- Email: [email protected]
Measuring the density and viscosity of H2S-loaded aqueous methyldiethanolamine solution
Accepted Manuscript Measuring the Density and Viscosity of H2S-Loaded Aqueous Methyldiethanolamine Solution Mohammad Shokouhi, Reza Ahmadi PII: DOI: Reference:
S0021-9614(16)30109-4 http://dx.doi.org/10.1016/j.jct.2016.06.007 YJCHT 4673
To appear in:
J. Chem. Thermodynamics
Received Date: Revised Date: Accepted Date:
2 August 2015 30 April 2016 13 June 2016
Please cite this article as: M. Shokouhi, R. Ahmadi, Measuring the Density and Viscosity of H2S-Loaded Aqueous Methyldiethanolamine Solution, J. Chem. Thermodynamics (2016), doi: http://dx.doi.org/10.1016/j.jct.2016.06.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Measuring the Density and Viscosity of H2S-Loaded Aqueous Methyldiethanolamine Solution
Mohammad Shokouhi,* Reza Ahmadi
Research Institute of Petroleum Industry (RIPI), National Iranian Oil Company (NIOC). Gas Research Division, Research Institute of Petroleum Industry (RIPI), P.O. Box 14665–137, Tehran, Iran.
* Corresponding author. E–mail: [email protected] ; [email protected] Tel: 98 21-48252466.
Page 1
ABSTRACT The density and viscosity of H2S-loaded aqueous 46.78 mass% methyldiethanolamine solution were experimentally measured accompanied with the solubility of H2S at temperatures (313.15, 328.15 and 343.15) K, pressures from vapor pressure of fresh solution up to 1.0 MPa and loadings up to 1.00 mol of H2S per 1 mol of amine. All experimental trials have been carried out using the new setup developed in our laboratory. It was observed that both density and viscosity of mixtures decrease by increasing temperature and density increase by increasing acid gas solubility (loading) by about 4.7%, whereas viscosity has a complicated behavior with H2S solubility. Viscosity decreases by increasing acid gas solubility (loading) at 313.15 K by about 20.6% and at 328.15 K by about 15.0%, but it is comparable at 343.15 K in terms of H2S solubility. Finally, the experimental density and viscosity data correlated using Modified Setchenow equation.
Keywords: Gas solubility, Chemical absorption, Transport property, Methyldiethanolamine, Density, Viscosity.
Page 2
1. Introduction Knowledge on the thermal and non-thermal properties of gas-loaded physical and chemical solvents such as density, viscosity, surface tension, enthalpy, heat capacity, thermal conductivity and thermal diffusivity are necessary for 1) the design of acid gas treatment equipment and subsequent measuring operation 2) the solvent regeneration at reduction stage 3) predicting and interpreting the laboratory data such as kinetic rates constants [1], diffusion coefficients [2] and mass-transfer characteristics of gas/liquid contacting devices. An enormously growing number of experimental results for thermophysical properties of pure solvents as well as those of their mixtures with other solvents have been reported in the literature. However, data regarding acid gas adsorbed in physical and chemical solvents are scarce. Jalili, et al. [3], have reported density and viscosity of sulfolane loaded by CO2 and H2S and sih et al. [4] have reported those properties of CO2-methanol system. Both reported data [3,4] have shown that density and viscosity of sulfolane and methanol loaded with CO2 reduce with increasing loading. Density and viscosity
of
methyldiethanolamine
(MDEA),
diethanolamine
(DEA),
monoethanolamine (MEA) and their mixture loaded with CO2 have recently been reported by Weiland et al. [5]. It is worth noting that those data reported by Weiland et al. [5] were prepared using quenching and separating method in which solutions were partially carbonated. Dell,Era et al. [6] reported the density of CO2 loaded solution of Diisopropanolamine (DIPA). Jayarathna et al. [7,8], measured the density and surface tension of CO2 loaded aqueous MEA solution at (303.15 to 343.15) K and loading from (0 to 0.5). Density and viscosity of aqueous MDEA solution (45 mass %), as well as its aqueous mixtures with 2-Amino-2-methyl-1-Propanol (AMP) (40 mass% MDEA + 5 mass% AMP) and DIPA (40 mass% MDEA + 5 mass% DIPA) under the CO2 loading at temperatures from (303.15 to 363.15) K and pressures up to 2.0 MPa were experimentally measured by our research group [9]. Freeman et al. [10] have measured the density and viscosity of aqueous Piperizine + CO2 solutions at temperature range from (293.15 to 333.15) K and the concentration ranges from (2 to 20) molal Piperazine and CO2 loading from (0 to 0.4). Some other documented literature data for carbonated alkanolamine solutions have been reported including MEA, DEA, MDEA, AMP, 2-dimethylaminoethanol
Page 3
(DMAE), 2-diethylaminoethanol (DEAE) and 1-dimethylamino-2-propanol (1DMA2P) [11,12,13,14,15,16,17,18]. All reported data show that density, viscosity and surface tension measurements of carbonated aqueous alkanolamine solutions increase with increasing CO2 loading and decrease with temperature. A literature survey on H2S solubility in MDEA and mixed MDEA with the other alkanolamines aqueous solution has been done by Shokouhi et al. [19]. None of the experimental works collected in that reference measured the density and viscosity of H2S-loaded solution. Making further efforts show that, except for Rinker et al. [20], there isn’t any experimental data on density and viscosity of H2S loaded alkanolamine solution. In Rinker et al. [20], measurements at 25 °C were carried out for (10-50 mass %) MDEA aqueous solution at loadings 0.1, 0.3, 0.4 and 0.5 (mol of H2S/mol of MDEA) and measurements at 40 °C were made for 50 mass % MDEA aqueous solutions at H2S loadings of 0.1, 0.3, 0.4 and 0.5 mol of H2S/mol of MDEA. They have shown that, the density of aqueous MDEA increases with increasing H2S loading while the viscosity decreases with increasing H2S loading which this result is completely opposite of carbonated alkanolamine solutions. In this work, we focus on the measurement of density and viscosity of 46.78 mass% methyldiethanolamine (MDEA) loaded with H2 S using a setup developed in our laboratory without any quenching procedure. All experimental trials are carried out on stand situation at temperatures 313.15, 328.15 and 343.15 K, and pressure from vapor pressure of uncharged solution up to about 1.00 MPa and loadings from zero up to about 1.00 mol of H2S per 1 mol of amine. Gas loading was systematically measured by gravimetric method. Finally, the experimental density and viscosity data correlated using Modified Setchenow equation [21,22,23,24]. 2.
Experimental Section
2.1.
Materials:
Hydrogen sulfide [c.p. grade 99.95 % min and CAS registry number 7783-06-4] was obtained from Roham Gas Company. N-Methyldiethanoleamine [CAS registry number 105-59-9] was purchased from Sigma-Aldrich with purity >99% and used as purchased without further purification. Water used as solvent was distilled deionized, which was degassed in an ultrasonic bath (FUNGILAB, model UA10MFD) at temperature 343 K
Page 4
and wave frequency 50 kHz for about 1 hour prior to use. All solutions were prepared by measurement of mass of the solutes and solvent on a calibrated balance (Mettler model AE 200) with an uncertainty ± 0.0001 g and measuring the volume of the solution by a standard volumetric flask up to 250 mL. The specifications of the chemicals used in this work contain source, purification method, purity base (mass or mole) are summarized in table 1. 2.2. Apparatus and procedure: The details of the experimental method for the measurement of gas solubility, density and viscosity have previously been presented [3] and only a short description will be provided here. The experimental setup as having shown in figure 1, comprised of four separate apparatus, each one considered to perform a specific task: (a) the main equilibrium cell (EC), (b) a gas injection system (GIS), (c) a vibrating tube densimeter (VTD), and (d) a falling weight viscometer (FV). The EC was equipped with a magnetically driven mechanical stirrer to accelerate equilibrium state and a side glass window (SW) to allow one to monitor both the liquid level and the foaming of the solution in the EC. The EC was double wall jacketed to facilitate the circulation of water from water recirculation bath (WCB) to control the temperature of the solution in the cell. The temperature of EC was monitored using a Lutron model TM-917 digital thermometer with a 0.01 K resolution using a PT-100 sensor inserted into the thermowell of the EC (not shown in figure 1). The pressure of the system was measured using a Keller model PA-33X pressure transmitter sensor (P2) in the range of (0 to 4) MPa with an uncertainty within ±0.01% of full scale. The gas injection system (GIS) consists of a gas container (GR) with constant known volume equipped with a Keller model PA-33X pressure transmitter sensor (P1) in the range of (0 to 4) MPa with an uncertainty within ±0.01% of full scale. The GIS, which is connected to the EC through V3 valve, is directly fed from gas cylinders through V1 valve. To measure the solubility of gases in the liquid solvent, known quantities of gaseous solute and degassed solvent are brought into contact at a constant temperature inside the measurement system (EC + VTD + FV) with known volume. On reaching thermodynamic equilibrium, the pressure above the liquid solution, which is measured
Page 5
by the P2 pressure transmitter, is constant and directly related to the solubility of the gas in the liquid. The quantity of solute present in the liquid solution is calculated by the difference between two PVT measurements: first on introduction of the gas from the container GR of known volume into the measurement system (EC + VTD + FV) containing the solvent and secondly after reaching thermodynamic equilibrium, i.e. when autoclave pressure, PT , remains fixed and no longer changes with time. Using the procedure adopted by Park and Sandall [25] and Hosseini-Jenab et al. [26]:
nag =
Vgc Pi Pf − RTa Z i Z f
(1)
where Vgc denotes the volume of the gas container, Zi and Zf are the compressibility factors corresponding to the initial and final pressures, Pi and Pf, respectively, in the gas container before and after transferring gas, and Ta is the ambient temperature, which is equal to that in the gas container. Compressibility factors were calculated using NIST [27]. Equilibration between liquid and vapor phases inside the cell were normally achieved within about 2 h after beginning of stirring and the partial pressure of gas at equilibrium state in the equilibrium cell, Page , was calculated as follow:
Page = PT − PVP
(2)
where PT and PVP denote the total pressure and vapor pressure of solution. A key issue is the determination of the vapor pressure of the mixture solution since this value must be subtracted from the total pressure to obtain the partial acid gas pressure. The moles of remaining acid gas in the gas phase, nagg , was determined from: g ag
n =
Vg Page Z ag RT
= Vg .ρ ag
(3)
where Vg is the gas phase volume, T is the equilibrium temperature of the cell, and Zag and ρ ag are the compressibility factor and density of acid gas at Page and T, respectively. The quantity of gas in the liquid phase was then determined from: nagl = nag − nagg
(4)
The molality and loading of charged gas in the liquid phase is defined as:
Page 6
m ag =
α ag
nagl ( mol) wSolvent (g)
(5)
× 1000
nagl (mol) = w (g) ∑i Mi (g) i
(6)
wi is the mass of comprised amines in fresh solvent in g, and Mi is molar mass of pure
comprised amines in g.mol-1. The isochoric saturation solubility method may be addressed some drawbacks such as; a) it serves static situation which don’t never fulfill dynamic situation in absorption/stripping towers b) The amounts of acid gas which are present in the liquid phase should be approximated using in the vapour phase. In this method, it has been assumed that the partial pressure of acid gas in the vapour phase is the difference between their experimental results for the total pressure (above the charged liquid mixture) and the saturation pressure of the uncharged solvent c) There is a more serious problem in which by dissolving acid gas in the solvent, the volume of the liquid phase remains unchanged, i.e., this method neglects any volume expansion or compression caused by acid gas in the liquid phase. Our new developed set-up has paved a rout to tackle the third drawback [9]. As mentioned above the amount of dissolved gas is calculated from the amount of gas charged into the cell by subtracting a correction for the amount of gas that is still present in the vapor phase. That correction is calculated assuming that the volume of the vapor phase is the difference between the cell volume and the volume of the unloaded solvent (which is known from the preparation of the fresh uncharged solution). However, that procedure neglects a volume change (mostly a volume expansion) when a gas is dissolved in a liquid. To improve further our measurement processes, equation (3) may be corrected as (7). mag ⋅ wSolvent + wSolvent 1000 g = ρ ag . Vauto − nag ρ
(7)
mag is the molality of acid gas obtained from iteration technique or as an approximate estimation obtained from molality calculated with equations (3) through (6). ρ is the
Page 7
charged solution density measured on stand (without quenching method) using procedure described in following section. Vauto is the volume of auto clave which is equal to volume of EC + VTD + FV. The error propagation theory was used to estimate the uncertainties of final results [28]. In the base of this theory, the uncertainty δq of the interest variable q(r … u) is given by equation (8): 2
∂q ∂q U (q ) = ± .dr + ... + .du ∂r ∂u
2
(8)
The measured quantity q is dependent upon the variables r… u which fluctuates in a random and independent manner (dr… du). The uncertainties of all parts of the instruments used in the measurements were considered for estimating the uncertainty of the H2S loading in the liquid phase. The main contributions to the uncertainty of the solubility are attributed to errors in the pressure sensor for equilibrium cell and gas container (both are equal to ±4×10-4 MPa), temperature sensors (±0.2 K), and scale for the amount of solvent in equilibrium cell (±0.0001 g). The volumes of the different parts of the set up were obtained by performing pressure swing experiments [9,19, 29,30,31]. Several measurements using this method were done to obtain the volumes and uncertainties of the volumes of different parts of the setup. According to equations (6) and (8), the gas solubility is related to the amount of absorbed gas in the liquid phase, and that of solvent. The average calculated uncertainties of these two parameters at T=328.15 K were calculated to be U (nHl 2S ) = ±0.003 mole and U (wSolvent ) = ±0.0001 g, therefore the average uncertainty of loading, ± U (α ) , is equal to ±0.006.
