Water in Petroleum and Petroleum Products

Abstract

The chapter presents experimental data, published in numerous source materials and reviews, on the mutual solubility of water in hydrocarbons and hydrocarbons in water, relationship of water solubility in hydrocarbons depending on the structure of the organic compound molecule and the change of solubility as a function of temperature. Possibilities of water solubilization, dissolved and dispersed in hydrocarbons, their mixtures as well as fractions, and petroleum products were analyzed. With the help of turbidimetric measurements, surfactants and mixtures of surfactants with the highest water solubilization capacity in fuels were selected. Different methods of dewatering of distillate fractions, gasolines, and diesel fuels with the use of coalescing partitions were investigated. A number of barrier materials, methods of modifying the structure of the partition, and the hydrophilicity of glass fibers were tested, obtaining very good results in industrial applications.

1. Introduction

Water dissolved in hydrocarbons forms molecular solutions and does not dissociate into ions. Water shows a high value of the dipole moment, a tendency to form hydrogen bonds, and a low molar volume. On the other hand, hydrocarbons are characterized by a low or no dipole moment of molecules, no tendency to form hydrogen bonds, and a large molar volume. Due to the significant differences in the molecular structure, water shows limited solubility in hydrocarbons.

The key properties of water that determine its solubility in hydrocarbons are the dipole moment, the dielectric constant, and the ability to form hydrogen bonds between its own molecules and form hydrogen bonds with π electrons of unsaturated and aromatic hydrocarbons.

1.1 General characteristics of the liquid-liquid equilibrium in the hydrocarbon-water system

The coexistence of two liquid phases characterizes thermodynamic equilibrium in the water-hydrocarbon system. One of these liquids is almost pure water saturated with hydrocarbon; the other liquid contains dissolved water. Pressure has little effect on solubility, so this discussion is about solubility under saturated vapor pressure as measurements of this type are usually reported in the literature. Under the conditions of saturated steam pressure, the compositions of both phases in a given system depend only on temperature. The temperature dependence of benzene solubility in water, Figure 1, shows a clear minimum at the temperature of 290 K. Such minima are also observed for other hydrocarbons. At higher temperatures, the solubility of hydrocarbons increases, but the growth is slow. Even at 550 K, the mole fraction of benzene in water is less than 0.001.

Figure 1.

The solubility of benzene in water as a function of temperature [1]. Where: x_b—Mole solubility of benzene in water. Reprinted with permission from: [2]. Copyright 2004, American Institute of Physics.

The curve of water solubility in benzene behaves completely differently—Figure 2. The solubility of water in hydrocarbons does not show a minimum and increases rapidly with increasing temperature. At room temperature, the mole fraction of water is less than 0.001, and as the temperature increases, it becomes comparable to the mole fraction of hydrocarbon.

Figure 2.

Water solubility in benzene as a function of temperature [3]. Where: x_w—Mole solubility of water in benzene. Reprinted with permission from: [2]. Copyright 2004, American Institute of Physics.

Both liquid phases coexist with each other up to the critical solubility temperature. If the hydrocarbon is an alkane, alkene, or alkyl derivative of benzene, the composition of the two liquid phases does not equalize at the critical solubility temperature. At this temperature, the hydrocarbon phase disappears, while water saturated with hydrocarbons exists in equilibrium with saturated vapor, also above the critical temperature.

The difference between the two equilibrium curves can be explained by the differential equation describing the changes in solubility caused by temperature changes:

$$\partial ln\left(\mathit{x\gamma }\right)/\partial \left(1/T\right)=-{\mathrm{\Delta }}_{\mathrm{sln}}h/\mathrm{R}$$E1

where

x—mole fraction of the solute,

γ—solute activity coefficient, standardized for an infinitely dilute solution,

Δ_slnh—partial molar enthalpy of dissolution of a substance in an infinitely dilute solution—thermal effect accompanying the transfer of 1 mole of a substance from a pure liquid to an infinitely dilute solution.

1.2 The curve of the solubility of hydrocarbons in water

Applying the formula (1) to describe the solubility of hydrocarbons in water, it can be assumed that γ = 1, because the corresponding mole fraction has a very small value in the entire temperature range.

The calorimetric data show that the formula can approximate the heat of dissolution of hydrocarbons:

$${\mathrm{\Delta }}_{sln}h/R=C\left(T-T_{\min }\right)$$E2

where:

C—coefficient related to heat capacity,

temperature T = T_min, at which Δ_slnh = 0 corresponds to the zero of the derivative in Eq. (1), which gives the minimum on the solubility curve.

Integrating Eq. (1) with using Eq. (2) gives the Eq. (3), describing the solubility of the hydrocarbon as a function of temperature:

$$lnx=ln{x}_m+C\left[T_m/T+ln\left(T/T_m\right)-1\right]$$E3

where:

x_m = x (T_m) is the mole fraction in the minimum solubility when T = T_m.

As it results from the works [3, 4], T_m in Eq. (3) is the same for a given class of systems (e.g., n-alkanes or isoalkanes with water). While the other two coefficients of the equation: ln x_m and C depend linearly on the volume of solute molecules in water (parameter b) [5]. An example of a relationship is shown in Figure 3.

Figure 3.

The dependence of log x (T_m) on the volume of the hydrocarbon molecule (b) [5]. Reprinted with permission from Springer science and business media from [6]. Copyright 2003, Springer Wien New York.

The corresponding equations for the parameters ln x (T_m) and C in Eq. (3) for the different classes of hydrocarbons, such as alkanes, unsaturated, and aromatic hydrocarbons, are very accurate, easy to use, and are applicable over the entire temperature range over which the data are reported experimental, i.e., from 273 K to a temperature exceeding the critical temperature of dissolution. These equations also make it possible to predict the solubility of hydrocarbons in systems not tested experimentally. The solubility of hydrocarbons of the same type may vary by several orders depending on the size of the molecules.

The second factor influencing the solubility of hydrocarbons in water is the intermolecular interactions, particularly the hydrogen bonds between water and π bonds in unsaturated hydrocarbons or in the aromatic ring. This is visible when comparing the increase in solubility in the series: alkanes, alkenes, alkadienes, aromatics. The increase in solubility results from an increase in the degree of co-association between water and hydrocarbons of the given series.

The solubility of hydrocarbons in water is essential, among others, for the protection of the environment. However, since this work concerns technological problems related to water in fuels, the issue of hydrocarbon solubility in water has been briefly discussed here.

1.3 Thermodynamic description of the liquid-liquid equilibrium in the water-hydrocarbons system

The relationship between the compositions of the coexisting phases is described by the phase equilibrium conditions, which show that the chemical potential of the i-th component must be the same in all phases of the system. Hence, the equilibrium between two liquids in a two-component system is described by two equations of chemical potentials (μ):

$${{\mu }_1}^{\left(1\right)}={{\mu }_1}^{\left(2\right)}$$E4

$${{\mu }_2}^{\left(1\right)}={{\mu }_2}^{\left(2\right)}$$E5

where the subscript numbers the components (1 or 2) and the superscript numbers the phase.

A computational model is needed to use this equation, which shows the dependence of chemical potentials on concentration, temperature, and pressure. Pressure slightly affects solubility, so it is enough to estimate it approximately for a given temperature. To determine the composition of phases in a two-component system, it is sufficient to provide one-mole fraction for each phase x (1), x (2). These values for a given temperature can be calculated by solving the system of Eqs. (4) and (5). The problem is to find a model that properly describes the tested system. Several types of models can be distinguished here:

1. Two-parameter cubic equations, which are modifications of the Van der Waals state equation. For these equations to apply to the mixture, it is assumed that the equation parameters, commonly determined symbols A and B, depend on concentration. These relationships are called mixing rules. After deciding the mixing rules, you can derive the patterns to the chemical potentials of the ingredients in the mixture and insert them into conditions (4) and (5). In the classic version, mixing rules and the parameters of pure ingredients introduce additional stick parameters (binary coefficients), the energy-dependent physical impact of both solution ingredients. Such an equation does not include the phenomenon of hydrogen bonds. It is not suitable for the description of the liquid-liquid balance in aquatic systems. Therefore, it fails to choose a stick inserted into Eqs. (4) and (5) which gives valid values x(1), x(2) for a two-component system. Adjusting the parameter to the composition of one phase gives an incorrect composition of the second phase. However, this model can calculate the balance between the liquid, consisting of alkanes saturated with water and a steam in a balance with this liquid. In low and moderate temperatures, water dissolved in the alkanes is diluted. Therefore, the association can be omitted with such a solution. This explains why according to the authors [7], a good description of this balance can be achieved using one parameter independent from temperature.

   In numerous cases, the liquid-vapor balance in the said system concerns increased temperature and high-pressure conditions, for example, in rock gas tanks on the mining field. At that time, cubic equations of condition are used, usually two parameter, with a non-classical Huron and Vidal mixing rule [8]; the rule, based on redundant free enthalpy $G^E$, enables the correct description of the layout. The work results [9] indicate that at 200°C and under high pressure (700–1000 bar), water solubility in a hydrocarbon phase decreases with increasing pressure.

2. If the two-component system model comprises two parameters, it is always possible to accept both phases for known and solve the system of Eqs. (4) and (5) relative to these two parameters. They usually depend on the temperature. Temperature coefficients are determined using the experimental data from the entire temperature range for calculations. After optimizing the model parameters, you can, from Eqs. (4) and (5), count the compositions of coexisting liquids as a temperature function. In this way, the set of numeric data is reduced to designs specifying the model parameters. For this purpose, each two-parameter model is suitable. However, no such model allows for anticipating solubility in multi-component systems based on data for two-component mixtures.

   By increasing the number of model parameters used in the operation [10], we extended the applicability of a cubic state equation to aquatic systems, by modifying the energy parameter and applying classic mixing rule of Van der Waals:

   $$a={x_1}^2{a}_{11}f\left(x_2\right)+{x_2}^2{a}_{22}+2x_1{x}_2{\left(a_{11}{a}_{22}\right)}^{0,5}\left(1-k_{12}\right)$$E6

   where: function f(x₂) describes changes in the structure of water molecules around hydrocarbon.

   Thanks to introducing more data parameters, the mixing rule (6) obtained compliance with experimental data, except for water solubility with higher hydrocarbons, particularly at low temperatures. This method of approaching phase balance also describes less exactly the experimental data for the water-cyclohexane system—Figure 4.

   The authors of the paper [11] introduced a parameter describing changes in water structure caused by the presence of hydrocarbons and interactions between water molecules bound between themselves and their molecules. They assume that m-molecules of water form a complex with one hydrocarbon molecule as a result of a quasi-chemical reaction:

   $$\mathit{mW}+\mathit{HC}\to {\mathit{HCW}}_{\mathit{M}}$$E7

   A small part of hydrocarbon molecules takes part in this reaction, and the quasi-chemical equilibrium constant of the reaction depends on the temperature. This computational method of describing the water-hydrocarbon system shows good agreement with the measurements of water solubility in hydrocarbons and hydrocarbons in water.

   Both described methods ignore such an important phenomenon as the formation of hydrogen bonds between water molecules and between water molecules and π bonds of hydrocarbons. Therefore, such an approach is purely formal. It is intended only to approximate the experimental data. And it does not make it possible to predict the solubility in multicomponent mixtures on the basis of given binary systems.

3. In contrast to the calculation methods discussed above, some methods consider the formation of hydrogen bonds in water-hydrocarbon mixtures. There are two problems here: creating a sufficiently realistic association model and relating this model to chemical potentials. The association model must have a physical sense, which means, for example, that the enthalpy of hydrogen bonding and the degree of association must be consistent with spectroscopic data and data from other sources. It also means that the same water association model and the same model parameter values should be used when describing different water-hydrocarbon systems.

Figure 4.

Mutual solubility of water and cyclohexane [11]. This article was published in [11], copyright Elsevier (2002).

Despite the passage of several years and the successes in applying the Statistical Associating Fluid Theory (SAFT) equation in many other fields, no progress has been made in the application to aqueous solutions. Chapter [12] describes the correlation method (EoSC—Equation of State + Chemical term), based on the division of the chemical potential into a physical part and a chemical part, depending on the association. The chemical element was derived in the general form, depending on the degree of association of the mixture components. A general algorithm was developed to calculate the degree of association in various mixtures, including multicomponent mixtures. The EoSC method rejects some approximations in the treatment of associations, which improves the accuracy of correlation. For example, the association mixture was not treated as a perfect mixture, described by molar concentrations. Instead, the activities calculated from the equation of state were used. Despite this, the EoSC method does not pose any accounting difficulties.

Concerning aqueous systems, the EoSC method was initially used to calculate the solubility of water in n-alkanes, isoalkanes, and cycloalkanes. The fact that the composition of the second liquid phase, i.e., the solubility of alkanes in water, can be accurately calculated from the formulas described in 1.3 was used. The calculated solubilities of alkanes were used as input data for the calculations of the two-component liquid-liquid system using Eqs. (4) and (5). Two unknowns determined from these equations were: the parameter of physical interactions and the fraction of dissolved water in the alkane. Water solubility curves ranging from 273 K to over 500 K were determined by repeating the calculations at various temperatures. The data available in the literature for water systems with n-alkanes, isoalkanes, and cycloalkanes were calculated using the same water association model defined by three constants. Once these constants are optimized, no experimental solubility data is needed to perform the calculations described. Therefore, they can also be performed for systems for which there are no solubility measurements.

In subsequent works [3, 4, 5], the above method of predicting solubility was extended to water systems containing other types of hydrocarbons. It required taking into account the possibility of creating hydrogen bonds between water hydrogen atoms and π electrons of unsaturated and aromatic hydrocarbons in the association model. The existence of such bonds causes a marked increase in mutual solubility concerning water-alkanes systems. Comparing the calculations to the experimental points suggests that the EoSC method does not bring any noticeable systematic errors. An illustration is the diagram of water solubility in n-octane - Figure 5.

Figure 5.

Water solubility in n-octane as a function of temperature [1]. Was reprinted with consent from: [1]. Copyright 2004, American Institute of Physics.

The described method of solubility calculation was used in a series of publications under the common title “IUPAC-NIST SOLUBILITY Data Series, Hydrocarbons with Water and Seawater - Revised and Updated” [13].

1.4 Water solubility in individual hydrocarbons

At a temperature of 20°C, the solubility of water in hydrocarbons does not exceed 0.0582% (582 ppm m/m). It increases rapidly with temperature, depending on the type of hydrocarbon [14]. Near the three-phase critical point, the mole fraction of water becomes comparable to the mole fraction of hydrocarbon.

The solubility of water in hydrocarbons depends on their structure. The ability to co-associate increases the solubility, which is clearly visible when comparing the solubility of water in alkanes, alkenes, alkadienes, and aromatics. Aromatic hydrocarbons, especially benzene, have the greatest ability to dissolve water. Within a given homologous series, the solubility of water depends on the number of carbon atoms (Figure 6).

Figure 6.

The molar solubility of water in hydrocarbons at 25°C depending on the number of carbon atoms in the molecule [15, 16].

Water solubility in hydrocarbons, depending on their structure, can be described by the following empirical equation:

$$ln{x}_w=\frac{A+B\cdot N}{C+N}$$E8

where:

x_w—water solubility expressed as a mole fraction.

N—number of carbon atoms in the hydrocarbon molecule.

A, B, C—parameters characteristic for a given group of hydrocarbons (Table 1).

Hydrocarbons	A	B	C
Alkylbenzenes	−78.1518	−7.9107	15.7423
Olefins	4.6649	−7.3894	−0.3834
Alkylcyclohexanes	−102.4415	6.7228	11.9077
Paraffins	−79.6677	−6.6547	9.5470

Table 1.

Parameters of the equation describing the solubility of water in hydrocarbons at 25°C [15, 16].

These articles were published in: [12], Copyright Elsevier (1999) and [13], Copyright Elsevier (2001).

Graphs showing water solubility in hydrocarbons of various homologous series are presented in Figures 6 and 7.

Figure 7.

Mass solubility of water in hydrocarbons at 25°C depending on the number of carbon atoms in the molecule [15, 16].

Different groups of hydrocarbons show the various influences of molecular weight size on molar water solubility. Thus, in the case of aromatic hydrocarbons and olefins to a lesser extent, a decrease in molar solubility is observed as the number of carbon atoms in the molecule increases. Paraffins and naphthenes show the opposite tendency. The mass solubility of water in hydrocarbons decreases with increasing molecular weight irrespective of the nature of the compound.

The dependence of the solubility of water in hydrocarbons on their molecular weight is much smaller than the dependence of the solubility of hydrocarbons in water. The molar solubility of water in n-hexadecane is only 60% higher than in n-pentane. In comparison, the molar solubility of n-hexadecane in water is four orders of magnitude lower than that of n-pentane [7].

Figure 8 shows the temperature dependence of the mass solubility of water in paraffinic hydrocarbons containing 6, 10, and 16 carbon atoms in the straight chain (a) and in hydrocarbons with six carbon atoms of different homologous series (b). Differences between the solubility curves of saturated hydrocarbons (a) and between hexane and cyclohexane (b) result from the difference in physical interactions. The order of the curves results from the degree of water-hydrocarbon co-association, which is the highest in the water-benzene system.

Figure 8.

The dependence of water solubility on temperature in paraffinic hydrocarbons of different molecular weight and in various C₆ hydrocarbons [15, 16].

The mutual solubility of water and hydrocarbons depends on temperature and pressure. The system pressure has little effect on the mutual solubility, in particular, on the mutual solubility of water and petroleum solids.

Various types of empirical formulas are used for engineering calculations; temperature dependence of water solubility (x_w) in hydrocarbons is expressed by the equation [17]:

$$lg{x}_w=a_1+\frac{a_2}{T}+a_{3}T$$E9

where:

x_w—water solubility, mole fraction;

a₁, a₂, a₃—constant values for a given component;

T—temperature, °C.

