Sage Journals: Discover world-class research

Abstract

Equilibrium and kinetic adsorption behaviours of hydrogen fluoride (HF) onto zeolite 3A were investigated under different temperatures (298–338 K) and pressures (0.15–1.1 bar). The HF adsorption isotherms on zeolite 3A were well described by the Dubinin–Astakhov (DA) model. The isosteric heat of HF adsorption was calculated based on the DA model and Clausius–Clapeyron equation, and different kinetic models were used to analyze the HF adsorption kinetic. The results showed HF adsorption isotherms are type I of the IUPAC classification. The maximum adsorption capacities by the DA model are 0.443–0.5631 mg/g. The $H_{s t}$ values decrease with increasing surface loading indicating the surface energy heterogeneity of zeolite 3A. The HF adsorption kinetics on zeolite 3A follows simple first-order kinetics. The calculated film diffusion parameter $B_{t}$ values show nonlinear characteristics with time. The order of magnitude of interparticle HF diffusion coefficient $D_{c}$ is 10⁻⁹ m²/s. The higher pressure and temperature are in favour of HF diffusion. The intraparticle diffusion curves for HF adsorption on zeolite 3A show quart-linear characteristics, indicating the presence of four consecutive HF adsorption steps including film diffusion, interparticle diffusion, intraparticle diffusion, and surface adsorption. The intraparticle diffusion is a rate-controlling stage.

1. Introduction

Fluorine-containing chemicals are widely used in atomic energy, metallurgy, machinery, chemical industry, and other industries because of their unique physical and chemical characteristics [1]. As a basic fluorine chemical product, hydrogen fluoride (HF) is widely used in producing refrigerants, fluorine resin, fluorine rubber, and fluorine-containing intermediate and fine chemicals [2]. HF is corrosive and toxic, and it poisons humans and animals through the food chain, destroys the normal metabolism of calcium and phosphorus, inhibits enzyme activity, and affects the nervous system [1]. In the use of fluorine (F₂), e.g., uranium enrichment, optics, and material fluorination, HF is often accompanied by F₂, which reduces the purity of fluorine [3–5]. Therefore, HF purification treatment and recovery are indispensable to save resources and protect the environment.

The common methods of collecting and recovering HF are compression condensation, absorption, and adsorption [6]. The compression condensation method can purify HF gas well in theory, but this method needs low temperature and consumes more energy [7]. The HF amount treated by the absorption method is huge. However, absorption purification is wet with high requirements for equipment and chemical treatment problems [8]. Collecting HF by adsorption method cannot only remove HF gas but also reduce the emission of toxic gas HF and energy consumption. The adsorption method is mainly based on the HF adsorption performance. Thus, the basic HF adsorption data is required [9]. The frequently used adsorbents for HF adsorption are porous, e.g., sodium fluoride, activated alumina, activated carbon, and silica gel [2]. Zeolite is a commonly used adsorption material in gas separation and gas purification [6]. The application of zeolite can be extended to HF treatment.

The studies on HF treatment by adsorption have been investigated by some scholars. Sodium fluoride (NaF) is a commonly used adsorption material for HF [10–12]. NaF reacts with HF to produce NaHF₂. When the temperature is higher than 250°C, NaHF₂ can be decomposed into NaF and HF. Afzal et al. conducted experiments on HF adsorption with NaF by the fixed bed, and the influence of concentration, temperature, and other parameters on the surface loading has been studied in detail [10–12]. The experimental results show that Langmuir and Temkin can better describe the HF adsorption isotherm than Freundlich [11]. Moreover, Liu et al. used NaF to adsorb and remove trace HF in the production of high-purity tungsten hexafluoride, and the HF adsorption by NaF under different temperatures, pressures, and gas flows was measured by orthogonal test [13]. The results showed the suitable adsorption temperature was 20–80°C, and the flow rate has a significant effect on adsorption efficiency. Activated alumina (Al₂O₃) is also used for HF adsorption. Activated alumina reacts with HF to produce AlF₃ and H₂O. In addition, Wu et al. studied the adsorption characteristics and influencing factors for HF adsorption by Al₂O₃, and the relationships between crystal structure, the specific surface area of Al₂O₃, and HF adsorption have been considered [14]. Sheng studied the HF adsorption behaviour on activated carbon, calcium oxide, activated alumina, iron oxide, and fly ash, and the Freundlich model described the adsorption isotherms well [15]. In addition to NaF and Al₂O₃, some other porous materials are also gradually applied to HF adsorption. Che et al. studied the HF adsorption behaviour on DB18C6/SiO2 composite [16]. Siahooe et al. studied the HF adsorption on carbon nanotubes and found that the HF adsorption density decreased with temperature increase [17]. Wang et al. investigated the HF adsorption behaviour on three metal–organic frameworks (MOFs) and found that MOFs had better HF adsorption performance than commercial Al₂O₃ and 13X zeolite [18]. Bahrami et al. investigated the HF gas adsorption by three kinds of activated carbon under vacuum and described the experimental data by isotherm adsorption models [19].

Zeolite 3A is an excellent performance adsorbent for gas storage and separation [20]. However, the studies about HF adsorption on zeolite 3A are very lacking. Therefore, the study of the adsorption characteristics and influencing factors of HF on zeolite 3A is necessary. In this study, the equilibrium and kinetic adsorption curves of HF on zeolite 3A at different temperatures and pressures were measured. The HF adsorption isotherms on zeolite 3A were described by the three isotherms model. The DA model and the Clausius–Clapeyron equation were used to calculate the isosteric heat of HF adsorption, and different kinetic models were used to analyze the kinetic process for HF adsorption on zeolite 3A. The adsorption mechanism of HF onto zeolite 3A was evaluated based on kinetics and thermodynamics.

2. Materials and Methods

2.1. Materials

Zeolite 3A (8(K₁₂[(AlO₂)₁₂(SiO₂)₁₂]·24H₂O) used in this study was commercial zeolite manufactured by Lu Xinyuan Company (Henan, China). In this investigation, spherical 3A particles of 0.2 cm radius were studied. Hydrogen fluoride and helium with a purity of 99.99% were provided by China Shipbuilding (Handan) Perry Special Gas Co., Ltd.