2.2.1. Density measurement
Measurement of the density of fresh and charged liquid solutions prepared in equilibrium cell, EC, at each specified temperature, T and pressure, p was carried out by the vibrating tube densimeter (VTD). The VTD was an Anton-Paar model DMA HPM, equipped with a liquid jacket by the supplier, the temperature of which was thermostatted through circulation of water from the external water recirculation bath, WCP. The DMA HPM density measuring cell is a U-shaped Hastelloy C-276 tube
Page 8
electronically excited to vibrate at its characteristic frequency, which in turn depends on the density of the sample. The DMA HPM is connected to the interface module (IM), which generates and measures the period of oscillation and is also responsible for the temperature measurement. The IM is in turn connected to an mPDS 2000 V3 evaluation unit (EU), which converts the measured raw data (oscillation period, temperature and pressure) to the density of the sample through the appropriate calibration equations. The VTD density measuring device was calibrated by using distilled deionized water, pure methanol and pure sulfolane in the temperature range (298.15 to 363.15) K and from atmospheric pressures up to about 7.0 MPa employing equation (9) as suggested by the supplier:
ρ=
∑a
ijk
(9)
⋅τ i ⋅ T j ⋅ P k
i, j ,k
where aijk, τ , T, and P stand for apparatus constant, period of oscillation, temperature, and absolute pressure, respectively, with i = 0, 2, 4, j = 0, 1, 2, and k = 0, 1, 2. Uncertainties of density U (ρ ) using equation (9) obtained about U( ρ )= 2.5×10-3g.cm-3.
2.2.2. Viscosity measurement
The viscosity of pure as well as gas saturated liquid solutions was measured by the falling weight viscometer (FV). FV, which was a cylindrical steel tube with 85 cm length and 6.8 mm inner diameter, was equipped with a liquid jacket to be thermostatted by an external water recirculation bath (WRB). The FV was mounted vertically and consisted of a cylindrical sinker. The sinker with 6.4 mm outer diameter and 20.4 mm length, and hemispherical ends was an iron magnetic bar coated with Teflon PTFE to prevent corrosion and its density was obtained 2880.3 kg · m-3. A working equation suggested by Daugé et al. [32] was used to measure the viscosity of solutions.
η (T , p ) = K a (T ) + K b (T ) ⋅ (ρ s − ρ ) ⋅ t
(10)
In this equation, the viscosity of unknown solution is considered to be linearly dependent on the difference between the density of the sinker, ρs , and that of the solution, ρ . t is the time for the sinker descending between two sensors with a specified and adjustable distance. The Ka(T) and Kb(T) parameters were determined
Page 9
through calibration of the viscometer by liquids with known viscosity at a specified temperature. Distilled demonized water, pure sulfolane and pure methanol were used as the reference fluids for calibration of the apparatus at various temperatures and pressures. The digital timer is capable of registering the falling time of the sinker with a minimum reproducibility ± 0.03 s. After the introduction of the known amounts of liquid and gas into the EC and adjustment the temperature of the system, the centrifuge pump was turned on to circulate the liquid thoroughly in the system. This operation helps achieve two important goals during measurement: (1) increasing the rate of equilibrium approach through mixing and homogenizing the liquids and more importantly (2) starting the measurement of the viscosity of the mixtures whereby lifts the sinker up the vertical cylinder without the need for any additional external/internal apparatus. By closing V4 valve, the drag force ceases and the sinker falls through the liquid down from the top to the base of the vertical tube. By passing through the two S1 and S2 sensors, the falling time is registered and the viscosity is calculated by using Eq. (10). The speed of the centrifuge pump was adjusted in such a way that a laminar flow regime of the liquid always prevailed in the FV vertical cylinder. A laminar flow ensured a uniform liquid phase without any bubble formation during falling of the sinker in the cylinder. 3. Results and discussion As reported in our previous paper [3] the accuracy and reliability of the experimental data produced by the apparatus developed in this work was checked through comparison of the measured solubility, density and viscosity of pure sulfolane, pure methanol and the (CO2 - methanol) binary mixture with the available experimental data reported in the literature [4,33]. The Sih et al. [4] and Chiehming et al. [33] methods are in principle similar to the method used in this work. The maximum per cent deviation of the density values reported by Sih et al. [4] from this work is less than 0.50%. The maximum per cent deviation of the viscosity values reported by Sih et al. [4] from this work is less than 1.0 % with average deviation of 0.39 %. The two sets of solubility measurements are also in quite good agreement with each other. This is also the case for the solubility and density data from (298 to 313) K reported by Chiehming et al. [33].
Page 10
The PTx solubility data obtained using our setup for (CO2 - methanol) binary mixture are also consistent with the corresponding values reported by Weber et al. [34] at 298 K (AAD% = 3.7% and MAD% = 6.0%), with those reported by Ohgaki and Katayama [35] at T = (298 and 313) K (AAD% = 4.8% and MAD% = 6.8%), and also with those reported by Xia et al. [36] at 313.75 K (AAD% = 7.2% and MAD% = 8.4%), in which, AAD% and MAD% stand for the average of absolute deviations and maximum of
absolute deviations of mole fraction, density or viscosity. Experimental solubility of H2S in 46.78 mass% aqueous MDEA solution and vapor pressure of gas free solution as well as fresh and loaded solution density and viscosity accompany with their uncertainties have been reported in table 2. Temperature and pressure dependency of H2S solubility graphically has been shown in figure 2. Data from this work (at T = 313.15 K) has been compared with those obtained using the computer-operated apparatus reported by Sidi-Boumedine et al. [37] in 46.78 mass% aqueous MDEA and also data reported by Kamps et al. [38] in 48.8 mass% aqueous MDEA and Li et al. [39] as well. Vapor pressure of gas free MDEA solution obtained from this work graphically compared with those reported by Xu et al. [40], kim et al. [41] and Shokouhi et al. [9] in figure 3. One of the main assumptions in equilibrium state was given with equation (2) in which vapor pressure of solution was kept constant with the acid gas loading. To confirm the validity of equation (2), vapor pressures of H2S-loaded solution at different temperatures and loading were tried with values given by Results law. All loading were recalculated using this more sensible assumption and results show that all recalculated loading values are the same as those reported in table 2 with three digits after the point. Comparisons of density and viscosity of unloaded 46.78 mass% MDEA aqueous solutions with documented published data has been reported in table 3. The average deviation of density reported by Bernal-Garcia [42] and Hsu-Li [43] from this work is less than 0.3%, in which both reported their experimental data as Redlich-Kister type equation. The average deviation of viscosity reported in this work from those data obtained from correlation in Bernal-Garcia [44]is less than 7.0%, where as that is less than 0.8% in comparison with data reported by Teng et al. [45]. Deviation of density and viscosity of unloaded MDEA solution obtained from this study were schematically compared to other reported ones [42-45,46,47] in figures 4 and 5, respectively.
Page 11
Dependence of H2S solubility on density and viscosity of H2S-loaded solution has been shown in figures 6 and 7. As may be seen from figures 6 and 7, both density and viscosity of mixtures decrease by increasing temperature, and density increases by increasing acid gas solubility (loading) by about 4.7%, whereas viscosity has complicated behavior with H2S solubility. Rinker et al. [20] have measured the variation of density and viscosity of aqueous MDEA with addition of hydrogen sulfide. Their measurements were carried out at 298.15 K for loading up to 0.5 mol H2S per 1 mol of amine for amine concentrations up to 50 mass % and also at 313.15 K for amine concentration of 50 mass % and for H2S loading up to 0.5. They have found that, the density of aqueous MDEA increases with increasing H2S loading while the viscosity decreases with increasing H2S loading, and also MDEA concentration in unloaded solution has increasing effect on density and viscosity of solution. In table 4, values of percent rate variation for density, ∆ρ%, and viscosity, ∆η%, of loaded solution in comparison of unloaded one at two loading values (α = 0.5 and 1.0) have been reported and compared with those values obtained from Rinker et al. [20]. From table 4, both sets of data (this work and Rinker et al.) show that, percent rate increase of density, ∆ρ%, is roughly independent of temperature, and also according to Rinker et al. it depends on MDEA concentration in unloaded solution. There is an agreement with the measurements of density reported by Rinker et al. and this work. Rinker et al. [20] found that for the case of (50 mass % MDEA solution, T = 313.15 K, α = 0.5) the density increases about 2.8%, and in this work (46.78 mass % MDEA
solution, T = 313.15 K and α = 0.5) the density increases about 3.2%, while this value for α = 1.0383 mol H2S/mol MDEA is about 4.5%. For the case of viscosity, slop of viscosity decrease for Rinker et al. is higher than that reported in this work, for instance, Rinker et al. found that the viscosity decreases about 16.1% (50 mass % MDEA, T = 313.15 K and α = 0.5), whereas in this work (46.78 mass % MDEA and T = 313.15 K) viscosity decreases about 12.8% for the H2S loading 0.5, and about 20.6 % for H2S loading 1.0988 mol H2S/mol MDEA.