Various empirical formulas were used to determine the dependence of water solubility on hydrocarbon structure. The simplest and sufficiently accurate formula to calculate the water solubility in paraffinic, naphthenic, and aromatic hydrocarbons or in their mixture expresses the water solubility depending on the hydrogen-carbon ratio in the molecule:

$$lg{x}_w=-\left(4200\times \frac{H}{C}+1050\right)\times \left(\frac{1,8}{T}-0,0016\right)$$E10

where:

x_w—water solubility, mole fraction;

H/C—the ratio of the mass of hydrogen to the mass of carbon in the molecule;

T—absolute temperature, K.

1.5 Water solubility in hydrocarbon mixtures

Water solubility in hydrocarbon mixtures, petroleum fractions, or fuels can be calculated using the formula [17]:

$${\mathrm{lgx}}_{\mathrm{w}}={\mathrm{A}}_1+\frac{{\mathrm{A}}_2}{\mathrm{T}}$$E11

where:

x_w—water solubility in the mix of hydrocarbons;

A₁, A₂—system constants;

T—temperature, °C.

Small amounts of impurities, additives, or components with a significantly different structure can change the solubility of water in the hydrocarbon mixture. For example, a small amount of aromatic hydrocarbons with a high affinity for water can greatly increase the solubility of water in a mixture of paraffinic hydrocarbons.

Petroleum fuels are a particular case of mixtures of hydrocarbons. The content of aromatic hydrocarbons and the fractional composition determine the solubility of water in fuels. Fuels with a similar fractional composition show the higher the water solubility, the more aromatic hydrocarbons they contain. Fuels with a comparable content of aromatic hydrocarbons dissolve a decreasing amount of water with increasing fuel boiling point range.

2. Water dispersions in hydrocarbons

In individual hydrocarbons and their mixtures, water dissolves only in a limited amount. A saturated solution of water in hydrocarbons, as a result of lowering the temperature, forms a two-phase dispersion system. Part of the water in the system, determined by the solubility at a given temperature, remains in molecular dispersion, and its excess is associated with the formation of a new liquid phase.

Forming a new phase requires creating an interface made up of molecules with greater energy than inside the phase. Hence, the interface is endowed with surface energy, called interfacial or surface tension. The surface phase, interface, is a real area with a thickness of 0.5 to several nm. The size of the interface area depends on the size and number of particles of the dispersed phase.

The thermodynamic description of the phase boundary characterizes the energy state of the system. The change in the free enthalpy of the system, called the Gibbs energy, can be described by the following general equation:

$$\mathit{dG}=-\mathit{SdT}+\mathit{VdP}+\mathit{Qd\psi }+\mathit{BdM}+\sum _i{\mu }_i{\mathit{dn}}_i+\mathit{\sigma dA}+\dots$$E12

where:

S—entropy, T—temperature, V—volume, P—pressure, Q—charge, ψ—electric potential, B—magnetic induction, M—magnetic moment, μ_i—the chemical potential of i component, n_i—quantity of i component (mol), σ— surface tension, A— interface area [18].

Assuming that T, P, ψ, M are constant, we get the equation in the form

$${\mathit{dG}}_i={\mu }_i{\mathit{dn}}_i+\mathit{\sigma dA}$$E13

In the equation, n_i and μ_i refer to the interface. The integral form of this equation is as follows:

$$G_i-G_o={\mu }_i\left(n_i-n_o\right)+\sigma \left(A-A_o\right)$$E14

After differentiation, you get

$${\mathit{dG}}_i=n_{i}d{\mu }_i+{\mu }_i{\mathit{dn}}_i+\mathit{\sigma dA}+\mathit{Ad\sigma }$$E15

Starting from the thermodynamic equilibrium condition, where dG = 0, and comparing Eqs. (13) and (15), the following equation is obtained in the form

$$0=n_{i}d{\mu }_i+\mathit{Ad\sigma }$$E16

If the independent variable is only an area, then you get

$$\mathit{dG}=\mathit{\sigma dA}$$E17

or in integral form

$$\mathrm{\Delta }G=\mathrm{\Delta }\mathit{A\sigma }$$E18

This equation determines the amount of energy required to create or enlarge the surface phase under isothermal and isobaric conditions. It is equal to the magnification of the surface area and the surface tension (surface energy) at the interface.

The basic parameters determining the thermodynamic state of the surface phase are: surface tension—σ, surface area–A, surface charge density—q, electric potential—ψ, and surface concentration - Γ = n_i/A. The formula determines the entropy of creating a new phase

$$S^A=-A{\left(\frac{\mathit{\delta \sigma }}{\mathit{\delta T}}\right)}_A$$E19

where the index A means that the function relates to the area.

The enthalpy of creating a new phase is described, respectively, by the formula

$$H^A=A\left[\sigma -T{\left(\frac{\mathit{\delta \sigma }}{\mathit{\delta T}}\right)}_A\right]$$E20

The formation of the interface and the thermodynamically stable dispersion may occur spontaneously. It takes place when the work of creating (enlarging) the surface phase or interface is sufficiently small and is compensated by the increase in the system’s entropy. The amount of surface energy must then be less than 0.01 J/m² [19].

It is known from the physical chemistry of inhomogeneous systems that there is a critical minimum particle size of the new phase, determined by the magnitude of the interfacial energy. The new phase particles, smaller than the critical size, must break down into individual molecules. There is a thermodynamic forbidding on the permanent existence of new phase particles smaller than the critical ones.

A very large interface characterizes the formed dispersion system of water in hydrocarbons. This is the case, for example, when water-saturated crude oil distillates are cooling.

This chapter presents the possibility of technological use of the developed interface to improve the quality of petroleum distillates, which are raw materials of catalytic processes. The phenomena accompanying the formation of the water phase create model conditions for the extraction of mineral salts and other polar compounds of metals from hydrocarbon fractions into water. The studies described in this chapter specifically pertain to the purification of the diesel fuel (gas oil) fraction.

The formation of water dispersion in hydrocarbons, fractions, and petroleum products requires a significant supersaturation of the solution. Also, for the stable existence of water microdroplets, an increased concentration of dissolved water is needed. The smaller the radius of curvature of a single water droplet (interface), the higher the molecular concentration of water. Kelvin’s law quantifies this phenomenon:

$$ln\frac{p_r}{p_{\infty }}=\frac{2\sigma V_{\mathit{mol}}}{\mathit{RTr}}$$E21

where:

p_r—vapor pressure (concentration) of water dissolved in a hydrocarbon in the vicinity of a droplet with radius r,

p_∞—vapor pressure (concentration) above a flat surface,

σ—interfacial tension on the water-hydrocarbon interface.

The difference in water resilience (concentration), Δp = p_r - p_∞, for water droplets with a radius of 1 μm is only 0.1% of the p_∞ value, and for droplets with dimensions of 10 nm, it is 11.4%.

The phenomenon of changing the vapor pressure over the curved interface can be and is used in the technology of petroleum products. This chapter presents the research results on the concept of a movable nitrogen blanket used for dewatering fraction of jet fuel in a tank.

The presence of dispersed water in petroleum products may cause many unfavorable phenomena during the product’s transport, storage, and use. It usually causes visible turbidity; water phase can be the environment for the development of biological life, and it increases the risk of corrosion and can form ice crystals at low, sub-zero temperature. Suppose for various reasons, it is impossible to remove the dispersed water phase. In that case, its presence can be masked by taking advantage of the phenomenon of solubilization. The addition of suitable surfactants to gasoline or diesel fuel converts the dispersion with water to a clear solution; at the same time, it protects the fuel against formation of humid dispersion due to lowering the temperature or allows for adding an additional amount of water.

The results of the research on the possibility of water solubilization in fuels were described on the examples of motor gasolines, diesel fuels, and biofuels containing ethanol.

2.1 Water solubilization in petroleum fuels

The phenomenon of solubilization involves increasing the solubility of substances in a system with limited mutual solubility. Obtaining the solubilization effect requires the introduction of appropriate surfactants in appropriate concentrations into the system. The use of this phenomenon takes place in many areas of life, ranging from washing processes, through oil extraction, healing to the so-called micellar catalysis. The analysis of literature data on water solubilization in hydrocarbons [20] shows that this phenomenon has been of interest to researchers for many years.

Surfactants, dissolved in hydrocarbons in a concentration at least equal to the critical micellization concentration (CMC), form micellar solutions containing the so-called reverse micelles - Figure 9.

Figure 9.

Structure of reverse micelle of anionic surfactant [21]. Reprinted with permission from springer science and business media from: [18], 2003.

The interior of the reverse micelle, the core, made of hydrophilic surfactant fragments, can attach molecules of polar substances: water, alcohols, or nitrogen compounds. In this way, a significant amount of water can be “dissolved” in a micellar solution of surfactant in hydrocarbons. Systems containing water bound in this way remain clear and are treated as single phase [22, 23]. Water microemulsions in hydrocarbons exhibit similar characteristics. The difference between the micellar solution and the microemulsion comes down to the different water/surfactant ratios. Suppose the entire amount of water is bound by the hydrophilic fragments of the surfactant molecule in the micelle. In that case, we are dealing with a micellar solution. Increasing the amount of water leads to the formation of droplets in the micelle core and the entire system becomes an isotropic, thermodynamically stable microemulsion.

Ionic surfactants show a greater ability to solubilize water compared to non-ionic surfactants. Amphiphilic surfactants change from hydrophilic to lipophilic with increasing temperature. At the equilibrium temperature, called the HLB temperature, solubilization is maximum and a microemulsion phase is formed.

In addition to the microemulsion and micellar solution, there is a third way to increase the solubility of water in hydrocarbons. It involves the binding of water molecules by polar, hydrophilic fragments of a single molecule of a surfactant. Hydration of surfactant molecules is also a preliminary step in water solubilization in micellar solutions [24].

In petroleum fuels, water accompanies distillates and petroleum components in the step of separation and in refining processes. Its amount is so small that it does not fall into the concentration area characteristic of microemulsions. Thus, the mechanism of microemulsion formation cannot be taken into account for the solubilization of water in gasolines or diesel fuels. The concentration of surfactants required to form micellar solutions is generally so high that this method of water solubilization should be considered as inappropriate from the point of view of refining technology. Only the third way remains, in which individual surfactant molecules bind the water molecules. As a result, the concentration of a suitably selected surfactant or a mixture of surfactants can be minimized.

It is known from the literature that a mixture of surfactants, especially when their hydrophilic-hydrophobic properties are similar, significantly increases the ability to solubilize the system. There is also a synergism of the action of non-ionic surfactants, resulting in the maximum water solubilization ability [25]. Taking into account the above-mentioned properties of surfactants, their behavior in relation to water contained in a mixture with hydrocarbons and the level of water failure of petroleum fuels, a search for a method of water solubilization in motor gasolines was undertaken.

2.1.1 Research on water solubilization in gasoline

The waterlogging of motor gasoline occurs at the stages of technological process in the refinery, during storage and distribution [26].

The production of motor gasolines consists of a series of processes for producing components and their subsequent blending. The schematic diagram of gasoline technology is shown in Figure 10.

Figure 10.

A simplified diagram of gasoline technology [26].

For technological and economic reasons, the main components of gasolines in the refinery are reformates—products of catalytic reforming of straight-run naphtha and cracking gasoline—fraction from the fluid catalytic cracking (FCC) process. Due to the high-octane number and ecological features, the magnitude of alkylate—the product of C₄ fraction alkylation and isomerate—the product of the isomerization process of the C₅-C₆ fraction in the pool of hydrocarbon components, are constantly growing. Contemporary unleaded gasolines are also composed of oxygen compounds—alkyl ethers and alcohols, most often ethanol. The mixture of hydrocarbon components and oxygen compounds, supplemented with appropriate additives, constitutes the finished motor gasoline.

The presence of water in the hydrocarbon components of gasolines results mainly from the widespread use of superheated steam (stripping) in refinery processes. The water content in the component is usually relatively low and depends on the origin, chemical composition, and boiling range. Therefore, the water content level in gasolines results from the failure level of the components and their share in the gasoline pool (Table 2).

Component	Water content, ppm m/m				Average value, ppm m/m
Light naphtha	55	46	15	95	53
Reformate	74	31	22	54	45
Cracking gasoline	131	153	92	102	120

Table 2.

Typical water content in selected components of gasoline pool [26].

The relatively high water content in FCC gasoline resulted from “sweetening”—catalytic oxidation of mercaptans to disulfides in the Merox process. The dehydration of cracked gasoline is discussed in more detail in the study of coalescence methods of water separation from dispersion.

Additional amounts of water may get into gasoline during storage and transport, through leaks in tanks and cisterns, during rainfall, washing, and disinfection.

Due to day and night temperature fluctuations, the suction of moist air to storage tanks and transport cisterns during their emptying or “breathing” causes water condensation on the cold walls of the tank or on fuel surface. If the near-surface layer of fuel is not saturated with water, water absorption can occur directly from the vapor phase. The degree of waterlogging of the gasoline in the tank depends on many other factors and is changing dynamically—Figure 11.

Figure 11.

Water circulation in the fuel tank [26]. 1—Dissolution of water molecules (vapor) in fuel (fuel watering). 2—Formation of dispersion, condensation. 3—Flocculation, coalescence, Ostwald ripening. 4—Sedimentation. 5—Emulsifying water droplets from the bottom water phase. 6—Dissolving water droplets in the fuel. 7—Transfer of water molecules from the bottom water phase to the fuel phase. 8—Water evaporation (fuel dehydration).

As a result of water droplet sedimentation, a water layer is formed at the bottom of the gasoline tank, which is periodically discharged into the sewage system. Tests of water samples from the drainage of component tanks, tanks of base gasoline, and commercial unleaded gasoline showed significant differences in pH and the degree of contamination with organic substances. One of the commonly accepted characteristics of the pollution level of wastewater is Chemical Oxygen Demand—COD (Table 3).

Water from	Average value pH	COD, mg O2/l
Reformate	3,88	2.550
Light naphtha	8,75	7.800
Cracking gasoline	11,20	13.200
Base gasoline BB	9,86	12.570
Unleaded gasoline ES-95	10,02	27.000

Table 3.

COD and pH values of water from gasoline dehydration and its components [26].

The alkalinity of the water layer, separated from cracking gasoline, is caused by the content of a certain amount of sodium hydroxide from the Merox process. In turn, the acidic pH of the water separated from the reformate is related to the platforming process catalyst matrix (chlorinated alumina).

The significant COD difference between base gasoline and ES-95 gasoline is due to methyl t-butyl ether (MTBE), which dissolves much better in the water phase than hydrocarbons. This is confirmed by the data presented in Table 4, where the COD values of the water layer, remaining in equilibrium with the base gasoline and the base gasoline compositions with various amounts of MTBE, are compared. The presented data indicate the diversified nature of water pollutants contained in commercial gasoline and its main components. These factors determine the degree of nuisance of water from the drainage of reservoirs from the point of view of sewage treatment plants operation.

Concentration of MTBE in gasoline, % v/v	COD value of water layer, mg O2/l
0	3100
5,5	9600
11,0	12900

Table 4.

The COD value of the water layer, in equilibrium with the base gasoline with different contents of methyl t-butyl ether [26].

Water’s high COD value is primarily due to the dissolution of oxygen compounds and other polar organic compounds in gasoline. An additional factor is the alkalinity of water, which increases the solubility of phenol derivatives in the form of phenates, metal sulfides, and mercaptides. These substances can affect the conditions and method of binding water accompanying gasoline. The influence of electrolytes on the phenomenon of water solubilization in hydrocarbons is well known [27]. The presence of the electrolyte, due to the reduction of electrostatic repulsion (screening) forces between polar groups of surfactant, increases water solubilization capacity.

The content of organic substances: hydrocarbons, oxygen components of gasoline, and additives components affect the effectiveness of water-binding by surfactants to a lesser extent. Most of the listed organic contaminants of the water phase are completely redissolved in gasoline when the dispersed water is solubilized.

From the point of view of gasoline technology, using an effective solubilizing additive at the stage of blending means that the fuel is retained in the technological amount of water and there is no wastewater usually accompanying of gasoline blending.

The search for a suitable water solubilizing agent in gasoline began with analyzing literature data and information from companies, offering surfactants and potential additives. The basic parameter determining the use of a given surfactant is the value of the HLB parameter, which determines the hydrophilic-hydrophobic balance of the compound or mixture. Water emulsifiers in oils (W/O) show the value of the HLB parameter in the range of 3–6. The -CH₂-CH₂-O- group brings the value of 0.33 to the HLB parameter, which means that increasing the number of oxyethylene groups in the molecule increases the HLB, i.e., the hydrophilicity of the surfactant. One of the methods for synthesizing surfactants with a specific HLB parameter is, in addition to introducing ethoxy groups, propoxylation with propylene oxide. The introduction of the -CH₂-CH₂-CH₂-O- segment to a compound molecule lowers its HLB parameter by 0.15. The synthesis of surfactants based on copolymers of ethylene oxide and propylene oxide, taking into account different sequences and number of monomers, allows obtaining products with a wide range of molecular weight and different values of the HLB parameter. The length of hydrocarbon chains and the number of oxyethylene or oxypropylene units have a significant effect on the solubility and CMC size of the surfactant and the efficiency of water-binding [28].

According to Robbins [29], each drop of water in a W/O emulsion is surrounded by a monolayer of surfactant molecules, whose hydrophilic (smaller volume) groups are oriented toward the water and the hydrophobic (larger volumes) toward the hydrocarbon. The curvature of the interface results from the tendency of water to penetrate between the polar groups and the hydrocarbon group to penetrate between the hydrophobic groups. The resulting pressure difference determines the radius of curvature of the interface, and therefore the size of the water-in-oil microdroplets. The value of pressure gradient is expressed by parameters such as: the ratio of volume of the hydrophilic and hydrophobic parts of the surfactant molecules, the compressibility of the individual elements of the molecule and the interfacial tension. The influence of various parameters on the behavior of water-oil-surfactant systems is presented in Figure 12.

Figure 12.

Influence of selected parameters on the behavior of water systems-oil-surfactant (ethoxylated alkylphenol) [29].

Tests and selection of surfactants capable of solubilizing a small amount of water dispersed in gasolines required testing dozens of substances and products. In our research, the first criterion for the preliminary selection of potential additives was their solubility in toluene at a concentration of 10% m/m at room temperature. It was assumed that highly soluble products should form stable, close to molecular, solutions in gasoline, where the expected surfactant concentrations would be up to several hundred mg/kg (0.1%). Non-ionic and ionic surfactants representing various groups of chemical compounds were selected for the tests—Table 5.

