Journal of Natural Gas Science and Engineering 73 (2020) 103053

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering

journal homepage: http://www.elsevier.com/locate/jngse

Effect of polymer aggregation on the kinetics of hydrate formation

�rbara Louise Lemos Drumond Silva a, b, Isabella Leite Ferraz b, Jos�e Adilson de Castro a,
Ba
Letícia Vitorazi b, *
a
Laborat�
orio Multiusu�
ario de Caracterizaç~ ao Macrosc�
opica de Fluídos e Sistemas Multif�
asicos, Programa de P�
os-Graduaç~ao em Engenharia Metalúrgica, Universidade
Federal Fluminense, Volta Redonda, Rio de Janeiro, 27255-125, Brasil
b
Laborat�
orio de Materiais Polim�ericos, Departamento de Engenharia Metalúrgica e Materiais, Universidade Federal Fluminense, Volta Redonda, Rio de Janeiro, 27255-
125, Brasil

A R T I C L E I N F O A B S T R A C T

Keywords: Previous studies have highlighted the importance of low-dosage hydrate inhibitors on the kinetic inhibition of
Kinetic hydrate inhibitor gas hydrates. Here, we investigated the polymers poly(diallyldimethylammonium chloride) (PDADMAC) and
Polyelectrolyte poly(acrylic acid) (PAA) as well as their mixtures on the kinetics of hydrate formation. The results show that the
Polyelectrolyte complex
polymer mixtures considerably improved the inhibition of hydrate formation. The best inhibition occurred for
the mixtures with an excess of PAA2k containing aggregates of around 100 nm in size. The fraction of hydrates
formed over time was strongly reduced in the presence of the mixtures, reaching values of accumulated gas
consumption of 0.0831 mol for a molar charge ratio Z-/þ ¼ 0.67 compared with 0.6025 mol for water experi­
ments after 15 h. The Johnson–Mehl–Avrami–Kolmogorov (JMAK) phenomenological model was implemented
to predict the fraction of transformed hydrate and determine the crystal growth dimensionality. The range of
values for the Avrami exponent was 0–1.36, indicating diffusion-controlled one-dimensional growth for methane
characterized by the thickening of particles with an appreciable initial volume.

1. Introduction hydrates and the related deferments may account for up to 8% of their
operational expenditure. Thus, understanding the kinetics of hydrate
Gas hydrates are crystalline compounds similar to ice in which nat­ formation has great importance for the oil industry as it is needed to
ural gas is trapped within the water molecular structure at high pres­ maintain the unrestricted flow of oil and gas into pipelines (Tang et al.,
sures and relatively low temperatures (S. Xu et al., 2016). Hydrates can 2017).
be classified into three main structures: types I, II, and III. In general, it Traditionally, hydrate formation has been prevented by using ther­
has been shown that the crystal structure type mainly depends on the modynamic inhibitors that can shift the equilibrium conditions of hy­
size of the gas molecule located at the core of the hydrate (Ripmeester drate formation by i) lowering the system temperature or ii) increasing
and Alavi, 2016). The gas hydrate structures formed by methane have the system pressure. Methanol, glycol, or ethanol at high dosages
been predominantly observed as type I (Xu et al., 2010). (Daraboina et al., 2011) are examples of thermodynamic inhibitors.
In oil and gas systems, gas hydrate particles can agglomerate, grad­ However, these chemicals are disadvantageous economically and envi­
ually accumulate, and form hydrate plugs, which can block flow lines, ronmentally, and there are technical problems related to their use at
valves, chokes, and other production equipment (Sadeq et al., 2017). high dosages. As alternatives, there are kinetic inhibitors (KIs), which
Furthermore, gas hydrate formation often leads to production defer­ are low-dosage hydrate inhibitors that delay the formation of hydrates
ment, environmental damage, and process safety concerns (AlHarooni and crystal growth; that is, KIs increase the time required for hydrate
et al., 2015). According to a previous study (AlHarooni et al., 2015), the formation beyond the residence time of the gas in hydrate-prone sec­
offshore oil and gas industry expenditures due to the formation of tions of the transport facility (Daraboina et al., 2013). KIs are

Abbreviation: JMAK, Johnson–Mehl–Avrami–Kolmogorov; KI, kinetic inhibitor; PAA, poly(acrylic acid); PANa, poly(sodium acrylate); PDADMAC, poly(dia­
llyldimethylammonium chloride); PEC, polyelectrolyte complex.
* Corresponding author.
E-mail addresses: barbara.drumond@ubm.br (B.L. Lemos Drumond Silva), isabellaleiteferraz@id.uff.br (I. Leite Ferraz), joseadilsoncastro@id.uff.br (J.A. de
Castro), leticiavitorazi@id.uff.br (L. Vitorazi).

https://doi.org/10.1016/j.jngse.2019.103053
Received 21 May 2019; Received in revised form 8 October 2019; Accepted 2 November 2019
Available online 5 November 2019
1875-5100/© 2019 Elsevier B.V. All rights reserved.
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