4. Density and Viscosity modeling:
Page 12
The experimental density and viscosity of solution are correlated using modified Setchenow type equation which is recently used for correlating of density, viscosity, surface tension, heat capacity, thermal conductivity and thermal diffusivity of binary and ternary solutions [21,22,23,24]. The working equation is represented by equation (11): 2 p ln = ∑ k j .α p r j =1
(11)
j
where p and pr stand for the physical property of the mixture and reference substance, respectively; for example p represents the density / viscosity of the (H2S - alkanolamine solution) mixture and pr is the density / viscosity of uncharged alkanoleamine solution at the same temperature as the mixture. The kj’s are temperature dependent parameters, which are known as the Setchenow coefficients, and α is the concentration of H2S dissolved in the liquid phase at the specified temperature and pressure. The final working equation for correlation of density and viscosity are; ρ ln ρr
= (k 0,0, ρ + k 0,1, ρ ⋅ T ). ⋅ α + (k1,0, ρ + k 1,1, ρ ⋅ T ) ⋅ α 2
η ln ηr
= (k 0,0,η + k 0,1,µ ⋅ T ). ⋅ α + (k1,0,η + k 1,1,η ⋅ T ) ⋅ α 2
(12)
(13)
The optimized values of adjustable parameters, k i , j ,ρ s and k i , j ,η s, were obtained using minimizing the objecting function expressed as sum of the relative deviation values and all adjustable numerical values are tabulated in table 5. The average of relative deviations, ARD%, defined by equation (14) and maximum of relative deviations, MRD%, defined by equation (15) for a number of N experimental points which both
may be regarded as the quality of correlation. cal exp 100 N pi − pi ARD % = ∑ P exp N i =1 i
(14)
pical − piexp MRD% = Maximum ⋅ 100 exp pi
Page 13
(15)
p ical and piexp in equations (14) and (15) stand for the calculated and experimental value
of the physical property of interest, respectively. The average and maximum per cent deviations of the correlated from experimental values together with the corresponding experimental uncertainties (ARD%)exp and (MRD%)exp, defined in equations (16) and (17), are given in table 5.
(
exp 100 N U pi ( ARD %) exp = ∑ exp N i =1 pi
)
(
(16)
)
U piexp (MRD%)exp = Maximum exp ⋅ 100 pi
(17)
In which U ( piexp )is the uncertainty of physical property which was obtained using 4 or 5 times measuring the viscosity and for the case of density, it was obtained using experimental values of pure methanol, water and sulfolane in the temperature range (298.15 to 363.15) K and atmospheric pressures up to about 7.0 MPa employing equation (9). It can be observed from table 5, ARD% and MRD% values of loading and viscosity are smaller than (ARD%)exp and (MRD%)exp, i.e., the correlation represents the density and viscosity values within experimental uncertainty. It means that, modified Setchenow equation is able to describe the temperature and loading dependency of the density and viscosity of the system studied in this work.
5. Conclusion
The physical equilibrium and transport properties, i.e. H2S loaded density and viscosity of MDEA solutions which are in use in the natural gas sweetening processes, were measured in this work. The experimental setup developed in our laboratory is able to measure simultaneously the density and viscosity with the solubility of acid gases in both physical and chemical solvents. The reliability of the experimental values generated by the setup was assessed through comparison of the measured solubility, density and viscosity of pure sulfolane, pure methanol and the (CO2 + methanol) binary mixture with the available experimental data reported in the literature [3]. All measurements at 313.15, 328.15 and 343.15 K were carried out for H2S loadings from zero to about 1.0 (mol H2S / mol MDEA) in 46.78 mass% MDEA aqueous solution.
Page 14
These data have been shown in figures 5 and 6, and also listed in table 2. As may be seen from those figures and table, the density of aqueous MDEA increases with increasing H2S loading by about 4.7%, while the viscosity decreases with increasing H2S loading at 313.15 K by about 20.6% and at 328.15 K by about 15.0%, but it is roughly unchanged with loading at 343.15 K.
NOMENCLATURE
NIST;
National Institute of Standards and Technology
R; Universal gas constant MDEA; N-Methyldiethanolamine EC;
The main equilibrium cell
GIS;
Gas injection system
VTD; Vibrating tube densimeter FV;
Falling weight viscometer
GR;
Gas container
Vgc ;
Volume of the gas container (GR)
Vg;
Gas-phase volume in the equilibrium cell
ρ ag ; Density of acid gas in equilibrium situation in gas phase Zi and Zf; Compressibility factors of the initial and final state in the gas container T a;
Ambient temperature
P0 ;
Initial pressure of solution
PT ;
Total absolute pressure
PVP ; Vapor pressure of solution which is equal to P0 PH 2S or PHe2S ;
Partial pressure of H2S at equilibrium state
Vauto ; Volume of auto clave wi ; Mass of pure species i
Mi ; Molar mass of pure species i mag ; Molality of acid gas
ρ charged solution ; Charged solution density p;
Physical property of mixture
pr;
Physical property of reference system
Page 15
nHg 2S ; Amount of H2S in the gas phase at equilibrium state nHl 2S ; Amount of H2S in the liquid phase at equilibrium state n Amine ; Amount of total amine in the liquid phase
( )
U nil ; Uncertainty of amount of species i in liquid phase.
α CO ; Loading of H2S in solution 2
U(α); Uncertainty of loading U(m); Uncertainty of molality of MDEA in aqueous solution AAD; Average of absolute deviations MAD; maximum of absolute deviations ARD; Average of relative deviations MRD; Maximum relative deviation
ρ s ; Density of sinker (g/cm3)
ρ ; Density of solution (g/cm3) aijk; Apparatus (Anton-Paar) constant
τ ; Period of oscillation Ka(T), Kb(T); Viscometer calibration constants t; Time in second
η ; Viscosity of solution η r ; Viscosity of reference system kj; Temperature dependent parameters of Setchnow equation ∆ρ%; Percent rate variation of solution’s density ∆η%; Percent rate variation of solution’s viscosity
ACKNOWLEDGEMENT
We are thankful to the Research Council of the Research Institute of Petroleum Industry (RIPI) and the Research and Development of the National Iranian Oil Company (NIOC) for their support of this work.
REFERENCES
Page 16
[1] E.B. Rinker, S.S. Ashour, O.C. Sandall, Chem. Eng. Sci. 50 (1995) 755-768. [2] A. Tamini, E.B. Rinker, O.C. Sandall, J. Chem. Eng. Data 39 (1994) 330-332. [3] A.H. Jalili, M. Shokouhi, F. Samani, M. Hosseini-Jenab, J. Chem. Thermodyn. 85 (2015) 13-25. [4] R. Sih, F. Dehghani, N.R. Foster, J. Supercrit. Fluids 41 (2007) 148–157. [5] R.H. Weiland, J.C. Dingman, D.B. Cronin, G.J. Browning, J. Chem. Eng. Data 43 (1998) 378-382. [6] C. Dell,Era, P. Uusi-Kyyny, J.–P. Pokki, M. Pakkanen, V. Alopaeus, Fluid Phase Equilb. 293 (2010) 101-109. [7] S.A. Jayarathna, A. Weerasooriya, S. Dayarathna, D.A. Eimer, J. Chem. Eng. Data 58 (2013) 986-992. [8] S.A. Jayarathna, C.K. Jayarathna, D.A. Kottage, S. Dayarathna, D.A. Eimer, M.C. Melaaen, J. Chem. Eng. Data 58 (2013) 343-348. [9] M. Shokouhi, A.H. Jalili, F. Samani, M. Hosseini-Jenab, Fluid Phase Equilibria 404 (2015) 96–108 [10] S.A. Freeman, G.T. Rochelle, J. Chem. Eng. Data 56 (2011) 574-581. [11] T.G. Amundsen, L.E. Øi, D.A. Eimer, J. Chem. Eng. Data 54 (2009) 3096−3100. [12] D. Fu, L. Chen, L. Qin, Fluid Phase Equilib. 319 (2012) 42−47. [13] D. Fu, Z. Li, F. Liu, J. Chem. Thermodyn. 68 (2014) 20−24. [14] A. Hartono, O.E. Mba, H.F. Svendsen, J. Chem. Eng. Data 59 (2014) 1808−1816. [15] J. Han, J. Jin, D.A. Eimer, M.C. Melaaen, J. Chem. Eng. Data 57 (2012) 1843−1850. [16] D.D.D. Pinto, J.G.M.S. Monteiro, B. Johnsen, H.F. Svendsen, H. Knuutila, Int. J. Greenhouse Gas Control 25 (2014) 173−185. [17] J. Zhang, P.S. Fennell, J.P.M. Trusler, J. Chem. Eng. Data In Press [18] Z. Idris, L. Ang, D.A. Eimer, J. Ying, J. Chem. Eng. Data 60(5) (2015) 1419-1425. [19] M. Shokouhi, H. Bozorgzade, P. Sattari, J. Chem. Eng. Data 60 (2015) 2119-2127. [20] E.B. Rinker, A.T. Colussi, N.L. McKnight, O.C. Sandall, J. Chem. Eng. Data 45 (2000) 254-256.
Page 17
[21] J.M. Prausnitz, R.N. Lichtenthaler, E. Gomes de Azevedo, Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed.; Printice Hall: Englewood Cliffs, NJ, 1999. [22] S. Amad Kelayeh, A.H. Jalili, C. Ghotbi, M. Hosseini-Jenab, J. Chem. Eng. Data 56 (2011) 4317-4324. [23] M. Shokouhi, A.H. Jalili, A. Mohammadian, M. Hosseini-Jenab, S. Sadraei-Nouri, Thermochem. Acta 560 (2013) 63-70. [24] M. Shokouhi., A.H. Jalili, M. Vahidi, M. Hosseini-Jenab, J. Mol. Liq. 186 (2013) 142-146. [25] M.K. Park, O.C. Sandall, J. Chem. Eng. Data 46 (2001) 166-168. [26] M. Hosseini-Jenab, M. Abedinzadegan Abdi, S.-H. Najibi, M. Vahidi, N.-S. Matin, J. Chem. Eng. Data 50 (2005) 583-586. [27] NIST Scientific and Technical Databases, Thermophysical Properties of Fluid Systems. http://webbook.nist.gov/chemistry/fluid/ (accessed Sep. 2014). [28] D.P. Shoemaker, C.W. Garland, J.I. Steinfeld, J.W. Nibler, Experiments in Physical Chemistry, fourth ed., McGraw-Hill, New York, 1981.
[29] M. Shokouhi, A.T. Zoghi, M. Vahidi, M. Moshtari, J. Chem. Eng. Data 60 (2015) 1250-1258. [30] M. Shokouhi, H. Farahani, M. Hosseini-Jenab, Fluid Phase Equilib. 367 (2014) 2937. [31] M. Shokouhi, H. Farahani, M. Hosseini-Jenab, A.H. Jalili, J. Chem. Eng. Data 60 (2015) 499-508. [32] P. Daugé, A. Baylaucq, L. Marlin, C. Boned, J. Chem. Eng. Data 46 (2001) 823 – 830. [33] J.C. Chiehming, C. Kou-Lung, D. Chang-Yih, J. Supercrit. Fluids 12 (1998) 223– 237. [34] W. Weber, S. Zeck, H. Knapp, Fluid Phase Equilib. 18 (1984) 253–278. [35] K. Ohgaki, T. Katayama, J. Chem. Eng. Data 21 (1976) 53–55. [36] J. Xia, M. Jödecke, Á. Pérez-Salado Kamps, G. Maurer, J. Chem. Eng. Data 49 (2004) 1756–1759.