No	Products	Characteristics of chemical composition
1.	Rokaphenols	Ethoxylated and propoxylated nonylphenols
2.	Rokanols	Ethoxylated saturated and unsaturated fatty alcohols
3.	Rokacets	Ethoxylated fatty acids
4.	Rokamins and Rokamids	Ethoxylated fatty amines and fatty amides
5.	Rokwins and Rokwinols	Monoesters of 1,4-sorbitol, or ethoxylated
6.	Rokopols	Block copolymers of ethylene oxide and propylene oxide
7.	Polikols	Polioksyethyleneglycols
8.	ABS	Alkylbenzenesulfonic acid
9.	AOT	Sodium salt of di-(2-ethylheksanoic) ester of sulfosuccinic acid

Table 5.

Characteristics of the chemical composition of the groups of tested surfactants [20].

The vast majority of tested surfactants came from the polish chemical factory “Rokita” S.A., (PCC); some of the tested preparations were synthesized in the research laboratory of this factory.

Ethoxylated surfactants show exciting properties from the point of view of the problem studied. The amount of water in the micellar solution indicates that each ethoxy (EO) group binds approximately 3 water molecules. Presumably, they form octagonal rings by connecting two hydrogen atoms to the oxygen of the EO group. This creates a 5 Å thick layer around the hydrophilic fragments of the non-ionic surfactant molecule [30] —Figure 13.

Figure 13.

The probable structure of the ethoxy groups of the surfactant bindings with water molecules [30]. This article was published in: [27], copyright Elsevier (1996).

The water-binding energy of the ethoxy group increases as the temperature decreases, which is of particular importance for maintaining gasoline clarity and preventing the formation of ice.

To obtain an effective additive, increasing the number of ethoxy groups in the surfactant molecule would be desirable. However, this would increase the HLB parameter and lower the solubility of this compound in gasoline. The HLB parameter of about seven characterizes a substance with balanced hydrophilic-hydrophobic properties, which means relatively good solubility in both phases—hydrocarbons and water. Therefore, the toluene-soluble formulations were subjected to the water-solubility test. In this way, eight products of comparable water and hydrocarbon solubility were selected: Sulforokaphenol N660, Rokanol L80, Rokanol Ł10, Rokamin K15, Rokopol 30p27, Rokopol RF55, Rokopol RF551, and Rokwinol 80.

To test the effectiveness of selected products, measurements of gasoline turbidity were carried out in a specially built measuring set—Figure 14.

Figure 14.

Laboratory kit for turbidimetric flow measurements [20]. 1—Peristaltic pump; 2—Gasoline container; 3—Cooler; 4—Electronic thermometer; 5—Hach 2100 AN turbidimeter; 6—Measuring cuvette.

Turbidimetric measurements in the visible light area show an advantage over other methods, such as microscopy, laser light scattering, or dynamic light scattering [31] (Figure 15). The size range of the recorded particles ranges from a few nanometers to several tens of millimeters. Therefore, the results obtained with the Hach 2100 AN turbidimeter and light filters with a light beam length of 560 to 860 nm give a reliable picture of the formation, transformation, and disappearance of the water phase dispersed in hydrocarbons.

Figure 15.

Comparison of particle sizes, measured with different optical techniques [31]. Published with permission from “Formulaction”.

A 5 dm³ gasoline sample was pumped constant through the cooler and the connected flow cell. The cooling medium, flowing through the radiator jacket, was cooled from room temperature to −30°C at a rate of 1,0°C/min, comparable in all tests. By registering the systematically decreasing (down to about −25°C) temperature of the tested sample, every 1°C, the indications of the turbidimeter were recorded. After cooling medium flow was turned off, the flowing gasoline was gradually heated to room temperature.

The tested base gasoline consists only of hydrocarbon components and is saturated with water. In the form of 10% m/m solutions in toluene, the tested additives were added to gasoline at a concentration of 50, 200, and 500 ppm m/v (calculated on the substance). A typical plot of sample turbidity changes with temperature is shown in graph Figure 16.

Figure 16.

A typical course of changes in base (hydrocarbon) gasoline turbidity (NTU - Nephelometric Turbidity Unit) as a result of cooling and then heating the flowing sample [20].

The turbidity of the gasoline sample during cooling is initially slight, but after exceeding a specific temperature, it increases. This is caused by the condensation of water dissolved in gasoline and the formation of a dispersion system consisting of tiny droplets evenly dispersed throughout the entire volume. The temperature of the appearance of a marked increase in turbidity indicates the system’s tendency to release water in the form of a separate phase (equivalent to the cloud point). After reaching the lowest temperature, the heating of gasoline maintains high turbidity, higher than that of cooled gasoline, at the same temperature. The cause of the turbidity hysteresis loop is the appearance of ice crystals in the gasoline and deposited on the cooler’s walls. Against the background of gasoline turbidity curves, the effect of the additive is visible, which shows the ability to bind water—Figure 17.

Figure 17.

Comparison of the turbidity curves of gasoline and gasoline with additive as a result of cooling and then heating the samples [20].

The turbidity of gasoline with additive is slightly lower at positive temperature, and the increase in turbidity is observed at a lower temperature than gasoline without additive. The maximum turbidity of gasoline with additive is also significantly lower than that of gasoline. The difference between the turbidity of the cooled and heated sample (hysteresis loop) is also reduced. By comparing the magnitude of turbidity changes, the water-binding capacity of the gasoline additive and the effect of the surfactant concentration can be assessed.

The results of turbidimetric measurements of the effectiveness of water-binding, present in hydrocarbon gasolines, made it possible to select three preparations for further research: Rokamin K15, Rokopol 30p27, and Rokanol Ł10. Similar composition of non-ionic surfactants has been presented in literature [32].

A review of the extensive literature data [33], particularly the patents, shows that mixtures of surfactants are almost always used to emulsify water in hydrocarbons.

In the search for the possibility of cooperation and optimization of the additive composition, the mixture of three selected surfactants was investigated on the low-temperature properties (turbidity) influence of the base gasoline. The program was based on the method of sequential simplex planning [34]. The half-factor plan from experiment 23 was implemented (3 variables—3 surfactants, 2—two concentrations of each surfactant)—Table 6.

	Rokopol 30p27	Rokamin K15	Rokanol Ł10
Concentration, mg/dm^3	150	75	50
	100	50	30

Table 6.

Starting concentrations of additives [20].

Considering the relatively highest efficiency of Rokopol, it was decided that the share of this product in the mixture would be the highest.

A plan of experiments was prepared (Table 7); the initial simplex was experimented on 1–4. The turbidity curves of the base gasoline samples containing the appropriate mixture of surfactants were determined twice.

No	Rokopol 30p27	Rokamin K15	Rokanol Ł10	NTU 1 (-18°C)	NTU 2 (-18°C)
1.	150	50	30	5.28	4.38
2.	150	75	50	3.66	3.41
3.	100	50	50	7.45	6.74
4.	100	75	30	6.30	5.42
5.	167	83	23	2.70	2.95
6.	212	64	39	2.10	2.90

Table 7.

Simplex experiment plan and experimental results [20].

The optimizing parameter was the turbidity value of the gasoline sample at temperature − 18°C. Comparison of obtained data shows that the worst result in the first series was obtained in Experiment 3 (the turbidity values at −18°C were 7.45 and 6.74 NTU, respectively). The parameters of experiment No 3 were replaced with experiment No 5. The composition of the new surfactant mixture was calculated based on the appropriate formulas:

$$\frac{1\left(K-15\right)+2\left(K-15\right)+4\left(K-15\right)}{3}={\left(K-15\right)}^{\ast }$$E22

$$2{\left(K-15\right)}^{\ast }-3\left(K-15\right)=5\left(K-15\right)$$E23

where:

5 $\left(K-15\right)$-concentration of the K-15 component in mixture No 5.

These formulas were also applied to the remaining components of the mixture. The procedure was repeated until the target function was achieved; in this case, it was the lowest turbidity value of the gasoline sample at −18°C.

It should be added that the dispersion turbidity of less than 3 NTU is invisible to the naked eye, which means that the last two surfactant compositions keep the gasoline visually clear down to −18°C.

The second remark concerns a particular feature of the simplex method of experiment planning, which leads solutions not being fully optimal. The comparison shows that the improvement of gasoline quality is achieved without taking into account the total concentration of the additive. In such a situation, both the above remarks should be combined, and it may be concluded that the intended goal was achieved in experiment No 5. In this case, the concentration of surfactants is lower than the concentration in experiment No. 6, and the effect is comparable and entirely satisfactory [35].

High octane unleaded, reformulated gasolines contain oxygen components, mainly alkyl ethers. To check the influence of methyl- t-butyl ether (MTBE) on the gasoline turbidity, two compositions of the base hydrocarbon gasoline with the addition of 4 and 10% v/v MTBE were prepared—Figure 18.

Figure 18.

Influence of the oxygen component (MTBE) on the turbidity of the base (hydrocarbon) gasoline change with temperature [36].

MTBE is entirely miscible with gasolines, while it has limited miscibility with water. Its solubility in water at 20°C is 6.5% m/m, while the solubility of water in MTBE is 1.5% m/m. The dissolution of MTBE ether in groundwater, which occurs due to a failure of pipelines, tanks, or leaks in tanks, leads to poisoning of drinking water. Ethers also facilitate the migration of hydrocarbons from gasoline to groundwater. These are some of the reasons why MTBE is limited and eliminated from the composition of gasoline.

The successor of MTBE in motor gasolines is ethyl t-butyl ether—ETBE. The solubility of ETBE in water is much lower than MTBE and amounts to 2.4% m/m.

Alcohols, especially ethanol, are becoming a vital oxygen component of gasolines. Oxygen compounds, as components of engine gasolines, have many advantages, among which one should mention their excellent resistance to detonation combustion. The octane index of the listed ingredients, defined as the arithmetic mean of the research (RON) and motor octane numbers (MON), is respectively:

MTBE - 110, ETBE – 112, and ethanol - 115.

The presence of oxygen compounds in gasolines is subject to quantitative restrictions and also has some negative consequences.

The increased water dissolving capacity of gasolines containing ethanol requires special protection against its excessive content. The addition of 5% m/m of ethanol increases the solubility of water in gasoline up to 3000–3500 ppm m/m. As the temperature of the mixture is lowered, the mutual solubility of the components decreases, leading to the formation of two separate phases. The upper layer of lower density is made of ethanol-saturated gasoline. The lower layer is the water phase, containing up to 75% m/m of ethanol and a certain amount of hydrocarbons, mainly aromatic. The resistance of gasolines containing ethanol to phase separation under the influence of water is determined by the phase separation temperature (PST).

The water-solubilizing additive in gasolines is a mixture of surfactants with optimal composition, which allows the use of the lowest possible concentrations and obtaining the full effect. A commercial product called Aquasol (Aq) can clarify cloudy, waterlogged gasoline and can also be used prophylactically. Adding an additive to the clear fuel should protect it against water release when lowering the temperature. The additive should also allow the gasoline to absorb some water, leaving it clear [36].

To check the effectiveness of the prophylactic action of the additive, some tests were carried out, for example, consisting in the selection of the concentration effectively solubilizing water added to gasoline—Figure 19.

Figure 19.

Changes in turbidity of gasolines with water-solubilizing additive (Aq) [35].

Measurements of changes in turbidity of commercial gasoline (ES - unleaded petrol, ether) with lowering the temperature allow for optimization of the dose of the water-binding additive. The defective (200 ppm v/v) unleaded gasoline requires the addition of max. 150 ppm of water solubilizing additive to restore the original turbidity at sub-zero temperature.

Compared to hydrocarbon (BB) and ether (ES) gasoline, commercial gasoline, composed with ethanol (E-94A), is able to dissolve a much larger amount of water without any signs of phase separation—Figures 20 and 21.

Figure 20.

Turbidity of E-94A gasoline with the addition of various amounts of water [36].

Figure 21.

Dependence of the phase separation temperature (PST) on the amount of water introduced into ES-95E gasoline with ethanol and the effect of a solubilizing additive at a concentration of 100 ppm [36].

The additive, containing a mixture of selected surfactants, increases the dissolving power of ethanol-gasoline without any signs of phase separation—water tolerance. A solubilizing additive in which a co-solvent, for example, higher aliphatic alcohols, is used to dissolve the surfactant composition is particularly effective. Thus, the prepared additive, when applied at a concentration of 100 ppm m/m, lowers the phase separation temperature of gasoline containing about 5% v/v ethanol by many degrees.

The additive also enables to compose of non-dehydrated ethyl alcohol—rectified, into the fuel. Introducing into gasoline 5.0% v/v of the rectified ethanol, containing about 4% m/m of water, gives about 2000 ppm of water in such a fuel. The diagram in Figure 11 shows that the phase separation temperature of such fuel is over 20°C. Introducing a 100 ppm water solubilization additive into the biofuel lowers the phase separation temperature (PST) to a level below 0°C.

2.1.2 Research on water solubilization in diesel fuels

Diesel fuels, dedicated for compression ignition engines, are produced by blending the hydrotreated products of medium petroleum distillates (boiling range 180–360°C), appropriate fractions from the vacuum distillate hydrocracking process, biocomponent and improving additives. A schematic diagram of diesel fuel technology is shown in Figure 22.

Figure 22.

Simplified diagram of diesel fuel technology [26]. Where: 1—Jet fraction. 2—Light diesel fraction. 3—Heavy diesel fraction.

Deep desulfurization of distillates in the hydrotreating process is necessary to lower the sulfur content in the hydrotreating material to the level of 50 or even 10 mg/kg (ppm m/m). Some types of diesel fuel, particularly those intended for operation in winter conditions, require maintaining appropriate low-temperature properties. The reduction of the cloud point of diesel fuel is realized in the hydroisomerization process. The highest quality diesel components are obtained in the hydrocracking process of vacuum distillates.

Diesel fuels dissolve significantly smaller amounts of water, compared to hydrocarbon gasolines. The lower water solubility is due to the lower aromatic hydrocarbon content and the higher molecular weight of the diesel fuel components. Similar to gasoline, diesel fuel produces a water phase when a water-saturated fuel is cooled. The turbid dispersions formed in this way are more stable due to diesel fuel’s higher viscosity than gasoline.

In diesel fuel technology, several methods of component dehydration are used, including fuel filtration through a layer of mineral salts (salt filters), stripping with an inert gas, e.g., nitrogen, vacuum evaporation.

The water level in diesel fuel components depends on the method of water removal, and, in the tested streams, it ranged from several dozen to over 300 ppm m/m.

The influence of water present in diesel fuel components on the formation of the dispersion system while lowering the temperature required examining the low-temperature properties, cloud point (CP), and cold filter plugging point— CFPP. The results of CP and CFPP measurements show a very diverse content of high-boiling paraffins, which is related to the boiling point range, origin, and purpose of the tested components.

The turbidity of the tested fractions was measured with a Hach 2100AN turbidimeter in a set prepared for gasoline tests (Figure 23).

Figure 23.

The change in HDF turbidity with the temperature [37].

Changes in turbidity are caused by the formation of water dispersion in the hydrotreated fraction (HDF); as a result of lowering temperature and subsequently heating the cold stream (return), a hysteresis loop appears—Figure 23. This proves the condensation of water and the formation of ice on the walls of fuel cooler during the measurements.

HDF shows very low cloud point and CFPP values as well as moderate water content. Nevertheless, an apparent increase in turbidity at 6–7°C indicates the formation of water dispersion. The tested component shows a significant difference between the temperature of the appearance of visible n-paraffin hydrocarbon crystals, the cloud point (−35°C), and the temperature of water dispersion formation (6–7°C). Due to this temperature difference, this gas oil component is a good base material for testing water-binding additives [37].

Using the experience from research on water solubilization in petrol, tests of the one component of the additive, Rokamin K-15 to HDF, were conducted—Figure 24.

Figure 24.

The effect of Rokamin K-15 additive on the change in HDF turbidity with temperature [37].

The addition of Rokamin K-15 effectively reduces the turbidity of HDF. The dose of 50 ppm m/m of the additive allows maintaining the clarity of the tested component down to the temperature of –20°C. Increasing the concentration to 100 ppm m/m practically does not change the course of turbidity changes with temperature.

Optimization of surfactant mixture composition as additive solubilizing water in HDF was carried out using the simplex method. The complete stabilization turbidity of the tested component was achieved with the use of a mixture of surfactants at a concentration of 60 ppm m/m. The main ingredients of the supplement remained Rokopol 30p27 and Rokamin K-15 in slightly changed proportions. The product Aquasol2 (Aq2) has been tested for effectiveness against various types of diesel fuel.

The tested summer diesel fuel (SDF) maintains a stable turbidity value when the concentration of the solubilizing additive is 50 to 60 ppm m/m. The remaining grades of diesel fuel, spring (autumn) and winter, are similar sensitive to the Aq2 additive.

The water-solubilizing additive in diesel fuel is intended primarily for use in the autumn and winter season, when there are significant temperature drops, causing the formation of water (ice) dispersion in fuels. Aquasol2 has been successfully used in the production of diesel fuel in the refinery and in the secondary fuel market.

The organoleptic properties of diesel fuels were improved, and most of all, the difficulties in transport (unloading tanks), storage, and use of fuels for diesel engines were eliminated.

It turned out that the water-solubilizing additive in diesel fuels was also used to improve the quality of summer types of the fuel. One of the essential parameters of diesel fuel quality is the cetane number. This property determines the engine’s combustion course and emission of toxic components in exhaust gases. The carrier of high cetane number of diesel fuel is paraffin hydrocarbons, especially of normal structure.