2.2. Characterization of the Zeolite 3A

The compositions were measured by X-ray diffraction (XRD) using a Rigaku Smartlab 9kW with Cu Kα radiation (40 kV and 40 mA). The Fourier transform-infrared (FT-IR) of zeolite 3A spectra was obtained from an FT-IR spectrometer (Thermo Nicolet iS5, USA) using pellets of zeolite/KBr. The surface morphology was obtained by SEM (Zeiss sigma500) at an accelerating voltage of 10 kV. The pore structure distribution (PSD) of zeolite 3A was measured by the low-temperature CO₂ adsorption method at 273 K using 3H-2000 PM.

2.3. Adsorption Isotherm and Kinetics Measurements

Equilibrium and kinetics measurements of HF adsorption on zeolite 3A were performed in a static volumetric instrument. The detailed composition of the experimental instrument is shown in Figure 1. The pressures of the experimental system were recorded using a precision pressure sensor (Keller, Switzerland, ±0.01% precision). An automatic thermostatic oil bath with ±0.01 K accuracy was used to keep the system temperature constant.

Figure 1

A schematic of experimental setup for HF adsorption.

The main experimental procedures for equilibria and kinetics adsorption are: (1)

Sample pretreatment and filling. The 3A sample used in the adsorption experiment was first dried in the drying oven at 385 K for 24 h, weighed, and then quickly transferred to the sample cell. The constant temperature of the oil bath was set to 358 K and vacuumed for 12 h to remove the water and other impurities adsorbed in the sample transfer process

(2)

Leakage test for adsorption system. After the sample pretreatment, the temperature of the oil bath was set to 338 K. After the system temperature is constant, helium (He) was fed into the reference and adsorption cells to a certain value (maximum pressure required for adsorption). The leakage testing time was 12 h. The system was heated and vacuumed for 4 h after the leakage test

(3)

Measurement of void volume. The void volume of the adsorption cell filled with adsorbent was measured by the helium expansion method. The temperature for the helium expansion test was set to the adsorption experimental temperature. The helium density used in void volume measurement is calculated by the BWR equation [21]. Helium expansion tests were repeated at least four times and the values were averaged to minimize error

(4)

Equilibrium adsorption experiment. A certain HF amount was injected into the reference cell when the temperature of the adsorption system was constant with the experimental setting value. The gas pressure for the reference cell was recorded until it is stable. The valve between the reference and sample cells was opened, the HF flowed into the sample cell from the reference cell and the adsorption began. The system pressure was recorded when the system pressure was constant for 5 h. The above operations were repeated until the maximum pressure point of the equilibrium adsorption is reached

(5)

Adsorption kinetics experiment. HF was injected into the reference cell to a calculated pressure value when the temperature of the adsorption system was constant with the experimental set value. The valve between the reference and sample cells was then opened, the HF flowed into the adsorption cell, and the system pressure was recorded in real time by the high-precision pressure transmitter until the final system pressure was stabilized at the desired value. The kinetic experimental conditions were 0.15, 0.35, 0.5, 0.70, 0.95, and 1.10 bar at 298 K and 298, 308, 318, 328, and 338 K at 0.5 bar

During the experiment, the calculation process for HF adsorption capacity was based on the mass conservation under the ideal gas state, so the adsorption capacity is [19]: $\begin{matrix} (1) & q = \frac{1}{R T} [p^{0} V_{R} - p^{1} (V_{R} + V_{v})], \end{matrix}$ where $q$ is the adsorption capacity, $p^{0}$ is the gas pressure in the reference cell before adsorption and $p^{1}$ is the system gas pressure during the adsorption process. $R$ and $T$ are the ideal gas constant and the system temperature, respectively.

2.4. Adsorption Theory

2.4.1. Isotherm Models

The relationship between adsorption capacity, temperature, and pressure in the adsorption process is usually described by the adsorption model. The Langmuir and Freundlich models had been successful in adapting the described HF adsorption process [11, 12]. The DA model is often used to describe the gas adsorption in micropores [22]. The DA model is expressed as follows: $\begin{matrix} (2) & q = q_{m} \exp (- {(\frac{R T}{E} \ln (\frac{p_{0}}{p}))}^{n}), \end{matrix}$ where $q$ and $q_{m}$ are the adsorption capacity and saturated adsorption capacity of gas, $p_{0}$ is the saturated vapor pressure of gas, and E is the adsorption characteristic energy.

The saturated vapor pressure $p_{0}$ of HF is calculated by the Antoine equation with extension [23]: $\begin{matrix} (3) & \log_{10} p_{0} = A + \frac{B}{T} + C \log_{10} T + D T + {E T}^{2}, \end{matrix}$ where the values of $A$ , $B$ , $C$ , $D$ , and $E$ are 23.7347, −1799.6, −6.1764, 5.0046 × 10⁻¹⁰, and 6.15 × 10⁻⁶, respectively. The unit of $p_{0}$ is mmHg in the calculation formula, which needs to be multiplied by 0.0013332 to convert to bar. The calculated saturated vapor pressures of 298, 308, 318, 328, and 338 K are 1.215, 1.678, 2.223, 3.130, and 4.149 bar, respectively.

2.4.2. Heat of Adsorption

The isosteric heat of adsorption is an important thermodynamic parameter which indicates that heat released since the adsorbed state is more stable than the unadsorbed state for adsorbate molecule [24]. The Clausius–Clapeyron equation was used to calculate the isosteric heat at constant adsorption capacity, and its expression is as follows [24]: $\begin{matrix} (4) & H_{s t} = {- R \frac{\partial \ln p}{\partial (1 / T)}|}_{q}, \end{matrix}$ where $H_{s t}$ is the calculated isosteric heat at the adsorption capacity $q$ .

The isosteric heat of adsorption ( $H_{s t}$ ) can be obtained by combining the DA model with the Clausius–Clapeyron equation [24]: $\begin{matrix} (5) & H_{s t} = E [{(- \ln θ)}^{1 / n} + \frac{α T}{n} {(- \ln θ)}^{1 - n / n}] + h_{f g}, \end{matrix}$ where $α$ is the thermal expansion coefficient of adsorbed phase and $α$ is calculated by $1 / T$ . $h_{f g}$ is the enthalpy of evaporation, which is expressed as follows: $\begin{matrix} (6) & h_{f g} = {- R \frac{\partial \ln p_{0}}{\partial (1 / T)}|}_{C} . \end{matrix}$

2.4.3. Adsorption Kinetics

(1) Adsorption Rate Models. The pseudo first-order (PFO) model assumes that the adsorption rate is directly proportional to the number of active sites. The rate expression is given by [25, 26]: $\begin{matrix} (7) & d q / d t = k_{1} (q_{1} - q), \end{matrix}$ where $q_{1}$ and $q$ represent the adsorption capacity at equilibrium and time $t$ , respectively, and $k_{1}$ is the first-order rate constant.