water-soluble polymers, and their induction time or delay time usually addition of a PDADMAC100k solution (by titration) into PANa2k (both
depends on the degree of subcooling. Subcooling is the temperature solutions at pH 10), first, negatively charged aggregates form at a low
difference between the hydrate dissociation temperature and operating molar ratio (Zþ/- < 1) with a medium diameter of 150 nm. At around
temperature at a given pressure (Kelland, 2009). Some examples of KIs stoichiometric quantities of charges (Zþ/- ~ 1), a phase separation called
are the natural polymers chitosan (Xu et al., 2010), pectin (P. Xu et al., coacervation occurs with micro-sized droplets. Finally, at higher molar
2016; S. Xu et al., 2016), tapioca (Lee et al., 2007), amino acids (Bhat­ ratios (Zþ/- > 1), positively charged aggregates form that also contain
tacharjee et al., 2016; Sa et al., 2016, 2013), ice-structuring proteins micro-sized droplets. On the other hand, adding PANa2k to a PDAD­
(Jensen et al., 2011), and antifreeze proteins (Walker et al., 2015; Per­ MAC100k solution forms positively charged aggregates that become a
feldt et al., 2014; Daraboina et al., 2015, 2011), which are currently the coacervate phase, followed by the formation of negatively charged
focus of various studies. These natural products contribute to the inhi­ species (Vitorazi et al., 2014).
bition of hydrate formation. Synthetic polymers can also act as KIs, and Investigations conducted using one-shot mixing procedures have
examples include polyvinylpyrrolidone (Abrahamsen and Kelland, shown that the aggregation and coacervation processes of PANa2k/
2018; Choudhary et al., 2016; Ke et al., 2016), polyvinyl caprolactam PDADMAC100k (using solutions at pH 10) are not completely indepen­
(Abrahamsen and Kelland, 2018; Jensen et al., 2011), and poly dent and that some coacervates occur below a molar ratio of 1, with
(3-methylene-2-pyrrolidone) (Abrahamsen et al., 2017). formation of soluble PECs at Z � 0.6, insoluble 100-nm PECs at Z > 0.6,
In recent years, many studies have attempted to elucidate the ther­ and droplets of coacervates at Z~1.1 (Liu et al., 2016, 2017).
modynamic and kinetic conditions of the hydrate formation process. It is worth noting that pH is another important parameter to be
This increase in research interest may be due to the broad applicability considered regarding weak polyelectrolyte complexation. According to
of such knowledge to new technologies. For example, gas traps for Alonso et al. (2013), the complexation between PAA and PDADMAC is
carbon dioxide, which are motivated by environmental issues, represent an enthalpically driven process in acidic media at pH 3 and 6 and an
an important topic that may even become a new method to exploit entropically driven process in basic media at pH 10. Because the ener­
natural gas. Hydrogen storage and transportation are other examples getics of the process are dependent on pH, it is expected that pH also
where hydrates can be technologically used (Shi et al., 2009; Taheri influences the structures of the complexes. To the best of our knowledge,
et al., 2014). no study has described the use of these polymer mixtures in the form of
Despite the various efforts toward studying KIs, little is known about polymer aggregates in the field of gas hydrates or to inhibit or store gas
the influence of polymer aggregation state on methane hydrate forma­ hydrates.
tion. Thus, we investigated the effects of mixtures of PAA2k/PDAD­
MAC100k [PAA, poly(acrylic acid); PDADMAC, poly 2. Materials and methods
(diallyldimethylammonium chloride)] on gas hydrate formation to un­
derstand and quantify the formation mechanism. Additionally, the effect 2.1. Materials
of the size of the PAA and PDADMAC100k polymer chains was evaluated.
With our results, we intend to establish a link between polymer aggre­ PAA2k (powder, MW ¼ 2100 g mol 1), PAA100k (aqueous solution
gation state and hydrate formation. 35 wt%, MW ¼ 100000 g mol 1), and PDADMAC100k (MW <
PDADMAC is a strong cationic polyelectrolyte, whereas PAA and 100000 g mol 1) were purchased from Sigma-Aldrich and used without
PANa [poly(sodium acrylate)] are weak polyelectrolytes, where the further purification. All solutions were prepared in distilled water, fol­
degree of protonation is related to the aqueous solution pH (Fig. 1). The lowed by pH correction with NaOH for experiments performed at pH 10.
protonated form, PAA, predominates in acidic conditions, whereas the Methane gas (code N45) was acquired from Air Liquide.
carboxylate form, PANa, predominates in basic conditions. The sodium
salt is shown in Fig. 1 to represent a hypothetical case where the solution
pH is corrected using sodium hydroxide. The pKa value of PAA is around 2.2. Polymer solution and one-shot mixing
pH 4.5; at this pH, there is an equilibrium between equivalent amounts
of acid and acrylate groups (Swift et al., 2016). Polymer solutions: PAA2k as a powder and PAA100k or PDAD­
Association structures between PANa and PDADMAC can form so­ MAC100k as solutions were dissolved in or diluted by distilled water to
lutions containing both species through the weak electrostatic forces of achieve 1% (wt./vol) solutions. The solutions were used in water alone
ion-pairing interactions since they are oppositely charged poly­ without pH adjustment or were pH corrected by the addition of a
electrolytes. These association structures are called polyelectrolyte concentrated NaOH solution for experiments performed at pH 10.
complexes (PECs), and different types of PECs can be achieved in Mixture of polymer solutions using one-shot mixing protocol:
aqueous solutions depending on, for example, the molar ratio of the Appropriate volumes of stock solutions of PDADMAC100k at 1% (wt./
polymer mixture. The molar ratio is normally expressed as the ratio of vol) and pH 5.7 were added via one-shot mixing to solutions of the
the molar concentration of the negatively charged species to that of the oppositely charged polymer, PAA2k, also at 1% (wt./vol) and pH 2.9 to
positively charged species, i.e., Z-/þ ¼ [PANa]/[PDADMAC], or vice prepare mixtures with Z-/þ and Zþ/- molar ratios, where Z-/þ indicates
versa, Zþ/- ¼ [PDADMAC]/[PANa]. In this expression, the molar mass of the addition of PAA2k into PDADMAC100k and Zþ/- indicates the addition
the repeat unit n-mer is used for the molar concentration of PANa or of PDADMAC100k into PAA2k as described in Fig. 2. Mixtures at Z-/
PDADMAC (Vitorazi et al., 2014). þ ¼ [PAA]/[PDADMAC] ¼ 0.5, 1, and 1.5 and Zþ/- ¼ [PDADMAC]/
Additionally, the preparation of PAA and PDADMAC mixtures plays [PAA] ¼ 0.67, 1, and 2 were prepared, where [PAA] is the molar con­
an important role in the formation of their PECs. For the stepwise centration of the negatively charged PAA n-mer and [PDADMAC] is the
molar concentration of the positively charged PDADMAC n-mer.
Table 1 lists the volumes of solutions used for each polymer solution
(VPDADMAC100k or VPAA2k) at 1% (wt./vol) to obtain a total volume of
260 mL, the final concentration of each polymer (CF, PDADMAC100k and CF,
PAA2k), and the total polymer concentrations (CF, total (%)) for the
different Z values tested.
Mixture of polymer solutions in aqueous NaCl using one-shot
mixing protocol: Mixtures of PAA2k and PDADMAC100k at 1% (wt./
vol) were prepared in aqueous NaCl (3.5% wt./vol) by one-shot addition
Fig. 1. Chemical structures of PAA, PANa, and PDADMAC. of PAA2k into PDADMAC100k at a molar ratio of Z-/þ ¼ 1.