In a modern refinery, during hydrocracking of vacuum distillates, fractions of diesel fuel are obtained with a saturated character and a very high cetane number. They include the so-called heavy diesel fraction (HvDF) with a cetane number above 70 and a pour point in the range of 10–20°C. Due to the high cetane number, very low sulfur, and aromatic hydrocarbons content, HvDF can be a valuable component of diesel fuel. The use of such a component allows obtaining the fuel cetane number required by the standard without using special additives—cetane improvers. The fuel of natural cetane number, resulting from the hydrocarbon composition, shows a significant quality advantage over the ones, obtained using cetane additives (artificial cetane number). For these reasons, the use of HvDF for composing diesel fuel is quite justified. This component is especially suitable for producing summer types of diesel fuel with a high cetane number and relatively mild low-temperature parameters. However, summer diesel fuel, composed with HvDF, is permanently turbid, which does not cause the release of water and does not change the standard properties.

Attempts were made to determine the effect of the end point of the HvDF distillation on the turbidity of summer diesel fuel. The share of HvDF fraction in diesel fuel changes proportionally to the amount of heavy components collected, but it slightly varies the course of fuel turbidity with temperature. The effect of the water-solubilizing additive in diesel fuel, even at a low concentration, causes visible changes in dependence of turbidity and temperature [38]. The base diesel fuel, designated as SDF (summer diesel fuel), contains 9.6% m/m of the HvDF fraction with a distillation end temperature equal to 371°C. Compared to the other compositions, the SDF base fuel composition contains n-paraffin hydrocarbons with the highest pour point and the highest amount. Hence, it shows an increase in turbidity at a relatively high temperature of +15°C.

Partial solubilization of water, visible after adding 25 ppm m/m of Aq2 additive, reduces the temperature to which the cooled fuel remains clear. As the concentration of the additive increases, the turbidity of the fuel is clearly reduced, and the hysteresis of turbidity is reduced.

However, it is of practical importance to test diesel fuel commercially, i.e., with a package of improvers. For this purpose, 550 ppm m/m of the package was added to the base fuel, followed by the solubilizing additive Aq2. The comparison of the results of turbidity measurements proves that the water contained in the amount of 150 ppm m/m in the tested fuel is fully bound—Figure 25. The water-solubilizing additive at a concentration of 50 ppm m/m restores the clarity of the turbid fuel. It maintains it to the temperature of 3°C, below which precipitation of solid petroleum waxes begins.

Figure 25.

Influence of improvers package and Aq2 on the turbidity of base fuel SDF, composed with HvDF fraction [39].

In this way, the effectiveness of water-binding by solubilizing additive was confirmed in base diesel fuel and the fuel with a package of improvers [39].

The suitability of the newly developed additive for use in motor fuels, in addition to laboratory tests, requires the performance of engine tests as outlined in appropriate procedures. In the case of the water solubilizing additive in diesel fuels, the ecological properties of the fuel were assessed using the XUD9 engine.

The scope of work included:

1. measurements of gaseous emissions of toxic substances: carbon monoxide, hydrocarbons, and nitrogen oxides;

2. measurements of particulate emissions and assessment of the degree of coking of the injectors under the CEC PF-023 procedure.

The tests were carried out in the laboratories of the Institute of Vehicles, Automotive and Construction Machinery Engineering Faculty of the Warsaw University of Technology, comparing commercial summer diesel fuel and the same fuel with water solubilizing additive (Aq2) at a concentration of 50 ppm m/m (Table 8) [40].

Emission, g/(kWh)
Substance	Commercial fuel	Commercial fuel +Aq2
Carbon monoxide, CO	3.69	3.65
Hydrocarbons, HC	0.081	0.079
Nitrogen oxides, NOx	3.05	3.06
Particulate matters, PM	0.362	0.351

Table 8.

Comparison of the emission of toxic substances in the exhaust of the Peugeot XUD9 test engine, powered by commercial diesel fuel and the same fuel with a water-solubilizing additive [40].

The results of the exhaust gas analysis, in terms of the emission of toxic substances, indicate that the addition of Aq2 does not deteriorate the parameters of commercial fuel. On the other hand, the research results on the level of contamination with coke deposits on the injector nozzles indicate a beneficial effect of the additive—Figure 26. The examination of the coking tendency of diesel engine fuel atomizers is one of the indicators of the cleaning capacity of diesel fuels. The permissible limit for reducing the flow of injectors after the engine test is 80%, which means that commercial diesel fuel meets the above requirements. This results from the fuel’s cleanliness, but most importantly, it is attributed to the inclusion of an appropriate detergent additive in the improvers package.

Figure 26.

Comparison of airflow reduction through nozzles after the CEC PF023 test, an engine powered by commercial diesel fuel (DF) and the same fuel with a water-solubilizing additive [40].

The reduced tendency to contaminate the fuel atomizers proves that the water-solubilizing additive in diesel fuel enhances the cleaning effect of the package of improvers.

In summary, it can be stated that the developed composition of surfactants with the ability to solubilize water in diesel fuels effectively prevents and removes fuel turbidity caused by dispersed water. In particular, the additive is suitable for preserving the quality of diesel fuel with a high natural cetane number. An additional feature of the additive is supporting the detergency properties of the fuel, which improves the ecology of diesel engine operation.

2.2 Dispersion dewatering by coalescence

In refining technology, dispersion water systems in hydrocarbons are created, among other things, by cooling water-saturated crude oil distillates. Vapor is commonly used in the distillation process, both as an aid in the distillation of the fractions in the atmospheric tower, stabilizing the distillates in the stripping columns, and increasing the turbulence of the atmospheric residue in tubes of furnace. The amount of water vapor used for the distillation of crude oil is up to 2% m/m to the raw material. As a result, hot atmospheric distillates are saturated with water, which condenses when the fractions are cooled, to form the dispersive phase of the dispersion.

The resulting dispersion system is characterized by a huge interface, although the water content in the oil distillates does not exceed fractions of a percent. Under the conditions of intensive mixing in the flow and large contact surface of the aqueous and organic phases, the extraction process of mineral salts and other polar compounds of metals contained in the distillates occurs. As a result, the gasoline and gas oil fractions are purged of mineral contaminates. The technological use of the division of mineral pollutants between dispersed water and hydrocarbons requires effective phase separation of the dispersion.

2.2.1 Kinetic properties of water dispersion in hydrocarbons

The particles of the dispersed phase, with a diameter of less than 1 μm, perform Brownian motion, characterized by disordered and intense movements in all directions due to collisions with molecules of the continuous dispersion phase. These movements significantly contribute to the stability of the dispersion system, preventing sedimentation. Brownian motion and interactions between particles also have a significant influence on the coalescence process. The coalescence rate of the dispersed phase particles depends on the size of the forces acting between them: gravitational—FG, Brownian force—FB, Van der Waals attraction—FV, repulsive forces of electric double layers surrounding the particles – FE, and external forces - FZ, which include hydrodynamic and caused by the presence of a magnetic or electric field.

The first attempt to describe the rate of particle coalescence was made by Smoluchowski [41]. The equation for the coalescence rate of spherical particles of a monodisperse colloid under the effect of shear stress with a gradient of G has the following form:

$$\frac{\mathit{dn}}{\mathit{dt}}=-\frac{16}{3}{\mathit{Gn}}^2{r}^3$$E24

where: n—total number of particles per unit volume.

r—particle radius.

t—time.

This equation describes the rate of particle fusion—perikinetic coalescence. The influence of attraction and repulsion forces between particles under the influence of liquid movement and colloidal interactions was considered [42], introducing the orthokinetic coalescence coefficient into the above equation. Theoretical analysis and experimental data show that the coalescence of particles of similar size is statistically favored over the coalescence of spherical particles of different radii [43]. In turn, for the monodisperse system of colloidal particles, the efficiency of orthokinetic coalescence decreases with the increase of the particle radius, and for the critical value, it reaches zero. As a result of fusing, the growing particles undergo more and more deformations during collisions. Thus, the effectiveness of coalescence is reduced. The capillary number characterizes the deformability of the particle, Ca

$$C_a=\mathit{Gr\mu }/\gamma .$$E25

describing the ratio of shear forces (Grμ) to the amount of interfacial tension (γ).

Additives and impurities significantly influence the effectiveness of orthokinetic coalescence in the colloidal emulsion of the W/O type. Pollutants, including solid particles, usually adsorb at the interface, reducing the interfacial tension of the system. As a result, the dispersed phase particles are more deformed, and the critical radius of coalescence is reduced. They significantly reduce the interfacial tension of solid particles, such as asphaltenes and waxes [44], where an important role is played by the size of particles accumulating at the interface—the Pickering effect.

Also, electrolytes present in the water dispersed in oil, facilitate deformation of the droplets and the corresponding increase in the stability of the emulsion. As a result, the mean coalescence time of the two particles lengthens, and thus its efficiency decreases. Electrolytes reduce orthokinetic coalescence in various ways; KCl, for example, has a much less effect than AlCl₃ [45, 46].

2.2.2 Coalescence on barrier

Separation of dispersion into phases by sedimentation is generally ineffective. Droplets of dispersed phase with diameters below 1 μm are subject to Brownian motion, and their gravitational separation is very slow. The Stokes’ law shows that the time needed for a drop of water with a diameter of 0.1 μm to travel a distance of 1 m in jet fuel is approximately 6 years, and a drop of water with a diameter of 1 μm is 1 day, and droplets with a size of <0.05 μm remain in the fuel until they dissolve [47].

The diameter of particles dispersed in emulsions is usually in the range of 0.1–10 μm, while acceptable dispersion emulsions are the most difficult to separate. These are traditionally systems with low water content.

A practical method of accelerating the coalescence, so that it is possible to separate the dispersed phase in the form of large droplets, is passing the dispersion through an appropriately selected fibrous or porous barrier. In practice, this method is most often used to separate water dispersed in aviation fuels and oils.

The course of coalescence depends on many factors, including:

• the amount of the dispersed phase (concentrations),

• fiber diameter and thickness of the coalescing barrier,

• linear dispersion flow rate.

The following steps of the coalescence process are distinguished:

• droplet capture,

• droplet growth,

• emission (passing) of drops or streams into the barrier,

• droplet release.

During capture, the dispersed particles stick to the fibers. It can occur in several different ways, such as direct collisions, by diffusion of particles to the fiber’s surface, by the forces of inertia of the particle, and by an interference mechanism. Different capture methods have a different proportion depending on the particle size of the dispersed phase.

The diffusion mechanism dominates in the case of particles with dimensions below 0.1 μm. It occurs in laminar flow conditions, mainly due to Brownian motion. The smaller the size of the particles, the more intensively they fall out of the streamline, increasing the probability of adhering to the surface of the flowing fiber.

The particle catching on the fiber due to the inertia forces is significant for particles with a size of 0.1–1 μm and a linear flow velocity above 1 m/s. The greater the linear speed and the greater the particle diameter (mass), the higher the effectiveness of this mechanism.

The mechanism of direct hooking depends on the relationship between the dimensions of the particles and the fiber; it does not depend on the dispersion flow velocity.

The interference mechanism always occurs in fibrous septa, regardless of the size of the dispersed particles. The efficiency of depositing particles on a single fiber in a barrier differs from deposition on insulated fiber. The average fluid flow velocities are higher in the fence, and the velocity field around the individual fibers affects adjacent fibers. The presence of adjacent fibers increases the capture efficiency of the dispersed phase particles. The catching efficiency of the particles by the interference mechanism depends on the thickness of the barrier and the packing density (porosity).

The essential factor determining the efficiency of the coalescence process is the linear speed of the dispersion stream.

The interfacial tension at the water-hydrocarbon interface and the wettability of the barrier material by both liquids significantly impact the emulsion coalescence process. The presence of surfactants causes the formation of an adsorptive protective layer around the water particles. Its destruction is necessary for water particles to be caught by the fiber or the water droplet on it. Surfactants permanently reduce interfacial tension, increasing the flexibility of the boundary layer and facilitating the deformation of the droplets as they approach them.

The results of numerous studies and operational experiments [48, 49, 50, 51, 52, 53] allow for the formulation of a few general principles of designing fibrous coalescing barrier:

1. the coalescing bed is generally multilayered;

2. the speed of flow through the bed is 1.25–5×10⁻³ m/s;

3. the bed is made of fibers with a diameter from 1 to 25 μm;

4. barrier porosity should increase in the direction of flow;

5. the dispersed phase should well wet fibers.

The good wettability of the barrier material by water favors the efficient capture of the particles of the dispersed phase, but may hinder the transport of trickles or droplets along the fiber and may also cause the water to redisperse in the release stage. The hydrophobization of a part of the barrier material has a beneficial effect on increasing the diameter of the released water droplets—Figure 27.

Figure 27.

The influence of fiber hydrophobization on the coalescence process [54]. Published with the permission of Gulf Publishing.

The factors that determine the stability of water-in-hydrocarbon emulsions do not have to choose the course and effectiveness of coalescence on fibrous bed, as the process takes place under conditions of significant external shear stresses. The barrier permeability coefficient, k_r, shows a high value when the dispersed liquid wets the fibers well.

$$k_r=\frac{\mathrm{\Delta }{P}_i}{\mathrm{\Delta }{P}_e}$$E26

where:

ΔP_i—initial resistance of the dispersion flow through the barrier;

ΔP_e—resistance to the dispersion flow through the barrier saturated with the dispersed phase.

The flow resistance of the dispersion through the bed is due to the number of droplets of the dispersed phase retained on the fiber, the space occupied by the capillary transported aqueous phase, and the accumulation before the release of the aqueous phase at the layer boundary. The dispersed phase’s weaker wettability of this part of the bed significantly lowers the flow resistance and increases the size of the droplets released. Chemical modification of the surface of the coalescing bed fibers can change the mechanical strength of the barrier, wet by the dispersed phase and the relative permeability factor, and hinder the adsorption of harmful pollutants.

The analysis of the literature data on the material properties of fibrous coalescing barriers [55] indicates a controversy regarding the importance of fiber wettability and the nature of the dispersion interface. Therefore, the construction of a coalescing bed requires an individual approach, which must consider the chemical composition, properties, and parameters of the dispersion system.

The design of the barrier usually takes place in three consecutive stages: laboratory tests, pilot-scale trials, and tests under the operating conditions of the device. Two examples of the coalescing method of dispersion separation investigated and described in this paper cover all stages of partition design. However, they are mainly based on the results of experiments on a semi-technical scale conducted using real industrial streams. The main reason for choosing such a technology development mode is the instability and variability of the tested water dispersions in petroleum products.

2.2.3 Research on coalescence dehydration of gasoline from FCC process

One of the main components of motor gasolines is gasoline from the Fluid Catalytic Cracking (FCC) process. Cracking gasoline is characterized by an effective content of aromatic, isoparaffinic, and olefinic hydrocarbons. The average group composition of such gasoline, based on chromatographic analysis, is presented in Table 9.

Group hydrocarbon content, % v/v	No of series
	I	II
Paraffinic	4.4	3.8
Isoparaffinic	31.9	33.8
Naphtenic	9.6	9.7
Aromatic	18.6	18.7
Olefinic	33.1	31.5

Table 9.

Group composition of cracking gasoline [56].

Stabilized cracking gasoline is split into two streams: light gasoline with a boiling range of 36–38°C start and 90–98°C end, and naphtha with a boiling range of 85–90°C start and 198–215°C boiling end.

As a result of chemical transformations of the raw material of the FCC process in cracking gasoline, there are mercaptans (R-SH) and hydrogen sulfide. These compounds are highly corrosive to copper and its alloys and, therefore, should be removed from cracked gasoline or transformed into corrosion-neutral products. Until recently, in the refining industry, the process of “sweetening” gasoline, called Merox, is widely used to reduce the corrosiveness of gasolines from destructive processes. A schematic diagram of the cracking gasoline “sweetening” unit is shown in Figure 28.

Figure 28.

Schematic diagram of the Merox plant for cracked gasoline [56]. A—Sand filter, B—Settler, C—Mixer, 1—“acidic” naphtha, 2—Air, 3—Merox catalyst, 4—Fresh lye addition, 5—Sweeted gasoline, 6—Oxidation inhibitor, 7–circulating lye.

Light and heavy gasoline flow through the technological line consisting of three apparatuses: a mixer-reactor, a lye decanter, and a sand filter. The process is carried out at a temperature of 38–45°C and a pressure of 0.22–0.24 MPa. Before entering the mixer, the gasoline stream is combined with the circulating soda lye and air. The mixture is directed to the apparatus through a bubbler to mix the reactants thoroughly. Gasoline extraction with a 2–7% m/m sodium hydroxide solution cleans gasoline from residual hydrogen sulfide. In the presence of a cobalt catalyst, the air (oxygen) oxidizes mercaptans to disulfides (RS-SR). The mixture coming out from the top of the reactor goes to the clarifier, separated by gravity into two phases: the lower-lye, recycled pump as circulation lye to the reactor, and the upper-gasoline, directed to the sand filter. The gasoline with the entrained lye droplets flows down the filter through the aggregate bed where it agglomerates. After the filter passes, the water droplet size allows gravity to fall to the bottom of the apparatus. At the end, an oxidation inhibitor is added to the gasoline stream.

As a result of the process, the content of sulfur compounds in cracked gasolines changes—Table 10.

Gasoline.	Sulfur compound content, ppm m/m
	H2S	R-SH
Before Merox
light	4.3	55.6
heavy	4.8	16.8
After Merox
light	absence	2.9
heavy	absence	3.8

Table 10.

An example of the content of sulfur compounds in cracking gasolines [56].

Cracking gasoline, treated in the Merox plant, meets the standard’s requirements in terms of corrosion to copper (Doctor’s test).

The effective removal and conversion of sulfur compounds in the Merox process require creating conditions for fully developing the contact surface of the gasoline, soda lye, and air phases. The solid catalyst of the oxidation process accumulates at the interface of the liquids, water, and gasoline. A greater degree of dispersion of the water phase (soda lye) in the hydrocarbon phase results in a more significant physicochemical change and an improvement in the quality of gasoline. On the other hand, it is difficult to separate the lye, which coalesces and settles only in the cracking gasoline tank. Samples of water from the drainage of cracking gasoline tanks are black in color and show a pH of 10.5–11.5, which proves the presence of soda lye.