After integrating Eq. (7) with the boundary conditions ( $t$ =0, $q$ =0 and $t = t_{\infty}$ , $q = q_{1}$ ), $\begin{matrix} (8) & q = q_{1} (1 - e^{- k_{1} t}) . \end{matrix}$

The pseudo second-order model (PSO) assumes the adsorption rate is proportional to the square of number of active sites; therefore, the rate model is given by [25, 26]: $\begin{matrix} (9) & d q / d t = k_{2} {(q_{2} - q)}^{2}, \end{matrix}$ where $q_{2}$ represents the equilibrium adsorption capacity, and $k_{2}$ is the second-order rate constant.

With the boundary conditions ( $t$ =0, $q$ =0 and $t = t_{\infty}$ , $q = q_{2}$ ), the PSO is given by [25, 26]: $\begin{matrix} (10) & q = k_{2} q_{2}^{2} t / (1 + k_{2} q_{2} t) . \end{matrix}$

(2) Diffusion Models. (1)

Film diffusion

Boyd’s film diffusion model assumes that the main diffusion resistance is boundary film around the adsorbent particles [26, 27]. Therefore, the model was used to distinguish whether the film diffusion was a rate-controlling step in HF adsorption on 3A. The expression of this model is given by [26]: $\begin{matrix} (11) & F = 1 - \frac{6}{π^{2}} \sum_{n = 1}^{\infty} \frac{1}{n^{2}} \exp (- n^{2} B_{t}), \end{matrix}$

where $F$ is the fraction at times $t$ and can be expressed using adsorption capacity: $\begin{matrix} (12) & F = \frac{q_{t}}{q_{m}} . \end{matrix}$

$B_{t}$ can be expressed in $F$ [28]:

when $F > 0.85$ , $\begin{matrix} (13) & B_{t} = - 0.4977 - \ln (1 - F), \end{matrix}$ when $F < 0.85$ , $\begin{matrix} (14) & B_{t} = {(\sqrt{π} - \sqrt{π - (\frac{π^{2} F}{3})})}^{2} . \end{matrix}$ (2)

Interparticle diffusion

The model assumes interparticle diffusion control the adsorption rate, the model is given by the following expression [26, 27]: $\begin{matrix} (15) & \frac{q_{t}}{q_{e}} = 1 - \frac{6}{π^{2}} \sum_{n = 1}^{\infty} \frac{1}{n^{2}} \exp (- \frac{n^{2} π^{2} D_{c} t}{r_{p}^{2}}), \end{matrix}$ where $q_{t}$ is adsorption amount at time $t$ , $q_{e}$ is adsorption amount at infinite time, $D_{c}$ is interparticle diffusion coefficient, and $r_{p}$ is the radius of the particle. When $q_{t} / q_{e}$ value is >70%, Eq. (15) can be simplified to [26]: $\begin{matrix} (16) & \frac{q_{t}}{q_{e}} = 1 - \frac{6}{π^{2}} \exp (- \frac{π^{2} D_{c} t}{r_{p}^{2}}) . \end{matrix}$

By plotting $\ln [1 - q_{t} / q_{e}]$ versus $t$ , the linear slope is $- π^{2} (D_{c} / {r^{2}}_{p})$ and the intercept is $ln 6 / π^{2}$ . (3)

Intraparticle diffusion

If the HF adsorption process on zeolite 3A is controlled by intraparticle diffusion, the HF adsorption ( $q_{t}$ ) should change linearly with time square root ( $t^{1 / 2}$ ) [26, 27]. The intraparticle diffusion model assumes three steps typically in adsorption: bulk or film diffusion, gradual adsorption, and the final adsorption equilibrium [26]. The intraparticle diffusion model can be used for determining the adsorption mechanism, and the model is given by [29]: $\begin{matrix} (17) & q_{t} = k_{i} t^{1 / 2} + C, \end{matrix}$ where $k_{i}$ is the intraparticle diffusion constant and $C$ is the intercept. The intercept value is related to the film thickness. A large intercept value suggests a great film effect.

3. Results and Discussion

3.1. Characterizations of the Zeolite 3A

The powder XRD pattern of zeolite 3A is shown in Figure 2. The XRD characteristic diffraction peaks are in accordance with 3A crystals by Wang [30], indicating that the sample is a typical 3A zeolite. The FT-IR of zeolite 3A spectra is shown in Figure 3. The broad bands around 3,447.9 cm⁻¹ belong to the stretching vibration of the O-H group. The bands at around 1,659.3 cm⁻¹ belong to the bending and tensile vibrations of the O-H and Si-OH groups, respectively [31]. At about 1,005.5, 665.8, and 462.9 cm⁻¹, three typical absorption peaks formed by amorphous aluminosilicate structure were observed, which belong to the asymmetric tensile vibration of the tetrahedron in the sample, the symmetric tensile vibration of the Si-O-Si bond, and the bending vibration of T-O (T is Si or Al) bond, respectively [31]. The same absorption peaks have also been found by Cao et al. [31]. The absorption peak at 554.6 cm⁻¹ belongs to the vibration of double binary ring D4R, the main subunit structure unit of zeolite 3A.

Figure 2

XRD pattern of zeolite 3A.

Figure 3

IR spectra of zeolite 3A.

The surface morphology of zeolite 3A is shown in Figure 4, which indicates that the surfaces of the samples are not homogeneous. Zeolite 3A has a large grain size, good dispersibility, and many surface fractures, which has also been found in the literature [30]. The low-temperature CO₂ isotherm and PSD of zeolite 3A are shown in Figure 5. The pore size of zeolite 3A is mainly concentrated at ~0.3 and 0.37 nm. The specific surface area of zeolite 3A and the pore volume are 402.81 m²/g and 0.1325 cm³/g, respectively.

