontaining oils on human health, which is undesirable and may be subject to stringent regulatory controls and the environment. Desulfurization can be classified according to its techniques to Hydrodesulfurization HDS and Non-Hydro desulfurization Non-HDS, (Betiha, et al., 2018). (Marc, et al., 2005) showed application of modified Hydro desulfurization HDS as a new approach to the ultra-deep desulfurization of liquid fuel, with high selectivity towards the refractory substituted dibenzothiophenes that are very difficult to remove by standard hydro desulfurization techniques. The new method is based on the selective adsorption of dibenzothiophenes by the formation of charge-transfer complexes with immobilized p-acceptor molecules. The application of very mild process conditions (low pressure, ambient temperature, no hydrogen consumption) is an additional advantage of this new approach in comparison to traditional HDS. AcOH as a polar solvent in the presence of H2SO4 . The effectively reduction of the sulfur content of naphtha and adsorption with silica gel further reduced the sulfur content to below 0.5 mass ppm where the oxidation proceeded in the AcOH phase, and most of the oxidized sulfur compounds resided in this phase, resulting in the successive removal of the sulfur compounds from the octane phase. (Mohammed and AbdulWahhab, 2014) showed desulfurization of sulfur by ion-exchange using NaY zeolite from kaolin clay origin the results were shown as promoted adsorbent gives a higher percentage of sulfur removal percentage (82.15%) after 10 minutes while removing records 40.15% after 2 hrs. (Ibrahim and Jabbar, 2015) examined the effects of the operating conditions (contact time, temperature, mixing speed, and sorbent dose) on the desulfurization efficiency, where they examined two different oxidative desulfurization strategies based on oxidation/adsorption or oxidation /extraction were evaluated for the desulfurization of AL-Ahdab (AHD) sour crude oil(3.9wt% sulfur content). (Mohammed, et al., 2015) showed that desulfurization by oxidation process using hydrogen peroxide as an oxidant and acetic acid as a homogeneous catalyst. The solvent extraction process used acetonitrile (ACN) and N-methyl-2-pyrrolidone (NMP) as extractants. These best conditions were applied upon real diesel fuel (produced from Al-Dora refinery) with 1000 ppm sulfur content. It was found that sulfur removal was 64.4% using ACN solvent and 75% using NMP solvent. Desulfurization by adsorption using Activated carbon and modified activated carbon by metals as oxides or salts to enforce the process of removal of sulfur compounds from petroleum products as an alternative and low-cost method in order to enhance the adsorbent surface to increase the ability of adheres the adsorbate in fluid at the surface of activated carbon, (Sarahm, 2017) examined different metal zinc , cobalt and nickel for loading at activated carbon AC and the samples of experiment includes ZnO/AC, ZnO/NiO/AC and ZnO/NiO/CoO/AC where, the loading was kept constant at 10%. Characterization revealed that the three types are highly crystalline with distinct peaks, of nanoparticle size below 100nm, and have high surface area., (Suryawanshi, et al., 2018) studied desulfurization under batch experiment conditions temperature 28 o C, initial sulfur concentration (25-600 ppm S), and equilibrating with a known weight of the adsorbent (0.05-0.75 g per 10mL of the model fuel) for single and double metal modifications were studied using zinc, cobalt, nickel, and copper. (Abbas and Ibrahim, 2020) studied modified desulfurization of light naphtha LN to remove the best removal of sulfur content from light naphtha in batch mode using Activated carbon and white eggshells WES. The operating conditions studied were hydrogen peroxide to LN ratio, pH of the solution, agitation speed, temperature, contact time, and the catalyst weight. The aim of this study is the evaluation of sulfur removal using modified activated carbon at different conditions. Equilibrium and kinetic studies are done for explaining the nature of the adsorption of sulfur.

EXPERIMENTAL WORK 2.1 Materials
Light naphtha was supplied from Al-Dora refinery in Baghdad with specific gravity 0.64-69, initial boiling point(30-40)C o and sulfur content 155 ppm(mg/l) which are shown in Table 1, granular activated carbon AC from coconut shell origin is produced by Unicrbo Co., Italy with particle size (1.18)mm, surface are (1100)m 2 /gm and particle density (1100-1200) kg/m 3 which are shown in  available adsorption sites where the results show loaded concentrations of Ni and Cu were 12.4 and 11.6 mg/g respectively.