Page 18
[37] R. Sidi-Boumedine, S. Horstmann, K. Fischer, E. Provost, W. Fürst, J. Gmehling, Fluid Phase Equilib. 218 (2004) 149-155. [38] A. P. –S. Kamps, A. Balaban, M. Jödecke, G. Kuranov, N. A. Smirniva, G. Maurer, Ind. Eng. Chem. Res. 40, 2001, 696-706. [39] M. -H. Li, K. -P. Shen, J. Chem. Eng. Data 38 (1993) 105-108. [40] S. Xu, S. Qing, Z. Zhen, C. Zhang, Fluid Phase Equilib. 67 (1991) 197-201. [41] I. Kim, H. F. Svendsen, E. Brrensen, J. Chem. Eng. Data, 53 (2008) 2521-2531. [42] J. M. Bernal-Garcia, J. Chem. Eng. Data, 48, 2003, 1442-1445. [43] C. -H. Hsu, M. -H. Li, J. Chem. Eng. Data, 1997, 42, 502-507. [44] J. M. Bernal-Garcia, J. M. Bernal-Garcia, L. A. Galicia-Luna, K. R. Hall, M. Ramos-Estrada, G. A. Iglesias-Silva, J. Chem. Eng. Data, 49, 2004, 864-866. [45]T. T. Teng, Y. Maham, L. G. Hepler, A. E. Mather, J. Chem. Eng. Data 1994,39, 290-293 [46] M. -H. Li, Y. -C. Lie, J. Chem. Eng. Data, 39, 1994, 444-447. [47] H. A. Al-Ghawas, P. Hagewlesche, G. Rulz-Ibanez, O. C. Sandall, J. Chem. Eng. Data, 34, 1989, 385-391.
Figure Captions: FIGURE 1. Experimental setup for the simultaneous measurement of gas loading,
density and viscosity; V1–V6: valves; p1, p2: pressure transmitter sensors; EC: main liquid tank; SW: side glass window; GIS: gas injection system; VTD: Anton-Paar vibrating– tube densimeter; IM: interface module of densimeter; EU: evaluation unit of densimeter; CP: high-pressure centrifuge pump; FV: viscometer tube; S1, S2: magnetic sensors; TM: electronic time registration device; WRB: water recirculation bath; GR: constant volume gas container; WCP: water circulation path. FIGURE 2. Partial pressure – loading curves at different temperatures for 46.78 mass%
MDEA solution compared with data reported in literature at T=313.15 K, SidiBoumedine et al. [37] 46.78 mass% MDEA), Kamps at al. [38] (48.8 mass% MDEA) and Li et al. [39] (2.57 kmol.m-3 MDEA).
Page 19
FIGURE 3. Vapor pressures of 46.78 mass% MDEA solutions obtained from this work
compared with literature data (shokouhi et al. [9], Xu et al. [40], Kim et al. [41]) at different temperatures. FIGURE 4. Comparison of unloaded MDEA aqueous solutions density obtained from
this work with those reported data, Hsu et al. [43], Bernal-Garcia et al. [42], Shokouhi et al. [9], Rinker et al. [20], Al-Ghawas et al. [47] at the given temperatures. FIGURE 5. Comparison of unloaded MDEA aqueous solutions viscosity obtained from
this work with those reported data, Teng et al. [45], Bernal-Garcia et al. [44], Shokouhi et al. [9], Rinker et al. [20], Li et al. [46], Al-Ghawas et al. [47] at the given temperatures. FIGURE 6. Plot of measured density (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (lines) as a function of the loading (mol H2S/mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%. FIGURE 7. Plot of measured viscosity (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (lines) as a function of the loading (mol H2S /mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%.
Page 20
FIGURE 1. Experimental setup for the simultaneous measurement of gas loading,
density and viscosity; V1–V6: valves; p1, p2: pressure transmitter sensors; EC: main liquid tank; SW: side glass window; GIS: gas injection system; VTD: Anton-Paar vibrating– tube densimeter; IM: interface module of densimeter; EU: evaluation unit of densimeter; CP: high-pressure centrifuge pump; FV: viscometer tube; S1, S2: magnetic sensors; TM: electronic time registration device; WRB: water recirculation bath; GR: constant volume gas container; WCP: water circulation path.
Page 21
FIGURE 2. Total pressure – loading curves at different temperatures for 46.78 mass%
MDEA solution compared with data reported in literature at T=313.15 K, SidiBoumedine et al. [37] (46.78 mass% MDEA), Kamps at al. [38] (48.8 mass% MDEA) and Li et al. [39] (2.57 kmol.m-3 MDEA).
Page 22
FIGURE 3. Vapor pressures of 46.78 mass% MDEA solutions obtained from this work
compared with literature data (shokouhi et al. [9], Xu et al. [40], Kim et al. [41]) at different temperatures.
Page 23
FIGURE 4. Comparison of unloaded MDEA aqueous solutions density obtained from
this work with those reported data, Hsu et al. [43], Bernal-Garcia et al. [42], Shokouhi et al. [9], Rinker et al. [20], Al-Ghawas et al. [47] at the given temperatures.
Page 24
FIGURE 5. Comparison of unloaded MDEA aqueous solutions viscosity obtained from
this work with those reported data, Teng et al. [45], Bernal-Garcia et al. [44], Shokouhi et al. [9], Rinker et al. [20], Li et al. [46], Al-Ghawas et al. [47] at the given temperatures.
Page 25
FIGURE 6. Plot of measured density (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (solid lines) as a function of the loading (mol H2S/mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%.
Page 26
FIGURE 7. Plot of measured viscosity (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (lines) as a function of the loading (mol H2S /mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%.
Page 27
TABLE 1
Specifications and sources of chemicals used in this work. Chemical name
CAS registry number
Hydrogen sulfide
[7783-06-4]
N-Methyldiethanoleamine
[105-59-9]
Purity – analysis method 99.95 % (mol%) GC >99 % (mass%) GC
Source Roham Gas Company Sigma-Aldrich Company
TABLE 2
Total pressure, PT, partial pressure of H2S, PH2S , experimental solubility ( α = mol H2S /mol amine and xH2S = mole fraction), density, ρ , and viscosity, η , of (H2S – 46.78 mass% MDEA) system. (Or H2S – 7.377±0.003 mole MDEA/mass of water (kg)). T/K
α
± U (α )
xH2S ±U(x)
PT/MPa
PH2S /
ρ
MPa
(g/cm3 )
0.000 0.006 0.011 0.036 0.052 0.097 0.146 0.303 0.435 0.643 0.773 0.919
1.0314 1.0390 1.0435 1.0545 1.0605 1.0684 1.0721 1.0779 1.0794 1.0799 1.0800 1.0800
η / mPa.s ± U (η )
Pvp = 0.0063 MPa 313.15
0.000 0.080 0.155 0.349 0.465 0.648 0.695 0.866 0.936 1.019 1.057 1.099
0.007 0.008 0.009 0.010 0.011 0.012 0.014 0.015 0.016 0.018 0.020
0.0000 0.0093±0.0005 0.0178±0.0007 0.0393±0.0010 0.0517±0.0011 0.0706±0.0013 0.0754±0.0014 0.0923±0.0016 0.0989±0.0018 0.1067±0.0019 0.1103±0.0020 0.1142±0.0021
0.006 0.012 0.018 0.042 0.059 0.103 0.152 0.309 0.441 0.649 0.779 0.925
Page 28
4.23 4.20 4.06 3.76 3.71 3.65 3.59 3.55 3.54 3.46 3.41 3.36
0.16 0.16 0.16 0.05 0.16 0.08 0.16 0.16 0.16 0.12 0.16 0.12
Pvp =0.0137 MPa 328.15
0.000 0.080 0.154 0.346 0.461 0.641 0.685 0.850 0.919 0.997 1.038
0.004 0.006 0.009 0.011 0.013 0.015 0.018 0.019 0.020 0.021
0.0000 0.0092±0.0005 0.0178±0.0007 0.0391±0.0010 0.0513±0.0012 0.0699±0.0013 0.0744±0.0015 0.0907±0.0016 0.0973±0.0017 0.1047±0.0019 0.1086±0.0020
0.014 0.021 0.030 0.067 0.098 0.171 0.247 0.461 0.603 0.855 0.971
0.000 0.007 0.016 0.054 0.085 0.157 0.233 0.447 0.589 0.842 0.957
1.0216 1.0294 1.0341 1.0455 1.0515 1.0595 1.0638 1.0686 1.0699 1.0708 1.0708
2.82 2.81 2.78 2.71 2.65 2.60 2.54 2.51 2.43 2.41 2.40
0.14 0.09 0.10 0.10 0.11 0.15 0.14 0.07 0.14 0.14 0.14
Pvp =0.0274 MPa 343.15
0.000 0.079 0.153 0.343 0.455 0.630
0.004 0.006 0.009 0.010 0.011
0.0000 0.0092±0.0005 0.0176±0.0007 0.0386±0.0010 0.0507±0.0011 0.0689±0.0013
0.027 0.038 0.054 0.116 0.168 0.279
0.000 0.011 0.027 0.089 0.140 0.252
1.0097 1.0186 1.0230 1.0350 1.0405 1.0488
1.83 1.88 1.83 1.85 1.92 1.90
0.04 0.04 0.07 0.13 0.04 0.08
0.671 0.836 0.899 0.975
0.014 0.016 0.018 0.020
0.0730±0.0015 0.0893±0.0016 0.0954±0.0017 0.1026±0.0019
0.380 0.617 0.810 1.093
0.353 0.589 0.782 1.065
1.0527 1.0568 1.0583 1.0594
1.82 1.86 1.83 1.84
0.10 0.04 0.13 0.09
Expanded Uncertainties U at 95% confidence interval are U (T) = 0.02 K, U (PVP) = 0.0006 MPa, U (PT) = 0.0004 MPa, U ( ρ ) = 2.5×10-3 g.cm3, U (η ) obtained from standard deviation of 5 measurements and uncertainty of MDEA concentration in molality base, U (m) = 0.003.
TABLE 3
Comparison of density and viscosity values obtained from this work and some other published ones for 46.78 mass% MDEA aqueous solutions.
Page 29
Density (g.cm-3) Hsu and Bernal-Garcia Li [43] [42] 1.0309 1.0306 1.0206 1.0207 1.0097 1.0099
T/K
313.15 328.15 343.15
Viscosity (mPa.s) Bernal-Garcia Teng et al. [44] [45] 4.552 4.212
This work 1.0314 1.0216 1.0097
1.819
This work 4.228 2.82 1.834
TABLE 4
Percent rate variation of density and viscosity of loaded solution in comparison of fresh one at (α = 0.5 and 1.0) in this work and reported literature data. MDEA Solution (mass%)
T/K
∆α
∆ρ%
∆η%
46.78%
0 - 0.5 0 – 1.0988 0 - 0.5 0 – 1.0383 0 - 0.5 0 – 0.9748
This work 313.15 313.15 328.15 328.15 343.15 343.15
3.5 4.7 3.3 4.8 3.5 4.9
12.8 20.6 6.0 15.0 ~0 ~0
10% 20% 30% 40% 50% 50%
0 - 0.5 0 - 0.5 0 - 0.5 0 - 0.5 0 - 0.5 0 - 0.5
Ref. [20] 298.15 298.15 298.15 298.15 298.15 313.15
0.5 0.6 1.1 2.4 2.7 2.8
3.3 6.2 23.3 23.1 26.9 16.1
Page 30
TABLE 5 Numerical values of the parameters of the modified Setchenow equation for correlation of density and viscosity for the given solution.
46.78 mass% MDEA k 0 , 0, ρ
-0.01903
k 0 ,1,ρ
0.000293
k1,0 ,ρ
0.028774
k1,1,ρ
-0.00018
k 0 ,0 ,η
-3.66042
k 0 ,1,η
0.010809
k1, 0,η
1.419516
k 1,1,η
-0.00432
ARD% ( ρ ) ( ARD % ) exp (ρ ) MRD% ( ρ ) (MRD % ) exp (ρ ) ARD% (η ) ( ARD % ) exp (η ) MRD% (η ) (MRD %) exp (η )
0.06% 0.23% 0.25% 0.26% 1.48% 4.17 % 4.93% 7.06%
Page 31
In the name of god Highlights Measurement solubility of H2S in 46.78 mass% MDEA aqueous solutions. Measurement density of H2S loaded of MDEA aqueous solution. Measurement viscosity of H2S loaded of MDEA aqueous solution. Correlation of the density and viscosity of H2S loaded of MDEA aqueous solution using modified setchenow equation.
