Universal Correlation For Gas Hydrates Suppression Temperature of Inhibited Systems: I. Single Salts

Thermodynamics and Molecular-Scale Phenomena AIChE Journal
DOI 10.1002/aic.15846
Universal Correlation for Gas Hydrates Suppression Temperature

of Inhibited Systems: I. Single Salts
Yue Hu1, Bo Ram Lee1† and Amadeu K. Sum1*
1
Hydrates Energy Innovation Laboratory, Chemical & Biological Engineering Dept.,
Colorado School of Mines, Golden, CO 80401
*Corresponding author e-mail: asum@mines.edu
†
Currently at Dept. of Chemical Engineering, Pohang University of Science & Technology, Cheongam-Ro,
Nam-Gu, Pohang, Gyongbuk, 37673 – Korea
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/aic.15846
© 2017 American Institute of Chemical Engineers (AIChE)
Received: Apr 21, 2017; Accepted: Jun 16, 2017
This article is protected by copyright. All rights reserved.

AIChE Journal Page 2 of 72
Abstract
Reliable prediction of hydrate suppression temperature in presence of inhibitors, such as salts,
over a wide range of pressures (up to 200 MPa) is critically important, especially in the area of
deepwater oil and gas production. However, the existing models and correlations that account for
salts have severe limitations and deficiencies in estimating the hydrate suppression temperature.
Herein, we propose a new correlation, to be called Hu-Lee-Sum correlation, that significantly
improves the predictions of the hydrate suppression temperature by considering the salt species
and concentrations, system temperature and pressure, and hydrate structure. This paper
represents part I of this work and it will detail the development of the correlation and
demonstrate the generality and universality of the correlation to predict the hydrate suppression
temperature for any single salt system. Specifically, we show accurate predictions for the hydrate
suppression temperature for a number of chloride and bromide salt brine systems.
Topical Heading: Thermodynamics
Keywords: gas hydrates, suppression temperature, salt, brine, correlation
2
AIChE Journal
Page 3 of 72 AIChE Journal
Introduction
Gas hydrates are formed by the inclusion of light guest molecules inside cavities consist of
strongly hydrogen-bonded water molecules.1 The guest molecules that interact with water
molecules through van der Waals type dispersion forces generally consist of light hydrocarbons,
such as methane (CH4), ethane (C2H6) and propane (C3H8), size of which determines the crystal
structure of the hydrate formed.2 The unit cell of structure I (e.g., methane hydrates) is
constructed from 46 water molecules to make 2 small and 6 large cavities where the gas
molecules can occupy, while structure II consists of 136 water molecules to make 16 and 8 of
small and large cavities, respectively.2
Hydrates have unique physical properties, such as, structures are solid solution of water and
gas, the dissociation temperature is higher than ice melting temperature at same pressure, they
are non-stoichiometry,3 and they have high energy density.1 Therefore, gas hydrates, especially
methane hydrates have been extensively studied and investigated both for scientific and practical
interests. As a resource, methane hydrates are considered a huge source for methane gas trapped
in naturally occurring deposits in the deep ocean sediments due to both their widespread
existence and their concentrated form.2 Besides, it is found that utilization of hydrate formation
as a separation process for the desalination of water could be feasible and efficient.4 On the other
hand, gas hydrates are ever-concerning in oil and gas production since they are a nuisance and
responsible for reduced/lost production due to accumulation and plugging of flowlines resulting
in pressure build up in the flowlines.5
3
AIChE Journal
As the development and production of oil/gas fields continue to delve into deeper water
depths, increasingly harsh conditions in terms of pressure and water salinity are expected, posing
significant flow assurance challenges, especially due to gas hydrate formation and scale
precipitation in flowlines.6 The pre-salt reservoirs, a layer of oil bearing rock of carbonate
composition, are positioned below sea level with depths between 5,000 m and 6,000 m, in
ultra-deep water (1,900 m to 2,400 m) beneath a thick salt layer (in some areas, up to 2,000 m),
resulting in reservoir pressures higher than 100 MPa and salt concentrations more than 200,000
ppm (20 wt%) corresponding to near-saturated/saturated conditions.7 In addition, heavy brines
are used during well completion to offset the high reservoir pressures with the brine hydrostatic
pressure in order to avoid a well blowout. At the same time completion brines must remain
hydrate-free at the mudline (seabed) conditions in case light hydrocarbon gases migrate from the
bottom of the well up the wellbore and come in contact with cold (~277 K) brine, if a well is
drilled in the offshore environment.8
It is well known that inorganic salts, such as sodium chloride, calcium chloride, etc., are
hydrate inhibitors. Due to strong electrostatic forces, water tightly bounds to the dissolved salt
ions (electrostatic forces)9 and becomes unavailable to hydrogen bond to other water molecules
and form the hydrate. As such, salts suppress the hydrate formation temperature at a given
pressure. The degree of the suppression temperature closely depends on the salt type and its
concentration in solution. Therefore, it is considerably important and necessary to estimate the
hydrate suppression temperature accurately, especially for hydrate management in oil and gas
production, transport and processing, as unreliable predictions of hydrate dissociation condition
4
AIChE Journal
could lead to serious safety consequences of hydrate plug and over/under use of hydrate
inhibitors (e.g., methanol or glycol) increasing the operational cost as well as the possibility of
salt precipitation and corrosion. Accurate estimate of the hydrate suppression temperature with
salts is also critical in the use of hydrates in the desalination of water,4 an area increasingly
important in the energy application of hydrates.
Review
Correlations to Estimate Hydrate Suppression Temperature in Brine Systems
While the prediction of hydrate dissociation conditions with fresh water has been well developed,
the estimation of the hydrate suppression temperature with thermodynamic inhibitors is limited.
There are two main methods to study the hydrate dissociation conditions in inhibited systems.
One of the methods is based on the van der Waals-Platteeuw model10 for gas hydrate equilibrium
(equality of chemical potential of water in aqueous and hydrate phase), and simultaneously,
different models are employed to calculate the water activity in electrolyte solutions which
contribute to the chemical potential of water in the aqueous solution.2,11-19 This method assumes
that the chemical potential of water can be accurately calculated in both liquid and hydrate
phases. The major deficiency of this method for inhibited systems is the inaccuracy of activity
models used to account for the salts and organic inhibitor over a wide range of conditions (type,
concentration, temperature, and pressure). The other method for estimating the hydrate
suppression temperature for inhibited systems is more empirical but simpler. Due to the
empirical nature of correlations, the estimated suppression temperature is often less reliable and
5
AIChE Journal
is not readily transferable for conditions beyond those considered in the development of the
correlation. As a background to the development of the new correlation proposed herein, we
briefly discuss the existent correlation in the literature for the hydrate suppression temperature.
The Hammerschmidt equation20 was the very first correlation proposed to approximate the
hydrate suppression temperature. It is based on the freezing point depression principle stating
that the depression in freezing point is directly proportional to the weight of a dissolved
substance in a given amount of solvent.21
kHW
∆T = (1)
M (100 − W )
where ∆T is the temperature depression in °C, M and W are the molar mass of the inhibitor in
g/mol and the concentration of the inhibitor in weight percentage in aqueous phase, and kH is a
constant with a value of 1297. However, it is limited to concentrations up to about 30 wt% of
methanol and ethylene glycol and only to about 20 wt% for other glycols.22 Over years,
modifications to the Hammerschmidt equation have been proposed to improve its predictability
by introducing parameters dependent on the inhibitor type.22 While Hammerschmidt suggested a
modification to the molecular weight of the inhibitor when considering associating compounds
(e.g., salts), there is nothing in the correlation to account for the gas composition.1,22
As the produced water from oil/gas reservoirs tend to contain brines, determining the effect
of salts on hydrate inhibition is equally important. McCain23 developed a correlation for
estimating the effect of brine on the hydrate suppression temperature, which is related to gas
specific gravity and salt concentration.
6
AIChE Journal
∆T = AS + BS 2 + CS 3 (2)
where ∆T is the temperature depression in °F, S is the salinity in weight percentage, and the
coefficients A, B and C are functions of the gas gravity, γ, which are included in the
Supplementary Information. Although McCain pointed out that the predictions agree with
existing data from Deaton et al.,24 it is limited for gas specific gravities smaller than 0.65 and
salinity lower than 20 wt%.23 Moreover, this correlation implicitly assumed that the hydrate
suppression temperature is independent of salt species, which can be inappropriate for the
systems with high salinity, in particular to aqueous solutions containing divalent ions, such as
Ca2+ or Mg2+ of which electrostatic forces with water molecules are stronger than most
monovalent ions.25,26
Gas hydrate formation in deepwater offshore drilling is a well recognized operational hazard.
At water depth exceeding 300 m, the seabed ambient pressure and temperature becomes
conducive for hydrate formation, imposing serious well control difficulties.27 Yousif and Young27
developed a simple correlation to predict hydrate suppression temperature for systems with salts
and glycerol that are the main hydrate inhibitors in water-based drilling muds.
∆T = 112 .3 x + 2011 .6 x 2 − 6505 .0 x 3 (3)
where ∆T is the temperature depression in °F and x represents total mole fraction of the hydrate
inhibitors in the solutions, which is related to the apparent molecular weight of salt in aqueous
solutions that contains other salts and glycerol, as well as the degree of ionization (details are
included in the Supplementary Information). This correlation estimates the hydrate suppression
temperature assuming that the gas composition has negligible effect, which is similar to the ice
7
AIChE Journal
depression temperature. The correlation is developed based on experimental data of drilling
fluids containing NaCl, KCl, NaBr, CaCl2 and glycerol. If other salts are included in the system,
experimental data are needed to calculate the degree of ionization of specific salts, which is
costly and time consuming.
In order to overcome the deficiencies in the Yousif and Young correlation,27 Mohammadi
and Tohidi21 developed a predictive technique for estimating the hydrate inhibition effects of
single and mixed thermodynamic inhibitors, given by
[
∆T = −a ln(1 − xsolute ) + bxsolute
2
(
+ ∑ c1,iWi + c2 ,iWi + c3,iWi