2
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

is composed of a cylindrical body made of Inconel alloy C-276

(~290 cm3), a stirring paddle that maintains contact between the water
and gas, and a cooling jacket that maintains a stable temperature inside
the cell (Fig. 3). The gas was supplied by a high-pressure vessel con­
nected to the cell. Coupled to the cell, there is also a control tower for
regulating the temperature, pressure, and stirring.
The polymers solutions (PAA2k, PAA100k, and PDADMAC100k with or
without pH correction), mixtures of polymers at different charge ratios
(Zþ/- or Z-/þ), and aqueous NaCl polymer mixture (Z-/þ ¼ 1) were added
directly to a high-pressure cell for hydrate experiments. The experiments
were conducted using methane as the gas for hydrate formation at a
stirring rate of 500 RPM, temperature of 4 � C, and pressure of 100 bar.
Tests were performed by the isochoric method: after temperature sta­
Fig. 2. Representative one-shot mixing schemes for the studied mixtures.
bilization of the solution, the cell was pressurized to 100 bar and closed.
Hydrate formation was identified by a pressure decrease and tempera­
Table 1 ture increase because methane is consumed to form hydrate cavities,
Parameters for one-shot mixing. and thus it is an exothermic process. Data were acquired every 10 s over
Z VPDADMAC100k, VPAA2k, CF, PDADMAC100k/CF,
a period of 12 h. The hydrates formed in the presence of the methane gas
1% 1% PAA2k
(mL) (mL) (%) are expected to have a type I structure (Moon et al., 2003).
¡/þ 0.5 213 47 0.18/0.82
1 180 80 0.69/0.31 3. Results and discussion
1.5 156 104 0.60/0.40
þ/¡ 2.0 213 47 0.18/0.82 3.1. Hydrate formation in the presence of synthetic polymers
1 180 80 0.69/0.31
0.67 156 104 0.60/0.40
Some solutes in water can act as kinetic hydrate inhibitors, which
modify the hydrate formation kinetics by delaying the process without
2.3. Light scattering measurements of PAA2k/PDADMAC100k mixtures changing the thermodynamics of the formation process (Ribeiro and
Lage, 2008). Fig. 4 shows typical pressure versus time curves, where
Light scattering measurements were performed using a NanoZS hydrate formation is indicated by the decrease in pressure owing to the
Zetasizer (Malvern Instruments) at a detection angle of 173� . Measure­ gas that is consumed to accumulate in the bulk in the form of “ice.”
ments were performed in triplicate at 25 � C. For these experiments, the These curves were acquired for samples of pure water or water with 1%
total volume (Vtotal) from Table 1 was fixed at 2 mL, and the pH of the (wt./vol) PDADMAC100k, PAA2k, or PAA100k with or without pH
PAA2k and PDADMAC100k solutions before mixture was 2.86 and 5.7, correction, methane at an initial pressure of 100 bar, and at a temper­
respectively. Additionally, for comparison, the aqueous NaCl polymer ature of 4 � C.
solution of PAA2k added to PDADMAC100k at a molar ratio of Z-/þ ¼ 1 In all curves shown in Fig. 4, the pressure decreases as a function of
was also tested. The results are expressed as size distribution curves time due to the incorporation of gas molecules into water driven by the
according to intensity and Zaverage (nm). gas hydrate formation. The results for pure water showed that the
pressure stabilized at 40 bar after 4 h. The solutions containing the
studied polymers resulted in lesser pressure decreases, reaching around
2.4. Experimental study of hydrate formation 60–80 bar for the same period of 4 h. These experimental results
demonstrate that hydrate formation is inhibited in the presence of the
The hydrate formation experiments were performed using a high- studied polymers. The use of PAA2k and PAA100k solutions at pH 10
pressure cell (Hydrafact, United Kingdom) with a coupled refriger­ resulted in a greater reduction in hydrate formation than the use of a
ated/heating circulator F25 (JULABO, Germany). The high-pressure cell

Fig. 3. Schematic diagram of the high-pressure cell.

3
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

Fig. 4. Pressure versus time plots for pure water and aqueous solutions of (a) PDADMAC100k, (b) PAA2k, and (c) PAA100k at 4 � C.

PDADMAC100k solution at the same pH. to the deprotonated COO- groups as they are more solvated by water
For PDADMAC, the pH of the solution does not affect the charge of than the COOH groups. The results for PAA100k showed the greatest
the polymer. The quaternary ammonium groups are completely disso­ difference in degree of inhibition between the pH-adjusted and unad­
ciated in solution since it is a strong polyelectrolyte. A pH value of 5.7 justed solutions for all observed times. This may be associated with the
was observed for the PDADMAC solution at 1% (wt./vol) before molar mass in addition to the COO-/COOH group protonation state
adjustment. In contrast, PAA is a weak polyelectrolyte, and thus the (Alonso et al., 2013).
degree of ionization is directly affected by the pH of the medium. Before A comparison between solutions containing PDADMAC100k and PAA
adjustment, the pH values of the PAA2k and PAA100k solutions at 1% with different molar masses (2 k and 100 k) in distilled water and with
(wt./vol) were 2.9 and 2.5, respectively. pure water is shown in Fig. 5-a, and the same polymers at pH 10.0 in
Comparing the PDADMAC100k and PAA2k curves (Fig. 4), at the distilled water are compared in Fig. 5-b. PAA100k resulted in the greatest
beginning of the process up to 3 h, there is no difference in the behaviors impact on hydrate formation, followed in sequence by PDADMAC100k
of the PDADMAC100k curves at pH 10 and 5.7. In the same interval, there and PAA2k. PAA100k at pH 10 resulted in a lower pressure drop as it
is a greater difference in inhibition for PAA2k, which may be attributable maintained the pressure above 80 bar up to approximately 5 h. These

Fig. 5. Pressure versus time plots for pure water and aqueous solutions of polymers with (a) unadjusted pH and (b) pH 10 at 4 � C.