S0021-9614(16)30109-4 http://dx.doi.org/10.1016/j.jct.2016.06.007 YJCHT 4673
To appear in:
J. Chem. Thermodynamics
Received Date: Revised Date: Accepted Date:
2 August 2015 30 April 2016 13 June 2016
Please cite this article as: M. Shokouhi, R. Ahmadi, Measuring the Density and Viscosity of H2S-Loaded Aqueous Methyldiethanolamine Solution, J. Chem. Thermodynamics (2016), doi: http://dx.doi.org/10.1016/j.jct.2016.06.007
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Measuring the Density and Viscosity of H2S-Loaded Aqueous Methyldiethanolamine Solution
Mohammad Shokouhi,* Reza Ahmadi
Research Institute of Petroleum Industry (RIPI), National Iranian Oil Company (NIOC). Gas Research Division, Research Institute of Petroleum Industry (RIPI), P.O. Box 14665–137, Tehran, Iran.
* Corresponding author. E–mail: [email protected] ; [email protected] Tel: 98 21-48252466.
Page 1
ABSTRACT The density and viscosity of H2S-loaded aqueous 46.78 mass% methyldiethanolamine solution were experimentally measured accompanied with the solubility of H2S at temperatures (313.15, 328.15 and 343.15) K, pressures from vapor pressure of fresh solution up to 1.0 MPa and loadings up to 1.00 mol of H2S per 1 mol of amine. All experimental trials have been carried out using the new setup developed in our laboratory. It was observed that both density and viscosity of mixtures decrease by increasing temperature and density increase by increasing acid gas solubility (loading) by about 4.7%, whereas viscosity has a complicated behavior with H2S solubility. Viscosity decreases by increasing acid gas solubility (loading) at 313.15 K by about 20.6% and at 328.15 K by about 15.0%, but it is comparable at 343.15 K in terms of H2S solubility. Finally, the experimental density and viscosity data correlated using Modified Setchenow equation.
Keywords: Gas solubility, Chemical absorption, Transport property, Methyldiethanolamine, Density, Viscosity.
Page 2
1. Introduction Knowledge on the thermal and non-thermal properties of gas-loaded physical and chemical solvents such as density, viscosity, surface tension, enthalpy, heat capacity, thermal conductivity and thermal diffusivity are necessary for 1) the design of acid gas treatment equipment and subsequent measuring operation 2) the solvent regeneration at reduction stage 3) predicting and interpreting the laboratory data such as kinetic rates constants [1], diffusion coefficients [2] and mass-transfer characteristics of gas/liquid contacting devices. An enormously growing number of experimental results for thermophysical properties of pure solvents as well as those of their mixtures with other solvents have been reported in the literature. However, data regarding acid gas adsorbed in physical and chemical solvents are scarce. Jalili, et al. [3], have reported density and viscosity of sulfolane loaded by CO2 and H2S and sih et al. [4] have reported those properties of CO2-methanol system. Both reported data [3,4] have shown that density and viscosity of sulfolane and methanol loaded with CO2 reduce with increasing loading. Density and viscosity
of
methyldiethanolamine
(MDEA),
diethanolamine
(DEA),
monoethanolamine (MEA) and their mixture loaded with CO2 have recently been reported by Weiland et al. [5]. It is worth noting that those data reported by Weiland et al. [5] were prepared using quenching and separating method in which solutions were partially carbonated. Dell,Era et al. [6] reported the density of CO2 loaded solution of Diisopropanolamine (DIPA). Jayarathna et al. [7,8], measured the density and surface tension of CO2 loaded aqueous MEA solution at (303.15 to 343.15) K and loading from (0 to 0.5). Density and viscosity of aqueous MDEA solution (45 mass %), as well as its aqueous mixtures with 2-Amino-2-methyl-1-Propanol (AMP) (40 mass% MDEA + 5 mass% AMP) and DIPA (40 mass% MDEA + 5 mass% DIPA) under the CO2 loading at temperatures from (303.15 to 363.15) K and pressures up to 2.0 MPa were experimentally measured by our research group [9]. Freeman et al. [10] have measured the density and viscosity of aqueous Piperizine + CO2 solutions at temperature range from (293.15 to 333.15) K and the concentration ranges from (2 to 20) molal Piperazine and CO2 loading from (0 to 0.4). Some other documented literature data for carbonated alkanolamine solutions have been reported including MEA, DEA, MDEA, AMP, 2-dimethylaminoethanol
Page 3
(DMAE), 2-diethylaminoethanol (DEAE) and 1-dimethylamino-2-propanol (1DMA2P) [11,12,13,14,15,16,17,18]. All reported data show that density, viscosity and surface tension measurements of carbonated aqueous alkanolamine solutions increase with increasing CO2 loading and decrease with temperature. A literature survey on H2S solubility in MDEA and mixed MDEA with the other alkanolamines aqueous solution has been done by Shokouhi et al. [19]. None of the experimental works collected in that reference measured the density and viscosity of H2S-loaded solution. Making further efforts show that, except for Rinker et al. [20], there isn’t any experimental data on density and viscosity of H2S loaded alkanolamine solution. In Rinker et al. [20], measurements at 25 °C were carried out for (10-50 mass %) MDEA aqueous solution at loadings 0.1, 0.3, 0.4 and 0.5 (mol of H2S/mol of MDEA) and measurements at 40 °C were made for 50 mass % MDEA aqueous solutions at H2S loadings of 0.1, 0.3, 0.4 and 0.5 mol of H2S/mol of MDEA. They have shown that, the density of aqueous MDEA increases with increasing H2S loading while the viscosity decreases with increasing H2S loading which this result is completely opposite of carbonated alkanolamine solutions. In this work, we focus on the measurement of density and viscosity of 46.78 mass% methyldiethanolamine (MDEA) loaded with H2 S using a setup developed in our laboratory without any quenching procedure. All experimental trials are carried out on stand situation at temperatures 313.15, 328.15 and 343.15 K, and pressure from vapor pressure of uncharged solution up to about 1.00 MPa and loadings from zero up to about 1.00 mol of H2S per 1 mol of amine. Gas loading was systematically measured by gravimetric method. Finally, the experimental density and viscosity data correlated using Modified Setchenow equation [21,22,23,24]. 2.
Experimental Section
2.1.
Materials:
Hydrogen sulfide [c.p. grade 99.95 % min and CAS registry number 7783-06-4] was obtained from Roham Gas Company. N-Methyldiethanoleamine [CAS registry number 105-59-9] was purchased from Sigma-Aldrich with purity >99% and used as purchased without further purification. Water used as solvent was distilled deionized, which was degassed in an ultrasonic bath (FUNGILAB, model UA10MFD) at temperature 343 K
Page 4
and wave frequency 50 kHz for about 1 hour prior to use. All solutions were prepared by measurement of mass of the solutes and solvent on a calibrated balance (Mettler model AE 200) with an uncertainty ± 0.0001 g and measuring the volume of the solution by a standard volumetric flask up to 250 mL. The specifications of the chemicals used in this work contain source, purification method, purity base (mass or mole) are summarized in table 1. 2.2. Apparatus and procedure: The details of the experimental method for the measurement of gas solubility, density and viscosity have previously been presented [3] and only a short description will be provided here. The experimental setup as having shown in figure 1, comprised of four separate apparatus, each one considered to perform a specific task: (a) the main equilibrium cell (EC), (b) a gas injection system (GIS), (c) a vibrating tube densimeter (VTD), and (d) a falling weight viscometer (FV). The EC was equipped with a magnetically driven mechanical stirrer to accelerate equilibrium state and a side glass window (SW) to allow one to monitor both the liquid level and the foaming of the solution in the EC. The EC was double wall jacketed to facilitate the circulation of water from water recirculation bath (WCB) to control the temperature of the solution in the cell. The temperature of EC was monitored using a Lutron model TM-917 digital thermometer with a 0.01 K resolution using a PT-100 sensor inserted into the thermowell of the EC (not shown in figure 1). The pressure of the system was measured using a Keller model PA-33X pressure transmitter sensor (P2) in the range of (0 to 4) MPa with an uncertainty within ±0.01% of full scale. The gas injection system (GIS) consists of a gas container (GR) with constant known volume equipped with a Keller model PA-33X pressure transmitter sensor (P1) in the range of (0 to 4) MPa with an uncertainty within ±0.01% of full scale. The GIS, which is connected to the EC through V3 valve, is directly fed from gas cylinders through V1 valve. To measure the solubility of gases in the liquid solvent, known quantities of gaseous solute and degassed solvent are brought into contact at a constant temperature inside the measurement system (EC + VTD + FV) with known volume. On reaching thermodynamic equilibrium, the pressure above the liquid solution, which is measured
Page 5
by the P2 pressure transmitter, is constant and directly related to the solubility of the gas in the liquid. The quantity of solute present in the liquid solution is calculated by the difference between two PVT measurements: first on introduction of the gas from the container GR of known volume into the measurement system (EC + VTD + FV) containing the solvent and secondly after reaching thermodynamic equilibrium, i.e. when autoclave pressure, PT , remains fixed and no longer changes with time. Using the procedure adopted by Park and Sandall [25] and Hosseini-Jenab et al. [26]:
nag =
Vgc Pi Pf − RTa Z i Z f
(1)
where Vgc denotes the volume of the gas container, Zi and Zf are the compressibility factors corresponding to the initial and final pressures, Pi and Pf, respectively, in the gas container before and after transferring gas, and Ta is the ambient temperature, which is equal to that in the gas container. Compressibility factors were calculated using NIST [27]. Equilibration between liquid and vapor phases inside the cell were normally achieved within about 2 h after beginning of stirring and the partial pressure of gas at equilibrium state in the equilibrium cell, Page , was calculated as follow:
Page = PT − PVP
(2)
where PT and PVP denote the total pressure and vapor pressure of solution. A key issue is the determination of the vapor pressure of the mixture solution since this value must be subtracted from the total pressure to obtain the partial acid gas pressure. The moles of remaining acid gas in the gas phase, nagg , was determined from: g ag
n =
Vg Page Z ag RT
= Vg .ρ ag
(3)
where Vg is the gas phase volume, T is the equilibrium temperature of the cell, and Zag and ρ ag are the compressibility factor and density of acid gas at Page and T, respectively. The quantity of gas in the liquid phase was then determined from: nagl = nag − nagg
(4)
The molality and loading of charged gas in the liquid phase is defined as:
Page 6
m ag =
α ag
nagl ( mol) wSolvent (g)
(5)
× 1000
nagl (mol) = w (g) ∑i Mi (g) i
(6)
wi is the mass of comprised amines in fresh solvent in g, and Mi is molar mass of pure
comprised amines in g.mol-1. The isochoric saturation solubility method may be addressed some drawbacks such as; a) it serves static situation which don’t never fulfill dynamic situation in absorption/stripping towers b) The amounts of acid gas which are present in the liquid phase should be approximated using in the vapour phase. In this method, it has been assumed that the partial pressure of acid gas in the vapour phase is the difference between their experimental results for the total pressure (above the charged liquid mixture) and the saturation pressure of the uncharged solvent c) There is a more serious problem in which by dissolving acid gas in the solvent, the volume of the liquid phase remains unchanged, i.e., this method neglects any volume expansion or compression caused by acid gas in the liquid phase. Our new developed set-up has paved a rout to tackle the third drawback [9]. As mentioned above the amount of dissolved gas is calculated from the amount of gas charged into the cell by subtracting a correction for the amount of gas that is still present in the vapor phase. That correction is calculated assuming that the volume of the vapor phase is the difference between the cell volume and the volume of the unloaded solvent (which is known from the preparation of the fresh uncharged solution). However, that procedure neglects a volume change (mostly a volume expansion) when a gas is dissolved in a liquid. To improve further our measurement processes, equation (3) may be corrected as (7). mag ⋅ wSolvent + wSolvent 1000 g = ρ ag . Vauto − nag ρ
(7)
mag is the molality of acid gas obtained from iteration technique or as an approximate estimation obtained from molality calculated with equations (3) through (6). ρ is the
Page 7
charged solution density measured on stand (without quenching method) using procedure described in following section. Vauto is the volume of auto clave which is equal to volume of EC + VTD + FV. The error propagation theory was used to estimate the uncertainties of final results [28]. In the base of this theory, the uncertainty δq of the interest variable q(r … u) is given by equation (8): 2
∂q ∂q U (q ) = ± .dr + ... + .du ∂r ∂u
2
(8)
The measured quantity q is dependent upon the variables r… u which fluctuates in a random and independent manner (dr… du). The uncertainties of all parts of the instruments used in the measurements were considered for estimating the uncertainty of the H2S loading in the liquid phase. The main contributions to the uncertainty of the solubility are attributed to errors in the pressure sensor for equilibrium cell and gas container (both are equal to ±4×10-4 MPa), temperature sensors (±0.2 K), and scale for the amount of solvent in equilibrium cell (±0.0001 g). The volumes of the different parts of the set up were obtained by performing pressure swing experiments [9,19, 29,30,31]. Several measurements using this method were done to obtain the volumes and uncertainties of the volumes of different parts of the setup. According to equations (6) and (8), the gas solubility is related to the amount of absorbed gas in the liquid phase, and that of solvent. The average calculated uncertainties of these two parameters at T=328.15 K were calculated to be U (nHl 2S ) = ±0.003 mole and U (wSolvent ) = ±0.0001 g, therefore the average uncertainty of loading, ± U (α ) , is equal to ±0.006.