2 3
)] (4)
where a, b, c1, c2 and c3 are constants, xsolute and Wi are mole fraction of organic inhibitor and
weight percentage of salt i in the aqueous solutions, respectively. This correlation is a
combination of the Yousif and Young,27 Nielsen and Bucklin28 and Carroll correlations,22 which
shows acceptable agreement with experimental data, in particular at high concentrations of salts
and organic inhibitors. However, this correlation does not take in consideration the effect of
pressure on the hydrate suppression temperature, which could lead to large deviation between
experimental data and predicted results at higher pressure.
In yet another attempt to correlate the hydrate suppression temperature, Østergaard et al.29
noted discrepancies between the experimental data and the prediction from the Yousif and Young
correlation, and proposed a modified correlation considering the pressure dependence,29,30
( )
∆T = c1W + c2W 2 + c3W 3 (c4ln(P) + c5 )× (c6 (P0 − 1000) + 1) (5)
where c1, c2, c3, c4, c5 and c6 are parameters related to hydrate inhibitors, P and P0 are the hydrate
dissociation pressures in unit of kPa with aqueous solution and fresh water, respectively. This
8
AIChE Journal
was the first correlation that accounted for the effect of pressure on the hydrate suppression
temperature with the additional expense of introducing a total of six adjustable parameters per
inhibitor system. As such, the use of this correlation is relatively difficult and limited to inhibitor
systems for which parameters available, although it is the most reliable among the existing
correlations.
In addition, Najibi et al.31 developed a systematic method to predict hydrate suppression
temperature of fluids along the pipeline and downstream conditions from information of freezing
point depression of aqueous solutions containing different concentration of inhibitors:
∆T = 0.6825∆TF (6)
where ∆T and ∆TF are the hydrate temperature depression and freezing point of depression of the
aqueous solution, respectively. This method does not need the analytical composition of the
aqueous solution,31 however, the measurement or calculation of the freezing point for the
aqueous solution is necessary for use of the correlation.
To compare the performance of these existing correlations, Figure 1 shows results from the
correlations to measured experimental data for structure I and II hydrates for CaCl2 solutions. As
it is clearly seen, the Yousif and Young correlation27 does not capture the difference between
hydrates structure, which depends on the gas composition, as well as the pressure. Moreover, it
considerably underestimates the suppression temperature. Although the McCain correlation23
estimates the hydrate suppression temperature based on the gas composition as indicated in
Figure 1, the predicted results are inconsistent with experimental data for the CaCl2 system. In
addition, even though the correlation by Østergaard et al.30 performs much better than the Yousif
9
AIChE Journal
and Young correlation and McCain correlation, the deviation between the predicted suppression
temperature and experimental data can be relatively large.
As it is clearly evident from Figure 1, the existing correlations for hydrate suppression
temperature are inaccurate and have a number of limitations in their applicability. As such, we
proposed herein a new correlation that not only significantly improves the prediction of hydrate
suppression temperature but is also solidly based on the fundamental principles of
thermodynamics. The proposed correlation takes into account the effect of pressure, hydrate
structure, types and concentration of hydrate inhibitors to give accuracy estimation of hydrate
inhibition effect for different salts.
Literature Data
As a starting point in the development of the correlation, we only considered the effect of
pressure, hydrate inhibitor species (salt) and concentration based on collected data for methane
hydrate phase equilibria with single chloride salt solutions, as shown in Table 1. These data span
a range of conditions in terms of temperature and pressure, and a wide range of salt
concentration, which for NaCl and KCl approaches the saturation concentration.
While the new correlation will be developed based on the methane hydrate phase
equilibrium data shown in Table 1, its validation and applicability will be tested with other salts
or hydrates system. Tables 2 to 4 list the data for methane, carbon dioxide, and ethane hydrates,
respectively, that will be used to evaluate and compare the new correlation.
10
AIChE Journal
Method
Derivation of the Hu-Lee-Sum Correlation
The goal of developing a new correlation for the hydrate suppression temperature is to capture
the proper physics and thermodynamic consideration for the effect of salts in the hydrate phase
equilibria. The Hu-Lee-Sum correlation is based on the fundamental principle of freezing point
depression, which for hydrate is equivalent to the suppression temperature from the uninhibited
(salt-free) system. At the equilibrium hydrate dissociation, the equality of fugacity of water (w)
in the solid (hydrate) phase and the liquid phase has to be equal50, that is,
f ws (T , P) = f wL (T , P) = xwγ w (T , P, x) f wL0 (T , P) (7)
where xw and γw are the mole fraction and activity coefficient, respectively, of water, f represents
the fugacity for solid (S), liquid (L) and pure water (L0). The assumptions for expressing the
equality of fugacity in Eq. 7 are: i) the system pressure is constant; ii) the hydrate is a pure phase,
that is, the inhibitor (salt) is not part of the hydrate structure; iii) the composition of the
hydrocarbon-rich vapor or liquid phase is constant; and iv) the composition of the hydrate is
constant.50
Since the solid is a pure phase, we can relate the ratio of the pure phases fugacity to the
Gibbs energy of fusion,
∆ fusG (T , P ) f wL0 (T , P )
= ln s (8)
RT f w (T , P )
which can then be expressed in terms of the enthalpy and entropy of fusion (∆G = ∆H – T ∆S),
given by,
11
AIChE Journal
T
( )
∆ fus H (T , P ) = ∆ fus H T 0 + ∫ ∆C p dT (9)
T0
T
∆C p
∆ fus S (T , P ) = ∆ fus S (T0 ) + ∫ dT (10)
T0
T
where ∆Cp is the heat capacity difference of pure water and solid crystal and T0 represents the
normal freezing temperature. Equations 9 and 10 relate the enthalpy and entropy change of
fusion at temperature T to those changes at the freezing point T0 at the same pressure. When the
temperature T equals the freezing temperature T0, ∆fusG(T0) equals to zero, allowing to rewrite
Eq. 10 as:
∆ fus H (T0 )
T
∆C p
∆ fus S (T , P ) = +∫ dT (11)
T0 T0
T
Substituting the expression of the enthalpy and entropy of fusion into Eq. 8 gives
f wL0 (T , P ) 1   T  T 
T
∆C p
ln =  ∆ H (T )
0  1 −  ∫
 + ∆C dT − T ∫ T  dT (12)
f ws (T , P ) RT 
fus p
 T0  T0 T0 
The heat capacity terms are usually relatively small contributions and if they are neglected, we
can combine Eqns. 7 and 12 to obtain
f wS (T , P ) 1   T 
ln = ln( xwγ w ) = ln(aw ) = − ∆ fus H (T0 )1 −  (13)
f w (T , P )
L0
RT   T0 
where aw represents the water activity.
For gas hydrates, where gas and water combine to form the solid crystal, Gas + nH2O →
Gas.nH2O, where n is the hydration number, the enthalpy of fusion ∆fusH(T) can be replaced by
hydrate dissociation enthalpy (∆Hdiss = Hgas + nHH2O – Hgas.nH2O). Therefore, the hydrate
depression temperature (∆T = T0 – T) can be expressed by51
12
AIChE Journal
∆T nR
=− lna w (14)
T0T ∆H diss
where T0 and T are the hydrate dissociation temperature with fresh water and aqueous solution,
respectively, at the same pressure, as illustrated in Figure 2. The hydrate dissociation temperature
with fresh water is assumed to be known and readily available.
The right side of Eq. 14 has two contributions: nR/∆Hdiss and ln aw. The first term takes into
account the heat of hydrate formation/dissociation, which depends on the hydrate forming
system, that is, gas composition. Hu et al.52 pointed that the hydrate dissociation enthalpy (∆Hdiss)
can increase along the phase boundary for pressures higher than 50 MPa, and the hydration
number (n) accordingly decreases with increasing pressure to approach the full occupancy limit
(n = 5.75 for structure I). However, the combined changes of ∆Hdiss and n, which are a function
of pressure, are relatively small compared with the absolute value of the quantity, and as such, it
is reasonable to assume that the term nR/∆Hdiss remains constant for any given hydrate forming
system. In addition, if we assume that the activity of water is independent of pressure, each term
on the right side of the equation is independent of pressure, which indicates that ∆T/T0T as a
whole is also independent of pressure, though the hydrate suppression temperature (∆T), hydrate
dissociation temperature with fresh water (T0) and salt solution (T), all depend on the system
pressure. Since hydrates are formed from water and gas molecules, the hydration number (n) and
hydrate dissociation enthalpy (∆Hdiss) depends on the gas composition, regardless of salts added
to the solution, while the activity coefficient depends on the system temperature, salt species and
concentration.
13
AIChE Journal
In order to simplify the correlation, we first only consider methane hydrate (sI), and as such,
the hydration number and hydrate dissociation enthalpy are constant so that the value of ∆T/T0T
is only a function of system temperature (T) and salt species and concentration (xi), given by
∆T
= β × lnaw ( xi ,T ) (15)
T0T
where β is a constant which only depends on the hydrate former and type. Based on this equation,
the value of ∆T/T0T can be calculated by knowing the water activity.
Temperature Dependence
To better assess and address the effect of temperature on the values of ∆T/T0T, all literature data
for methane hydrate at all salt concentrations for NaCl, CaCl2, KCl and MgCl2 solutions,
separately, are plotted in Figure 3 in the form of ∆T/T0T versus hydrate dissociation temperature
T. As clearly seen in the plots, the quantity ∆T/T0T is nearly independent of temperature T.
Therefore, it is reasonable to assume that the value of ∆T/T0T, or equivalent the water activity
(Eq. 15) is independent of temperature, which is one of the key findings and assumptions in the
Hu-Lee-Sum correlation. The average value of ∆T/T0T for each system is summarized in Tables
S2-S5 in the Supplementary Information. We note that some data reported in the literature show
a strong temperature dependence for ∆T/T0T, suggesting the data are unreliable, and as such,
these have not been included in Figure 3 (see details of these data in the Supplementary
Information).
14
AIChE Journal
Salt Species and Concentration Dependence
The water activity in Eq. 15 is strongly dependent on the salt species and concentration, and it is
the key quantity for determining the hydrate suppression temperature. Therefore, the goal was to
develop a simple and consistent correlation describing the relationship between water activity
and salt concentration for different chloride salts. There are a number of different ways used to
express the salt concentration in solutions, such as weight fraction, mole fraction, molality, and
ionic strength. Each of these concentration variables were tested to correlate the water activity,
however, none resulted in a form that combined the data for monovalent and divalent salts
together. One of the best, if not the best, activity coefficient model for electrolyte systems is the
eNRTL53, which uses the effective mole fraction, X = ∑z