Experiments Condition
2.3.1 Type of adsorbent Four samples of 50 ml light naphtha were prepared with initial sulfur content 155 ppm where 1 gram of origin activated carbon ACnon modified-and modified activated carbon AC/Ni , AC/Cu and AC/Ni-Cu (MAC) were added to these samples respectively at normal atmospheric condition room temperature 25-+ 2 o C, agitation speed 500 rpm for 300 min (6 hrs). At the end of adsorption desulfurization the mixture is filtered by filter paper for sulfur content analysis.
2.3.2 Agitation speed experiment. Three sets each involve three samples of 50 ml light naphtha were prepared for each (MAC) with initial sulfur content of 155 ppm where 1 gram of modified activated carbon (MAC) AC/Ni, AC/Cu, and AC/Ni-Cu were added to these samples respectively at three agitation speed (200, 300 and 500) rpm, normal atmospheric condition at room temperature 25± 2 o C, for 300 min (6 hrs). At the end of adsorption desulfurization the mixture is been filtered for sulfur content analyses.

MAC dosage experiment.
Three sets involve samples of 50 ml light naphtha with 155 ppm initial concentration were prepared, and different dosage of each MAC ( 0.1, 0.3, 0.5, 0.7, and 1 gm ) were added to the light naphtha samples respectively at room temperature with agitation speed 500 rpm until equilibrium at 300 min, separation of filtrate is done for sulfur content analysis.

Contact time experiment.
Five samples of 250 ml light naphtha were prepared with initial sulfur content 155 ppm where 2 grams of MAC AC/Ni , AC/Cu and AC/Ni-Cu and were added to these samples respectively at normal atmospheric condition room temperature 25-+ 2 o C, agitation speed 500 rpm for 0-300 min (0-6 hrs).at the end of adsorption desulfurization the mixture is filtered for sulfur content analyses.

Analytical Procedure.
• Amount of loaded metal on AC by Atomic absorption spectrophotometer.
• Xrd analyses by X-Ray Diffraction.
• Sulfur content by using ANTEK 9000N/S analyzer A according to ASTM D-5453.

Data Analyses
Removal of sulfur content was shown as Desulfurization efficiency (DS) Eq.(1), which is calculated as the ratio of sulfur removed to that initially present in light naphtha. (3) Where qe adsorption capacity at equilibrium mg/g , qt adsorption capacity at sampling time mg/g, and Ct are sulfur content at intervals sampling time respectively ppm.

RESULTS and DISCUSSION
Loading of metals on AC was in order to modify the surface of the adsorbent to be more acceptable to adsorb sulfur compounds, after preparation of MAC by loading Ni and Cu by adsorption method to prepare MAC for adsorption desulfurization, where the amount of adsorbed Ni and Cu were 12.2 mg Ni/g MAC and 11.6 mg Cu/g MAC. In Fig.1, Fig.2, and Fig.3 XRD analyses for AC/Ni, AC/Cu, and AC/Ni-Cu, respectively. Recorded 2θ values were (26.2, 36.5, 43.5 and 48) for AC/Ni, (26, 34.5, 38.5, 48.4 and 53) for AC/Cu and(26.5,34.6,36.6,38.4,43.5,50 and 58) Fig.5 represents the efficiency of desulfurization DS % for activated carbon before and after modification by loaded metals where the DS% at initial sulfur content 155 ppm, adsorbent dosage 1 g, and agitation speed 500 rpm ( 29, 58, 59.3, and 66.45)% for AC, AC/Ni, AC/Cu, and AC/Ni-Cu, respectively which was explained by past works due to that modification play a major role in enhancing adsorption of the sulfur compound by improving their surface, where the chemistry of activated carbon surface is the main factor for the removal of sulfur content . (Gaddafi, et al., 2017). Furthermore, the results show increasing % removal of sulfur for modified activated carbon with multi-metals. It is concluded that the addition of di metal exhibited more efficiency than mono, where these metals act as active sites for interaction with cycle sulfur compounds (Al-Karkhi, 2017). Previous work showed a slightly decrease in surface area are reported for modified activated carbon. The results of Fig.5 show increasing in DS % with the best result at 500 rpm were recorded (58, 59.3, and 66.45)% for AC/Ni ,AC/Cu and AC/Ni-Cu, respectively may be due to enhance mass transfer of sulfur compounds from fluid phase to solid phase to be adsorbed on sites of adsorbent surface (MAC) where high agitation will give the sulfur molecules high energy to contact with solid surface at same time high agitation will reduce the thickness of layer between the fluid and the adsorbent surface (AlKarkhi, 2017). Figure 6. Effect of agitation speed on DS% for MAC at an initial sulfur concentration of 155 ppm and adsorbent dosage 1 gm.  Fig.8 shows the equilibrium isotherm curve as a relation between sulfur concentration Ce and adsorption capacity qe by Eq. (2)  = + 1 (7) By plotting Ce/qe vs. Ce and log qe vs. log Ce, the results show invalid fitting with Langmuir model due to negative slope while acceptable for Freundlich isotherm model as shown in Fig.9, and the obtained constants are shown in Table 3. The listed results which are shown in Table 3. best R 2 is (0.95) for AC/Ni-Cu even though the other value are accepted, which refers to good fitting with the Freundlich model for desulfurization of light naphtha by all type of MAC which agreement with the previous works (Gaddafi and Tawfik, 2017) but It doesn't indicate a finite uptake capacity of the adsorbent. It can thus only be reasonably for low and intermediate concentration range (Ahmed and Ahmaruzzaman, 2015) where the Freundlich model doesn't show maximum adsorption capacity as by the Langmuir model.