Analysis of water content in gasolines from FCC indicates insufficient ability to separate the circulating lye in the settling tank and low agglomeration efficiency of water phase particles in sand filters (Table 11). Among many factors hindering the rapid coalescence of water phase in cracking gasoline, polydisperse nature of the emulsion and saturation of the mixture with air should be mentioned. Oxygen and nitrogen dissolve better in water than in gasoline and thus reduce the difference in density of dispersion phases.

Heavy gasoline			Light gasoline
Temperature, °C	Water content, ppm		Temperature, °C	Water content, ppm
	before filter	after filter		before filter	after filter
46	419	391	42	169	150
42	393	387	43	175	170
36	353	338	38	152	143

Table 11.

Water content in cracked gasoline [57].

Due to the quality of gasoline and the limits of the amount of wastewater discharged from production plants, it was necessary to look for more effective methods of separating the water phase within FCC gasolines.

Research in the pilot plant was carried out on the original cracking gasoline stream for almost 2 years. The essential element of the research set was a vertical filter, imitating the industrial sand filter of the Merox plant (Figure 29).

Figure 29.

Diagram of a vertical coalescing filter [57]. 1—Watered cloudy petroleum fraction, 2—Dehydrated petroleum fraction, 3—Aqueous phase drain, 4—Baffle, 5—Bumper, 6,7—Steel flange, 8—Sampling valve, 9—Distribution plate, 10 —Air vent.

The coalescing filter is a cylindrical, vertically mounted tank made of carbon steel. The cylindrical part of the apparatus, 300 mm in diameter and 600 mm high, ended at the oval bottom and at the top with a cover fixed with bolts to the cylinder flange. The linear flow rate in the empty apparatus is 0.4–0.5 cm/s. Cracking gasoline was introduced from the top through a unique distributor to ensure an even gasoline flow across the entire cross-section of the filter. The gas cushion formed above the distributor was periodically released into the atmosphere without disturbing liquid flow through the coalescing barrier.

In one-third of the height of the filter, there is a connector for draining the dehydrated gasoline. In the middle of the camera’s height, a net supports the bed and the cover of the petrol outlet. The separated water phase was drained through the drain valve located in the lowest part of the apparatus. Above and below the filtering bed, pressure gauges are installed to control pressure in filter and its drop on the barrier. The entire apparatus with pipelines and accessories was insulated to minimize the temperature drop of the tested streams.

In the second phase of the research, a horizontal coalescer was added to the set. The apparatus is made of three cylindrical segments with a diameter of 300 and a length of 530 mm. Each segment of the clarifier is equipped with connectors for draining the water phase (from the bottom); connected by screws, they form an apparatus almost two meters long. The whole thing was carefully insulated, and the apparatus was used as a settler, working in series after the vertical filter. The apparatus operates in a set tilted a few degrees from the horizontal in the flow direction—Figure 30. In other studies, a metal insert was mounted between the segments—a bimetallic coalescing barrier.

Figure 30.

Schematic diagram of a horizontal coalescer [57]. A—Watered cloudy naphtha fraction, B—Dehydrated naphtha fraction, C—Water, 1—Steel flange, 2—Baffle.

The primary material for coalescing barriers was low-alkali glass fiber mats and nonwovens with a fiber diameter from 5 to 12 μm, domestic and foreign production. The thickness of the bed was varied from 5 to 15 cm, using uniform or mixed fibrous materials, supplemented in some barriers with a bimetallic mesh; an internal cooler in the baffle was also tested.

Eight different variants of the barrier structure and the method of supplying the apparatus (set) were tested. Several dozen to several hundred cubic meters of gasoline were passed through each of the tested barriers [57].

The variable parameter determined during the tests was the gasoline flow rate through the filter. The coalescence efficiency criterion was the share of the water phase separated from the tested stream (ppm v/v).

As a reference point for comparing the effectiveness of the tested barriers, measurements of the water phase separated by the aggregate bed used in the sand filter were taken. The test results, including the vertical separator (P) and clarifier (L) set, are presented in Table 12. The obtained results are challenging to interpret unequivocally; the fraction of the volume of the water phase separated from the cracking gasoline stream is very variable and shows no correlation with the volumetric velocity of the flow. Also, the amount of water phase collected in the horizontal decanter does not correlate with the degree of gasoline dehydration. However, assuming three flow rate ranges—small - up to 15, medium—15 to 25, and high–over 25 dm³/min—a particular trend of changes in the amount of drainage can be seen. The higher the volume velocity of the dispersion flow, the smaller the proportion of the separated water phase was.

No	Time, h	Full volume of gasoline, dm^3	Flow rate, dm^3/min	Volume of separated water phase, dm^3				Dewatering, ppm v/v	Share L/P, %
				P	L1	L2	L3
1.	24	15.503	10.8	1.72	0.11	0.05	0.03	123	10
2.	54	43.939	13.6	3.40	0.08	0.05	0.03	81	5
3.	24	19.316	13.4	2.10	0.06	0.04	0.02	115	6
4.	24	23.519	16.3	3.00	0.10	0.09	0.06	138	8
5.	24	25.884	18.0	1.50	0.05	0.05	0.05	64	10
6.	24	45.821	31.8	2.80	0.10	0.05	0.05	65	7
7.	24	40.610	28.2	3.10	0.10	0.05	0.05	81	6
8.	20	26.291	21.9	2.00	0.05	0.05	0.05	82	8

Table 12.

Statement of the results of the work of the set; vertical (P) filter filled with sand, horizontal (L)—Empty [57].

The average volume of the water phase separated in the set of vertical coalescer and horizontal settler, with the use of aggregate as a barrier, is:

• at the lowest flow (up to 15 dm³/min)—106 (99) ppm v/v;

• with an average flow (up to 25 dm³/min)—95 (87) ppm v/v;

• with the high flow (over 25 dm³/min)—73 (69) ppm v/v.

(Drainage in the vertical filter is given in brackets).

The heavy cracking gasoline flowing into the test filter carries a variable amount of water (sodium hydroxide solution). In addition, the process catalyst particles and an anti-corrosion additive are present in the air-saturated stream. Changes in the dispersion composition resulting from the process conditions and parameter fluctuations make it difficult to compare the obtained test results. Considering the circumstances mentioned above and limitations, we received a relatively high coalescence efficiency on the barriers made only of glass fiber. The drainage results with the use of the “B” bed, data in Table 13, may serve as an example.

No	Time, h	Full volume of gasoline, dm^3	Flow rate, dm^3/min	Volume of separated water phase, dm^3	Dewatering, ppm v/v
1.	10	17.402	29.0	4.18	240
2.	9	15.668	29.0	4.10	262
3.	8	10.561	22.0	2.44	231
4.	9	11.883	22.0	2.65	223
5.	15	15.321	17.0	3.31	216
6.	12	12.304	17.1	3.04	247
7.	6	6.114	17.0	1.38	226

Table 13.

Results of dehydration of cracked gasoline in a vertical filter equipped with a glass fiber barrier, “B” [57].

The coalescing barrier, marked as “B”, was made of two types of nonwoven fabric. It consisted of a 12-cm-thick layer of nonwoven fabric I and a 3 cm thick layer of nonwoven fabric II, both supported by a mesh. The bed was situated horizontally, i.e., perpendicular to the direction of gasoline flow. The partition area was about 700 cm², and the volume after laying was about 10.600 cm³.

The volume of separated water is slightly dependent on the gasoline flow rate through the filter. Only the first two measurements show a somewhat higher degree of dispersion dehydration. Compared to coalescence in aggregate bed, an average increase in drainage is obtained from about 90 to over 230 ppm v/v. The thickness of the partition does not affect the efficiency of water separation, as the “B” barrier is initially 15 cm and the “F” partition is only 5 cm (nonwoven fabric I).

The average share of the water phase volume separated with the use of the “F” partition is:

• at the lowest flow (up to 15 dm³/min) - 253 ppm v/v;

• at the highest flow (over 25 dm³/min) - 228 ppm v/v.

All the results mentioned above were obtained as a result of dewatering dispersion in heavy cracked gasoline, directed to the research filter from an industrial decanter. Comparative measurements were carried out by supplying the experimental installation with gasoline from an industrial sand filter. In this way, we checked the validity of introducing an additional degree of gasoline dehydration. By comparing the amount of dehydration of gasoline from an industrial decanter and gasoline from an industrial sand filter, data on the efficiency of water coalescence in the sand filter were obtained. The average fraction of the water phase volume separated with the use of the “F” barrier, fed from the sand filter is:

• at the smallest flow (up to 15 dm³/min) - 178 ppm v/v;

• with an average flow (up to 25 dm³/min) - 129 ppm v/v;

• at the high flow (over 25 dm³/min) - 118 ppm v/v.

Based on the obtained results, it can be estimated that the industrial sand filter reduces gasoline water failure by approx. 80–110 ppm v/v.

The individual barriers were usually tested for 2–3 weeks in the conditions of different dispersion flow rates and two different power sources of the test installation. It was noted that a longer measurement time resulted in better repeatability of the obtained results.

The assessment of the condition of the nonwoven fabric used to build the barrier shows no noticeable wear of the material due to contact with water-loaded gasoline. The nonwoven fabric structure has loosened, probably as a result of washing away the binder sticking the fibers. The material turns blue as a result of the adsorption of Merox catalyst particles. On the surface of the barrier, small amounts of corrosive impurities carried by the gasoline stream were observed.

The choice of materials for coalescing barriers was dictated by the chemical composition of the drained stream. Due to the high content of aromatic hydrocarbons and olefins, cracked gasoline is a good solvent for many organic substances. The soda lye accompanying the gasoline contains dissolved sulfur compounds and phenols. Altogether, these factors create an aggressive environment, and the only material that could function well under these conditions was glass fibers.

Glass has a high surface energy value, about 100 mJ/m², and is well wetted by water. Due to the Si-OH groups on the surface, it can form hydrogen bonds with polar liquids such as water. In all types of W/O dispersions, water coalesces on the surface of the glass fibers.

Two types of fiberglass nonwoven fabric were used to make the partitions:

Nonwoven fabric (I) is a glass fiber mat, type EM 1002/600/125, manufactured by KHS Krosno S.A., made of glass fibers with a low-alkali content. The monofilament diameter of the mat is 12 μm, the surface weight is 600 g/m², polyvinyl acetate is used as a binder of the fibers, and the thermal resistance of the material is up to 200°C.

The knit (II) of glass fiber Polmaterm - DK, produced in Spain (Montero Fibras y Elastomeros S.A.), is an insulating fabric with a thickness of 10 mm, made of E-glass fibers with a thickness of 6 μm, subjected to special chemical treatment. The surface weight of the knitted fabric is about 2.2 kg/m², and its thermal resistance reaches 1200°C.

In the Merox plant of the FCC plant, it was not about creating an additional filter or water phase separator, but about a possible modernization of the functioning cracked gasoline sand filter.

A sufficient improvement in the degree of dehydration of cracked gasoline from the lye decanter is achieved by replacing the standardized aggregate (sand) with a coalescing glass fiber bed. The average level of reduction of gasoline waterlogging in the sand filter is up to 110 ppm v/v and with the use of an effective coalescing barrier about 230 ppm v/v. This means that it is possible to increase the amount of water phase (diluted soda lye) recovered from heavy cracked gasoline in the Merox plant of the FCC by 100%.

In conclusion, it should be added that the improvement of the lye separation rate from the cracking naphtha does not solve the problem of water formation in the storage tank of gasoline blending department for two reasons.

First, gasoline leaves the installation at a temperature above 30°C; lowering the temperature will cause water to separate again. Cracking gasoline, as a mixture of e.g., aromatic and unsaturated hydrocarbons, dissolves water relatively well and shows a strong dependence of water solubility on temperature.

Second, even under the best conditions of gasoline coalescing dehydration, a clear product was not obtained. The main reason for this result was an anti-corrosion additive used in the process. After excluding the dosing of this product for research purposes, practically clear gasoline was obtained.

The effect of long-term and labor-intensive industrial research was applying the results of the work by replacing the aggregate with a glass fiber barrier. A layer of two types of nonwoven fabric, about 15 cm thick, was covered with aggregate (about 10 cm) to protect the material from liquid impacts while filling the filter. The filter with the changed baffle operated without reservations over 6 years from year 2000, until the Merox plant was shut down and replaced with catalytic desulfurization plant.

It should be added that before research was undertaken to improve the dehydration of cracking gasoline, no ready-made solution to the problem was found. Literature reports also indicate that sand filters were used in approximately 1500 operating Merox installations. The report on the application of coalescing dehydration of cracking gasoline was published in 2001 [58].

The use of a more efficient method of water coalescence in a cracked gasoline filter mainly brought about ecological benefits. The amount of hazardous sewage from dehydration of cracking gasoline tanks and gasoline composition has decreased. In addition, emergency stopping of the filter operation, quite frequent in the original filling, has been completely eliminated. After 6 years of operation, the barrier changed significantly without losing the water dispersion coalescence ability (photo, Figure 31).

Figure 31.

Photograph of the material of the coalescing barrier, installed in the cracked gasoline filter and removed after 6 years of operation (dark one) [55].

From the initial thickness of approximately 15 cm (8 layers of nonwoven fabric), a sand-like board about 1.5 cm thick with visible glass fiber weaves remained.

While testing various coalescing barriers made of glass fibers, in a horizontal coalescer, the effectiveness of water coalescence in cracked gasoline was tested using meshes made of two metals—iron and titanium. Very promising coalescence results were obtained for a short time, but after about half an hour, the barrier’s effectiveness decreased to the fibrous barrier level. After disassembling the barrier, it turned out that the mesh elements were contaminated with sediment, which prevented the conductive contact between the two metals.

The concept of a bimetallic coalescer was returned to in laboratory studies on dehydration of jet fuel.

2.2.4 Coalescence studies on metallic barriers

The research on the possibility of using the electric field to accelerate water coalescence began with laboratory tests of dehydration of jet fuel.

Fuels for turbojet engines (kerosene) are produced from petroleum fractions with a boiling range of 150–250°C, deeply refined in the processes of hydrotreating or hydrocracking.

The basic requirements for this type of petroleum products include, but are not limited to:

• chemical composition ensuring complete combustion giving a high amount of heat;

• high thermo-oxidative stability;

• suitable low-temperature properties;

• electrical conductivity.

The low crystallization temperature and good filterability of aviation fuel require deep water removal. For this reason, various methods of jet fuel dehydrating are used in technology and applications. The most commonly used are coalescence using baffles, adsorption methods, inert gas stripping, and water removing additives—dehazers.

JET A-1 aviation fuel, containing an antioxidant and antistatic additive in the hydrotreated kerosene fraction, was used for coalescing water removal laboratory tests. 500 or 1000 ppm m/m of water was added to the fuel, which was dispersed using an ultrasonic disintegrator. The dispersion samples, 1.0 dm³ in volume, were passed through a bimetallic or metal-carbon bed placed on the bottom of the glass column. The bed material was iron and aluminum filings and crushed coal (coke), all sieved to separate fractions with grain sizes up to 0.75 mm and 0.75–1.5 mm. The effectiveness of the dispersed water coalescence was assessed on the basis of the turbidity of the initial sample and the dehydrated one. The product obtained after passing through the coalescing baffle column was, in each case, visually apparent (NTU less than 3).

The average values of turbidity reduction of samples of water-bearing JET A-1 fuel on Al-Fe and C-Fe beds with the grain size of 0.75–1.5 mm are from 80 to 85% of the initial turbidity. Beds of powder mixtures with a grain size of 0.75 mm exhibit higher dispersion dewatering efficiency, resulting in 93 to 98% turbidity reduction.

This method of accelerating the coalescence of emulsified water particles in petroleum products is based on galvanic microcells formed at the contact of materials with different electrochemical potentials [59]. In emulsions, the dispersed phase particles are often charged relative to the continuous phase and are attracted in an electric field by electrodes with opposite charges.

The speed of movement of spherical particles in an electric field (electrophoretic velocity) is given by the formula:

$$\nu =\frac{\zeta \mathit{EKC}}{6\mathit{\pi \eta }}$$E27

where:

ζ—(zeta potential) difference of potentials between particles of the dispersed phase and the liquid is the continuous phase,

η—the viscosity of the dispersing phase,

E—electric field strength,

K—dielectric constant of the medium,

C—constant depending on the electric hydrodynamic radius of the particle, the value of which is from 1.0 to 1.5.

In stable O/W emulsions, the potential is equal to 25 mV or more, resulting in the electrophoretic velocity of the hydrocarbons being on the order of 0.2 × 10^–4 m/s, in a field of 1 V/cm. The dielectric constant of water, and therefore the electrophoretic velocity of the particles, is about 40 times higher.

One of the conditions for high coalescence efficiency in an electric field is the low electrical conductivity of the continuous phase.

The intensity of the electric field generated in such cells depends on the distance between the electrodes. For example, electrodes 0.1 mm apart with a potential difference of 1 V (Fe-Al) create an electric field of 100 V/cm. Under these conditions, the emulsion particle, characterized by an electrophoretic velocity of 10⁻⁴ cm/s travels the distance between the electrodes for 1 s. The shape and size of the material granules thus determine the average value of the electric field intensity and, consequently, the coalescing capacity of the bed.

The time of gravitational flow of the tested samples of water dispersion in jet fuel through the barriers made of materials with smaller grain sizes is 1200–1400 s and is half as long as the time of flow through the partitions made of larger grains (800–900 s). Very low turbidity of the dehydrated fuel, of the order of 0.24–0.26 NTU, is obtained here.

The choice of material of coalescing bed depends on the required efficiency and the nature of the environment. Elements with a significant difference in electrochemical potential, such as magnesium and gold, give a cell potential difference of 3 V; less noble carbon in combination with iron gives 1 volt and aluminum 2 volts. The ionic reactions on the electrodes in the bed cause gradual wear of more reactive electrode material.

To verify the laboratory tests’ results, tests for the dehydration of cracking gasoline in a vertical coalescing filter were carried out. The bed of mixed iron and aluminum filings was heaped on a nonwoven layer I, about 2 mm thick, placed on a supporting mesh made of acid-resistant steel. The volume of the swarf mixture was 200 cm³, which gave a layer about 3 mm thick; gasoline was drained from the sand filter—Table 14.