4
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

results indicated that the carboxylic groups of PAA are more effective for complexes, resulting in species of approximately 20 nm in diameter as
hydrate inhibition than the quaternary ammonium groups of PDAD­ expected due to the salt shielding the electrostatic interactions.
MAC, although molar mass also plays an important role. Table 2 shows the size data expressed as Zaverage (d.nm) for the
The presence of a kinetic inhibitor disturbs the system due to the mixtures shown in Fig. 7. Zaverage varied with molar ratio, with a
structure of the macromolecule and consequently decreases the methane maximum of 110.9 nm for Zþ/- ¼ 1 and minimum of 55.1 nm for Z-/þ ¼ 1
consumption relative to the experiments in distilled water. PAA contains prepared in aqueous NaCl at 3.5%. The latter resulted in a high standard
carboxylic groups that can disturb the water hydrogen bridges around error due to the presence of few aggregated species, which increased the
hydrates as described previously by S. Xu et al. (2016) using pectin. average size values. The values of Zaverage for triplicate measurements of
PDADMAC contains quaternary ammonium groups that are positively Z-/þ ¼ 1 (NaCl at 3.5%) were 132.7, 16.45, and 16.29 nm.
charged regardless of the solution pH. In contrast, PAA in basic condi­ For the Z-/þ experiments, the aggregate size increased with
tions is negatively charged due to the deprotonation of carboxylic acid increasing addition of PAA2k, and the aggregates formed by the addition
groups. Both PAA and PDADMAC have solvation water that disturbs the of PDADMAC100k to PAA decreased in size with increasing Zþ/-. Note
organization of water molecules around the hydrates. The process of that aggregates were not observed on a micrometric scale which did not
PAA and PDADMAC association studied by calorimetry appeared to be cause a strong liquid-liquid phase separation for these studies at acidic
event-driven by an increase in entropy due to the liberation of solvation pH. This result is important because the size of the aggregates in the
water around the polyelectrolytes in basic conditions (Alonso et al., range used in this study did not promote faster phase separation with the
2013). formation of a gel in the bottom of the cell after exposition at 100 bar.
The gas uptakes for the samples containing PDADMAC100k, PAA2k, This phase separation previously occurred for this system when the
and PAA100k at 1% (wt./vol) and pure water at 4 � C without and with pH experiment was performed at pH 10, and the gel phase can be separated
adjustment to 10 are shown in Fig. 6-a and b, respectively. The amounts by centrifugation, mainly at Z ~1 (Vitorazi et al., 2014).
of gas consumed during methane hydrate formation also demonstrated In this study, the formation of methane hydrate in the presence of the
that PAA100k is more effective as an inhibitor than the other polymers PAA2k and PDADMAC100k mixtures was significantly reduced compared
and that adjustment to pH 10 intensified this effect. with the hydrate formation in pure water (Fig. 8-a and b). The results
show that the inhibition was more efficient for the PAA2k and PDAD­
MAC100k mixtures at molar ratios of Z-/þ ¼ 1.5 and Zþ/- ¼ 0.67, as
3.2. Hydrate formation in the presence of PDADMAC100k/PAA2k observed at the expanded regions from the respective graphics. These
mixtures mixtures maintained the pressure above 88 bar and 92 bar, respectively,
for up to approximately 10 h.
Here, we observed that the complexation of PAA2k and PDAD­ A comparison of the PAA2k/PDADMAC100k mixtures and pure solu­
MAC100k at pH ~3 obtained by a one-shot mixing procedure yielded tions of PDADMAC100k, PAA2k, and PAA100k is presented in Fig. 9. The
complexes that did not separate into a gel phase at the bottom of the PAA2k/PDADMAC100k sample afforded a better inhibition effect and a
pressure cell during gas hydrate formation experiments performed at lower gas uptake. The PAA2k and PDADMAC100k mixtures formed rela­
100 bar. However, the mixtures formed by PAA100k and PDADMAC100k tively large particle aggregates with hydrophobic nuclei. Both of these
at pH ~3 separate into a gel phase at the bottom of the pressure cell characteristics favor disturbance of the system, which decreases the
during the experiment at 100 bar and, this mixture was not able to be formation rate of crystalline structures that encapsulate the gas to form
evaluated as hydrate inhibitor. stable complexes.
The polymer mixtures of PAA2k/PDADMAC100k were characterized These results suggest PAA2k and PDADMAC100k mixtures as potential
at different charge ratios by light scattering. These experiments were new hydrate inhibitors. Interestingly, all samples containing PAA in
performed to understand the effect of polymer charge ratio on the excess were better kinetic inhibitors. It was observed that the surface of
structure of the complexes at the investigated pH values. It is known that the aggregates were negatively charged, indicating that carboxylic
the pH of the solution can affect the charge of the polymer, thermody­ groups exist on the aggregate surfaces at an excess of PAA (Vitorazi
namics of association (Alonso et al., 2013) and consequently on the et al., 2014). These results corroborate with those of S. Xu et al. (2016),
aggregates formed in solution. The PAA2k/PDADMAC100k mixtures were where the carboxylic groups were responsible for preventing hydrate
characterized by dynamic light scattering, and the size distribution formation.
curves according to intensity are shown in Fig. 7. The aqueous solutions Aggregates are formed by electrostatic interactions and are dissolved
of polymers at different molar charge ratios (Z) yielded complexes in the in the presence of high salt concentrations (Yan et al., 2011), which
size range of 50–100 nm (Fig. 7-a). For the mixture in aqueous NaCl at discourages the use of such mixtures as kinetic hydrate inhibitors for
3.5% and a molar ratio of Z-/þ ¼ 1 (Fig. 7-b), the salt disintegrated the

Fig. 6. Gas uptake curves during methane hydrate formation in aqueous solutions at 1% (wt./vol), 100 bar, and 4 � C for (a) unadjusted pH and (b) pH 10.