Figure 4

SEM images of zeolite 3A.

Figure 5

(a) Low-temperature CO₂ isotherm. (b) PSD of zeolite 3A.

(b)

3.2. Adsorption Equilibrium

3.2.1. Adsorption Isotherms

The HF adsorption isotherms on zeolite 3A at 298 to 338 K are shown in Figure 6. All adsorption isotherms are type I according to the IUPAC classification. The adsorption amount of HF on zeolite 3A is close to that on activated carbon by Shen et al. [15] but less than that on activated carbon by Bahrami et al. [19] and MOF [18], which is mainly due to the pore characteristics and surface physicochemical properties of the adsorbent. The adsorption capacity increases with increasing pressure at the same temperature. The adsorption capacity decreases with the temperature increase because the binding strength of the zeolite 3A surface decreases with temperature, indicating that the HF adsorption on the zeolite 3A surface is exothermic. The Langmuir, Freundlich, and DA models were used to correlate the HF adsorption data using regression analysis [19]. Three statistical functions, including SSE, RMSE, and $χ^{2}$ , were used to estimate the goodness of fit [24].

Figure 6

Adsorption isotherm and DA fitting of HF on zeolite 3A.

The fitting parameter values for the three models are listed in Table 1. The Langmuir parameters $q_{0}$ and $b_{L}$ all decreased with the increasing temperature, which implied that the bonding of HF to zeolite decreased. The Freundlich parameter $k_{F}$ represents the adsorption capacity, and $n$ shows the adsorption driving force and indicates the favourability of adsorption [10]. The $k_{F}$ values showed a negative correlation with temperature. The $n$ values (1.271–1.368), determined from the Freundlich model suggested the moderately difficult nature of HF adsorption on zeolite 3A [10]. The same results for the Freundlich parameter $n$ have been found in HF adsorption on produced NaF pellets and activated carbon [10–12].

Table 1

Fitting results of the studied isotherm model.

Model	Fitting parameters	Temperature (K)
Model	Fitting parameters	298	308	318	328	338
Langmuir	$q_{0}$ /mg·g^-1	1.027	0.988	0.921	0.875	0.827
	$b_{L}$ /bar^-1	0.618	0.571	0.554	0.505	0.481
	SSE/×10^-5	3.753	4.301	3.862	4.153	3.922
	RMSE/%	0.367	0.435	0.374	0.432	0.389
	$χ^{2}$ /%	0.015	0.027	0.022	0.019	0.023

Freundlich	$k_{F}$ /mg·g^-1	0.394	0.359	0.331	0.292	0.270
	$n$	1.360	1.368	1.302	1.373	1.271
	SSE/×10^-5	4.273	4.681	4.381	4.872	4.396
	RMSE/%	0.374	0.424	0.395	0.433	0.416
	$χ^{2}$ /%	0.017	0.024	0.023	0.020	0.021

DA	$q_{m}$ /mg·g^-1	0.4330	0.4913	0.5236	0.5452	0.5631
	$E$ /J·mol^-1	3472.04	3753.22	4146.69	4523.53	5005.89
	$n$	1.1926	1.2186	1.2423	1.2947	1.3568
	SSE/×10^-5	3.085	2.723	2.951	2.843	3.021
	RMSE/%	0.306	0.355	0.393	0.347	0.316
	$χ^{2}$ /%	0.014	0.017	0.016	0.014	0.014

The values of SSE, RMSE, and $χ^{2}$ in Table 1 for the DA model are the smallest among the three adsorption models, which indicates that the DA model predicted the adsorption data better than the other models. Figure 6 also shows the predicted HF adsorption data by the DA model. The DA model can describe well the HF adsorption curve, indicating that micropore filling existed in HF adsorption on zeolite 3A. Both characteristic energy $E$ and DA parameter $n$ increased with the increase of temperature. The DA parameter $n$ can be used to describe the surface heterogeneity of the adsorbent [13, 24]. The study shows that the closer the $n$ value is to 1, the higher the heterogeneity of the zeolite 3A surface, which is more related to the active adsorption site at low temperatures. The n value shows a positive correlation with the characteristics energy value $E$ (Figure 7). The lower $n$ value results in $E$ value.

Figure 7

Linear relationship between DA parameter $n$ and $E$ .

The regeneration of 3A zeolite affects the reuse of materials [18]. The adsorption capacity after regeneration was tested at 298 K. The zeolite 3A after HF adsorption was treated at 393 K in the muffle furnace for 12 h and then quickly transferred and sealed to the adsorption equipment (sample cell in Figure 5) and then vacuumed at 373 K for 10 h to reactivate its capacity. The HF repeatability experiments on zeolite 3A at 298 K are shown in Figure 8(a). The remaining HF capacity after regeneration was slightly reduced, which suggested the high stability of exposure to dry HF gas. The IR spectra of the zeolite 3A before and after HF adsorption are shown in Figure 8(b), suggesting no change in the surface functional group of the zeolite 3A molecular sieve after HF adsorption.

Figure 8

(a) Repeatability experiments of HF on zeolite 3A at 298 K. (b) IR spectra of the zeolite 3A before and after HF adsorption.

(b)

3.2.2. Heat of Adsorption

The isosteric heat of adsorption ( $H_{s t}$ ) calculated using Eq. (5) is plotted with the HF surface load ( $q / q_{m}$ ), as shown in Figure 9. $H_{s t}$ values of HF on zeolite 3A decrease with the increase of surface loading, which indicates the surface energy heterogeneity of zeolite 3A. A large number of vacant adsorption sites exist on the zeolite 3A surface at the initial stage, which results in strong interaction between the HF molecules and zeolite 3A. Therefore, the $H_{s t}$ value is high at the initial adsorption stage. With further adsorption capacity increase, the adsorption sites are occupied, resulting in the weakening of the interaction between the HF molecules and zeolite 3A. Therefore, $H_{s t}$ decreases with the surface loading.

Figure 9

The isosteric heat of adsorption ( $H_{s t}$ ) of HF.