2.2.1. Density measurement
Measurement of the density of fresh and charged liquid solutions prepared in equilibrium cell, EC, at each specified temperature, T and pressure, p was carried out by the vibrating tube densimeter (VTD). The VTD was an Anton-Paar model DMA HPM, equipped with a liquid jacket by the supplier, the temperature of which was thermostatted through circulation of water from the external water recirculation bath, WCP. The DMA HPM density measuring cell is a U-shaped Hastelloy C-276 tube
Page 8
electronically excited to vibrate at its characteristic frequency, which in turn depends on the density of the sample. The DMA HPM is connected to the interface module (IM), which generates and measures the period of oscillation and is also responsible for the temperature measurement. The IM is in turn connected to an mPDS 2000 V3 evaluation unit (EU), which converts the measured raw data (oscillation period, temperature and pressure) to the density of the sample through the appropriate calibration equations. The VTD density measuring device was calibrated by using distilled deionized water, pure methanol and pure sulfolane in the temperature range (298.15 to 363.15) K and from atmospheric pressures up to about 7.0 MPa employing equation (9) as suggested by the supplier:
ρ=
∑a
ijk
(9)
⋅τ i ⋅ T j ⋅ P k
i, j ,k
where aijk, τ , T, and P stand for apparatus constant, period of oscillation, temperature, and absolute pressure, respectively, with i = 0, 2, 4, j = 0, 1, 2, and k = 0, 1, 2. Uncertainties of density U (ρ ) using equation (9) obtained about U( ρ )= 2.5×10-3g.cm-3.
2.2.2. Viscosity measurement
The viscosity of pure as well as gas saturated liquid solutions was measured by the falling weight viscometer (FV). FV, which was a cylindrical steel tube with 85 cm length and 6.8 mm inner diameter, was equipped with a liquid jacket to be thermostatted by an external water recirculation bath (WRB). The FV was mounted vertically and consisted of a cylindrical sinker. The sinker with 6.4 mm outer diameter and 20.4 mm length, and hemispherical ends was an iron magnetic bar coated with Teflon PTFE to prevent corrosion and its density was obtained 2880.3 kg · m-3. A working equation suggested by Daugé et al. [32] was used to measure the viscosity of solutions.
η (T , p ) = K a (T ) + K b (T ) ⋅ (ρ s − ρ ) ⋅ t
(10)
In this equation, the viscosity of unknown solution is considered to be linearly dependent on the difference between the density of the sinker, ρs , and that of the solution, ρ . t is the time for the sinker descending between two sensors with a specified and adjustable distance. The Ka(T) and Kb(T) parameters were determined
Page 9
through calibration of the viscometer by liquids with known viscosity at a specified temperature. Distilled demonized water, pure sulfolane and pure methanol were used as the reference fluids for calibration of the apparatus at various temperatures and pressures. The digital timer is capable of registering the falling time of the sinker with a minimum reproducibility ± 0.03 s. After the introduction of the known amounts of liquid and gas into the EC and adjustment the temperature of the system, the centrifuge pump was turned on to circulate the liquid thoroughly in the system. This operation helps achieve two important goals during measurement: (1) increasing the rate of equilibrium approach through mixing and homogenizing the liquids and more importantly (2) starting the measurement of the viscosity of the mixtures whereby lifts the sinker up the vertical cylinder without the need for any additional external/internal apparatus. By closing V4 valve, the drag force ceases and the sinker falls through the liquid down from the top to the base of the vertical tube. By passing through the two S1 and S2 sensors, the falling time is registered and the viscosity is calculated by using Eq. (10). The speed of the centrifuge pump was adjusted in such a way that a laminar flow regime of the liquid always prevailed in the FV vertical cylinder. A laminar flow ensured a uniform liquid phase without any bubble formation during falling of the sinker in the cylinder. 3. Results and discussion As reported in our previous paper [3] the accuracy and reliability of the experimental data produced by the apparatus developed in this work was checked through comparison of the measured solubility, density and viscosity of pure sulfolane, pure methanol and the (CO2 - methanol) binary mixture with the available experimental data reported in the literature [4,33]. The Sih et al. [4] and Chiehming et al. [33] methods are in principle similar to the method used in this work. The maximum per cent deviation of the density values reported by Sih et al. [4] from this work is less than 0.50%. The maximum per cent deviation of the viscosity values reported by Sih et al. [4] from this work is less than 1.0 % with average deviation of 0.39 %. The two sets of solubility measurements are also in quite good agreement with each other. This is also the case for the solubility and density data from (298 to 313) K reported by Chiehming et al. [33].
Page 10
The PTx solubility data obtained using our setup for (CO2 - methanol) binary mixture are also consistent with the corresponding values reported by Weber et al. [34] at 298 K (AAD% = 3.7% and MAD% = 6.0%), with those reported by Ohgaki and Katayama [35] at T = (298 and 313) K (AAD% = 4.8% and MAD% = 6.8%), and also with those reported by Xia et al. [36] at 313.75 K (AAD% = 7.2% and MAD% = 8.4%), in which, AAD% and MAD% stand for the average of absolute deviations and maximum of
absolute deviations of mole fraction, density or viscosity. Experimental solubility of H2S in 46.78 mass% aqueous MDEA solution and vapor pressure of gas free solution as well as fresh and loaded solution density and viscosity accompany with their uncertainties have been reported in table 2. Temperature and pressure dependency of H2S solubility graphically has been shown in figure 2. Data from this work (at T = 313.15 K) has been compared with those obtained using the computer-operated apparatus reported by Sidi-Boumedine et al. [37] in 46.78 mass% aqueous MDEA and also data reported by Kamps et al. [38] in 48.8 mass% aqueous MDEA and Li et al. [39] as well. Vapor pressure of gas free MDEA solution obtained from this work graphically compared with those reported by Xu et al. [40], kim et al. [41] and Shokouhi et al. [9] in figure 3. One of the main assumptions in equilibrium state was given with equation (2) in which vapor pressure of solution was kept constant with the acid gas loading. To confirm the validity of equation (2), vapor pressures of H2S-loaded solution at different temperatures and loading were tried with values given by Results law. All loading were recalculated using this more sensible assumption and results show that all recalculated loading values are the same as those reported in table 2 with three digits after the point. Comparisons of density and viscosity of unloaded 46.78 mass% MDEA aqueous solutions with documented published data has been reported in table 3. The average deviation of density reported by Bernal-Garcia [42] and Hsu-Li [43] from this work is less than 0.3%, in which both reported their experimental data as Redlich-Kister type equation. The average deviation of viscosity reported in this work from those data obtained from correlation in Bernal-Garcia [44]is less than 7.0%, where as that is less than 0.8% in comparison with data reported by Teng et al. [45]. Deviation of density and viscosity of unloaded MDEA solution obtained from this study were schematically compared to other reported ones [42-45,46,47] in figures 4 and 5, respectively.
Page 11
Dependence of H2S solubility on density and viscosity of H2S-loaded solution has been shown in figures 6 and 7. As may be seen from figures 6 and 7, both density and viscosity of mixtures decrease by increasing temperature, and density increases by increasing acid gas solubility (loading) by about 4.7%, whereas viscosity has complicated behavior with H2S solubility. Rinker et al. [20] have measured the variation of density and viscosity of aqueous MDEA with addition of hydrogen sulfide. Their measurements were carried out at 298.15 K for loading up to 0.5 mol H2S per 1 mol of amine for amine concentrations up to 50 mass % and also at 313.15 K for amine concentration of 50 mass % and for H2S loading up to 0.5. They have found that, the density of aqueous MDEA increases with increasing H2S loading while the viscosity decreases with increasing H2S loading, and also MDEA concentration in unloaded solution has increasing effect on density and viscosity of solution. In table 4, values of percent rate variation for density, ∆ρ%, and viscosity, ∆η%, of loaded solution in comparison of unloaded one at two loading values (α = 0.5 and 1.0) have been reported and compared with those values obtained from Rinker et al. [20]. From table 4, both sets of data (this work and Rinker et al.) show that, percent rate increase of density, ∆ρ%, is roughly independent of temperature, and also according to Rinker et al. it depends on MDEA concentration in unloaded solution. There is an agreement with the measurements of density reported by Rinker et al. and this work. Rinker et al. [20] found that for the case of (50 mass % MDEA solution, T = 313.15 K, α = 0.5) the density increases about 2.8%, and in this work (46.78 mass % MDEA
solution, T = 313.15 K and α = 0.5) the density increases about 3.2%, while this value for α = 1.0383 mol H2S/mol MDEA is about 4.5%. For the case of viscosity, slop of viscosity decrease for Rinker et al. is higher than that reported in this work, for instance, Rinker et al. found that the viscosity decreases about 16.1% (50 mass % MDEA, T = 313.15 K and α = 0.5), whereas in this work (46.78 mass % MDEA and T = 313.15 K) viscosity decreases about 12.8% for the H2S loading 0.5, and about 20.6 % for H2S loading 1.0988 mol H2S/mol MDEA.