5
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

Fig. 7. Size distributions by intensity of the PAA2k/PDADMAC100k mixtures.

Usually, Eq. (3) is also applied in its linearized form, which is given in
Table 2
Eq. (4).
Values of Zaverage (d.nm) from dynamic light scattering for different charge ratios
(Z). k ¼ k0 ⋅ expð Ea =RTÞ
; (3)
Z Zaverage (d.nm) Standard error (�)
lnðkÞ ¼ lnðk0 Þ Ea = RT; (4)
Z-/þ ¼ 0.5 58.6 1.4
Z-/þ ¼ 1 77.1 3.0
Z-/þ ¼ 1.5 107.6 3.5 where k0 is the frequency factor, Ea is the activation energy, R is the
Zþ/- ¼ 0.67 109.6 3.6 universal gas constant, and T is the temperature.
Zþ/- ¼ 1 110.9 1.2 According to Weinberg (1999), the JMAK equation makes explicit or
Zþ/- ¼ 2 65.2 0.6 implicit use of assumptions such as infinite system size, uniform nucle­
Z-/þ ¼ 1 (NaCl 3.5%) 55.1 67.2
ation, and growth restriction at points of impingement forming
dispersed particles. We assumed that these features are compatible with
offshore gas field applications. However, the results reported here are the formation mechanisms of different hydrate structures. The JMAK
promising for the development of a new system that can form nanoscale model is also suitable for determining the predominant mechanism and
aggregates by other types of weak interactions such as hydrophobic nuclei shape that is dynamically formed during the transformation in
interactions, crystalline structure, or even nanostructures formed by a addition to the growth rate.
reticulated polymer that can support the salt screening.
3.3.2. Analysis of fraction of transformed hydrate using JMAK model
The classical JMAK approach described in Section 3.3.1 was applied
3.3. Johnson–Mehl–Avrami–Kolmogorov (JMAK) model and analysis of to predict the fraction of gas hydrate formed at 277 K and 100 bar. The
fraction of transformed hydrate stages of hydrate crystal growth were analyzed, which are limited by the
dominant resistance as shown in Fig. 10-a and b.
3.3.1. JMAK model The parameters k and n of the JMAK equation (Eq. (1)) were obtained
The acquired data was analyzed using the JMAK model, which is a by piecewise linear adjustment and are listed in Tables 3 and 4 for the
phenomenological model based on the assumption of spatially random solutions. The n exponent values are in the range of 0–1.36, which imply
nucleation considered as a constant and a constant growth rate applied diffusion-controlled one-dimensional growth characterized by the
to describe the overall crystallization kinetics under isothermal and non- thickening of particles with an appreciable initial volume (Rios, 2007).
isothermal conditions. Crystal growth rates can be modeled in terms of Non-integer values of n were obtained in some cases because the
nucleation and diffusion-limited growth by parametrical models for nucleation of hydrates can be between instantaneous and sporadic
nucleation and fundamental models for growth that allow for the pre­ (Kumar et al., 2016). Suitable fits were confirmed by the standard error
diction of time-dependent crystallization (Rios, 2007; Song and Stewart, and adjusted R2. Fig. 11-a and b also illustrate the excellent fitting of the
2018). experimental data to the JMAK model for the PAA2k/PDADMAC100k 1%
The JMAK model has been widely used to interpret crystallization (wt./vol) mixtures at Z-/þ ¼ 1.5 and Zþ/- ¼ 0.67.
experiments and infer crystallization mechanisms due to its universality
and applicability to any material type (Weinberg, 1999). The essence of 4. Conclusion
the model, which was initially proposed for isothermal transformations
in solids, can be summarized by the relation in Eq. (1) (Rios, 2007). This In this work, we investigated the inhibitory effects of PDADMAC and
model follows a sigmoidal fit and is most commonly used in the line­ PAA and their mixtures on the formation of gas hydrates. For the PAA
arized form given in Eq. (2) (Rios, 2007). polymer alone, the molar mass affected the inhibition efficiency and this
behavior was improved with the deprotonation of the weak polymer.
f ðtÞ ¼ 1 expð ðk ⋅ tn ÞÞ
; (1)
This effect was weaker when the polymer was a strong polyelectrolyte.
Polymers complexes have great potential as kinetic hydrate in­
ln lnð1=ð1 f ðtÞ Þ Þ ¼ lnðkÞ þ n lnðtÞ; (2)
hibitors, but the complexes formed by electrostatic interactions did not
where t is the effective time, n is the Avrami exponent, and k is the ki­ support the presence of salt for application in saltwater. Thus, the
netic rate constant. polymer should be modified to improve the strength of the complex; this
The kinetic rate constant is dependent on temperature and is would also prevent the strong phase separation.
assumed to follow an Arrhenius-type relation as given by Eq. (3). The fraction of hydrate formed was greatly reduced in the presence

6
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

Fig. 8. Pressure versus time plots for PAA2k and PDADMAC100k mixtures at different molar ratios compared with water (a and b). On the right, are presented the
expanded regions from each graphic.