No	Gasoline	Volume of gasoline, dm^3	Volume of water phase, dm^3	Flow rate, dm^3/min	Part of water phase, ppm v/v
1.	From lye settler	18.229	6.20	11.9	340
2.		24.312	2.25	15.6	93
3.		19.388	2.25	14.0	116
4.	From sand filter	11.719	3.30	8.0	282
5.		17.649	0.90	12.3	51
6.		18.385	0.95	12.3	52

Table 14.

The results of dehydration of naphtha cracking in a vertical filter, containing a bed of iron and aluminum filings [56].

Regardless of water content in gasoline, the first day of operation of the bed gives a much higher degree of dehydration than the next ones. This indicates that the bimetallic bed significantly increases the coalescence of water dispersed in cracking gasoline. At the same time, the results show a complete disappearance of the deposit activity over a more extended period. It seems, therefore, that in the environment of gasoline containing soda lye, metal filings are passivated, preventing the permanent functioning of galvanic micro-cells.

This experience confirms the correctness of accelerating water coalescence in petroleum products straightforward, effective, and safe. On the other hand, it can be seen that a different method of joining metals should be sought to provide an electrically conductive contact in the cell.

The use of meshes made of various metals as the material of the coalescing barrier was analyzed. Materials of the “Sitodrut” company in Gliwice were used in the research on dehydration of hydrotreated fraction of jet fuel. The mesh partition can be folded using a horizontal coalescing filter.

Two compartments are made of mono and bimetallic nets with different weaves, mesh size, and wire thickness. Copper-zinc, woven, meshes wire 0.20 mm with a valuable clearance of 25% were separated by steel meshes made of 0H18N9 wire, F 0.20 mm, and F 0.10 mm (Cr, Ni), with a helpful clearance of 25% and different thickness. A set of 30 meshes, cut in the shape of a circle, were fixed in steel rings with a diameter of 295 mm, bolted together. The barrier thus formed had a thickness of about 30 mm. To tightly fill the cross-section of the apparatus, a gasket with an outer diameter of 300 mm was attached between the mesh and the ring.

The results of drainage of the tested stream using a bimetallic mesh barrier indicate a significant share of this material [60]—Table 15.

No	Volume of hydrotreated jet, dm^3	Volume of water phase, dm^3			Flow rate, m^3/min	Part of water phase, ppm v/v
		L1	L2	L3
1.	307	0.13	0.25	0	30.7	1238
2.	306	0.16	0.27	0	30.6	1405
3.	261	0.23	0.20	0	26.1	1648
4.	145	0.15	0.09	0	14.5	1655
5.	171	0.22	0.09	0	17.1	1813
6.	88	0.20	0.02	0	8.8	2500

Table 15.

Results of the measurement of water separated from the jet fuel hydrotreated using a bimetallic barrier [60].

For coalescing barriers, glass fiber meshes coated with Teflon were used in the subsequent tests. This material is characterized by the regularity of the holes between the fibers (150×200 μm) and is entirely hydrophobic. The device was equipped with two barriers: the first one with a water-repellent mesh in the metal ring, a packed knitted fabric (II) and fiberglass mats, and the second—a circle with bimetallic mesh and a hydrophobic mesh on the side of liquid inflow. A very high degree of dehydration of the studied stream was obtained. The system of baffles allows for the separation of a large amount of water from the hydrotreating product and, at the same time, produces a clear product, regardless of the flow rate. It is worth noting that the barrier made of metallic mesh makes a significant contribution to the total water separated. This was especially visible in the experiments where the high flow rate of the tested dispersion was used.

To compare the effectiveness of individual barriers, the average volumes of water released at low (about 10 dm³/min), medium (about 20 dm³/min), and high (about 30 dm³/min) volumetric flow of the hydrotreating agent through the apparatus were summarized - Figure 32.

Figure 32.

Comparison of the effectiveness of coalescing partitions in the dehydration of jet fuel hydrotreated stream depending on the dispersion flow rate [60]. Designations in the drawing: I—Empty apparatus, II—Bimetallic barrier, III—First nonwoven barrier, second—Bimetallic barrier, IV—Combined bimetallic barrier with nonwoven fabric.

The summary of the results of testing the effectiveness of selected barriers shows that they all offer a high degree of dispersion dewatering (above 2000 ppm v/v), especially under low flow conditions.

The system of two barriers, the second of which is made of bimetallic nets, is characterized by high water separation efficiency, almost independently of the dispersion flow rate through the apparatus. This means that such a combination of baffles in the water separator gives great flexibility in terms of the flow rate. An even better solution seems to be a combined barrier, made of closely packed fibrous materials, preceded by bimetallic nets. The common feature of suitable barriers, including III and IV, is the ability to separate over 2000 ppm v/v of water from the industrial hydrotreated material. The dehydrated petroleum stream is clear at the separation temperature.

Regardless of the coalescing barrier used, the water separated in the separator shows an almost neutral pH = 6.1–6.2 and an effective content of organic substances (COD about 1000 mg O₂/dm³).

Only water dispersed in the hydrocarbon phase can be removed using the coalescence method; even the clear and transparent petroleum stream contains dissolved water at a saturated concentration at the separation temperature. Therefore, it is necessary to prevent turbidity (separation of the water phase) during the cooling of the hydrotreating material down the separator.

2.2.5 Research on coalescence dewatering of diesel fuel fractions

Diesel fuel technology is heading toward more profound treatment with the hydrogen of fractions from the crude oil distillation. Limiting the content of sulfur, polycyclic aromatic hydrocarbons, and reducing the density require specific catalysts for hydrotreating processes. Thus, the requirements for raw materials are increasing, which prompts a deeper analysis of impurities contained in gas oil distillates, their sources, and reduction possibilities (Figure 33) [61].

Figure 33.

Schematic diagram of the crude oil distillation plant under atmospheric pressure [61]. G-1 dry gas and lpg, RN—Crude oil, A1—Gasoline, broad-spectrum, A2—Naphtha (130–180°C), A3—Light fuel oil (175–280°C), A4—Heavy fuel oil (240–360°C), PA—Atmospheric residue, KS—Stabilizing column, OP—Fuel oil, pw—Steam, sol—Brine; A—Furnace, B—Distillation column, C—Stripper, D—Condenser, E—Phase separator, Eh—Desalter.

One of the main raw materials for producing diesel fuel components is distillates, obtained by separating crude oil under atmospheric pressure and the top fraction from distillation of atmospheric residue under reduced pressure.

An essential criterion for the quality of the raw material for hydrogen processes is the content of metal compounds, which come mainly as impurities in crude oil. The process of desalination of crude oil before distillation significantly reduces but does not wholly remove metal compounds from the raw material—Table 16. An additional source of some impurities in the distillates may be the corrosion processes of apparatus and devices.

Oil	Metal content, mg/dm^3
	Na	K	Ca	Mg	Fe
Crude oil	44.7	4.2	67.0	7.2	8.27
Desalted oil	10.6	1.7	11.5	2.6	6.04

Table 16.

Metal content in crude oil and desalinated oil [61].

Catalysts for the hydrotreatment of gas oil distillates are sulfides of cobalt and molybdenum (Co/Mo) or nickel and molybdenum (Ni/Mo), deposited on a carrier, usually alumina. Mineral salts contained in the raw material form deposits on the surface of the catalyst bed. As a result, the flow resistance of the reactor increases, which eventually leads to the stoppage of the installation and the exchange of part of the catalyst load. Another hydrotreating catalysts (platinum metals), responsible for the hydrogenation of aromatic hydrocarbons and the isomerization of n-paraffin hydrocarbons, are particularly sensitive to poisoning. Compounds of some elements, such as copper, lead, and especially arsenic, cause irreversible deactivation of the catalyst.

In the crude oil distillation plant, diesel fractions are stabilized with steam in stripping columns. During the cooling of distillates saturated with water, a dispersion system in fractions is formed, characterized by a well-developed interface. This creates favorable or model conditions for extracting mineral salts and other polar compounds of metals from the hydrocarbon fractions into water droplets.

In practice, dehydration of petroleum fractions is usually carried out by water evaporation under reduced pressure (vacuum drying). This drying method causes the water to evaporate from the dispersion as steam, and the mineral impurities remain in the petroleum fraction.

In the research on the separation of water from dispersion in hydrocarbons, a thesis was formulated about the possibility of additional purification of diesel fuel fractions through coalescence dehydration of distillates. The separation of water as a liquid phase should reduce the amount of mineral substances remaining in the fraction.

The water content analysis in atmospheric distillates shows that the kerosene fraction (A2) contains more than 1000 mg/dm³ (0.1%) of water, and the fractions A3 and A4 - within the range of 800 to 900 mg/dm³. The water content in top fraction of vacuum distillation (P2), which is also raw material for diesel fuel production, is 200 ppm m/m.

A vertical filter was used for dewatering diesel fraction, earlier used for dehydration of cracking gasoline, in which the method of installing the coalescing barrier was changed - Figure 34. The barrier had a cylindrical shape and was placed on a metal scaffold attached to the cover of the apparatus. The cylinder, 17 cm in diameter and 32 cm high, gives 0.17 m² of the internal surface area of the barrier. In this way, an increase in the bed surface and an extension of the sedimentation path of water droplets were obtained. For technological reasons, the research focused mainly on the A2 fraction making it possible to feed the coalescing filter with an actual stream without restrictions.

Figure 34.

Diagram of a vertical filter for research on the coalescence dewatering of atmospheric distillates [62]. 1—Watered cloudy atmospheric fraction, 2—Water, 3—Atmospheric dehydrated fraction, 4—Sampling port, 5—Vent, A—Coalescing layer, B, C—Kerosene-water dispersion, D—Aqueous phase.

The following glass fiber barrier materials were used in the study of dewatering water dispersion in diesel fractions:

• glass fiber mat produced by KHS in Krosno (the same as in the cracking gasoline dewatering tests);

• Ws-20 nonwoven fabric produced by “Banex” Ltd. in Kutno (needle punched and thermally treated glass mats produced by KHS in Krosno);

• Ws-20 nonwovens modified with polymeric materials to reduce the hydrophilicity of the barrier.

To achieve a moderate reduction of the hydrophilicity of the barrier, materials, and technologies developed at the Department of Binding and Coating Polymers of the Institute of Industrial Chemistry in Warsaw were used. They included:

1. Polysiloxyurethanes

   The DPU-PJ 402 emulsion, diluted 10 times with water, was applied by spraying on the surface of the Ws-20 nonwoven in the amount of 200 cm³ per 0.5 m². The nonwoven fabric was dried for 1 h at 105°C, left in the air for 2 days, and used for research.

2. Silicone resins

About 10 cm³ of Siltex EP34 emulsion was mixed with 3 cm³ of the CL4 catalyst, and 187 cm³ of water was added. The preparation was applied by spraying on the Ws-20 nonwoven fabric with an area of 0.5 m². The material prepared in this way was heated for about 10 minutes. At 150°C, left for 2 days in the air and used for research.

About 30 cm³ of 3% Anhydrosil Z solution was mixed with 0.5 cm³ of 1% catalyst solution, diluted with 90 cm³ of toluene, and sprayed onto 0.5 m² of Ws-20 nonwoven fabric. It was left at room temperature for 2 days and used for research.

The water dispersion of Tarflen at a concentration of 55%, produced by factory in Tarnów, was also used to modify the surface of nonwoven fabric. The melt temperature of the polymer was determined derivatographically. About 200 cm³ portions of the preparation were applied in two ways to Ws-20 nonwoven fabric sheets with an area of 0.5 m² each: dipping the nonwoven fabric and spraying it. The material was dried at about 150°C and then baked in an oven at 340°C for 2 hours. The nonwovens prepared in this way were used for research.

To obtain a transparent fraction, A2 was selected as the essential criterion for assessing the effectiveness of the coalescing bed, regardless of any fluctuations in the water failure of the technological stream. Obtaining a clear distillate means complete separation of the dispersed water. The clear fraction of A2 at 20°C contains about 240 mg/dm³ of water dissolved. The water content in the heavier clear fraction A3 is lower, and at the temperature of 20°C, is about 180 mg/dm³.

The second criterion of the evaluation was the flow rate of the dewatered fraction through the coalescing barrier. The relatively high linear flow rate allows for the reduction of the baffle area and the volume of the apparatus.

Additionally, the appearance of the aqueous phase separated from the distillate was observed. Usually, good dehydration of the kerosene fraction gives an almost or apparent aqueous phase. The volume of water separated from the fraction flowing through the coalescer is additional information about the effectiveness of the process and the temporary failure of the tested stream.

The test results show that an apparent, dehydrated fraction is obtained only using a barrier made of packed Ws-20 nonwoven fabric when the volumetric flow rate does not exceed 40 dm³/min. Nonwoven fabric preparation using silicone resin emulsions is unstable in the conditions of A2 fraction dewatering. The products show slight turbidity, and the amount of water separated is slightly lower than the starting Ws-20 nonwoven fabric. The nonwoven fabric impregnated with Teflon on the surface creates resistance to the flow of the drained liquid, which makes it impossible to obtain a flow rate above 50 dm³/min. It allows for the separation of a large amount of water, although the dehydrated fraction is slightly turbid. A comparison of the proportion of water separated with individual materials as coalescing partitions is shown in Figure 35.

Figure 35.

Comparison of water separation efficiency from the technologically impaired A2 fraction through the Ws-20 nonwoven barrier and the hydrophobic barrier with Teflon and Anhydrosil resin [62].

Despite the fact that surface hydrophobized with Teflon separates the most significant amount of water from the stream of water-repellent fraction, its effectiveness is lower than the barrier made of the hydrophilic, initial nonwoven WS-20. The likely cause is the significant fluctuations in the water content of the A2 fraction.

As a result of over a year of research, a very effective coalescing barrier was selected. The use of a Ws-20 nonwoven barrier allows for complete separation of water dispersed in the A2 fraction, i.e., obtaining a clear product. Such a barrier effectively drains the stream in the range of linear flow velocities, up to 7 mm/s, while allowing for stable results of water separation in a long period of operation. It accumulates a relatively large amount of water and reaches equilibrium after about 15–20 minutes of work. There is no justification for increasing its thickness over 20 mm due to the deterioration of water separation caused by redispersion.

Barriers modified with polymers in aqueous suspensions show reduced hydrophilicity, retain less water, and reach equilibrium faster. Only the Teflon-coated bed drains the A2 fraction, which is comparable to the coalescence on the Ws-20 mat partition.

Optimal, from the point of view of water separation from the A2 fraction, coalescing barriers do not show full effectiveness against the higher boiling gas oil fractions - A3 and A4.

Partial hydrophobization of the outer layer of the Ws-20 nonwoven barrier is necessary to obtain a clear product from the A3 fraction. The laboratory measurements show that a barrier constructed in this way can effectively separate the water dispersed in this fraction at high linear flow rates—up to 11 mm/s. This indicates a significant share of the inertial water particle capture mechanism by the barrier material.

Relatively, most difficult is separating water from the A4 fraction—heavy diesel. The combined material of the partition, 4 cm thick, still leaves slight turbidity in the product. The reasons for this phenomenon can be found in the higher viscosity of the stream and the high content of n-paraffin hydrocarbons.

For each of the tested fractions, the separated water contains significant amounts of sulfides, chlorides, and, above all, metal compounds—Table 17.

Water separated from fraction	Metal content, mg/dm^3
	Na	K	Ca	Mg	Fe
A2	58,5	3,0	36,3	7,8	17,9
A3	31,2	1,8	24,0	4,3	14,1
A4	46,1	2,4	25,6	5,1	22,3

Table 17.

Metal content in water separated from diesel distillates [62].

Alkali metals such as Na, K, Ca, and Mg are present in crude oil brine. Most of these metal compounds are sparingly soluble in water, which means that they exist in the form of organic compounds, such as salts of naphthenic acids and phenates. These compounds can exist on their own or in resins and asphaltenes.

The compounds separated by extraction into the water layer reduce the pollution of the organic phase—the diesel fraction. The results of the fraction analysis confirm a significant reduction in the contamination of distillates with metal compounds after the separation of water by coalescence (Figures 36 and 37).

Figure 36.

Reduction of the metal content in diesel distillates dehydrated by coalescence as compared to vacuum dehydrated [62].

Figure 37.

Reduction of the content of microelements in diesel fuel distillates dehydrated by coalescing in relation to vacuum dehydrated [62].

Comparison of the content of metal compounds in the A3 fraction, which is dehydrated by vacuum and coalescence, shows a twofold reduction in the concentration of alkali metals and heavy metals (from 6.8 to 3.6 ppm m/m). Even greater reduction in metal content is provided by the coalescence dewatering of the A4 fraction (from 14.3 to 2.6 ppm m/m). Such a significant reduction in the content of metal compounds means a considerable reduction in the amount of potential deposits in the heat exchangers and the upper catalyst layer of the hydrotreating reactor. Equally crucial for the activity and duration of the catalyst operation may be a severe reduction of the content of lead compounds in the raw material. The remaining heavy metals and arsenic are subject to more negligible hydrolysis and extraction to the water phase.

Irreversible changes in the water phase, separated from atmospheric distillates, make it impossible to redisperse the water in the fraction. Therefore, only tests using original streams give an accurate picture of the state of water dispersion and the possibility of its separation.

Water separated from crude oil atmospheric distillates is usually slightly cloudy and blue-green in color. After a short time, about 1 hour, it becomes relatively clear and turns yellow-brown and the growth of brown flocculent sediments is observed. Bulk represents little pollution; filtration of 1 l of water separated by coalescing from about 800 l of kerosene fraction gives 3–7 mg of sludge. The derivatographic analysis shows that the sediment contains about 20% m/m of organic matter—Figure 38.

Figure 38.

Derivatogram of the sediment separated from the water phase [62].

The pH of the aqueous layer is initially neutral and becomes slightly acidic. The possible cause of these changes may be sulfide ion reactions leading to the formation of deposits:

Ions $S^{2-}$, from the dissociation of alkali metal sulfides contained in the separated water

$${\mathit{Na}}_{2}S\to 2{\mathit{Na}}^{+}+S^{2-}$$E28

undergo a hydrolysis reaction giving an alkaline pH

$$S^{2-}+2H_{2}O\rightleftarrows H_{2}S+2{\mathit{OH}}^{-}$$E29

Heavy metal sulfide precipitation reactions reduce the concentration of sulfide ions,

$$S^{2-}+{\mathit{Fe}}^{2+}\to \mathit{FeS}\downarrow$$E30

which leads to a reduction in the concentration of hydroxide ions and, consequently, lowering the pH of the environment.