i = ions
i x i , as a concentration variable,
where xi and zi are the mole fraction and charge of the ions of the salt, respectively. Using the
effective mole fraction, we can plot the water activity for all the salts and concentrations to
obtain Figure 4. Each point in Figure 4 represents the average value of ∆T/T0T (see Figure 3 and
Tables S2-S5) for all the data available in the open literature for chloride salts (NaCl, KCl, CaCl2,
MgCl2 – Table 1). We note that these data cover a wide range of salt concentrations, including
near saturation for NaCl and KCl. As can be seen, all the data converge to a curve, which can be
easily fit to a polynomial form relating the water activity to the effective mole fraction, given by
∆T
= β × lna w = C1 X + C 2 X 2 + C 3 X 3 (16)
T0T
where the coefficients C1, C2, and C3 are given in Table 5. This equation is the Hu-Lee-Sum
correlation for the hydrate suppression temperature.
15
AIChE Journal
Results and Discussion
With the development of the Hu-Lee-Sum correlation, it is simple and convenient to calculate the
hydrate suppression temperature at a given condition in terms of pressure, salt species and salt
concentration. For the purposes of this study, the hydrate dissociation temperature (T0) is
obtained from a polynomial regression of the methane hydrate data with fresh water (shown in
the Supplementary Information), however, it is possible to use any other source (data or
prediction) to reliably give T0 at any given pressure. The effective mole fraction can be easily
calculated for a given salt type and concentration, from which Eq. 16 is used to give ∆T/T0T.
Finally, the hydrate dissociation temperature for the corresponding pressure at T0 can be simply
calculated with
−1
  ∆T  
T = T0 1 +  T0  (17)
  T0T  
This correlation has characteristics that make it very desirable and useful to estimate the hydrate
inhibition strength of any single salt. In addition to its simplicity, it exhibits the correct limiting
behavior, that is, in the limit as the salt concentration approaches zero, the hydrate suppression
temperature also approaches zero.
The collected hydrate phase equilibria data listed in Table 1 are used to test the performance
of the Hu-Lee-Sum correlation, as wells as check the limitations of the correlation, in terms of
salt concentration and pressure for chloride systems. Figures 5 to 8 show the measured data and
predicted results from the Hu-Lee-Sum correlation for methane hydrate phase equilibria for
systems containing NaCl, CaCl2, KCl, and MgCl2, respectively. The Hu-Lee-Sum correlation is
16
AIChE Journal
shown to accurately predict the hydrate suppression temperature even for near-saturated chloride
solutions. Moreover, the results calculated with the Hu-Lee-Sum correlation show good
agreement with measured data for a wide range of pressures, ranging from 2 to 200 MPa.
To better evaluate the accuracy and reliability of the Hu-Lee-Sum correlation, its predictions
are compared to three other commonly used methods to predict the hydrate suppression
temperature: i) a hydrate prediction tool based on the van der Waals and Platteeuw model with
the Soave-Redlich-Kwong equation of state for fluid properties and the Bromley equation to
account for the water activity coefficients with electrolytes54 (to be denoted as P1), ii) a hydrate
prediction tool based on van der Waals and Platteeuw model with Peng-Robinson equation of
state for the fluid properties and the Pitzer’s activity coefficient model to account for the
electrolytes55 (to be denoted as P2), and iii) the Østergaard et al.30 correlation (to be denoted as
P3). The results comparing the methods will be reported as the average absolute deviation (AAD)
in temperature, defined as:
∑T pred − Texp
AAD = i =1
(18)
N
where N is the number of equilibrium data points, Texp and Tpred are the measured and predicted
hydrate dissociation temperature, respectively, for aqueous solutions with salt. All the results for
the AAD for methane hydrate phase equilibrium temperature with NaCl, CaCl2, KCl, and MgCl2
are summarized in Figures 9 to 12 and Tables S8 to S11 in Supplementary Information,
respectively.
17
AIChE Journal
According to Figure 9, method P2 fails to provide prediction for methane hydrate with 26 wt%
NaCl, since it estimates that the solution is over-saturated at these conditions. However, Hu et al.
33
verified that the solution with 26 wt% NaCl is under-saturated through measurements of the
quadruple points, where hydrate, solution, NaCl solid, and CH4 vapor coexist33. Moreover,
method P3 becomes unreliable as NaCl concentration increases higher than 20 wt%, especially at
higher pressure. The AAD for method P3 are 2.6 and 4.1 K compared with the experimental data
for 23 and 26 wt% NaCl, respectively, measured up to 200 MPa. Inversely, both method P1 and
the Hu-Lee-Sum correlation can provide comparable and accurate results of methane hydrate
phase equilibria with 26 wt% NaCl up to 200 MPa.
As shown in Figure 10, methods P1, P2, and P3 predict the methane hydrate dissociation
conditions in presence of 32 wt% CaCl2 with relatively large AAD of 5.2, 9.6 and 14.8 K
respectively, which can be attributed to the inaccurate activity model used for CaCl2 at high
concentration in P1 and P2 and the limited concentration in P3. In contrast, the Hu-Lee-Sum
correlation successfully provides accurate hydrate dissociation temperature at this condition.
Figure 6 shows that the data for methane hydrate with 14.5 wt% CaCl2 from Kharrat et al.35 are
inconsistent with the data measured by Mohammadi et al.39 at 15 wt% CaCl2. Kharrat et al.35
measured methane hydrate phase equilibria with CaCl2 using differential scanning calorimetry
(DSC) method with continuous heating, which can lead to inaccurate measurements due to the
continuous change of salt concentration during hydrate dissociation. Figure 10 shows that the
18
AIChE Journal
AAD for the system with 14.5 wt% CaCl2 is 1.5, 1.4 and 1.6 K for methods P1, P2, and the
Hu-Lee-Sum correlation, respectively.
Figures 7 and 11 show that Hu-Lee-Sum correlation provides reliable and accurate predicted
results of methane hydrate in presence of KCl aqueous solutions with concentrations from 5 wt%
(low) to 23 wt% (near-saturated). As shown in Figure 11, the AAD with the Hu-Lee-Sum
correlation is lower than 1.0 K for all the data sets available, while method P3 shows larger AAD
of 1.6 and 2.0 K for methane hydrate with 20 and 23 wt% KCl, respectively, at pressures up to
200 MPa. Methods P1 and P2 also give unreliable predictions at high KCl concentrations.
Methods P1, P2, and P3 do not include magnesium chloride (MgCl2) as an option of salts to
perform the calculation, as such, Figure 12 only shows the AAD results for the Hu-Lee-Sum
correlation, which successfully provide reliable predictions for methane hydrate up to 25 wt%
MgCl2, with maximum AAD of 1.0 K. Overall, there is a very limited experimental data and
model/correlation for hydrate phase equilibria with MgCl2, as such, the Hu-Lee-Sum correlation
becomes considerably valuable for the prediction of systems with this salt.
Figures 5 to 12 show that the Hu-Lee-Sum correlation gives excellent results in predicting
the methane hydrate phase equilibria for the single chloride salts used in developing the
correlation. In order to check the applicability and reliability of the Hu-Lee-Sum correlation in
predicting hydrate phase equilibria for other single salt systems, we tested the correlation to
predict the methane hydrate phase equilibria with ammonia chloride (NH4Cl), sodium bromide
(NaBr), potassium bromide (KBr), and calcium bromide (CaBr2), separately, systems for which
experimental data are listed in Table 2. For each of these salts, we use the same correlation given
19
AIChE Journal
by Eq. 16, which again, is only based on data for NaCl, KCl, CaCl2 and MgCl2. Equation 16 only
requires the effective mole fraction of the salt, which can then be used to obtain ∆T/T0T. The
results of these true predictions are shown in Fig. S3 in the Supplementary Information with the
AAD shown in Figure 13 (Table S12 in the Supplementary Information shows the actual AAD
values). We note that these predictions are not possible with methods P1 as none of these salts are
included in their model; NH4Cl is also not included in the database of P2; parameters for method
P3 are only available for NaBr and KBr. Moreover, for CaBr2, it was shown in a recent study45
that commonly used prediction tools were unable to provide reliable hydrate phase equilibria
predictions, whereas the AAD with the Hu-Lee-Sum correlation is only about 1.3 K. As can be
clearly seen in Figure 13, the Hu-Lee-Sum correlation can accurately predict the experimental
data for these few single salt systems.
All the results shown so far are for methane hydrate suppression temperature. The
correlation developed is applicable for any single salt system at any concentration. Comparing