4. Density and Viscosity modeling:
Page 12
The experimental density and viscosity of solution are correlated using modified Setchenow type equation which is recently used for correlating of density, viscosity, surface tension, heat capacity, thermal conductivity and thermal diffusivity of binary and ternary solutions [21,22,23,24]. The working equation is represented by equation (11): 2 p ln = ∑ k j .α p r j =1
(11)
j
where p and pr stand for the physical property of the mixture and reference substance, respectively; for example p represents the density / viscosity of the (H2S - alkanolamine solution) mixture and pr is the density / viscosity of uncharged alkanoleamine solution at the same temperature as the mixture. The kj’s are temperature dependent parameters, which are known as the Setchenow coefficients, and α is the concentration of H2S dissolved in the liquid phase at the specified temperature and pressure. The final working equation for correlation of density and viscosity are; ρ ln ρr
= (k 0,0, ρ + k 0,1, ρ ⋅ T ). ⋅ α + (k1,0, ρ + k 1,1, ρ ⋅ T ) ⋅ α 2
η ln ηr
= (k 0,0,η + k 0,1,µ ⋅ T ). ⋅ α + (k1,0,η + k 1,1,η ⋅ T ) ⋅ α 2
(12)
(13)
The optimized values of adjustable parameters, k i , j ,ρ s and k i , j ,η s, were obtained using minimizing the objecting function expressed as sum of the relative deviation values and all adjustable numerical values are tabulated in table 5. The average of relative deviations, ARD%, defined by equation (14) and maximum of relative deviations, MRD%, defined by equation (15) for a number of N experimental points which both
may be regarded as the quality of correlation. cal exp 100 N pi − pi ARD % = ∑ P exp N i =1 i
(14)
pical − piexp MRD% = Maximum ⋅ 100 exp pi
Page 13
(15)
p ical and piexp in equations (14) and (15) stand for the calculated and experimental value
of the physical property of interest, respectively. The average and maximum per cent deviations of the correlated from experimental values together with the corresponding experimental uncertainties (ARD%)exp and (MRD%)exp, defined in equations (16) and (17), are given in table 5.
(
exp 100 N U pi ( ARD %) exp = ∑ exp N i =1 pi
)
(
(16)
)
U piexp (MRD%)exp = Maximum exp ⋅ 100 pi
(17)
In which U ( piexp )is the uncertainty of physical property which was obtained using 4 or 5 times measuring the viscosity and for the case of density, it was obtained using experimental values of pure methanol, water and sulfolane in the temperature range (298.15 to 363.15) K and atmospheric pressures up to about 7.0 MPa employing equation (9). It can be observed from table 5, ARD% and MRD% values of loading and viscosity are smaller than (ARD%)exp and (MRD%)exp, i.e., the correlation represents the density and viscosity values within experimental uncertainty. It means that, modified Setchenow equation is able to describe the temperature and loading dependency of the density and viscosity of the system studied in this work.
5. Conclusion
The physical equilibrium and transport properties, i.e. H2S loaded density and viscosity of MDEA solutions which are in use in the natural gas sweetening processes, were measured in this work. The experimental setup developed in our laboratory is able to measure simultaneously the density and viscosity with the solubility of acid gases in both physical and chemical solvents. The reliability of the experimental values generated by the setup was assessed through comparison of the measured solubility, density and viscosity of pure sulfolane, pure methanol and the (CO2 + methanol) binary mixture with the available experimental data reported in the literature [3]. All measurements at 313.15, 328.15 and 343.15 K were carried out for H2S loadings from zero to about 1.0 (mol H2S / mol MDEA) in 46.78 mass% MDEA aqueous solution.
Page 14
These data have been shown in figures 5 and 6, and also listed in table 2. As may be seen from those figures and table, the density of aqueous MDEA increases with increasing H2S loading by about 4.7%, while the viscosity decreases with increasing H2S loading at 313.15 K by about 20.6% and at 328.15 K by about 15.0%, but it is roughly unchanged with loading at 343.15 K.
NOMENCLATURE
NIST;
National Institute of Standards and Technology
R; Universal gas constant MDEA; N-Methyldiethanolamine EC;
The main equilibrium cell
GIS;
Gas injection system
VTD; Vibrating tube densimeter FV;
Falling weight viscometer
GR;
Gas container
Vgc ;
Volume of the gas container (GR)
Vg;
Gas-phase volume in the equilibrium cell
ρ ag ; Density of acid gas in equilibrium situation in gas phase Zi and Zf; Compressibility factors of the initial and final state in the gas container T a;
Ambient temperature
P0 ;
Initial pressure of solution
PT ;
Total absolute pressure
PVP ; Vapor pressure of solution which is equal to P0 PH 2S or PHe2S ;
Partial pressure of H2S at equilibrium state
Vauto ; Volume of auto clave wi ; Mass of pure species i
Mi ; Molar mass of pure species i mag ; Molality of acid gas
ρ charged solution ; Charged solution density p;
Physical property of mixture
pr;
Physical property of reference system
Page 15
nHg 2S ; Amount of H2S in the gas phase at equilibrium state nHl 2S ; Amount of H2S in the liquid phase at equilibrium state n Amine ; Amount of total amine in the liquid phase
( )
U nil ; Uncertainty of amount of species i in liquid phase.
α CO ; Loading of H2S in solution 2
U(α); Uncertainty of loading U(m); Uncertainty of molality of MDEA in aqueous solution AAD; Average of absolute deviations MAD; maximum of absolute deviations ARD; Average of relative deviations MRD; Maximum relative deviation
ρ s ; Density of sinker (g/cm3)
ρ ; Density of solution (g/cm3) aijk; Apparatus (Anton-Paar) constant
τ ; Period of oscillation Ka(T), Kb(T); Viscometer calibration constants t; Time in second
η ; Viscosity of solution η r ; Viscosity of reference system kj; Temperature dependent parameters of Setchnow equation ∆ρ%; Percent rate variation of solution’s density ∆η%; Percent rate variation of solution’s viscosity
ACKNOWLEDGEMENT
We are thankful to the Research Council of the Research Institute of Petroleum Industry (RIPI) and the Research and Development of the National Iranian Oil Company (NIOC) for their support of this work.
REFERENCES
Page 16
[1] E.B. Rinker, S.S. Ashour, O.C. Sandall, Chem. Eng. Sci. 50 (1995) 755-768. [2] A. Tamini, E.B. Rinker, O.C. Sandall, J. Chem. Eng. Data 39 (1994) 330-332. [3] A.H. Jalili, M. Shokouhi, F. Samani, M. Hosseini-Jenab, J. Chem. Thermodyn. 85 (2015) 13-25. [4] R. Sih, F. Dehghani, N.R. Foster, J. Supercrit. Fluids 41 (2007) 148–157. [5] R.H. Weiland, J.C. Dingman, D.B. Cronin, G.J. Browning, J. Chem. Eng. Data 43 (1998) 378-382. [6] C. Dell,Era, P. Uusi-Kyyny, J.–P. Pokki, M. Pakkanen, V. Alopaeus, Fluid Phase Equilb. 293 (2010) 101-109. [7] S.A. Jayarathna, A. Weerasooriya, S. Dayarathna, D.A. Eimer, J. Chem. Eng. Data 58 (2013) 986-992. [8] S.A. Jayarathna, C.K. Jayarathna, D.A. Kottage, S. Dayarathna, D.A. Eimer, M.C. Melaaen, J. Chem. Eng. Data 58 (2013) 343-348. [9] M. Shokouhi, A.H. Jalili, F. Samani, M. Hosseini-Jenab, Fluid Phase Equilibria 404 (2015) 96–108 [10] S.A. Freeman, G.T. Rochelle, J. Chem. Eng. Data 56 (2011) 574-581. [11] T.G. Amundsen, L.E. Øi, D.A. Eimer, J. Chem. Eng. Data 54 (2009) 3096−3100. [12] D. Fu, L. Chen, L. Qin, Fluid Phase Equilib. 319 (2012) 42−47. [13] D. Fu, Z. Li, F. Liu, J. Chem. Thermodyn. 68 (2014) 20−24. [14] A. Hartono, O.E. Mba, H.F. Svendsen, J. Chem. Eng. Data 59 (2014) 1808−1816. [15] J. Han, J. Jin, D.A. Eimer, M.C. Melaaen, J. Chem. Eng. Data 57 (2012) 1843−1850. [16] D.D.D. Pinto, J.G.M.S. Monteiro, B. Johnsen, H.F. Svendsen, H. Knuutila, Int. J. Greenhouse Gas Control 25 (2014) 173−185. [17] J. Zhang, P.S. Fennell, J.P.M. Trusler, J. Chem. Eng. Data In Press [18] Z. Idris, L. Ang, D.A. Eimer, J. Ying, J. Chem. Eng. Data 60(5) (2015) 1419-1425. [19] M. Shokouhi, H. Bozorgzade, P. Sattari, J. Chem. Eng. Data 60 (2015) 2119-2127. [20] E.B. Rinker, A.T. Colussi, N.L. McKnight, O.C. Sandall, J. Chem. Eng. Data 45 (2000) 254-256.
Page 17
[21] J.M. Prausnitz, R.N. Lichtenthaler, E. Gomes de Azevedo, Molecular Thermodynamics of Fluid-Phase Equilibria, 3rd ed.; Printice Hall: Englewood Cliffs, NJ, 1999. [22] S. Amad Kelayeh, A.H. Jalili, C. Ghotbi, M. Hosseini-Jenab, J. Chem. Eng. Data 56 (2011) 4317-4324. [23] M. Shokouhi, A.H. Jalili, A. Mohammadian, M. Hosseini-Jenab, S. Sadraei-Nouri, Thermochem. Acta 560 (2013) 63-70. [24] M. Shokouhi., A.H. Jalili, M. Vahidi, M. Hosseini-Jenab, J. Mol. Liq. 186 (2013) 142-146. [25] M.K. Park, O.C. Sandall, J. Chem. Eng. Data 46 (2001) 166-168. [26] M. Hosseini-Jenab, M. Abedinzadegan Abdi, S.-H. Najibi, M. Vahidi, N.-S. Matin, J. Chem. Eng. Data 50 (2005) 583-586. [27] NIST Scientific and Technical Databases, Thermophysical Properties of Fluid Systems. http://webbook.nist.gov/chemistry/fluid/ (accessed Sep. 2014). [28] D.P. Shoemaker, C.W. Garland, J.I. Steinfeld, J.W. Nibler, Experiments in Physical Chemistry, fourth ed., McGraw-Hill, New York, 1981.
[29] M. Shokouhi, A.T. Zoghi, M. Vahidi, M. Moshtari, J. Chem. Eng. Data 60 (2015) 1250-1258. [30] M. Shokouhi, H. Farahani, M. Hosseini-Jenab, Fluid Phase Equilib. 367 (2014) 2937. [31] M. Shokouhi, H. Farahani, M. Hosseini-Jenab, A.H. Jalili, J. Chem. Eng. Data 60 (2015) 499-508. [32] P. Daugé, A. Baylaucq, L. Marlin, C. Boned, J. Chem. Eng. Data 46 (2001) 823 – 830. [33] J.C. Chiehming, C. Kou-Lung, D. Chang-Yih, J. Supercrit. Fluids 12 (1998) 223– 237. [34] W. Weber, S. Zeck, H. Knapp, Fluid Phase Equilib. 18 (1984) 253–278. [35] K. Ohgaki, T. Katayama, J. Chem. Eng. Data 21 (1976) 53–55. [36] J. Xia, M. Jödecke, Á. Pérez-Salado Kamps, G. Maurer, J. Chem. Eng. Data 49 (2004) 1756–1759.
Page 18
[37] R. Sidi-Boumedine, S. Horstmann, K. Fischer, E. Provost, W. Fürst, J. Gmehling, Fluid Phase Equilib. 218 (2004) 149-155. [38] A. P. –S. Kamps, A. Balaban, M. Jödecke, G. Kuranov, N. A. Smirniva, G. Maurer, Ind. Eng. Chem. Res. 40, 2001, 696-706. [39] M. -H. Li, K. -P. Shen, J. Chem. Eng. Data 38 (1993) 105-108. [40] S. Xu, S. Qing, Z. Zhen, C. Zhang, Fluid Phase Equilib. 67 (1991) 197-201. [41] I. Kim, H. F. Svendsen, E. Brrensen, J. Chem. Eng. Data, 53 (2008) 2521-2531. [42] J. M. Bernal-Garcia, J. Chem. Eng. Data, 48, 2003, 1442-1445. [43] C. -H. Hsu, M. -H. Li, J. Chem. Eng. Data, 1997, 42, 502-507. [44] J. M. Bernal-Garcia, J. M. Bernal-Garcia, L. A. Galicia-Luna, K. R. Hall, M. Ramos-Estrada, G. A. Iglesias-Silva, J. Chem. Eng. Data, 49, 2004, 864-866. [45]T. T. Teng, Y. Maham, L. G. Hepler, A. E. Mather, J. Chem. Eng. Data 1994,39, 290-293 [46] M. -H. Li, Y. -C. Lie, J. Chem. Eng. Data, 39, 1994, 444-447. [47] H. A. Al-Ghawas, P. Hagewlesche, G. Rulz-Ibanez, O. C. Sandall, J. Chem. Eng. Data, 34, 1989, 385-391.