The presented data show that the water separated by coalescing filters from middle atmospheric distillates carries significant amounts of impurities with it. The content of alkali metal compounds is high. The presence of copper—a powerful catalyst for the aging processes of petroleum products—and arsenic—the poison of hydrodesulfurization catalysts, should also be emphasized.

The analysis of the content of mineral impurities remaining in the middle distillates of crude oil indicates the possibility of a significant improvement in the purity of the diesel fractions through the coalescence method of distillate dewatering.

2.3 Tests of fuel dehydration from the tank

Removal of water dispersed in petroleum streams, using coalescence at fibrous barriers, allows to obtain a clear product at the separation temperature; the amount of water that remains in the transparent product corresponds to the saturation concentration. Lowering the temperature of the coalescing dehydrated product causes cloudiness again due to the condensation of excess water. If kept in a tightly closed vessel, the milky, fine dispersion remains stable without visible coalescence and sediment.

During the research on jet fuel hydrotreated material, the behavior of water dispersion from the hydrotreating installation was compared to the coalescence dehydrated product under the conditions of nitrogen flow over the surface of the liquid. The quick and very effective course of clarifying the hydrotreated fractions prompted the study of water evaporation from the kerosene fraction and the jet fuel hydrotreated.

It is known from physicochemical data that water dissolved in hydrocarbons up to saturation concentration exists as single molecules or small aggregates. So it behaves similar to a solution of the gas in a liquid and obeys Henry’s law:

$$C=K\ast p$$E31

where: C - water concentration in the liquid hydrocarbon; p - partial pressure of water vapor in the gas phase; K - constant for a given system, defined by the formula:

$$K=C_{\max }/p_{\mathit{nas}}$$E32

that is the ratio of the maximum concentration of water in the liquid (hydrocarbon) phase to the pressure of saturated water vapor at the same temperature.

Substituting this relationship into Eq. (32), we get:

$$C=C_{\max }\ast p/p_{\mathit{nas}}$$E33

where p/p_nas = ψ is the relative humidity of the gas phase, e.g., air, nitrogen, or saturated vapor of a liquid substance. Thus,

$$C=C_{\max }\ast \mathrm{\Psi }$$E34

The equilibrium concentration of water in the hydrocarbons is proportional to the relative humidity of the gas phase.

Due to the molecular dispersion of water dissolved in liquid hydrocarbons, water exchange between the phases occurs very quickly; an anhydrous layer of fuel, thickness of 5 mm, is saturated with water from humid air to equilibrium within 30 seconds. The opposite phenomenon—water escaping from waterlogged fuel into dry air (Ψ = 2%)—takes place just as quickly; from the 5 mm layer within 10–15 seconds and from 15 mm layer within 10 minutes.

The rate of dehydration of the fuel stationary layer depends on the diffusion rate of water molecules. In homogeneous systems, the diffusion rate (also water-fuel until saturated with water) is determined by Fick’s law:

$$J_i=-{D_i}^{\ast }{\mathit{dc}}_i/\mathit{dx}$$E35

where the diffusion coefficient D (at a temperature of 10°C, the diffusion coefficient of water in the jet fuel is 4.9×10⁻⁴ cm²/s).

The diffusion equalization of water concentrations occurs very slowly in a large volume of fuel. For example, in a vertical tank containing 17 m³ of fuel with a liquid layer height of 2.4 m, takes about 4 years to increase the water concentration from 30 to 70 ppm by diffusion from humid air. The actual time of reaching the equilibrium water concentration is much shorter and amount from a few to a dozen hours. The reason for this is intense convection currents at the walls of the tank—downward during cooling and upward when heating the fuel from the walls (night and day) [15].

The fraction of jet fuel dehydrated by the coalescence method on the barrier, requires additional protection against secondary turbidity and sedimentation of water. The solution to this problem may be applying a small nitrogen flow above the surface of the fraction stored in the tank—Figure 39 [63].

Figure 39.

Schematic diagram of dehydration of the hydrotreated jet fuel by using a coalescer and a movable nitrogen blanket [63]: 1—Coalescer, 2—Tank.

The proposed method is similar to vacuum dehydration and, because it does not require reduced pressure, it can be used in a typical petroleum product tank. A minimum overpressure in the tank (a few kPa) is sufficient to maintain a low nitrogen flow above the liquid surface.

As a result, losses due to the evaporation of most volatile hydrocarbons are reduced. The presence of nitrogen in the gas phase of the tank also protects the product against aging processes. During the contact of the fraction of jet fuel or its hydrotreated product with oxygen, the hydrocarbons most susceptible to oxidation undergo free radical condensation and polymerization reactions. As a result, there are compounds of the type of petroleum resins, which reduce the quality of the stock product.

The studies compared the physicochemical properties of the A2 fraction and the hydrotreated and the behavior of the dispersed water [64]. From the point of view of water dispersion changes, the petroleum product’s surface properties are significant - Figure 40.

Figure 40.

Graph of the dependence of surface tension on temperature [64].

Lower surface tension values characterize the A2 fraction compared to the hydrotreated material, which results from the content of a significant amount of polar sulfur compounds, acting as natural surfactants, adsorbing at the phase boundary. The difference between the surface tension of both fractions increases with increasing temperature.

The interfacial tension of both products at the border with water differs according to the principle that substances with similar surface tension exhibit lower interfacial tension. The value of interfacial tension determines the stability of water dispersion and affects the drainage kinetics of the tested fraction.

Interfacial tension of the raw material, measured with a Krűss K100 tensiometer, is 43,34 mN/m, while the product of the hydrotreating process measures 48,52 mN/m, respectively.

To calculate the fraction drainage rate, the diffusion coefficients of dissolved water and dispersed water particles should be determined. In the case of low dissolved water concentrations in hydrocarbons, the diffusion coefficient can be calculated using the empirical formula [65]:

$$\mathcal{D}=\mathcal{7.4}\cdot {\mathcal{10}}^{-\mathcal{8}}\cdot \frac{{\mathcal{M}}^{\mathcal{0},\mathcal{5}}\cdot \mathcal{T}}{\mathcal{\eta }\cdot {\mathcal{V}}_{\mathcal{m}}^{\mathcal{0},\mathcal{6}}}$$E36

where:

D—diffusion coefficient, cm²/s

M—an average molar mass of the product, g/mol

T—absolute temperature, K

η—product viscosity, cP

V_m—a molar volume of the solute, cm³/mol.

Taking into account the values of the average molar mass of the fraction A2 (162.0 g/mol) and the hydrotreated product (157.3 g/mol), as well as the molar volumes and dynamic viscosity, it is possible to calculate the temperature dependence of the diffusion coefficients of water dissolved in the tested fractions – Figure 41.

Figure 41.

The dependence of the molecular diffusion coefficients of water on temperature [64].

In turn, the diffusion coefficient of spherical droplets, not interacting with each other, is described by the Stokes-Einstein relationship [66]:

$$D=\frac{k\cdot T}{6\pi \cdot \eta \cdot r}$$E37

where:

k — Boltzman’s constant 1.38 · 10⁻²³, J/K.

T — absolute temperature, K.

η — dynamic viscosity of the continuous phase, Pa.

r — drop diameter, m.

The calculated values of the diffusion coefficients of water droplets with a diameter of 4 μm, dispersed in the tested fractions (4–5.10⁻¹⁰ cm²/s), compared to the value of the diffusion coefficients of molecular water indicate that during the dispersion dewatering, the droplet diffusion is involved in the transport of water to the interface to a very limited extent. This statement is confirmed by the calculated value of the droplet diffusion coefficients as a function of their radius—Figure 42.

Figure 42.

Water droplet diffusion coefficients in tested fractions depending on the droplet size (T = 293°C) [64].

The differences in the diffusion coefficients of the dispersed droplets and water in a molecular dispersion are 3–4 orders of magnitude, even for sub-micron-sized particles.

Water solubility in the tested fractions was determined on the basis of turbidimetric measurements. Samples of the tested fractions with different water content, previously heated to the temperature ensuring complete solubility, were cooled and continuously analyzing their turbidity. The temperature at which the system’s turbidity rapidly increases corresponds to the onset of dissolved water condensation and the formation of a two-phase system. The data obtained in this way were used to prepare water solubility curves in the tested fractions—Figure 43. The course of the solubility curves of the hydrotreated and raw fraction A2 shows that the products clear at 50°C may contain about 200 ppm m/m of water in dispersed form after cooling to 20°C.

Figure 43.

Water solubility in the tested fractions and a few selected hydrocarbons depending on the temperature [64].

Based on the measurements of light scattering with different wavelengths, the size and concentration of water droplets in the tested fractions were estimated. Using a spreadsheet or MathCad program, the system of Eq. (38) is solved, and the mean radius of water droplets (r) as well as their concentration (number of particles N per unit volume) are calculated.

$$\left\{\begin{array}{l}{A}_1=\begin{array}{l}\frac{\pi \cdot r^2\cdot N}{2,303}\cdot \{2-\frac{{\lambda }_1}{\pi \cdot r\cdot \left(\frac{\mu }{{\mu }_0}-1\right)}\cdot sin\left[\frac{4\pi \cdot r}{{\lambda }_1}\cdot \left(\frac{\mu }{{\mu }_0}-1\right)\right]+\frac{4}{{\left(\frac{4\pi \cdot r}{{\lambda }_1}\cdot \left(\frac{\mu }{{\mu }_0}-1\right)\right)}^2}\\ \cdot \left[1-cos\left(\frac{4\pi \cdot r}{{\lambda }_1}\cdot \left(\frac{\mu }{{\mu }_0}-1\right)\right)\right]\}\end{array}\\ A_2=\begin{array}{l}\frac{\pi \cdot r^2\cdot N}{2,303}\cdot \{2-\frac{{\lambda }_2}{\pi \cdot r\cdot \left(\frac{\mu }{{\mu }_0}-1\right)}\cdot sin\left[\frac{4\pi \cdot r}{{\lambda }_2}\cdot \left(\frac{\mu }{{\mu }_0}-1\right)\right]+\frac{4}{{\left(\frac{4\pi \cdot r}{{\lambda }_2}\cdot \left(\frac{\mu }{{\mu }_0}-1\right)\right)}^2}\\ \cdot \left[1-cos\left(\frac{4\pi \cdot r}{{\lambda }_2}\cdot \left(\frac{\mu }{{\mu }_0}-1\right)\right)\right]\}\end{array}\end{array}\right.$$E38

The calculations require data including the refractive indices of both dispersion phases (μ) and the amount of scattered light intensity (turbidity) with two wavelengths (λ). The ratio of the refractive index of water and hydrocarbons is in the range from 1.05 to 1.20, i.e., in the so-called Rayleigh distraction area. This equation is applied assuming that the water droplets are spherical, are of equal size, and the absorption of radiation by the dispersion system is much lower than the scattering (non-absorbing particles).

In the A2 fraction, the mean droplet radius is almost independent of the total water concentration (1.85–1.95 μm). As time goes by, the average radius of water droplets in fraction A2 increases. In the hydrotreating material, water condenses in the form of slightly larger droplets, the average radius of which is about 2.05 μm and hardly changes over time (up to 2 hours).

The initial concentration of water droplets in the tested samples is in the order of 4–8.10⁵/cm³, and it decreases differently, more slowly in the A2 fraction. Changes in the size and concentration of water droplets in dispersions prove the advantage of coalescence over sedimentation.

The dehydration kinetics of A2 fraction and hydrotreated jet fuel were investigated using the turbidimetric and gravimetric methods. The analyses of the kinetics of water evaporation concern the real, i.e., polydisperse system. The smallest droplets in the dispersion require a higher concentration of dissolved water, in line with the Kelvin effect. The instability of this part of the dispersion is undoubtedly an additional driving force for water evaporation.

Based on the graphs of the dependence of turbidity on the drainage time of samples with different initial water concentrations, the reliance of the drying time on the initial water content was determined—Figure 44.

Figure 44.

Dependence of the dewatering time of the A2 fraction and hydrotreated jet fuel on the initial water content [64].

In the tested range, the drainage time is directly proportional to the initial water concentration in the sample. The experimentally determined dewatering time of JET-A1 hydrotreated is on average 1.5 times longer than that of the A2 fraction. The main reasons for this seem to be the differences in the diffusion coefficient and water solubility in these fractions. These factors directly affect the limiting stage of the dewatering process, i.e., the diffusion of water molecules toward the interface.

Increasing the temperature favors the dehydration process, increasing both the diffusion coefficient and the solubility of water in hydrocarbons. On the other hand, high temperature results in increased hydrocarbon losses.

Based on the data of the gravimetric analysis, a plot of the weight loss of the sample over time, the anhydrous fraction A2 and the samples fraction A2 containing 300 and 500 ppm of water were plotted.

By calculating the difference between the weight loss of the watered fraction and the pre-dried fraction, the relationship between the mass of water absorbed by the flowing nitrogen and time was determined—Figure 45.

Figure 45.

Graph of the dependence of the mass of water removed from the watered fraction, from failure level, and time [64].

When analyzing the above diagram, it should be stated that the absorption of water in nitrogen allows for almost complete drying of the sample; the water content of the dehydrated A2 fraction is about 30 ppm m/m, irrespective of the initial water content level.

During the dehydration of the A2 fraction, by evaporation, other volatile components are also removed along with water and lighter hydrocarbons, mainly odorous gases. Sulfur compounds, especially hydrogen sulfide, strongly influence the corrosive effects of the petroleum fraction. Removing them is also beneficial from a technological point of view like prevention to oxidation.

Evaporation of water with nitrogen flow allows for quick and efficient dehydration of A2 fraction, a raw material for jet fuel, with relatively low losses of volatile organic compounds. A system in which the fraction pre-dehydrated on a coalescing filter undergoes further drying by the evaporative method seems to be particularly advantageous. The first dehydration stage ensures, in this case, the removal of the water suspended in the form of droplets together with the dissolved mineral impurities. The second stage, evaporation, makes it possible to dehydrate the fraction below the saturation concentration and prevents re-condensation of water when the temperature is lowered.

2.4 Water dispersion in crude oil

Water accompanies crude oil in the deposit, and after its extraction to the surface, it is present in the mine. Due to the fact that water contains dissolved salts, it is usually called brine. Depending on the composition of the geological deposit and the location of the production well, the brine may constitute from a fraction to several dozen percent of the extracted reservoir fluid. Separating the brine from the crude oil is one of the first oil-cleaning operations in the mine. Possibly profound dehydration, and thus desalination of crude oil, significantly impacts the quality of the raw material. The brine remaining in the crude oil constitutes the “ballast” for storage and transport and, above all, causes corrosion of logistics systems’ tanks, pipelines, and fittings.

Water has limited solubility in crude oil; in hydrocarbons, which constitute the essential part of conventional crude oil, water dissolves to a maximum concentration of about 0.06% m/m. These data refer to room temperature because the mutual solubility of water and hydrocarbons increases with increasing temperature. Non-hydrocarbon components of crude oil, compounds of sulfur, nitrogen, and oxygen increase the solubility of water in crude oil. However, more significant amounts of water in the oil, which do not dissolve, form a two-phase dispersion system, called a water-in-crude emulsion.

Water in crude oil forms the most difficult to control dispersion system. In addition to dissolved water containing mineral salts, the crude oil includes brine droplets and solid particles—asphaltenes, petroleum waxes, and solid impurities.

Crude oil dewatering and desalination, as part of the preparation of the raw material for transport, includes partial removal of dispersed water (brine). The remaining amount of water, in a concentration of up to 1% m/m, is usually a stable emulsion with a high degree of fragmentation. The degree of water dispersion depends on crude oil’s physicochemical properties and emulsion formation conditions. The diameter of water droplets in crude oil ranges from 0.2 to 100 μm.

Emulsion durability depends on many factors, first of all on the content of liquid substances in the oil with the characteristics of emulsifiers, such as naphthenic acids and resins, and solid substances—emulsion stabilizers, such as asphaltenes, petroleum waxes, and mineral impurities. Such a mixture is kinetically stable, which means that crude oil’s water (brine) content gradually decreases over time under the influence of external influences. The water separated from the emulsion forms a separate continuous phase. As mentioned above, the instability of the water-in-crude oil emulsion is a significant problem in mutual settlements between the participants of the raw material logistics chain.

An additional difficulty in settlements between the producer, the transport company, and the final recipient of the raw material is the variety of methods used to analyze the water content in crude oil. In business practice, two analytical techniques are used: electrochemical and distillation. Electrochemical methods are generally used to determine small amounts of water in crude oil and are precise, automated, and standardized in many countries. Based on Karl–Fisher reactions, these methods often give inflated water content results due to the interference of many types of chemicals that are generally always present in crude oil. Distillation methods, on the other hand, differ in the determination methodology and the variety of kits and solvents used. The aromatic hydrocarbons used in the assays show significant water solubility, especially at elevated temperatures close to the boiling point. Even if the solvent used for the analysis contains a negligible amount of water, there is a substantial amount of water in the Dean-Stark receiver in equilibrium with the aqueous phase. Thus, the amount of water determination error depends to a large extent on its content in crude oil [67]. If the water in the crude oil content range does not exceed 0.3%, the result is underestimated, and the determination error may exceed 30%. In crude oil containing up to 1% of water, this method’s relative underestimation of the result may range from 2 to 17%. An additional factor that affects the result of the determination is the method of collecting the oil sample for testing. All these factors can cause significant discrepancies in the determination of water content, which in the case of large amounts of raw material in tanks, ships, or pipelines, can cause substantial disparities in the net weight and, thus cause financial consequences.

Preparing crude oil for distillation in the refinery requires repeated desalination and dehydration, carried out in electrodehydrators. The level of purification of mineral salts, mainly inorganic chlorides, of the raw material for distillation has a powerful impact on the rate of corrosion of DR-W installation components as well as on the course of catalytic processes of refinement and processing of petroleum fractions. Desalination efficiency is related to optimal selection.