Eqs. 15 and 16, the inherent assumption is that the factor β is constant, meaning that the ratio of
the hydration number to the dissociation heat are also constant. If we consider the value of
nR/∆Hdiss for other structure I hydrate systems, we find that it is 0.8657 and 0.9143 for ethane
and carbon dioxide hydrates, respectively, which compare very closely to 0.9115 for methane
hydrates (details in Table S15 in the Supplementary Information). Therefore, if we assume that
the variation in β is negligible for any structure I hydrate, the Hu-Lee-Sum correlation (Eq. 16)
can be readily used predict the hydrate suppression temperature for other structure I systems
besides methane hydrates, as long as the fresh water hydrate dissociation temperature is available.
20
AIChE Journal
We tested the Hu-Lee-Sum correlation to the literature data available for carbon dioxide and
ethane hydrate phase equilibria in single salt systems, as summarized in Tables 3 and 4,
respectively. Figures 14 and 15 summarize the average absolute deviation between experimental
data and predicted results for ethane and carbon dioxide hydrate phase equilibria, respectively.
For the hydrate dissociation temperature for carbon dioxide and ethane in fresh water, prediction
P2 was used, which is accurate compared to experimental data. As shown in Figures. 14 and 15,
the performance of Hu-Lee-Sum correlation is consistent for structure I (sI) hydrate in presence
of different kinds of salts with salt concentration from relatively low to near-saturated. In
addition, compared with method P3, the Hu-Lee-Sum correlation is more general, which can be
applied to any salt without the necessity of time-consuming and costly measurements and
analysis to obtain the model parameters.
To understand why the Hu-Lee-Sum correlation can be generally applied to predict the
hydrate phase equilibria for any single salt system, we take a closer look at Eq. 15 and Figure 4.
For hydrate structure I, the value of β in Eq.. 15 is reasonably constant, and Eq. 15 simply states
that the suppression temperature is directly related to the water activity. Any salt dissolved in
water has an effective contribution to change the water activity, which is mainly determined by
the charge density of the ions. As such, it may not be surprising that Figure 4 and Eq. 16
represent a “universal” curve for the water activity. To further demonstrate the generality and
universality of the universal curve, Figure 16 shows Figure 4 replotted with the additional values
of ∆T/T0T based on the experimental data for NH4Cl, NaBr, KBr, and CaBr2. It is evident in
Figure 16 that any single salt at any concentration will fall on the universal water activity curve.
21
AIChE Journal
Another significant outcome of this study is that the water activity can be very well
approximated to be independent of temperature, as shown in Figure 3 by the nearly horizontal
alignment of the data at any salt concentration. This result may not be surprising as well, as the
temperature range for the hydrate phase equilibria considered is at most about 30 K. A
consequence of the temperature independence of the water activity is that plotting the data as in
Figure 3 is a simple and reliable way to test the thermodynamic consistency and reliability of
measured hydrate phase equilibria data with salts. If the alignment of the ∆T/T0T data shows to
be temperature dependent or deviates from a horizontal line, one should question the data (see
details in Supplementary Information).
In this study, the focus was to determine the effect of system temperature, salt species and
concentration on hydrate suppression temperature for structure I hydrates. In Part II of this paper,
the discussion will focus in adjusting the predictions to consider structure II hydrates and mixed
salt systems.
Conclusion
The Hu-Lee-Sum correlation was developed to predict the hydrate suppression temperature of
single salt systems. The correlation is based on the well established thermodynamic principle of
suppression temperature for additives added to solution, which in this particular case are
inorganic salts. An equation for the relationship between hydrate dissociation enthalpy and the
effect of salt on the hydrate suppression is developed for fast and accurate estimation of the
hydrate suppression temperature with chloride salts, but it shown that a universal curve can be
22
AIChE Journal
applied to any salt for concentrations up to saturation in solution. This simple, yet reliable
correlation is validated for predicting methane hydrate suppression temperature for systems
containing sodium chloride (NaCl), potassium chloride (KCl), calcium chloride (CaCl2) and
magnesium chloride (MgCl2), separately, over a wide pressure range from 2 to 200 MPa, by
comparing the experimental and predicted results. Not only the performance of Hu-Lee-Sum
correlation is as good as or better than other predictive methods for systems containing NaCl or
KCl, but it also provides reliable results for systems containing CaCl2 with concentration as high
as 32 wt% and the aqueous solutions with MgCl2, systems which other methods fail to give
reliable predictions. Moreover, the correlation is applied to systems with other salts not included
in the development of the correlation, such as ammonia chloride (NH4Cl), sodium bromide
(NaBr), potassium bromide (KBr) and calcium bromide (CaBr2), as well as other structure I
hydrate systems. Further application and validation of the correlation for much more complex
systems, such as mixed gas hydrates with mixed salt systems or mixed salts and organic
inhibitors, will be presented in follow-up papers.
Acknowledgments
We thank Prof. Chau-Chyu Chen from Texas Tech University for helpful discussions about
modeling of electrolyte systems. We also thank Dr. Prasad Karanjkar from ConocoPhillips for
helpful discussions over the course of this project.
23
AIChE Journal
Literature Cited
1. Sloan ED, Koh C. Clathrate hydrates of natural gases. Boca Raton, FL: CRC Press, 2007.
2. Holder G, Zetts S, Pradhan N. Phase behavior in systems containing clathrate hydrates: a
review. Rev. Chem. Eng. 1988;5(1-4):1-70.
3. Van der Waals J, Platteeuw J. Validity of Clapeyron's equation for phase equilibria
involving clathrates. Nature. 1959;183:462.
4. Barduhn AJ. Desalination by crystallization processes. Chem. Eng. Prog. 1967;63(1):98.
5. Lafond PG, Olcott KA, Sloan ED, Koh CA, Sum AK. Measurements of methane hydrate
equilibrium in systems inhibited with NaCl and methanol. J. Chem. Thermodyn.
2012;48:1-6.
6. Shaughnessy JM, Daugherty WT, Graff RL, Durkee T. More ultra-deepwater drilling
problems. Paper presented at: SPE/IADC Drilling Conference2007.
7. Estrella G. Pre-salt production development in Brazil. Energy solutions for all–Promoting
Cooperation, Innovation and Investment. 2011:96-99.
8. Arrieta VV, Torralba AO, Hernandez PC, García ERR, Maia CT, Guajardo M. Case history:
lessons learned from retrieval of coiled tubing stuck by massive hydrate plug when well
testing in an ultradeepwater gas well in Mexico. SPE Prod. Oper. 2011;26(4):337-342.
9. Lu H, Matsumoto R, Tsuji Y, Oda H. Anion plays a more important role than cation in
affecting gas hydrate stability in electrolyte solution?—a recognition from experimental
results. Fluid Phase Equilib. 2001;178(1):225-232.
10. Van der Waals J, Platteeuw J. Clathrate solutions. Adv. Chem. Phys. 1959;2:1-57.
11. Parrish WR, Prausnitz JM. Dissociation pressures of gas hydrates formed by gas mixtures.
Ind. Eng. Chem. Des. Dev. 1972;11(1):26-35.
12. Englezos P, Bishnoi P. Prediction of gas hydrate formation conditions in aqueous
electrolyte solutions. AIChE J. 1988;34(10):1718-1721.
13. Englezos P. Computation of the incipient equilibrium carbon dioxide hydrate formation
conditions in aqueous electrolyte solutions. Ind. Eng. Chem. Res. 1992;31(9):2232-2237.
14. Tohidi B, Burgass R, Danesh A, Todd A. Hydrate Inhibition Effect of Produced Water: Part
1—Ethane and Propane Simple Gas Hydrates. Paper presented at: Offshore Europe1993.
15. Nasrifar K, Moshfeghian M, Maddox R. Prediction of equilibrium conditions for gas
hydrate formation in the mixtures of both electrolytes and alcohol. Fluid Phase Equilib.
1998;146(1):1-13.
16. Javanmardi J, Moshfeghian M, Maddox RN. An accurate model for prediction of gas
hydrate formation conditions in mixtures of aqueous electrolyte solutions and alcohol. Can.
J. Chem. Eng. 2001;79(3):367-373.
17. Masoudi R, Tohidi B, Danesh A, Todd AC. A new approach in modelling phase equilibria
and gas solubility in electrolyte solutions and its applications to gas hydrates. Fluid Phase
Equilib. 2004;215(2):163-174.
18. Chin H-Y, Hsieh M-K, Chen Y-P, Chen P-C, Lin S-T, Chen L-J. Prediction of phase
equilibrium for gas hydrate in the presence of organic inhibitors and electrolytes by using
24
AIChE Journal
an explicit pressure-dependent Langmuir adsorption constant in the van der Waals–