3.3. Adsorption Kinetics

3.3.1. Effect of Pressure on Adsorption Kinetic

The adsorption kinetic curve of HF on zeolite 3A at 298 K is shown in Figure 10. The adsorption capacity continuously increases with the adsorption time. The adsorption capacity rapidly increases in the initial stage, which is mainly due to the high adsorption energy sites on the adsorbent surface and the rapid occupation of HF molecules onto the adsorption sites, so the adsorption rate is high. When large numbers of active adsorption sites are occupied, the adsorption sites available for HF molecules become less; hence, the adsorption rate becomes slow. When the active sites are almost occupied, HF molecules reach dynamic equilibrium on the surface of zeolite 3A. At this time, the number of HF molecules on the adsorbent surface will not increase; therefore, the total adsorption rate tends to zero and the adsorption capacity will not increase.

Figure 10

Dynamic curves of HF under different pressures at 298 K.

(1) Adsorption Rate. The HF adsorption kinetics curves and the corresponding curves predicted by the PFO model are shown in Figure 10. The kinetic constants calculated with the PFO and PSO models and the characteristic parameters of the kinetic model are listed in Table 2. The larger $R^{2}$ of the PFO model indicates a better predictivity of the adsorption kinetic process over the whole pressure range well, so the adsorption of HF by zeolite 3A follows simple first-order kinetics. The calculated $q_{1}$ values from the PFO model are more in accordance with the experimental data than the calculated $q_{2}$ values by the PSO model.

Table 2

The fitting kinetic parameters for PFO and PSO.

Pressure (bar)	Pseudo first-order			Pseudo second-order
Pressure (bar)	$q_{1}$ /mg·g^-1	$k_{1}$ /min^-1	$R^{2}$	$q_{2}$ /mg·g^-1	$k_{2}$ /g mg^-1.Min^-1	$R^{2}$
0.15	0.0744	0.0395	0.9999	0.0844	0.6352	0.9783
0.35	0.1768	0.0472	0.9998	0.1966	0.3485	0.9726
0.50	0.2580	0.0534	0.9998	0.2832	0.2859	0.9685
0.70	0.3153	0.0616	0.9997	0.3426	0.2863	0.9638
0.95	0.3746	0.0687	0.9998	0.4037	0.2806	0.9602
1.10	0.4206	0.0743	0.9997	0.4510	0.2775	0.9576

Both $q_{1}$ and $q_{2}$ increase with pressure, which is consistent with the high pressure is conducive to HF adsorption. Meanwhile, the PFO and PSO model parameters ( $k_{1}$ and $k_{2}$ ) increase with pressure because the velocity of the HF molecule will increase and the migration speed of HF molecules inside the pores will be faster at higher pressure, leading to an increase in the diffusion rate. The increased mass transfer coefficient makes the kinetic curve steeper. The $k_{1}$ values (0.0395–0.0743 min⁻¹) for HF on zeolite 3A are larger than that of H₂S (0.00177 min⁻¹) on zeolite 3A, which suggested that the HF adsorption rate is faster than that of H₂S [6, 32].

(2) Diffusion Mechanism. The HF adsorption process on zeolite 3A can be divided into five continuous processes, as shown in Figure 11: (a) HF molecules diffuse from free state to adsorbent gas film (bulk diffusion), (b) HF molecules diffuse in the film (film diffusion), (c) HF molecules diffuse in pores between particles (interparticle diffusion), (d) HF molecules diffuse in 3A crystal (intraparticle diffusion), and (e) surface adsorption. Generally, surface adsorption and bulk diffusion are very fast, so they are ignored in the analysis of the HF adsorption kinetic process, and the role of the other three processes in the adsorption process is mainly considered. (1)

Film diffusion

Figure 11

Adsorption mechanism of HF on 3A zeolite.

Equations (11)–(14) can be used to predict HF adsorption control by particle or film diffusion [26, 27]. By plotting $B_{t}$ vs. time, the mass transfer rate is controlled by pore diffusion if the plot is a straight line passing through the origin [26–28]. If the curve is nonlinear or does not pass through the origin, film diffusion can also be considered to affect the adsorption rate [26–28]. The calculated $B_{t}$ values as a function of time are shown in Figure 12, and the curves are not entirely linear. The slopes and intercepts of the film diffusion model at all pressures are given in Table 3. The existence of intercept values suggests that the rate-controlling process for HF adsorption on zeolite 3A is mainly controlled by external mass transfer [26]. The intercept increases close to zero with increasing pressure, which indicates that pore diffusion appears to be more important at slightly higher pressures. (2)

Interparticle diffusion

Figure 12

Plots of Boyd’s film model for HF adsorption on 3A.

Table 3

Kinetic parameters of linear fitting of film diffusion model.

Pressure (bar)	Slope	Intercept	$R^{2}$
0.15	0.03954	−0.42992	0.99765
0.35	0.04693	−0.42781	0.99872
0.50	0.05256	−0.40566	0.99926
0.70	0.06069	−0.40949	0.99933
0.95	0.0678	−0.41098	0.99936
1.10	0.07321	−0.41012	0.99936

The fitting results for the adsorption kinetics data with the interparticle diffusion linear form are shown in Figure 13 and Table 4. A good correlation between $\ln [1 - q_{t} / q_{e}]$ with $t$ indicates that the experimental data can be well described by the model. The $D_{c} / {r^{2}}_{p}$ and $D_{c}$ values calculated by the slope of the linear fit increase continuously with increasing pressure, indicating that high pressure favours the HF gas diffusion in zeolite 3A, which is consistent with the previous calculation that high pressure favours the adsorption rate. The order of magnitude of interparticle diffusion coefficient $D_{c}$ of HF is 10⁻⁹ m²/s. The intercept values obtained by linear fitting the experimental data are not close to −0.4977 ( $ln 6 / π^{2}$ ), which indicates that the HF adsorption rate on zeolite 3A is not completely controlled by the interparticle diffusion. The calculated interparticle diffusion coefficient $D_{c} / {r^{2}}_{p}$ versus first-order rate constant $k_{1}$ does not satisfy the 1/15 relationship and also reflects that the HF diffusion over zeolite 3A is not completely controlled by interparticle diffusion [26]. (3)

Intraparticle diffusion

Figure 13

Linear fitting for experimental data with interparticle diffusion model.

Table 4

The linear fitting results for interparticle diffusion model.