gas consumed in the experiments with water alone. The polymers PAA2k,
PAA100k, and PDADMAC100k also had efficient inhibitory effects, mainly
at pH 10, reaching values of accumulated gas consumption of
0.5285 mol, 0.2966 mol, and 0.4061 mol, respectively. These values
were determined by the amount of hydrate formed at the end of the
experiment.
The results also showed that the mixtures of PAA2k/PDADMAC100k at
1% (wt./vol), mainly at molar ratios of Z-/þ ¼ 1.5 and Zþ/- ¼ 0.67, are
the most promising kinetic hydrate inhibitors among the studied mix­
tures. Both systems resulted in polymer complexes of around 100 nm in
size, and, interestingly, both resulted in a more negatively charged
polymer in the mixture.
The JMAK model was applied to the experimental results to predict
the fraction of transformed hydrates. The Avrami exponent (n) provided
by the model fitting was used as an indicator of the homogeneity of the
nucleation conditions and of the crystal growth dimensionality. The n
values were in the range of 0–1.36, which implies diffusion-controlled
one-dimensional growth characterized by the thickening of particles
with an appreciable initial volume. Non-integer values of n were also
Fig. 9. Gas uptake curves during methane hydrate formation in pure solutions found in some cases, illustrating that the nucleation of the formed hy­
and mixtures at 1% (wt./vol), 100 bar, and 4 � C. drates occurs anywhere from instantaneously to sporadically. The model
fitting of the experimental results presented a good statistical correlation
of the investigated polymers, reaching values of accumulated gas con­ with the experimental data, which confirms the applicability of the
sumption of 0.0831 mol and 0.1367 mol after 15 h for the mixtures at Z-/ proposed method of analysis.
þ ¼ 0.67 and Zþ/- ¼ 1.5, respectively, compared with the 0.6025 mol of

7
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

Fig. 10. Two-stage Avrami plots for PAA2k/PDADMAC100k 1% (wt./vol) mixture at (a) Z-/þ ¼ 1.5 and (b) Zþ/- ¼ 0.67.

Table 3 Table 4
Quantitative parameters adjusted to describe the hydrate formation kinetics. Quantitative parameters adjusted to describe the hydrate formation kinetics.
Stage Parameter Value Standard Adj. Stage Parameter Value Standard Adj. R2
error R2 error

PAA2k 1% pH not 1 k 0.596 1.09E-02 0.996 Z-/þ ¼ 0.5 1 k 0.065 5.96E-02 0.991
adjusted n 0.929 5.04E-03 n 0.354 1.29E-02
2 k 0.280 1.15E-03 0.978 2 k 0.120 4.69E-03 0.996
n 0.287 9.65E-04 n 0.576 2.71E-03
1–2 k 0.246 4.22E-03 0.876 3 k 0.097 6.67E-04 0.995
n 0.410 3.32E-03 n 0.351 3.37E-04
PAA2k 1% pH 10 1 k 0.583 3.07E-02 0.995 1–3 k 0.093 1.06E-03 0.989
n 0.843 9.51E-03 n 0.373 5.37E-04
2 k 0.109 2.05E-03 0.984 Z-/þ ¼ 1.0 1 k 0.268 2.80E-01 0.915
n 0.113 1.42E-03 n 0.924 7.03E-02
3 k 0.331 1.02E-02 0.995 2 k 0.105 7.35E-03 0.996
n 1.357 1.35E-02 n 0.595 3.88E-03
4 k 0.186 1.21E-03 0.991 3 k 0.073 7.22E-04 0.989
n 0.474 1.01E-03 n 0.252 3.66E-04
1–4 k 0.184 1.67E-03 0.984 1–3 k 0.067 2.23E-03 0.927
n 0.484 1.32E-03 n 0.294 1.13E-03
PAA100k 1% pH not 1 k 0.526 5.55E-02 0.993 Z-/þ ¼ 1.5 1 k 0.170 3.90E-02 0.959
adjusted n 1.027 1.60E-02 n 0.802 1.64E-02
2 k 0.156 4.28E-04 0.999 2 k 0.064 1.41E-03 0.968
n 0.523 3.54E-04 n 0.285 7.15E-04
1–2 k 0.151 1.94E-03 0.984 1–2 k 0.061 2.22E-03 0.932
n 0.557 1.53E-03 n 0.305 1.12E-03
PAA100k 1% pH 10 1 k 0.794 6.41E-02 0.988 Z-/þ ¼ 1.5 (NaCl 1 k 0.433 6.67E-02 0.980
n 1.071 1.95E-02 3.5%) n 0.696 1.89E-02
2 k 0.124 4.11E-04 0.998 2 k 0.110 8.94E-03 0.918
n 0.352 3.41E-04 n 0.130 4.86E-03
1–2 k 0.119 2.56E-03 0.946 3 k 0.129 9.55E-04 0.980
n 0.392 2.02E-03 n 0.247 4.85E-04
PDADMAC100k 1% pH 1 k 0.272 5.10E-02 0.984 1–3 k 0.127 1.27E-03 0.967
not adjusted n 0.841 1.62E-02 n 0.255 6.44E-04
2 k 0.431 2.58E-02 0.992 Zþ/- ¼ 0.67 1 k 0.075 2.52E-02 0.945
n 1.106 1.51E-02 n 0.489 1.10E-02
3 k 0.180 5.91E-04 0.998 2 k 0.052 9.15E-04 0.964
n 0.531 4.98E-04 n 0.174 4.65E-04
1–3 k 0.171 1.99E-03 0.985 1–2 k 0.049 1.65E-03 0.917
n 0.584 1.56E-03 n 0.204 8.35E-04
PDADMAC100k 1% pH 10 1 k 0.423 1.51E-02 0.991 Zþ/- ¼ 1.0 1 k 0.158 4.52E-02 0.966
n 0.894 7.09E-03 n 0.709 1.63E-02
2 k 0.172 2.18E-04 0.942 2 k 0.059 1.00E-03 0.953
n 0.027 4.59E-04 n 0.167 5.09E-04
3 k 0.179 4.01E-03 0.948 1–2 k 0.056 2.04E-03 0.870
n 1.076 2.28E-02 n 0.196 1.03E-03
4 k 0.225 3.96E-04 0.998 Zþ/- ¼ 2.0 1 k 0.169 1.81E-02 0.993
n 0.280 3.07E-04 n 0.647 6.55E-03
1–4 k 0.187 3.72E-03 0.911 2 k 0.063 7.56E-04 0.950
n 0.436 2.93E-03 n 0.122 3.85E-04
1–2 k 0.060 1.83E-03 0.826
*n is the Avrami exponent and k is the kinetic constant. n 0.149 9.27E-04

*n is the Avrami exponent and k is the kinetic constant.