Platteeuw model. J. Chem. Thermodyn. 2013;66:34-43.
19. Osfouri S, Azin R, Gholami R, Moshfeghian M. Wilson Non-random Factor Reference
State Based Model for Prediction of Gas Hydrate Formation Conditions in the Presence of
Electrolyte and/or Alcohol in Solution. J. Solution Chem. 2015;44(7):1382-1406.
20. Hammerschmidt E. Gas hydrate formations: a further study on their prevention and
elimination from natural gas pipe lines. Gas. 1939;15(5):30-34.
21. Mohammadi AH, Tohidi B. A novel predictive technique for estimating the hydrate
inhibition effects of single and mixed thermodynamic inhibitors. Can. J. Chem. Eng.
2005;83(6):951-961.
22. Carroll J. Natural gas hydrates: a guide for engineers: Gulf Professional Publishing, 2014.
23. McCain Jr W. Reservoir-Fluid Property Correlations-State of the Art (includes associated
papers 23583 and 23594). SPE Reserv. Eng. 1991;6(02):266-272.
24. Deaton W, Frost Jr E. Gas hydrates and their relation to the operation of natural-gas pipe
lines: Bureau of Mines, Amarillo, TX (USA). Helium Research Center;1946.
25. Sabil KM, Román VR, Witkamp G-J, Peters CJ. Experimental observations on the
competing effect of tetrahydrofuran and an electrolyte and the strength of hydrate inhibition
among metal halides in mixed CO2 hydrate equilibria. J. Chem. Thermodyn.
2010;42(3):400-408.
26. Hribar B, Southall NT, Vlachy V, Dill KA. How ions affect the structure of water. J. Am.
Chem. Soc. 2002;124(41):12302-12311.
27. Yousif M, Young D. A simple correlation to predict the hydrate point suppression in drilling
fluids. Paper presented at: SPE/IADC Drilling Conference1993.
28. Nielsen R, Bucklin R. Gas processing developments. Why not use methanol for hydrate
control. Hydrocarbon Process.;(United States). 1983;62(4).
29. Østergaard K, Tohidi B, Danesh A, Todd A. Gas hydrates and offshore drilling: predicting
the hydrate free zone. Ann. N. Y. Acad. Sci. 2000;912(1):411-419.
30. Østergaard KK, Masoudi R, Tohidi B, Danesh A, Todd AC. A general correlation for
predicting the suppression of hydrate dissociation temperature in the presence of
thermodynamic inhibitors. J. Pet. Sci. Eng. 2005;48(1):70-80.
31. Najibi H, Mohammadi AH, Tohidi B. Estimating the hydrate safety margin in the presence
of salt and/or organic inhibitor using freezing point depression data of aqueous solutions.
Ind. Eng. Chem. Res. 2006;45(12):4441-4446.
32. Hu Y, Makogon TY, Karanjkar P, Lee K-H, Lee BR, Sum AK. Gas Hydrates Phase
Equilibria for Structure I and II Hydrates with Chloride Salts at High Salt Concentrations
and up to 200 MPa. J. Chem. Thermodyn. Forthcoming 2017.
33. Hu Y, Makogon TY, Karanjkar P, Lee BR, Sum AK. The Phase Behavior and Kinetics for
Gas Hydrates Formed from High Concentration NaCl Brine at Ultra-High Pressures. J.
Chem. Eng. Data. Forthcoming 2017.
34. Cha M, Hu Y, Sum AK. Methane hydrate phase equilibria for systems containing NaCl,
KCl, and NH 4 Cl. Fluid Phase Equilib. 2016;413:2-9.
25
AIChE Journal
35. Kharrat M, Dalmazzone D. Experimental determination of stability conditions of methane