2.5 The role of water in the combustion of petroleum fuels

The positive water influence on the combustion of fuels, including petroleum ones, has been known and used for many years. One of the first reports is about difficulties in dewatering heavy fuel oil, which resulted in significant changes in the heat of combustion of the fuel and even a flame failure in the burners. The solution, which involves thorough mixing and dispersing of the water, led to the unexpectedly stable operation of power boilers and, at the same time, a visible improvement in combustion. It manifests itself in better use of fuel energy, which reduces its consumption. Also, it reduces the emission of toxic exhaust components—carbon monoxide (II), nitrogen oxides—NOx—and unburned hydrocarbons—CH. In the case of heating oils and diesel fuels, water dispersed in the form of emulsion or microemulsion, in a concentration of 20–30%, causes micro-explosion in the flame [68]. This phenomenon promotes complete evaporation of even residual oils and lowers the temperature of the combustion process; both additives and equipment for the production of water-fuel emulsions in heating oils are offered [69]. Heavy crude oil or bottom residues when mixed with water in about 30% and with the addition of surfactants create ecological boiler fuels of reduced viscosity, such as Orimulsion [70].

In the sixties and seventies of the last century, many studies were carried out and described the use of water for motor fuels, primarily for diesel [71], and gasoline. Extremely environmentally friendly, often straightforward solutions of using different amounts of water for combustion processes have not found practical application. In practice, few companies declare the use of water emulsions as a fuel, for example, Caterpillar [72].

In several countries, including Poland, LPG is very popular alternative fuel for spark ignition (SI) engines. The results of many own studies have been described [73], indicating advantages of this fuel, especially the environment. Also, for this type of fuel, you can use water [74], or the so-called dry steam [75] with positive effects both on the side of reducing the emission of toxic exhaust components and fuel consumption.

3. Discussion of results and summary of research

This chapter presents the results of laboratory and industrial tests in the area of petroleum dispersions—water in fuels. Physicochemical tests of dispersion systems were aimed at solving specific problems of refining technology and the exploitation of petroleum products:

• solubilization of water in fuels and

• coalescence dewatering of the fractions.

The first problem solved and described in water-hydrocarbon systems research is water masking in petroleum fuels by solubilization. The research, conducted mainly on the solubilization of water in gasolines, led to the development of an effective method of binding water and preventing the formation of cloudy dispersion. The water remaining in the gasolines in the production process is solubilized by the additive, a composition of appropriately selected non-ionic surfactants. The water solubilizing additive in gasolines includes ethoxylated alcohols and fatty amines and a copolymer of ethylene oxide and propylene oxide. The solubilizing additive binds at least that part of the water contained in the gasoline, which is released during the cooling of the fuel due to the temperature difference in water solubility and forms the dispersed phase.

From the cognitive and, above all, application point of view, it is important to quantify the solubilizing ability of a mixture of surfactants. This value was determined based on the water content in gasoline and the minimum effective dose of the additive. The variety of chemical compositions of gasoline and changes caused by aging processes significantly affect the surface properties of fuel and the ability to dissolve water. For these reasons, several tests were carried out on surfactant solutions in mixtures corresponding to the composition of gasolines produced in the refinery.

The solubility of individual surfactants in the model hydrocarbon gasoline, gasoline with ETBE, gasoline with ethanol, and different water content was determined. Using an additive that solubilizes water in economic practice primarily required testing its effectiveness in a commercial product. A sample of retail gasoline EuroSuper95, containing 80 mg/kg of water, shows a cloud point of −8°C. At a temperature of +12°C, this gasoline can dissolve 620 mg/kg of water and at 25°C, it can dissolve 850 mg/kg of water. Relatively good water solubility in the tested gasoline is due to the high content of aromatic hydrocarbons (over 31% m/m), olefins (13% m/m), and ETBE ether (2.3% m/m). Gasoline-containing water at a 620 mg/kg concentration, to which 50 ppm m/m of the solubilizing additive was added, becomes cloudy at −18°C. Using the simplified formula for the temperature dependence of water solubility in a mixture of hydrocarbons, it was calculated that the water solubilization capacity of the additive is about 12 g water/g additive. The effectiveness of the water solubilizing additive is determined in mass proportions concerning gasoline, although in the practical application, the volume proportion of the preparation is more convenient.

Prophylactically, the additive should be introduced into the clear fuel, usually at room temperature or elevated, in refinery. Under these conditions, the dissolved water molecules readily form hydrogen bonds with the oxygen of the alkoxy groups of the surfactant. Another way to use the water solubilizing additive is to incorporate the formulation into a cloudy water dispersion in gasoline. The action of the additive is also based on the binding of dissolved water molecules, which consequently leads to the breakdown of water droplets and the gradual disappearance of the dispersed phase. The kinetics of gasoline clarification depends on the amount of dispersed water, fuel temperature, and mixing intensity. Binding by the addition of 200 to 300 ppm m/m of water, dispersed in gasoline, takes place practically within several dozen minutes to several hours.

The turbidimetric method is beneficial for testing water dispersion in hydrocarbons, gasolines, and diesel fuels. It is one of the most sensitive optical techniques, based on low-energy radiation that does not breach the dispersion system under study.

The turbidity measurements sufficiently characterize the behavior of water-saturated hydrocarbons, their mixtures, and petroleum fuels in terms of the water phase’s separation and the masking of its presence. Through turbidimetric measurements, surfactants were selected, which effectively act as an additive that prevents water condensation in the fuel and eliminates dispersion turbidity. Stable and low turbidity value and the disappearance of the cooled and then heated gasoline turbidity hysteresis prove that the addition binds all the water (Figure 46).

Figure 46.

Examples of changes in gasoline turbidity with temperature and the effect of a water solubilizing additive [35].

The measurement of turbidity enables the assessment of the size of the water phase particles in the fuel and the kinetics of changes in dispersion. The method was used to test the evaporation of water dispersed in the kerosene fraction and jet fuel. When solving the system of equations derived from the theory of light scattering, we calculate the average radius of water droplets and their concentration (number of water phase particles per unit volume of fuel). For calculations, it is sufficient to measure the refractive index of both dispersion phases and the intensity of the scattered light with two wavelengths. The ratio of the refractive index of water and hydrocarbons is in the range from 1.05 to 1.20, i.e., in the so-called Rayleigh distraction area. The equations are applied assuming that the water droplets are spherical, are of equal size, and the absorption of radiation by the dispersion system is much lower than the scattering (particles not absorbing light). In the kerosene fractions and the jet fuel hydrotreating material containing 300 ppm m/m of water, the droplets of the dispersed phase reach an average radius of 2 μm, and their concentration (amount) is about 10⁵ per cm³.

It should be added that it is possible to calculate the size and concentration of the particles of the dispersed phase without knowing their mass fraction in the dispersion with this method. Comparing the mass of particles in the dispersed phase, calculated based on turbidity measurements, with the actual water content in hydrocarbons gives a reasonably good agreement of the obtained results.

The water dispersion resulting from cooling the coalescing dehydrated product contains the minimum size—the seeds of a new phase. Their formation requires a significant supersaturation of the solution, described by the Kelvin rule. It may be an essential driving force of the transformation—evaporation of water remaining in the coalescence dehydrated kerosene fraction. The kerosene and jet fuel fractions are quickly dehydrated by evaporation to the flowing inert gas in the tank. Turbidimetric measurements confirm the kinetics of water evaporation.

The sensitive and accurate turbidimetric method is not suitable for the characterization of waterlogged gasoline containing a few (generally about 5) percent of ethanol. Adding water or lowering the temperature of such a biofuel does not cause cloudiness, but leads to the condensation of relatively large, single drops of a water-alcohol solution. The phase stability of gasoline with ethanol is measured by the phase separation temperature, which changes with the water content in the fuel. A slight improvement in the water tolerance of biofuels is achieved by adding the so-called co-solvent. An example of such a product recommended for use in gasoline, is a mixture of C₃-C₅ alcohols with the trade name Octamyl. Due to its low efficiency, the recommended dose of the preparation is from 2 to 5% v/v to the fuel. So it is more of a gasoline component than an additive.

The solution, tested and proven in practice, consists of an additive that solubilizes water, dissolved in a mixture of higher aliphatic alcohols. In this way, the range of applicability of the additive was extended, which considerably increases the water tolerance of gasolines composed with ethanol.

The composition of three domestic-produced non-ionic surfactants submitted for patenting was produced under a license agreement. The product called Aquasol has gained numerous customers in the country and abroad.

Water solubilization in diesel fuel occurs under the influence of a slightly changed composition of the surfactant mixture. The additive is usually used in a lower concentration than the level of dosing to gasoline. The change in composition is related to the lower solubility of one of the additive components in the fuel. In turn, the dose of the additive results from the smaller amount of water that is released during the cooling of diesel fuel. These relations are consistent with the water-hydrocarbons systems; aromatic and unsaturated hydrocarbons show the highest water solubility. These compounds are also characterized by the most significant increase in water solubility with temperature. Due to their high-octane number, motor gasolines contain substantial amounts of both aromatic and unsaturated hydrocarbons. The opposite relationship occurs in diesel fuels, where, due to the required high cetane number, saturated hydrocarbons, including paraffinic hydrocarbons, are particularly desirable components.

The weight fraction of water as a phase dispersed in diesel fuels at room temperature is in the order of 10⁻⁴ (100 mg/kg). Therefore, the sufficient concentration of the water solubilizing additive is up to 50 ppm m/m.

Engine tests of diesel fuel supplemented with a water solubilizing additive at a concentration of 50 ppm m/m showed that the preparation supports the detergent effect of the improving package. Increased purity of the fuel supply system has a positive impact on the ecology of the diesel engine. In practically proven examples, about 40 ppm m/m of additive is sufficient for diesel fuel with a high natural cetane number to protect the fuel against turbidity. Thus, the solubilization of the water allows the diesel to be formulated using a highly paraffinic component from the hydrocracking process.

The practical result of research on the solubilization of water in fuels is an invention, which has been used for years based on a patent license agreement. The solution was awarded at the international exhibition of inventions, and the creators were honored by the Rector of Warsaw University of Technology award. The additive with a slightly changed composition, designated in this chapter as Aq2, which has successfully passed industrial trials at the refinery, has been sold for years on the secondary fuel market also under the name Aquasol.

Because of the market’s interest, and specifically the LPG industry, in the problem of water in liquefied hydrocarbon gases, the authors made another change in the composition of the additive, both of the active substance and the solvent, by introducing a new product to the market under the name of Aquagasol. The product has been used many years in both refineries and LPG warehouses in Poland and abroad.

Thus, research initiated by the grant of the Rector of the Warsaw University of Technology was continued in industrial research in gasoline and diesel fuels and finally in LPG. The family of additives enjoying a good reputation among recipients in the economy of the fuel industry was created.

The second issue examined in the study in water-hydrocarbon systems is the coalescence dewatering of refinery streams. This method is experiencing an apparent renaissance in the oil industry, which results from its numerous advantages and progress in the technology of materials for the construction of coalescing barriers.

As a result of research on the coalescence of water in cracking gasoline, the efficiency of gasoline cleaning from diluted soda lye in the Merox unit of the Fluid Catalytic Cracking plant was significantly improved. Due to the working conditions of the coalescing barrier, the aggressive environment, and the relatively high dispersion temperature, fiberglass was selected as the best material for the barrier. A fiberglass nonwoven fabric with a low-alkali content, a monofilament diameter of 12 μm, and an area weight of 600 g/m² and a glass fiber knitted fabric with an area weight of approximately 2.2 kg/m², made of E-glass fibers with a diameter of 6 μm were used, subjected to special chemical treatment. After 6 years of operation, the coalescing bed underwent significant destruction without losing the ability to coalesce the water phase. At the stage of testing various types of glass fiber materials in the coalescing filter, it was impossible to obtain fully clear gasoline. As with an industrial sand filter, some of the water-leach phases remained in the gasoline. The glass fiber coalescing barrier removes about 230 ppm m/m of the aqueous phase from gasoline, while the filter filled with aggregate removes between 80 and 110 ppm m/m. Replacing aggregate with glass fiber material doubles the efficiency of gasoline dewatering. A clear product at the dehydration temperature could only be obtained after switching off the dosing of the anti-corrosive additive to the gasoline. Application of glass fibers coalescence bed in gasoline filter in Merox unit of FCC plant gave measurable economic and ecological effects.

The influence of surfactants on the water coalescence process in hydrocarbons is frequently researched and remains controversial. In general, surfactants harm coalescence efficiency. Lowering the interfacial tension does not favor the fusing of the dispersed phase particles due to the ease of deformation of the boundary layer. Moreover, the viscosity of the interface, which depends on the nature of the adsorbed organic and inorganic compounds, has a significant influence on the coalescence process. The third important factor influencing the mechanism and kinetics of dispersed water coalescence is the wetting of the barrier material.

In the research on the dehydration of heavy cracking gasoline, it was clearly found that a surfactant (anti-corrosive additive) harmed the effectiveness of the water phase coalescence. The use of an efficient partition in the sand filter of the Merox plant of the FCC installation ended after 6 years with the launch of a new cracking gasoline desulfurization node, which resulted from the need to reduce radically the sulfur content in this stream.

The research on coalescence dewatering of crude oil distillates was brought to the stage of designing a filter-water separator with a unique glass fiber partition. Dewatering the gasoline, kerosene, and diesel fractions through the coalescence of the dispersed water phase requires the selection and modification of the surface of the barrier material. The study investigated that, unlike in the case of cracked gasoline, the water phase present in the distillates is relatively slightly contaminated. Water droplets are formed due to water vapor condensation, and then the water dissolves some of the mineral compounds in the oil fractions. Due to the high temperature of water dispersion in petroleum distillates, some dissolved substances may undergo dissociation and hydrolysis reactions.

For effective coalescence of water dispersed in gasoline and kerosene fractions, glass fiber is a good material for the partition. A needled nonwoven with oriented fibers and subjected to thermal treatment allows obtaining clear fractions at relatively high linear flow velocities through the partition. The barrier should have the optimal thickness as well as the material packing degree (porosity), and the volumetric flow rate should ensure the laminar nature of the dispersion flow.

The hydrophilicity of the glass fibers causes a strong attraction of the dispersed water droplets and then their coalescence on the fiber’s surface. Coalescence occurs mainly according to the inertial mechanism, and other tools’ participation in interference and diffusion is insignificant. Under optimal conditions, fully clear in temperature of separation, dehydrated gasoline and kerosene distillates are obtained.

Dehydration of the diesel fraction with a boiling range of 180–280°C (MDF) requires modification of the glass fiber surface. Partial hydrophobization of the coalescing barrier material with polysiloxane or Teflon resins facilitates the transport and release of water in the form of large droplets. This prevents the redispersion of water in the dewatered fraction and allows to obtain an apparent petroleum product. The reduction of the hydrophilicity of the glass fiber also allows a higher linear velocity of liquid flow through the barrier to be used. The adequate preparation shows various durability in the working conditions of the partition, resistance to washing with water, and hydrocarbons.

The coalescence dehydration of diesel fractions is much more difficult with the boiling point range of 240–350°C, the highest boiling crude oil atmospheric distillate (HDF). The separation must be carried out at an elevated temperature due to the high viscosity of the fractions and the high content of n-paraffin hydrocarbons in the distillate. The glass fibers of the partition material require hydrophobization, and the separation does not result in a completely clear distillate. Water droplets favor the agglomeration of solid paraffinic hydrocarbons, and the presence of paraffin aggregates stabilizes a small amount of dispersed water. This phenomenon was observed in the studies of water solubilization in diesel fuel containing the heavy diesel (HDF) fraction from the hydrocracking process.

A common feature of the coalescing dehydration of diesel distillates is a significant reduction in the content of metal compounds in the fractions. The degree of extraction of the metal compounds increases with the boiling point of the distillate and ranges from about 50 to over 80%. The alkali metal compounds of sodium, calcium, and magnesium are removed to the greatest extent from petroleum distillates to the water phase. The use of such a method of dehydrating distillates from the atmospheric part of the AVD installation results in the reduction of the content of mineral impurities and catalyst deactivators in the fractions constituting the raw material for hydrogen processes. The presented solution, protected by a patent, has not been used in industry so far.

Coalescing dehydration of the jet kerosene hydrotreating product with evaporative dewatering through a moving nitrogen blanket in the tank was also not commercially applied. This method of product drying additionally prevents it against aging.

Concerning water dispersion in hydrocarbons, it is possible to apply coalescence in an electric field without supplying energy to the device. The coalescing barrier consists of mixed particles (grains) of two metals or metal and non-metal with the greatest possible difference in electrochemical potentials. The conductive contact of both materials in combination with water (electrolyte) creates a micro-cell with a specific electromotive force value, e.g., iron-aluminum combination gives a potential of 1 V. Depending on the distance between the electrodes, the intensity of the electric field in the slots of the barrier, through which the dispersion flows, can reach values of hundreds of V/cm, i.e., similar to those used in industrial electrodehydrators.

During the research on dehydration of cracked gasoline, this type of element was used in a combined partition, and the results were very promising. Attempts to dehydrate aviation fuel hydrotreated in laboratory conditions confirmed the effectiveness of this type of baffle. The use of such a device in the industry seems possible, although it requires much work and research. The main problem to be solved is to create a permanent partition by combining fine-grained materials or meshes. This difficult problem in the field of welding engineering has been attempted to solve, unsuccessfully, by explosively joining bimetallic meshes under water as well as by laser multi-point welding. The topic remains open, although it has probably already found a solution and application in the oil industry.

Acknowledgments

The author would like to thank PETROL Invention Enterprise for financial support for the publication of this work. Author wants to thank many co-workers, researchers, technical assistants, and students preparing their diploma in laboratory and industrial researches. Special thanks to Mr. Romuald Zębrzycki for his help in preparing the graphics of this work.

Additional information

Some parts of this chapter were originally published in Polish, in the work Physicochemistry of Petroleum Dispersions in Refining Technology Optimizing, dated year 2007.