8
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

Fig. 11. Fraction of hydrate formed for PAA2k/PDADMAC100k 1% (wt./vol) mixtures at (a) Z-/þ ¼ 1.5 and (b) Zþ/- ¼ 0.67.

Declaration of competing interest Lee, J.D., Wu, H., Englezos, P., 2007. Cationic starches as gas hydrate kinetic inhibitors.
Chem. Eng. Sci. 62, 6548–6555. https://doi.org/10.1016/j.ces.2007.07.041.
Liu, X., Chapel, J.-P., Schatz, C., 2017. Structure, thermodynamic and kinetic signatures
None. of a synthetic polyelectrolyte coacervating system. Adv. Colloid Interface Sci. 239,
178–186. https://doi.org/10.1016/j.cis.2016.10.004.
Liu, X., Haddou, M., Grillo, I., Mana, Z., Chapel, J.-P., Schatz, C., 2016. Early stage
Acknowledgment kinetics of polyelectrolyte complex coacervation monitored through stopped-flow
light scattering. Soft Matter 12, 9030–9038. https://doi.org/10.1039/c6sm01979j.
The authors would like to thank CAPES, Brazil and FAPERJ, Brazil Moon, C., Taylor, P.C., Rodger, P.M., 2003. Molecular dynamics study of gas hydrate
formation. J. Am. Chem. Soc. 125, 4706–4707. https://doi.org/10.1021/ja028537v.
for financial support. Silva, B. L. L. D thanks CAPES for the doctoral
Perfeldt, C.M., Chua, P.C., Daraboina, N., Friis, D., Kristiansen, E., Ramlov, H.,
scholarship and Ferraz, I. L. thanks UFF, Brazil for the scientific initia­ Woodley, J.M., Kelland, M.A., von Solms, N., 2014. Inhibition of gas hydrate
tion scholarship. The authors also thank Prof. Dr. Daniel Grasseschi nucleation and growth: efficacy of an antifreeze protein from the longhorn beetle
Rhagium mordax. Energy Fuels 28, 3666–3672. https://doi.org/10.1021/
(UFRJ, Brazil) for help with DLS measurements and Bonif� acio Fialho for
ef500349w.
help with high-pressure cell measurements. Ribeiro Jr., C.P., Lage, P.L.C., 2008. Modelling of hydrate formation kinetics: state-of-
the-art and future directions. Chem. Eng. Sci. 63, 2007–2034. https://doi.org/
10.1016/j.ces.2008.01.014.
References
Rios, P.R., 2007. Transformaç~ oes de Fase. Artliber, S~ao Paulo.
Ripmeester, J.A., Alavi, S., 2016. Some current challenges in clathrate hydrate science:
Abrahamsen, E., Heyns, I.M., von Solms, N., Pfukwa, R., Klumperman, B., Kelland, M.A., nucleation, decomposition and the memory effect. Curr. Opin. Solid State Mater. Sci.
2017. First study of poly(3-methylene-2-pyrrolidone) as a kinetic hydrate inhibitor. 20, 344–351. https://doi.org/10.1016/j.cossms.2016.03.005.
Energy Fuels 31, 13572–13577. https://doi.org/10.1021/acs.energyfuels.7b03006. Sa, J.-H., Kwak, G.-H., Lee, B.R., Park, D.-H., Han, K., Lee, K.-H., 2013. Hydrophobic
Abrahamsen, E., Kelland, M.A., 2018. Comparison of kinetic hydrate inhibitor amino acids as a new class of kinetic inhibitors for gas hydrate formation. Sci. Rep. 3,
performance on structure I and structure II hydrate-forming gases for a range of 2428. https://doi.org/10.1038/srep02428.
polymer classes. Energy Fuels 32, 342–351. https://doi.org/10.1021/acs. Sa, J.-H., Kwak, G.-H., Han, K., Ahn, D., Cho, S.J., Lee, J.D., Lee, K.-H., 2016. Inhibition
energyfuels.7b03318. of methane and natural gas hydrate formation by altering the structure of water with
AlHarooni, K., Barifcani, A., Pack, D., Gubner, R., Ghodkay, V., 2015. Inhibition effects of amino acids. Sci. Rep. 6, 31582. https://doi.org/10.1038/srep31582.
thermally degraded MEG on hydrate formation for gas systems. J. Pet. Sci. Eng. 135, Sadeq, D., Iglauer, S., Lebedev, M., Smith, C., Barifcani, A., 2017. Experimental
608–617. https://doi.org/10.1016/j.petrol.2015.10.001. determination of hydrate phase equilibrium for different gas mixtures containing
Alonso, T., Irigoyen, J., Iturri, J.J., Larena, I.L., Moya, S.E., 2013. Study of the multilayer methane, carbon dioxide and nitrogen with motor current measurements. J. Nat. Gas
assembly and complex formation of poly(diallyldimethylammonium chloride) Sci. Eng. 38, 59–73. https://doi.org/10.1016/j.jngse.2016.12.025.
(PDADMAC) and poly(acrylic acid) (PAA) as a function of pH. Soft Matter 9, Shi, G., Jing, Y., Zhang, X., Zheng, G., 2009. Prospects of natural gas storage and
1920–1928. https://doi.org/10.1039/C2SM26884A. transportation using hydrate technology in China. In: 4th IEEE Conference on
Bhattacharjee, G., Choudhary, N., Kumar, A., Chakrabarty, S., Kumar, R., 2016. Effect of Industrial Electronics and Applications, 2009. Xi’an. IEEE, Xi’an, China,