hydrate in aqueous calcium chloride solutions using high pressure differential scanning
calorimetry. J. Chem. Thermodyn. 2003;35(9):1489-1505.
36. Maekawa T, Itoh S, Sakata S, Igari S-i, Imai N. Pressure and temperature conditions for
methane hydrate dissociation in sodium chloride solutions. Geochem. J.
1995;29(5):325-329.
37. De Roo J, Peters C, Lichtenthaler R, Diepen G. Occurrence of methane hydrate in saturated
and unsaturated solutions of sodium chloride and water in dependence of temperature and
pressure. AIChE J. 1983;29(4):651-657.
38. Atik Z, Windmeier C, Oellrich LR. Experimental gas hydrate dissociation pressures for
pure methane in aqueous solutions of MgCl2 and CaCl2 and for a (methane+ ethane) gas
mixture in an aqueous solution of (NaCl+ MgCl2). J. Chem. Eng. Data.
2006;51(5):1862-1867.
39. Mohammadi AH, Afzal W, Richon D. Gas hydrates of methane, ethane, propane, and
carbon dioxide in the presence of single NaCl, KCl, and CaCl2 aqueous solutions:
Experimental measurements and predictions of dissociation conditions. J. Chem.
Thermodyn. 2008;40(12):1693-1697.
40. Hu Y, Lee K-H, Lee BR, Sum AK. Gas hydrate formation from high concentration KCl
brines at ultra-high pressures. J. Ind. Eng. Chem. 2017.
41. Haghighi H, Chapoy A, Tohidi B. Methane and water phase equilibria in the presence of
single and mixed electrolyte solutions using the cubic-plus-association equation of state.
Oil & Gas Science and Technology-Revue de l'IFP. 2009;64(2):141-154.
42. Kang S-P, Chun M-K, Lee H. Phase equilibria of methane and carbon dioxide hydrates in
the aqueous MgCl2 solutions. Fluid Phase Equilib. 1998;147(1):229-238.
43. Mohammadi AH, Kraouti I, Richon D. Methane hydrate phase equilibrium in the presence
of NaBr, KBr, CaBr2, K2CO3, and MgCl2 aqueous solutions: Experimental measurements
and predictions of dissociation conditions. J. Chem. Thermodyn. 2009;41(6):779-782.
44. Maekawa T, Imai N. Equilibrium conditions of methane and ethane hydrates in aqueous
electrolyte solutions. Ann. N. Y. Acad. Sci. 2000;912(1):932-939.
45. Hu Y, Makogon TY, Karanjkar P, Lee K-H, Lee BR, Sum AK. Gas Hydrastes Phase
Equilibrium with CaBr2 and CaBr2 + MEG at Ultra-High Pressures. J. Nat. Gas Sci. Eng. In
press 2017.
46. Tohidi B, Danesh A, Todd A, Burgass R. Hydrate-free zone for synthetic and real reservoir
fluids in the presence of saline water. Chem. Eng. Sci. 1997;52(19):3257-3263.
47. Dholabhai PD, Kalogerakis N, Bishnoi PR. Equilibrium conditions for carbon dioxide
hydrate formation in aqueous electrolyte solutions. J. Chem. Eng. Data. 1993;38(4).
48. Mallard W, Linstrom P. NIST chemistry webbook. NIST Standard Reference Database.
1998;69.
49. Englezos P, Bishnoi P. Experimental study on the equilibrium ethane hydrate formation
conditions in aqueous electrolyte solutions. Ind. Eng. Chem. Res. 1991;30(7):1655-1659.
26
AIChE Journal
50. Sandler SI. Chemical, biochemical, and engineering thermodynamics. Vol 4: John Wiley &
Sons Hoboken, NJ, 2006.
51. Pieroen A. Gas hydrates‐approximate relations between heat of formation, composition
and equilibrium temperature lowering by “inhibitors”. Recueil des Travaux Chimiques des
Pays-Bas. 1955;74(8):995-1002.
52. Hu Y, Lee BR, Sum AK. Insight into Increased Stability of Methane Hydrates at High
Pressure from Phase Equilibrium Data and Molecular Strcucture. Fluid Phase Equilib.
Forthcoming 2017.
53. Chen CC, Evans LB. A local composition model for the excess Gibbs energy of aqueous
electrolyte systems. AIChE J. 1986;32(3):444-454.
54. CSMGem [computer program]. Version 1.1: Colorado School of Mines.
55. PVTsim Nova [computer program]. Revision 1.0.176 Version: Calsep A/S.
27
AIChE Journal
Fig. 1 Experimental data for methane (■) and methane-ethane (●) hydrate suppression
temperature from Hu et al.32 and predicted results from the McCain23 (solid-red: methane;
dash-red: methane-ethane), Yousif and Young27 (solid-blue) and Østergaard et al.30 (solid-black:
methane; solid-black: methane-ethane) correlations for systems with (a) 20 wt% CaCl2 and (b)
32 wt% CaCl2.
28
AIChE Journal
Fig. 2 Illustrative plot for the hydrate suppression temperature (∆T) for brine solutions from
fresh water equilibrium dissociation temperature.
29
AIChE Journal
Fig. 3 Relationship between ∆T/T0T and temperature T for methane hydrate (a) in NaCl systems
for data from NaCl-1-26-C1 (▲-green), NaCl-1-23-C1 (▲-blue), NaCl-1-12-C1 (▲-red),
NaCl-2-10.84-C1 (△-orange), NaCl-2-5.44-C1 (△-pink), NaCl-3-11.9-C1 (☐-red),
NaCl-4-20-C1 (◇-green), NaCl-4-10-C1 (◇-crayon), NaCl-4-3.5-C1 (◇-brown),
NaCl-5-24.1-C1 (◯-yellow), NaCl-5-21.5-C1 (◯-violet), NaCl-5-17.1-C1 (◯-blue) and
NaCl-5-11.7-C1 (◯-black); (b) in CaCl2 systems for data from CaCl2-1-32-C1 (●-green),
CaCl2-1-23.8-C1 ( ● -blue), CaCl2-1-20-C1 ( ● -red), CaCl2-2-7.14-C1 ( ● -orange),
30
AIChE Journal
CaCl2-2-3.51-C1 ( ● -pink), CaCl2-3-14.5-C1 ( □ -black), CaCl2-3-10.1-C1 ( □ -brown),
CaCl2-3-5-C1 ( □ -violet), CaCl2-4-17.05-C1 (+-blue), CaCl2-5-15-C1 ( △ -red), and
CaCl2-5-5-C1 ( △ -violet); (c) in KCl systems for data from KCl-1-23-C1 ( ■ -green),
KCl-1-20-C1 (■-red), KCl-2-13.41-C1 (□-orange), KCl-2-6.83-C1 (□-pink), KCl-3-10-C1 (△
-red), KCl-3-5-C1 (△-black) and KCl-4-15-C1 (▽-blue); (d) in MgCl2 systems for data from
MgCl2-1-25 (◆-green), MgCl2-1-20-C1 (◆-blue), MgCl2-2-15-C1 (+-green), MgCl2-2-10-C1
(+-blue), MgCl2-2-5-C1 (+-black), MgCl2-3-10-C1 (▓-blue), and MgCl2-3-6-C1 (▓-red). The
dashed lines correspond to the best fit constant value.
31
AIChE Journal
Fig. 4 Universal curve relating the salt concentration, in terms of the effective mole fraction, to
the average value of ∆T/T0T (=β ln aw) for methane hydrate for single salt systems with NaCl (△),
KCl (□), CaCl2 (○) and MgCl2 (◇). Solid line is the fitted polynomial line of the data (R2 =
0.998) given by Eqn. 16.
32
AIChE Journal
Fig. 5 Experimental data of methane hydrate with NaCl aqueous solutions: NaCl-1-26-C1 (▲
-green), NaCl-1-23-C1 ( ▲ -blue), NaCl-1-12-C1 ( ▲ -red), NaCl-2-10.84-C1 ( △ -orange),
NaCl-2-5.44-C1 (△-pink), NaCl-3-11.9-C1 (□-red), NaCl-4-20-C1 (◇-green), NaCl-4-10-C1
(◇- crayon), NaCl-4-3.5-C1 (◇- brown), NaCl-5-24.1-C1 (○-yellow), NaCl-5-21.5-C1 (○
-violet), NaCl-5-17.1-C1 (○-blue) and NaCl-5-11.7-C1 (○-black), along with predicted results
from the Hu-Lee-Sum correlation in dash lines (---). The solid line represents the methane
hydrate phase equilibria with fresh water calculated by the polynomial regression given in the
Supplementary Information.
33
AIChE Journal
Fig. 6 Experimental data of methane hydrate with CaCl2 aqueous solutions: CaCl2-1-32-C1 (●
-green), CaCl2-1-23.8-C1 (●-blue), CaCl2-1-20-C1 (●-red), CaCl2-2-7.14-C1 (●-orange),
CaCl2-2-3.51-C1 ( ● -pink), CaCl2-3-14.5-C1 ( □ -black), CaCl2-3-10.1-C1 ( □ -brown),
CaCl2-3-5-C1 ( □ -violet), CaCl2-4-17.05-C1 (+-blue), CaCl2-5-15-C1 ( △ -red) and
CaCl2-5-5-C1 (△-violet), along with predicted results from the Hu-Lee-Sum correlation in
dash lines (---). The solid line represents the methane hydrate phase equilibria with fresh water
calculated by the polynomial regression given in the Supplementary Information.
34
AIChE Journal
Fig. 7 Experimental data of methane hydrate with KCl aqueous solutions: KCl-1-23-C1 (■
-green), KCl-1-20-C1 ( ■ -red), KCl-2-13.41-C1 ( □ -orange), KCl-2-6.83-C1 ( □ -pink),
KCl-3-10-C1 (△-red), KCl-3-5-C1 (△-black), and KCl-4-15-C1 (▽-blue), along with predicted
results from the Hu-Lee-Sum correlation in dash lines (---). The solid line represents the methane
hydrate phase equilibria with fresh water calculated by the polynomial regression given in the
Supplementary Information.
35
AIChE Journal
Fig. 8 Experimental data of methane hydrate with MgCl2 aqueous solutions: MgCl2-1-25-C1 (◆
-green), MgCl2-1-20-C1 ( ◆ -red), MgCl2-2-15-C1 (+-green), MgCl2-2-10-C1 (+-blue),
MgCl2-2-5-C1 (+-black), MgCl2-3-10-C1 (▓-blue) and MgCl2-3-6-C1 (▓-red), along with
predicted results from the Hu-Lee-Sum correlation in dash lines (---). The solid line represents
the methane hydrate phase equilibria with fresh water calculated by the polynomial regression
given in the Supplementary Information.
36
AIChE Journal
Fig. 9 Average absolute deviation of prediction P1 (black), P2 (blue), P3 (green) and Hu-Lee-Sum
correlation (gray) compared with experimental data of methane hydrate phase equilibria with
NaCl systems: 1. NaCl-1-12-C1, 2.NaCl-1-23-C1, 3. NaCl-1-26-C1, 4. NaCl-2-5.44-C1, 5.
NaCl-2-10.84-C1, 6. NaCl-3-11.9-C1, 7. NaCl-4-3.5-C1, 8. NaCl-4-10-C1, 9. NaCl-4-20-C1, 10.
NaCl-5-11.7-C1, 11. NaCl-5-17.1-C1, 12. NaCl-5-21.5-C1, 13. NaCl-5-24.1-C1. The symbol ☆
indicates that the method fails to estimate the hydrate phase equilibrium at those conditions.
37
AIChE Journal
Fig. 10 Average absolute deviation of prediction P1 (black), P2 (blue), P3 (green) and
Hu-Lee-Sum correlation (gray) compared with experimental data of methane hydrate phase
equilibria with CaCl2 systems: 1. CaCl2-1-20-C1, 2. CaCl2-1-23.8-C1, 3. CaCl2-1-32-C1, 4.
CaCl2-2-3.51-C1, 5. CaCl2-2-7.14-C1, 6. CaCl2-3-5-C1, 7. CaCl2-3-10.1-C1, 8. CaCl2-3-14.5-C1,
9. CaCl2-4-17.05-C1, 10. CaCl2-5-5-C1, 11. CaCl2-5-15-C1.
38
AIChE Journal
equilibria with KCl systems: 1. KCl-1-20-C1, 2. KCl-1-23-C1, 3. KCl-2-6.83-C1, 4.
KCl-2-13.41-C1, 5. KCl-3-5-C1, 6. KCl-3-10-C1, 7. KCl-4-15-C1.
39
AIChE Journal
Fig. 12 Average absolute deviation of Hu-Lee-Sum correlation (gray) compared with
experimental data of methane hydrate phase equilibria with MgCl2 systems: 1. MgCl2-1-20-C1, 2.
MgCl2-1-25-C1, 3. MgCl2-2-5-C1, 4. MgCl2-2-10-C1, 5. MgCl2-2-15-C1, 6. MgCl2-3-6-C1, 7.
MgCl2-3-10-C1. Methods P1, P2, and P3 fail to estimate the hydrate phase equilibrium at those
conditions.
40
AIChE Journal
equilibria with salt systems: 1. NH4Cl-1-10-C1, 2. NH4Cl-1-5-C1, 3. NaBr-1-10-C1, 4.
NaBr-1-5-C1, 5. NaBr-2-20-C1, 6. NaBr-2-10-C1, 7. KBr-1-10-C1, 8. KBr-1-5-C1, 9.
CaBr2-1-32-C1, 10. CaBr2-2-15-C1, 11. CaBr2-2-5-C1, 12. CaBr2-3-15-C1, 13. CaBr2-3-5-C1.
The symbol ☆ indicates that the method fails to estimate the hydrate phase equilibrium at those
conditions.
41
AIChE Journal
Hu-Lee-Sum correlation (gray) compared with experimental data of carbon dioxide hydrate
phase equilibria with salt systems: 1. NaCl-1-5-CO2, 2. NaCl-2-20-CO2, 3. NaCl-2-10-CO2, 4.
NaCl-3-20-CO2, 5. NaCl-3-15-CO2, 6. NaCl-3-10-CO2, 7. NaCl-3-5-CO2, 8. NaCl-3-3-CO2, 9.
KCl-1-10-CO2, 10. KCl-2-10-CO2, 11. KCl-3-15-CO2, 12. KCl-3-10-CO2, 13. KCl-3-5-CO2, 14.
KCl-3-3-CO2, 15. CaCl2-1-15-CO2, 16. CaCl2-2-20-CO2, 17. CaCl2-2-15-CO2, 18.
CaCl2-2-10-CO2, 19. CaCl2-2-5-CO2, 20. CaCl2-2-3-CO2, 21. MgCl2-1-10-CO2, 22.
MgCl2-1-5-CO2, 23. MgCl2-1-3-CO2. The symbol ☆ indicates the method fails to estimate the
hydrate phase equilibrium at those conditions.
42
AIChE Journal
Hu-Lee-Sum correlation (gray) compared with experimental data of ethane hydrate phase
equilibria with salt systems: 1. NaCl-1-5-C2, 2. NaCl-2-20-C2, 3. NaCl-2-10-C2, 4.
NaCl-2-3.1-C2, 5. NaCl-3-15-C2, 6. NaCl-3-10-C2, 7. NaCl-4-20-C2, 8. NaCl-5-20-C2, 9.
KCl-1-10-C2, 10. KCl-2-12.3-C2, 11. CaCl2-1-15-C2, 12. CaCl2-1-5-C2, 13. CaCl2-2-15-C2, 14.
MgCl2-1-20-C2, 15. MgCl2-1-10-C2. The symbol ☆ indicates the method fails to estimate the
hydrate phase equilibrium at those conditions.
43
AIChE Journal
Fig. 16 Universal curve to determine the hydrate suppression temperature based on the effective
mole fraction of any salt as shown in Fig. 4 with the additional values based on experimental
data for NH4Cl (★), NaBr (▲), KBr (●), and CaBr2 (■) showing the universality of the
correlation. See Fig. 4 caption for description of the open symbols.
44
AIChE Journal
Table 1. Collection of literature studies reporting methane hydrate phase equilibria with single
chloride salt solutions.
Salt Referencea Concentration Temperature Pressure Range