Pressure (bar)	Slope	Intercept	$D_{c} / {r^{2}}_{p}$ (10^-3 s^-1)	$D_{c}$ (10^-9 s^-1)
0.15	−0.04085	0.09071	4.143	1.036
0.35	−0.04809	0.06103	4.877	1.219
0.50	−0.05338	0.00001	5.414	1.354
0.70	−0.06156	0.00003	6.244	1.561
0.95	−0.06873	0.00003	6.971	1.743
1.10	−0.07424	0.00002	7.530	1.882

The intraparticle diffusion curves for HF adsorption on zeolite 3A are shown in Figure 14. The curves are observed to be nonlinear. According to the intraparticle diffusion model, the adsorption amount $q_{t}$ plot vs. $t^{1 / 2}$ should be a straight line and pass through the origin of the mass transfer rate solely controlled by intraparticle diffusion [26]. Multilinearity can be observed when different processes are related to the adsorption mechanism. The slope of each linear portion represents the corresponding adsorption rate [27]. The curves in Figure 14 show quart-linear characteristics, indicating the presence of four consecutive HF adsorption steps on zeolite 3A. The initial part, region I, was ascribed to the HF diffusion from the free state to and through the gas film of zeolite 3A, and the initial stage was associated with the HF film diffusion. The second and third parts, regions II and III describe the adsorption-progressive stage where diffusion occurs within the particle. Regions II and III are related to interparticle and intraparticle adsorptions, respectively. The fourth part, region IV, is the final equilibrium adsorption stage where the diffusion becomes slower owing to the active adsorption site saturation. Figure 14 shows that the extended line for straight lines corresponding to regions II and III does not pass through the origin, which may result from the different diffusion rates for the first and other stages of HF adsorption.

Figure 14

Prediction of intraparticle diffusion model for HF adsorption on 3A.

The adsorption rate of each HF molecule is controlled by four processes at a certain pressure. In the first stage, the adsorption rate is controlled by the film diffusion, as shown in segment I of Figure 14. The interparticle diffusion would be the rate-controlling process when the HF molecules diffuse into the particle pore. The intraparticle diffusion then controls the rate. Finally, the rate is governed by surface adsorption. The slope of each linear segment before region IV can be used to deduce which stage controls the whole HF adsorption rate [27]. Moreover, Figure 14 shows that the linear slope in region III is the smallest. The linear portion with the lowest slope is the rate-controlling step [27, 28]. This suggests that HF diffusion into the 3A particle is slower and the intraparticle diffusion stage is the rate-controlling stage. The linear slopes for regions II and III gradually increase with increasing pressure, which may be because the film thickness decreases as the pressure rises and the film diffusion rate will increase.

3.3.2. Effect of Temperature

Kinetic experiments at 298–338 K and 0.50 bar were conducted to investigate the effect of temperature on the HF kinetic adsorption on zeolites 3A. The HF adsorption kinetics curves at five temperatures are shown in Figure 15. The HF kinetic curves at five temperatures can be divided into three different stages: rapid adsorption, slow adsorption, and adsorption equilibrium stages. The HF adsorption capacity gradually decreases with increasing temperature, which is consistent with the discovery in literature [19]. High temperature enables the HF molecules to obtain more energy to overcome the intermolecular force and electrostatic interaction; thus, HF molecules remain in the free state. Simultaneously, the affinity between HF molecules and zeolite 3A surface will be weakened by the increasing temperature and will result in the decrease in adsorption capacity. In addition, high temperature is not conducive to the formation of a more orderly and stable configuration of adsorbed molecules, resulting in the reduction of adsorption capacity. Higher temperature is not conducive to improving HF adsorption recovery. It can also be found that the time to reach equilibrium decreases with the increasing temperature, which also shows that the temperature is conducive to the gas adsorption rate.

Figure 15

Dynamic curves of HF under different temperature at 0.5 bar.

(1) Adsorption Rate. The PFO and PSO models were used to predict the HF adsorption kinetic curves. The fitting results are shown in Figure 15 and Table 5. The adsorption rate constants $k_{1}$ and $k_{2}$ of HF increase with temperature, indicating that the HF adsorption rate shows a positive correlation with temperature, which is due to the acceleration of molecular motion caused by temperature increase. $q_{1}$ and $q_{2}$ decrease with increasing temperature, suggesting that higher temperature is conducive to gas diffusion; however, it will reduce the gas adsorption capacity.

Table 5

The fitting kinetic parameters for PFO and PSO.

Temperature (K)	Pseudo first-order			Pseudo second-order
Temperature (K)	$q_{1}$ (mg/g)	$k_{1}$	$R^{2}$	$q_{2}$ (mg/g)	$k_{2}$	$R^{2}$
298	0.2580	0.0534	0.9898	0.2832	0.2859	0.9685
308	0.2213	0.0592	0.9873	0.2411	0.3861	0.9726
318	0.1990	0.0633	0.9867	0.2157	0.4715	0.9724
328	0.1803	0.0681	0.9897	0.1944	0.5757	0.9675
338	0.1601	0.0733	0.9837	0.1718	0.7166	0.9523

(2) Diffusion Mechanism. (1)

Film diffusion

The calculated $B_{t}$ values are shown in Figure 16. Moreover, $B_{t}$ is not completely linear. The slope and intercept of Boyd’s film diffusion model are listed in Table 6. The existence of intercept value shows that the HF adsorption rate on zeolite 3A was mainly controlled by external mass transfer. The intercept values decrease slightly with increasing temperature, suggesting that pore diffusion become more important at high temperature. The $B_{t}$ values show a positive correlation with temperature, indicating that the film diffusion rate increases with temperature increase. (2)

Interparticle diffusion

Figure 16

Plots of Boyd’s film model for HF adsorption on 3A.

Table 6

Kinetic parameters of linear fitting of film diffusion model.

Temperature (K)	Slope	Intercept	$R^{2}$
298	0.0526	−0.4048	0.9792
308	0.0583	−0.4042	0.9872
318	0.0623	−0.4054	0.9893
328	0.0670	−0.4044	0.9892
338	0.0721	−0.4032	0.9867

The linear fitting results for the HF adsorption kinetic curve using the interparticle diffusion model at five temperatures and 0.5 bar are shown in Figure 17 and Table 7. A good correlation between $\ln [1 - q_{t} / q_{e}]$ shows that $t$ indicates that the experimental data can be well described by the model. In addition, Table 7 shows that $D_{c} / {r^{2}}_{p}$ and $D_{c}$ values increase with increasing temperature. Based on molecular dynamics theory, the essence of HF diffusion in zeolite 3A is due to the molecular irregular thermal movement. The average kinetic energy of HF molecules increases with temperature, which leads to an increase in amplitude and frequency of molecular vibration. The higher molecular irregular thermal movement intensity occurs at a higher temperature. Consequently, the HF diffusion rate in zeolite 3A becomes faster. (3)

Intraparticle diffusion

Figure 17

Linear fitting for experimental data with interparticle diffusion model.