the amino acid L-histidine on methane hydrate growth kinetics. J. Nat. Gas Sci. Eng. pp. 530–534.
35, 1453–1462. https://doi.org/10.1016/j.jngse.2016.05.052. Song, I.T., Stewart, R.J., 2018. Complex coacervation of Mg(II) phospho-
Choudhary, N., Das, S., Roy, S., Kumar, R., 2016. Effect of polyvinylpyrrolidone at polymethacrylate, a synthetic analog of sandcastle worm adhesive phosphoproteins.
methane hydrate-liquid water interface. Application in flow assurance and natural Soft Matter 14, 379–386. https://doi.org/10.1039/c7sm01654a.
gas hydrate exploitation. Fuel 186, 613–622. https://doi.org/10.1016/j. Swift, T., Swanson, L., Geoghegan, M., Rimmer, S., 2016. The pH-responsive behaviour
fuel.2016.09.004. of poly(acrylic acid) in aqueous solution is dependent on molar mass. Soft Matter 12,
Daraboina, N., Linga, P., Ripmeester, J., Walker, V.K., Englezos, P., 2011. Natural gas 2542–2549. https://doi.org/10.1039/C5SM02693H.
hydrate formation and decomposition in the presence of kinetic inhibitors. 2. Stirred Taheri, Z., Shabani, M.R., Nazari, K., Mehdizaheh, A., 2014. Natural gas transportation
Reactor Experiments. Energy Fuels 25, 4384–4391. https://doi.org/10.1021/ and storage by hydrate technology: Iran case study. J. Nat. Gas Sci. Eng. 21,
ef200813v. 846–849. https://doi.org/10.1016/j.jngse.2014.09.026.
Daraboina, N., Malmos, C., von Solms, N., 2013. Investigation of kinetic hydrate Tang, C., Zhao, X., Li, D., He, Y., Shen, X., Liang, D., 2017. Investigation of the flow
inhibition using a high-pressure micro differential scanning calorimeter. Energy characteristics of methane hydrate slurries with low flow rates. Energies 10, 145.
Fuels 27, 5779–5786. https://doi.org/10.1021/ef401042h. https://doi.org/10.3390/en10010145.
Daraboina, N., Perfeldt, C.M., von Solms, N., 2015. Testing antifreeze protein from the Vitorazi, L., Ould-Moussa, N., Sekar, S., Fresnais, J., Loh, W., Chapel, J.-P., Berret, J.-F.,
longhorn beetle Rhagium mordax as a kinetic gas hydrate inhibitor using a high- 2014. Evidence of a two-step process and pathway dependency in the
pressure micro differential scanning calorimeter. Can. J. Chem. 93, 1025–1030. thermodynamics of poly(diallyldimethylammonium chloride)/poly(sodium
https://doi.org/10.1139/cjc-2014-0543. acrylate) complexation. Soft Matter 10, 9496–9505. https://doi.org/10.1039/
Jensen, L., Thomsen, K., von Solms, N., 2011. Inhibition of structure I and II gas hydrates c4sm01461h.
using synthetic and biological kinetic inhibitors. Energy Fuels 25, 17–23. https:// Walker, V.K., Zeng, H., Ohno, H., Daraboina, N., Sharifi, H., Bagherzadeh, S.A., Alavi, S.,
doi.org/10.1021/ef100833n. Englezos, P., 2015. Antifreeze proteins as gas hydrate inhibitors. Can. J. Chem. 93,
Ke, W., Svartaas, T.M., Kvaloy, J.T., Kosberg, B.R., 2016. Inhibition-promotion: dual 839–849. https://doi.org/10.1139/cjc-2014-0538.
effects of polyvinylpyrrolidone (PVP) on structure-II hydrate nucleation. Energy Weinberg, M.C., 1999. A few topics concerning nucleation and crystallization in glasses.
Fuels 30, 7646–7655. https://doi.org/10.1021/acs.energyfuels.6b01321. J. Non-Cryst. Solids 255, 1–14. https://doi.org/10.1016/S0022-3093(99)00420-2.
Kelland, M.A., 2009. Production Chemicals for the Oil and Gas Industry. CRC Press. Xu, P., Lang, X., Fan, S., Wang, Y., Chen, J., 2016. Molecular dynamics simulation of
Kumar, A., Khatri, D., Lee, J.D., Kumar, R., 2016. Crystallization kinetics for carbon methane hydrate growth in the presence of the natural product pectin. J. Phys.
dioxide gas hydrate in fixed bed and stirred tank reactor. Korean J. Chem. Eng. 33, Chem. C 120, 5392–5397. https://doi.org/10.1021/acs.jpcc.5b10342.
1922–1930. https://doi.org/10.1007/s11814-016-0040-7.

9
B.L. Lemos Drumond Silva et al. Journal of Natural Gas Science and Engineering 73 (2020) 103053

Xu, S., Fan, S., Fang, S., Lang, X., Wang, Y., Chen, J., 2016. Pectin as an extraordinary Yan, M., Fresnais, J., Sekar, S., Chapel, J.-P., Berret, J.-F., 2011. Magnetic nanowires
natural kinetic hydrate inhibitor. Sci. Rep. 6, 23220. https://doi.org/10.1038/ generated via the waterborne desalting transition pathway. ACS Appl. Mater.
srep23220. Interfaces 3, 1049–1054. https://doi.org/10.1021/am101188y.
Xu, Y., Yang, M., Yang, X., 2010. Chitosan as green kinetic inhibitors for gas hydrate
formation. J. Nat. Gas Chem. 19, 431–435. https://doi.org/10.1016/S1003-9953
(09)60083-9.

10