(wt%) Range (K) (MPa)
1. Hu et al., 2017 33 12, 23, 26 275 - 315 20 - 200

2. Cha et al., 2016 34 5.44, 10.84 281 -275 4 -12
3. Kharrat et al., 2003 35 11.9 272 - 278 4-8
NaCl
4. Maekawa et al., 1995 36 3.5, 10, 20 263 - 283 3-8
5. De Roo et al., 1983 37 11.7, 17.1, 268 -278 2-8
21.5, 24.1
1. Hu et al., 2017 32 20, 23.8, 32 260 - 300 20 - 200
2. Hu-Lee-Sumb 3.51, 7.14 276 - 285 4 - 11
CaCl2 3. Kharrat et al., 2003 35 5, 10.1, 14.5 278 - 285 5 - 11
4. Atik et al., 2006 38 17.05 264 - 282 3 - 23
5. Mohammadi et al., 2008 39 5, 15 267 - 282 3-9
1. Hu et al., 2017 40 20, 23 282 - 300 18 -180
2. Cha et al., 2016 34 6.83, 13.41 275 -285 4 - 12
KCl
3. Mohammadi et al., 2008 39 5, 10 270 - 283 2.7 -10
4. Haghighi et al., 2009 41 15 276 - 285 6 - 18
1. Hu et al., 2017 32 20, 25 265 - 290 40 - 175
MgCl2 2. Atik et al., 2006 38 5, 10, 15 268 - 288 3 - 20
3. Kang et al., 1998 42 6, 10 270 - 285 2 - 14
a
In the remainder of the paper, a particular reference will be cited abbreviated as: salt-#-conc-gas.
For example: NaCl-1-23-C1 corresponds to the methane hydrate with 23 wt% NaCl data from
Hu et al., 2017.40 b
See Table S6 in the Supplementary Information.
45
AIChE Journal
Table 2. Collection of literature studies reporting methane hydrate phase equilibria with single
salt solutions.
Conc. Temperature Pressure Range