Table 7

The linear fitting results for interparticle diffusion model.

Temperature (K)	Slope	Intercept	$D_{c} / {r^{2}}_{p}$ (10^-3 s^-1)	$D_{c}$ (10^-9 s^-1)
298	−0.0534	0.00001	5.416	1.354
308	−0.0592	−0.00003	6.004	1.501
318	−0.0633	0.00008	6.420	1.605
328	−0.0681	−0.00001	6.907	1.727
338	−0.0733	0.00001	7.434	1.859

The quart-linear characteristics for intraparticle diffusion curves can be observed in Figure 18, which indicated that three different processes are involved in the adsorption mechanism. The curves of the five temperatures in the first stage I are relatively close, indicating that the influence of temperature on the HF diffusion velocity from the free state to and through the gas film of zeolite 3A is little. In the second and third stages, the difference between the five curves increases with time. The linear slope for each stage decreases with increasing temperature, suggesting that the pore diffusion rate in the particles decreases slightly at this stage, which may be due to the impediment of diffusion by the interaction of HF molecules under high pressure. The same trend is also found in CO₂ in porous MgO [27]. The extended line for straight lines corresponding to regions II and III does not pass through the origin. The linear portion for stage III with the lowest slope indicates that the diffusion rate is controlled by intraparticle diffusion.

Figure 18

Prediction of intraparticle diffusion model for HF adsorption on 3A.

4. Conclusion

HF adsorption isotherms data on zeolite 3A at 298–338 K were type I. The adsorption capacity increased with increasing pressure and decreases with increasing temperature. HF adsorption isotherms data on zeolite 3A can be described well by the DA model. The $H_{s t}$ values of HF on zeolite 3A decreased with the increase of surface load, which suggested the surface energy heterogeneity of zeolite 3A.

The HF adsorption by zeolite 3A at different adsorption conditions followed simple first-order kinetics. The first-order rate increased with pressure and temperature. The calculated $B_{t}$ values showed nonlinear characteristics with time. $B_{t}$ values increased with increasing pressure and temperature, indicating that high pressure and temperature favour film diffusion. The order of magnitude of the interparticle diffusion coefficient $D_{c}$ of HF on zeolite 3A was 10⁻⁹ m²/s. The higher pressure and higher temperature favour the HF diffusion process. The intraparticle diffusion curves for HF adsorption on zeolite 3A showed quart-linear characteristics, indicating the presence of three consecutive HF adsorption steps on zeolite 3A. In the initial adsorption stage, the film diffusion controlled the adsorption rate. The intraparticle diffusion would be the rate-controlling process after HF molecules diffuse through the film. The intraparticle diffusion then controls the rate. Finally, the rate was governed by surface adsorption. HF diffusion into the particle of zeolite 3A was the slowest and therefore, the intraparticle diffusion is the rate-controlling stage.

In this study, the HF adsorption mechanism onto zeolite 3A was evaluated based on kinetics and thermodynamics. However, only one adsorbent zeolite 3A was studied due to limited conditions, and other porous materials have not been analyzed in this paper. The adsorption kinetics and thermodynamics of HF on different materials will be compared in subsequent studies.

Footnotes

Data Availability

All data included in this study are available upon request by contact with the corresponding author.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

This work was supported by the Project of the State Key Laboratory of Laser Interaction with Matter (SKLLIM1813), Natural Science Foundation of Hebei Province (E2020402072), and The Handan Science and Technology Research and development projects (19422091008-31).

References

Kumar

Suganya

Srinivas

Priyadharshini

Lichtfouse

Treatment of fluoride-contaminated water: a review

Environmental Chemistry Letters 2019 17

5790408

4 1707 1726

10.1007/s10311-019-00906-9

2-s2.0-85069533087

Wang

Deng

Zhao

Analysis of energy and consumption reduction in adsorption recovery of electronic HF

Chemical Industry and Engineering Progress 2021 40

5790408

3645 3655

Zhou

Huang

Tang

Huang

Sheng

Luan

Purify technology of laser medium in non-chain pulsed HF laser

Acta Photonica Sinica 2016 45

5790408

3, article 314004

10.3788/gzxb20164503.0314004

2-s2.0-84963537888

Lucsa

Smektala

Adam

J. L.

Fluorine in optics

Journal of Fluorine Chemistry 2002 114

5790408

2 113 118

10.1016/S0022-1139(02)00016-7

2-s2.0-0036011206

Mcpake

C. B.

Sandford

Selective continuous flow processes using fluorine gas

Organic Process Research Development 2012 16

5790408

5 844 851

10.1021/op200331s

2-s2.0-84862068927

Starke

Pasel

Bläker

Eckardt

Zimmermann

Bathen

Impact of Na⁺ and Ca²⁺ cations on the adsorption of H₂S on binder-free LTA zeolites

Adsorption Science and Technology 2021 2021, article 5531974

5790408

10.1155/2021/5531974

Jing

Ying

Adsorption and recovery technology of excess hydrogen fluoride in the production of fluorine-containing chemicals

Organo-Fluorine Industry 2007 3

5790408

Luo

Wang

Yang

Progress in the techniques and applications of dealing with hydrogenl fluoride

New Chemical Materials 2011 39

5790408

33 36

Jiang

Hydrogen fluoride adsorption ability of some inorganic compounds

Advanced Materials Research 2011 412

5790408

1 4

10.

Tavakoli

Ghasemi

Equilibrium, kinetics and breakthrough studies for adsorption of hydrogen fluoride on sodium fluoride

Chemical Engineering and Processing 2010 49

5790408

4 435 440

10.1016/j.cep.2010.02.006

2-s2.0-77951104196

11.