Kinetic adsorption models.
The fitting of kinetic adsorption models pseudo-first-order, pseudo-second-order, and interparticle diffusion Eq.(8), Eq.(9), and Eq.(10) is applied to determine the equilibrium time, mechanism of adsorption and to determine the rate-controlling step linear form of the kinetic model were log (qeqt) vs. t for pseudo-first-order, (1/qt) vs. t for pseudo-second-order (Junxiong and Lan, 2009) and qt vs. t 0.5 for interparticle diffusion Fig.12, Fig.13 and = 0.5 + (10) Where: qe and qt are adsorption capacity at equilibrium and at time t, respectively (mg.g-1 ), K1 is the rate constant of pseudo-first-order adsorption (min -1 ), K2 is the rate constant of the pseudosecond-order adsorption (g.mg -1 .min -1 ), Kp is the intra-particle diffusion rate constant (mgg -1 s -1/2 ), and C (mg/g) is the boundary layer thickness for intra-particle diffusion Table 4. showes parameters and R 2 values for kinetic models from slope and intercept of the linear form of models. For AC/Ni R 2 values are (0.974 and 0.96) and value of qe(cal) (4.752 and 5.625) mg/g for pseudofirst and second-order respectively, whereby pseudo-first-order more nearest to the actual value qe(exp) (5.625) mg/ g, for AC/Cu R 2 values are (0.981 and 0.916) and value of qe(cal) (4.337 and 14.86) mg/g for pseudo-first and second-order respectively which show that pseudo-first-order more nearest to the actual value qe(exp) (5.125) mg/ g than one that calculated by second-order, while AC/Ni-Cu R 2 values are (0.95 and 0.909) and value of qe(cal) (4.79 and 11.84) mg/g for pseudo-first and second-order respectively where pseudo-first-order more nearest to the actual value qe(exp) (6.625) mg/ g than one that calculated by second-order The results in Table 4. show more fitting with pseudo-first-order even though there is an acceptable value of R 2 for pseudo-second-order, which disagreement with previous works (Ammar and Jaafar, 2017) due to the process of loading metal on activated carbon by this study follow the adsorption method which shows agreement with (Hua, et al., 2014). The intraparticle regression diffusion is not linear, and it did not pass through the origin (Ibrahim and Jabbar, 2015), suggesting that the intraparticle diffusion does not influence the adsorption and that it was not the rate controlling step, implying that more than one process is controlling the adsorption process (Hua, et al., 2014). The initial rapid phase may also be due to the increased number of vacant sites available at the initial stage, an increase in the driving force of the concentration gradient between fluid phase (adsorbate) and solid phase (adsorbent) (Ong,et al., 2010). Figure.12 Pseudo-first-order linear form for AC/Ni-Cu.

CONCLUSIONS
The following conclusions are pointed during this work: 1. The MAC gives higher removal of sulfur than original AC with order AC/Ni-Cu>AC/Ni>AC/Cu>AC. 2. Maximum DS % and adsorption capacities are 66.45% and 20 mg/g, respectively, with AC/Ni-Cu of 1 gm dosage. 3. DS% are increasing with increasing MAC dose, agitation speed, and contact time. 4. Equilibrium isotherm results show a good fitting Freundlich model with a maximum R 2 value of 0.95. 5. Kinetic study results show good fitting with pseudo-first-order model with range R 2 values (0.95-0.981) and suggest that the adsorption is not influenced by the intraparticle diffusion and that it was not the rate controlling step more than one process is controlling the adsorption process. The rate constant of pseudo-first-order adsorption (s -1 ) K1=

NOMENCLATURE
The rate constant of the pseudo-second-order adsorption (g.mg -1 . s -1 ) K2= Coefficient related to the affinity between the adsorbate and adsorbent for Langmuir model (l/mg) KL= The intra-particle diffusion rate constant(mg.g -1 s -1/2 ) Kp= Predicted adsorption capacity(mg/g) qcal= Adsorbent (i.e., solid) phase concentration after equilibrium mg adsorbed/g adsorbent qe= Maximum adsorption capacity in for complete monolayer on the surface (mg/g) qm= adsorption capacity at equilibrium and at time t, (mg.g -1 ) qt= Mass of adsorbent MAC (g) m= Time (min) t= Volume of light Naptha ml V=