Salt Referencea
(wt%) Range (K) (MPa)
NH4Cl 1. Cha et al., 2016 34 5,10 274 - 284 3 - 12
43
1. Mohammadi et al., 2009 5, 10 272 - 285 2 - 12
NaBr
2. Maekawa et al., 2000 44 10, 20 269 - 285 3 - 11
KBr 1. Mohammadi et al., 2009 43 5, 10 272 - 286 2 -11
1. Hu et al., 2017 45 32 278 -293 31 - 200
CaBr2 2. Hu-Lee-Sumb 5, 15 285 - 310 13 - 175
3. Mohammadi et al., 2009 43 5, 15 273 - 285 3 - 11
a
A particular reference will be cited abbreviated as: salt-#-conc-gas. For example:
CaBr2-1-32-C1 corresponds to the methane hydrate with 32 wt% CaBr2 data from Hu et al.,
2017.45 b
See Table S7 in the Supplementary Information.
46
AIChE Journal
Table 3. Collection of literature studies reporting carbon dioxide hydrate phase equilibria with
single salt solutions.
Temperature Pressure
a Conc.
Salt Reference Range Range
(wt%)
(K) (MPa)
1. Mohammadi et al., 2008 39 5 271 - 281 1-4
46
NaCl 2. Tohidi et al., 1997 10, 20 263 - 276 1-4
3. Dholabhai et al., 1993 47 3, 5, 10, 15, 20 263 - 281 1-4
39
1. Mohammadi et al., 2008 10 273 - 278 2-4
46
KCl 2. Tohidi et al., 1997 10 273 - 279 1-4
47
3. Dholabhai et al., 1993 3, 5, 10, 15 269 - 282 1-4
39
1. Mohammadi et al., 2008 15 270 - 275 2-4
CaCl2
2. Dholabhai et al., 1993 47 3, 5, 10, 15, 20 259 - 281 1-4
42
MgCl2 1. Kang et al., 1998 3, 5, 10 272 - 282 15 - 40
a
A particular reference will be cited abbreviated as: salt-#-conc-gas. For example: NaCl-1-5-CO2
corresponds to the carbon dioxide hydrate with 5 wt% NaCl data from Mohammadi et al.,
2008.39
47
AIChE Journal
Table 4. Collection of literature studies reporting ethane hydrate phase equilibria with single salt
solutions.
Temperature Pressure
Conc.
Salt Reference Range Range
(wt%)
(K) (MPa)
1. Mohammadi et al., 2008 39 5 272 - 285 0.5 - 3
38
2. Atik et al., 2006 3.1, 10, 20 268 - 287 0.7 - 3
46
NaCl 3. Tohidi et al., 1997 10, 15 272 - 280 0.8 - 3
4. Tohidi et al., 1993 48 20 266 - 272 0.6 - 2
5. Englezos et al., 1991 49 20 265 - 272 0.7 - 2
39
1. Mohammadi et al., 2008 10 270 - 282 0.5 - 3
KCl
2. Englezos et al., 1991 49 12.3 269 - 279 0.5 - 2
1. Mohammadi et al., 2008 39 5, 15 270 - 284 0.4 - 3
CaCl2 49
2. Englezos et al., 1991 15 267 - 275 0.5 - 2
NaBr 1. Atik et al., 2006 38 10, 20 269 - 285 0.6 - 3
a
A particular reference will be cited abbreviated as: salt-#-conc-gas. For example: NaCl-1-5-C2
corresponds to the ethane hydrate with 5 wt% NaCl data from Mohammadi et al., 2008.39
48
AIChE Journal
Table 5. Parameters for the Hu-Lee-Sum correlation relating the water activity to the effective
mole fraction in Eqn. 16.
Parameters C1 C2 C3
Value 9.377×10-4 -2.670×10-3 3.328×10-2
49
AIChE Journal

Universal Correlation For Gas Hydrates Suppression Temperature of Inhibited Systems: I. Single Salts

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Universal Correlation For Gas Hydrates Suppression Temperature of Inhibited Systems: I. Single Salts

Uploaded by

Copyright:

Available Formats

Thermodynamics and Molecular-Scale Phenomena AIChE Journal

Universal Correlation for Gas Hydrates Suppression Temperature

Yue Hu1, Bo Ram Lee1† and Amadeu K. Sum1*

Colorado School of Mines, Golden, CO 80401

*Corresponding author e-mail: asum@mines.edu

This article is protected by copyright. All rights reserved.

Reliable prediction of hydrate suppression temperature in presence of inhibitors, such as salts,

Herein, we propose a new correlation, to be called Hu-Lee-Sum correlation, that significantly

Topical Heading: Thermodynamics

Keywords: gas hydrates, suppression temperature, salt, brine, correlation

small and large cavities, respectively.2

in pressure build up in the flowlines.5

ppm (20 wt%) corresponding to near-saturated/saturated conditions.7 In addition, heavy brines

drilled in the offshore environment.8

concentration in solution. Therefore, it is considerably important and necessary to estimate the

production, transport and processing, as unreliable predictions of hydrate dissociation condition

important in the energy application of hydrates.

Correlations to Estimate Hydrate Suppression Temperature in Brine Systems

correlation. As a background to the development of the new correlation proposed herein, we

substance in a given amount of solvent.21

constant with a value of 1297. However, it is limited to concentrations up to about 30 wt% of

by introducing parameters dependent on the inhibitor type.22 While Hammerschmidt suggested a

of salts on hydrate inhibition is equally important. McCain23 developed a correlation for

specific gravity and salt concentration.

∆T = 112 .3 x + 2011 .6 x 2 − 6505 .0 x 3 (3)

depression temperature. The correlation is developed based on experimental data of drilling

costly and time consuming.

single and mixed thermodynamic inhibitors, given by

weight percentage of salt i in the aqueous solutions, respectively. This correlation is a

experimental data and predicted results at higher pressure.

correlation, and proposed a modified correlation considering the pressure dependence,29,30

In addition, Najibi et al.31 developed a systematic method to predict hydrate suppression

point depression of aqueous solutions containing different concentration of inhibitors:

aqueous solution is necessary for use of the correlation.

considerably underestimates the suppression temperature. Although the McCain correlation23

temperature and experimental data can be relatively large.

suppression temperature but is also solidly based on the fundamental principles of

inhibition effect for different salts.

Derivation of the Hu-Lee-Sum Correlation

f ws (T , P) = f wL (T , P) = xwγ w (T , P, x) f wL0 (T , P) (7)

Gibbs energy of fusion,

can combine Eqns. 7 and 12 to obtain

where aw represents the water activity.

depression temperature (∆T = T0 – T) can be expressed by51

with fresh water is assumed to be known and readily available.

the value of ∆T/T0T can be calculated by knowing the water activity.

Salt Species and Concentration Dependence

eNRTL53, which uses the effective mole fraction, X = ∑z

correlation for the hydrate suppression temperature.

Results and Discussion

temperature also approaches zero.

in temperature, defined as:

are summarized in Figures 9 to 12 and Tables S8 to S11 in Supplementary Information,

phase equilibria with 26 wt% NaCl up to 200 MPa.

correlation successfully provides accurate hydrate dissociation temperature at this condition.

Hu-Lee-Sum correlation, respectively.

data for these few single salt systems.

analysis to obtain the model parameters.

approximated to be independent of temperature, as shown in Figure 3 by the nearly horizontal

details in Supplementary Information).

inhibitors, will be presented in follow-up papers.

helpful discussions over the course of this project.

an explicit pressure-dependent Langmuir adsorption constant in the van der Waals–