Afzal

Rahimi

Ehsani

Tavakoli

Experimental study of hydrogen fluoride adsorption on sodium fluoride

Journal of Industrial and Engineering Chemistry 2010 16

5790408

1 147 151

10.1016/j.jiec.2010.01.004

2-s2.0-77249140449

12.

Afzal

Rahimi

Ehsani

Tavakoli

Modeling hydrogen fluoride adsorption by sodium fluoride

Journal of Industrial and Engineering Chemistry 2010 16

5790408

6 978 985

10.1016/j.jiec.2010.04.001

2-s2.0-77957890161

13.

Liu

Wang

Dong

The optimal technique parameter mensuration of sodium fluoride adsorb hydrogen fluoride

Ship Science and Technology 2009 31

5790408

109 111

14.

Adsorption characteristics of hydrogen fluoride on alumina

Environmental Pollution Control 1990 12

5790408

34 36

15.

Sheng

Theoretical and Experimental Study on Dry Adsorption of HF Gas 2004

Zhejiang, China

Zhejiang University

16.

Che

Studies on adsorption properties of DB18C6/SiO2 Composite material toward Zr (IV) AND Hf (IV)

Mining Metallurgy 2015 24

5790408

35 40

17.

Siahooei

M. A.

Bordbari

Adsorption of Fluoride Gases in Aluminum Production by Using of Nanotechnology 2019

Berlin

Springer International Publishing

10.1007/978-3-030-05861-6_11

2-s2.0-85064879836

18.

Wang

Liu

Sun

Xie

Pan

New challenge of microporous metal-organic frameworks for adsorption of hydrogen fluoride gas

Materials Letters 2017 197

5790408

175 179

10.1016/j.matlet.2017.03.111

2-s2.0-85016178351

19.

Bahrami

Safdari

Moosavian

M. A.

Torab-Mostaedi

Adsorption of hydrogen fluoride onto activated carbon under vacuum conditions: equilibrium, kinetic and thermodynamic investigations

Chemical Industry and Chemical Engineering Quarterly 2012 18

5790408

4-1 497 508

10.2298/CICEQ110920027B

2-s2.0-84871991748

20.

Dobladez

J. A. D.

Maté

V. I. Á.

Torrellas

S. Á.

Larriba

Muñoz

G. P.

Sánchez

R. A.

Comparative simulation study of methanol production by CO₂ hydrogenation with 3A, 4A and 5A zeolites as adsorbents in a PSA reactor

Separation and Purification Technology 2021 262, article 118292

5790408

10.1016/j.seppur.2020.118292

21.

Mccarty

R. D.

Arp

V. D.

A new wide range equation of state for helium

Advances in Cryogenic Engineering 1990 35

5790408

1465 1475

22.

Duong

D. D.

Adsorption Analysis: Equilibria and Kinetics 1998

London

Imperial College Press

23.

Yaws

C. L.

Satyro

M. A.

The Yaws Handbook of Vapor Pressure (Second Edition) 2015

Gulf Professional Publishing

24.

Duan

Tao

Huang

Adsorption characteristics and thermodynamic property fields of methane and Sichuan Basin shales

Adsorption 2022 28

5790408

1-2 41 54

10.1007/s10450-021-00352-6

25.

Rodríguez-Narciso

Lozano-Álvarez

Salinas-Gutiérrez

Castañeda-Leyva

A stochastic model for adsorption kinetics

Adsorption Science and Technology 2021 2021, article 5522581

5790408

10.1155/2021/5522581

26.

Loganathan

Tikmani

Edubilli

Mishra

Ghoshal

CO₂ adsorption kinetics on mesoporous silica under wide range of pressure and temperature

Chemical Engineering Journal 2014 256

5790408

1 8

10.1016/j.cej.2014.06.091

2-s2.0-84904105049

27.

Song

Zhu

Chen

Liao

Ding

Chen

An investigation of CO₂ adsorption kinetics on porous magnesium oxide

Chemical Engineering Journal 2016 283

5790408

175 183

10.1016/j.cej.2015.07.055

2-s2.0-84938319433

28.

Basu

Ghosh

Saha

Adsorption characteristics of phosphoric acid induced activation of bio- carbon: equilibrium, kinetics, thermodynamics and batch adsorber design

Process Safety and Environmental Protection 2018 117

5790408

125 142

10.1016/j.psep.2018.04.015

2-s2.0-85046635473

29.

Tan

K. L.

Hameed

B. H.

Insight into the adsorption kinetics models for the removal of contaminants from aqueous solutions

Journal of the Taiwan Institute of Chemical Engineers 2017 74

5790408

25 48

10.1016/j.jtice.2017.01.024

2-s2.0-85015798722

30.

Wang

Measurements and modeling of water adsorption isotherms of zeolite linde-type A crystals

Industrial & Engineering Chemistry Research 2020 59

5790408

17 8304 8314

10.1021/acs.iecr.9b06891

31.

Cao

Wang

Sun

Green synthesis of magnetic zeolite LTA using NaOH activated fly ash

Zeitschrift fitsanorganische und allgemeine Chemie 2020 646

5790408

20 1666 1670

10.1002/zaac.202000215

32.

Georgiadis

A. G.

Charisiou

N. D.

Gaber

Polychronopoulou

Goula

M. A.

Adsorption of hydrogen sulfide at low temperatures using an industrial molecular sieve: an experimental and theoretical study

ACS Omega 2021 6

5790408

23 14774 14787

10.1021/acsomega.0c06157

34151059

Equilibrium and Kinetic Adsorption of Hydrogen Fluoride onto Zeolite 3A

Abstract

1. Introduction

2. Materials and Methods

2.1. Materials

2.2. Characterization of the Zeolite 3A

2.3. Adsorption Isotherm and Kinetics Measurements

2.4. Adsorption Theory

2.4.1. Isotherm Models

2.4.2. Heat of Adsorption

2.4.3. Adsorption Kinetics

3. Results and Discussion

3.1. Characterizations of the Zeolite 3A

3.2. Adsorption Equilibrium

3.2.1. Adsorption Isotherms

3.2.2. Heat of Adsorption

3.3. Adsorption Kinetics

3.3.1. Effect of Pressure on Adsorption Kinetic

3.3.2. Effect of Temperature

4. Conclusion

Footnotes

Data Availability

Conflicts of Interest

Acknowledgments

References