Figure Captions: FIGURE 1. Experimental setup for the simultaneous measurement of gas loading,
density and viscosity; V1–V6: valves; p1, p2: pressure transmitter sensors; EC: main liquid tank; SW: side glass window; GIS: gas injection system; VTD: Anton-Paar vibrating– tube densimeter; IM: interface module of densimeter; EU: evaluation unit of densimeter; CP: high-pressure centrifuge pump; FV: viscometer tube; S1, S2: magnetic sensors; TM: electronic time registration device; WRB: water recirculation bath; GR: constant volume gas container; WCP: water circulation path. FIGURE 2. Partial pressure – loading curves at different temperatures for 46.78 mass%
MDEA solution compared with data reported in literature at T=313.15 K, SidiBoumedine et al. [37] 46.78 mass% MDEA), Kamps at al. [38] (48.8 mass% MDEA) and Li et al. [39] (2.57 kmol.m-3 MDEA).
Page 19
FIGURE 3. Vapor pressures of 46.78 mass% MDEA solutions obtained from this work
compared with literature data (shokouhi et al. [9], Xu et al. [40], Kim et al. [41]) at different temperatures. FIGURE 4. Comparison of unloaded MDEA aqueous solutions density obtained from
this work with those reported data, Hsu et al. [43], Bernal-Garcia et al. [42], Shokouhi et al. [9], Rinker et al. [20], Al-Ghawas et al. [47] at the given temperatures. FIGURE 5. Comparison of unloaded MDEA aqueous solutions viscosity obtained from
this work with those reported data, Teng et al. [45], Bernal-Garcia et al. [44], Shokouhi et al. [9], Rinker et al. [20], Li et al. [46], Al-Ghawas et al. [47] at the given temperatures. FIGURE 6. Plot of measured density (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (lines) as a function of the loading (mol H2S/mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%. FIGURE 7. Plot of measured viscosity (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (lines) as a function of the loading (mol H2S /mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%.
Page 20
FIGURE 1. Experimental setup for the simultaneous measurement of gas loading,
density and viscosity; V1–V6: valves; p1, p2: pressure transmitter sensors; EC: main liquid tank; SW: side glass window; GIS: gas injection system; VTD: Anton-Paar vibrating– tube densimeter; IM: interface module of densimeter; EU: evaluation unit of densimeter; CP: high-pressure centrifuge pump; FV: viscometer tube; S1, S2: magnetic sensors; TM: electronic time registration device; WRB: water recirculation bath; GR: constant volume gas container; WCP: water circulation path.
Page 21
FIGURE 2. Total pressure – loading curves at different temperatures for 46.78 mass%
MDEA solution compared with data reported in literature at T=313.15 K, SidiBoumedine et al. [37] (46.78 mass% MDEA), Kamps at al. [38] (48.8 mass% MDEA) and Li et al. [39] (2.57 kmol.m-3 MDEA).
Page 22
FIGURE 3. Vapor pressures of 46.78 mass% MDEA solutions obtained from this work
compared with literature data (shokouhi et al. [9], Xu et al. [40], Kim et al. [41]) at different temperatures.
Page 23
FIGURE 4. Comparison of unloaded MDEA aqueous solutions density obtained from
this work with those reported data, Hsu et al. [43], Bernal-Garcia et al. [42], Shokouhi et al. [9], Rinker et al. [20], Al-Ghawas et al. [47] at the given temperatures.
Page 24
FIGURE 5. Comparison of unloaded MDEA aqueous solutions viscosity obtained from
this work with those reported data, Teng et al. [45], Bernal-Garcia et al. [44], Shokouhi et al. [9], Rinker et al. [20], Li et al. [46], Al-Ghawas et al. [47] at the given temperatures.
Page 25
FIGURE 6. Plot of measured density (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (solid lines) as a function of the loading (mol H2S/mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%.
Page 26
FIGURE 7. Plot of measured viscosity (solid points) of 46.78 mass% MDEA, and
correlated values using the Setchenow equation (lines) as a function of the loading (mol H2S /mol amine) accompanied with data reported by Rinker et al. [20] at T=313.15 K and 50 mass%.
Page 27
TABLE 1
Specifications and sources of chemicals used in this work. Chemical name
CAS registry number
Hydrogen sulfide
[7783-06-4]
N-Methyldiethanoleamine
[105-59-9]
Purity – analysis method 99.95 % (mol%) GC >99 % (mass%) GC
Source Roham Gas Company Sigma-Aldrich Company
TABLE 2
Total pressure, PT, partial pressure of H2S, PH2S , experimental solubility ( α = mol H2S /mol amine and xH2S = mole fraction), density, ρ , and viscosity, η , of (H2S – 46.78 mass% MDEA) system. (Or H2S – 7.377±0.003 mole MDEA/mass of water (kg)). T/K
α
± U (α )
xH2S ±U(x)
PT/MPa
PH2S /
ρ
MPa
(g/cm3 )
0.000 0.006 0.011 0.036 0.052 0.097 0.146 0.303 0.435 0.643 0.773 0.919
1.0314 1.0390 1.0435 1.0545 1.0605 1.0684 1.0721 1.0779 1.0794 1.0799 1.0800 1.0800
η / mPa.s ± U (η )
Pvp = 0.0063 MPa 313.15
0.000 0.080 0.155 0.349 0.465 0.648 0.695 0.866 0.936 1.019 1.057 1.099
0.007 0.008 0.009 0.010 0.011 0.012 0.014 0.015 0.016 0.018 0.020
0.0000 0.0093±0.0005 0.0178±0.0007 0.0393±0.0010 0.0517±0.0011 0.0706±0.0013 0.0754±0.0014 0.0923±0.0016 0.0989±0.0018 0.1067±0.0019 0.1103±0.0020 0.1142±0.0021
0.006 0.012 0.018 0.042 0.059 0.103 0.152 0.309 0.441 0.649 0.779 0.925
Page 28
4.23 4.20 4.06 3.76 3.71 3.65 3.59 3.55 3.54 3.46 3.41 3.36
0.16 0.16 0.16 0.05 0.16 0.08 0.16 0.16 0.16 0.12 0.16 0.12
Pvp =0.0137 MPa 328.15
0.000 0.080 0.154 0.346 0.461 0.641 0.685 0.850 0.919 0.997 1.038
0.004 0.006 0.009 0.011 0.013 0.015 0.018 0.019 0.020 0.021
0.0000 0.0092±0.0005 0.0178±0.0007 0.0391±0.0010 0.0513±0.0012 0.0699±0.0013 0.0744±0.0015 0.0907±0.0016 0.0973±0.0017 0.1047±0.0019 0.1086±0.0020
0.014 0.021 0.030 0.067 0.098 0.171 0.247 0.461 0.603 0.855 0.971
0.000 0.007 0.016 0.054 0.085 0.157 0.233 0.447 0.589 0.842 0.957
1.0216 1.0294 1.0341 1.0455 1.0515 1.0595 1.0638 1.0686 1.0699 1.0708 1.0708
2.82 2.81 2.78 2.71 2.65 2.60 2.54 2.51 2.43 2.41 2.40
0.14 0.09 0.10 0.10 0.11 0.15 0.14 0.07 0.14 0.14 0.14
Pvp =0.0274 MPa 343.15
0.000 0.079 0.153 0.343 0.455 0.630
0.004 0.006 0.009 0.010 0.011
0.0000 0.0092±0.0005 0.0176±0.0007 0.0386±0.0010 0.0507±0.0011 0.0689±0.0013
0.027 0.038 0.054 0.116 0.168 0.279
0.000 0.011 0.027 0.089 0.140 0.252
1.0097 1.0186 1.0230 1.0350 1.0405 1.0488
1.83 1.88 1.83 1.85 1.92 1.90
0.04 0.04 0.07 0.13 0.04 0.08
0.671 0.836 0.899 0.975
0.014 0.016 0.018 0.020
0.0730±0.0015 0.0893±0.0016 0.0954±0.0017 0.1026±0.0019
0.380 0.617 0.810 1.093
0.353 0.589 0.782 1.065
1.0527 1.0568 1.0583 1.0594
1.82 1.86 1.83 1.84
0.10 0.04 0.13 0.09
Expanded Uncertainties U at 95% confidence interval are U (T) = 0.02 K, U (PVP) = 0.0006 MPa, U (PT) = 0.0004 MPa, U ( ρ ) = 2.5×10-3 g.cm3, U (η ) obtained from standard deviation of 5 measurements and uncertainty of MDEA concentration in molality base, U (m) = 0.003.
TABLE 3
Comparison of density and viscosity values obtained from this work and some other published ones for 46.78 mass% MDEA aqueous solutions.
Page 29
Density (g.cm-3) Hsu and Bernal-Garcia Li [43] [42] 1.0309 1.0306 1.0206 1.0207 1.0097 1.0099
T/K
313.15 328.15 343.15
Viscosity (mPa.s) Bernal-Garcia Teng et al. [44] [45] 4.552 4.212
This work 1.0314 1.0216 1.0097
1.819
This work 4.228 2.82 1.834
TABLE 4
Percent rate variation of density and viscosity of loaded solution in comparison of fresh one at (α = 0.5 and 1.0) in this work and reported literature data. MDEA Solution (mass%)
T/K
∆α
∆ρ%
∆η%
46.78%
0 - 0.5 0 – 1.0988 0 - 0.5 0 – 1.0383 0 - 0.5 0 – 0.9748
This work 313.15 313.15 328.15 328.15 343.15 343.15
3.5 4.7 3.3 4.8 3.5 4.9
12.8 20.6 6.0 15.0 ~0 ~0
10% 20% 30% 40% 50% 50%
0 - 0.5 0 - 0.5 0 - 0.5 0 - 0.5 0 - 0.5 0 - 0.5
Ref. [20] 298.15 298.15 298.15 298.15 298.15 313.15
0.5 0.6 1.1 2.4 2.7 2.8
3.3 6.2 23.3 23.1 26.9 16.1
Page 30
TABLE 5 Numerical values of the parameters of the modified Setchenow equation for correlation of density and viscosity for the given solution.
46.78 mass% MDEA k 0 , 0, ρ
-0.01903
k 0 ,1,ρ
0.000293
k1,0 ,ρ
0.028774
k1,1,ρ
-0.00018
k 0 ,0 ,η
-3.66042
k 0 ,1,η
0.010809
k1, 0,η
1.419516
k 1,1,η
-0.00432
ARD% ( ρ ) ( ARD % ) exp (ρ ) MRD% ( ρ ) (MRD % ) exp (ρ ) ARD% (η ) ( ARD % ) exp (η ) MRD% (η ) (MRD %) exp (η )
0.06% 0.23% 0.25% 0.26% 1.48% 4.17 % 4.93% 7.06%
Page 31
In the name of god Highlights Measurement solubility of H2S in 46.78 mass% MDEA aqueous solutions. Measurement density of H2S loaded of MDEA aqueous solution. Measurement viscosity of H2S loaded of MDEA aqueous solution. Correlation of the density and viscosity of H2S loaded of MDEA aqueous solution using modified setchenow equation.