LITERATURE REVIEW

Introduction
Literature review shows and depicts a core explanation on gas hydrates and
regenerated Mono Ethylene Glycol. The initial section has presented the overview
to gas hydrate structures and nucleation Mechanisms. An extensive depiction on the
kinetics of hydrate formation and growth has been discussed following to the first
section of this chapter, which highlights essential hydrate formation’s mechanisms
in gas and liquid systems. Additionally, thermodynamic hydrate inhibitors
(Methanol, Monoethylene glycol, Diethylene glycol and Triethylene glycol) along
with the low-dosage hydrate inhibitors (Kinetic Inhibitor and Anti-Agglomerants)
and others gas hydrate formation methods have also been discussed. Then,
hydrates formation in natural gas production and transport systems are presented
with safety impact while depressurising hydrate plugs in pipelines. Follwing section
is about an overview of Mono Ethylene Glycol, the property, production of MEG,
different models of MEG regeneration and reclamation systems and MEG
degradation will be presented. Finally, gap in literature will be discussed. Lastly,
the contextual background is endowed in the chapter to emphasize the overall
chapters of the thesis.

Just Dummy Figure Just Dummy Figure

(a) (b)

Just Dummy Figure Just Dummy Figure

(c) (d)

1
Examples of naturally-occurring gas hydrates: (a) massive; (b) nodules; (c) veins; (d)
laminae after [1]

Gas hydrate Structure and Formation Mechanism

Gas hydrate are ice like solid components, often emphasized as crystalline
compounds when compared to other molecules as determined by appropriate shape
and size in hydrogen-bonded water molecules cages. Clathrate hydrates is a Latin
root word where clatratus means barred or latticed. Gas hydrate structures forms
once molecules of water form a cage, comprising of small gas molecules ( < 0.9 nm)
such as ethane, methane or carbon dioxide, at adequate pressures and low
temperature conditions (typically < 27 oC and > 6 bar). It is important to note that
the gas molecule and the water cage are bonded chemically; however, the formation
of water cage is generated by hydrogen bonding of adjacent water molecules. The
repulsion force of the trapped gas in water cage what restricts it from collapsing.
The concentration value of gas molecules in the hydrate structure can be up to 180
times, that’s is, 180 m3 of gas molecules can be stored in just 1 m3 of hydrate. [1]. It
is important to note that formation of gas hydrates are not welcomed by all gas
molecules as guests molecules are constrained to molecules with low water
solubility attributes and particular molecular size [2, 3]. The size of the gas
molecule is the reliant variable in distinctive hydrate crystal structures and classified
as cubic structures I, II and H [4].

2.2.1 Cubic structure I (Figure 2-1)

Cubic structure I (Figure 2-1) are outweighed in the natural environment of the
earth, comprising small guests molecules of 0.4 to 0.55 nm such as; C1, C2 (Figure
2-2) and CO2.

Just Dummy Figure

Figure 2-1 Crystal structures of sI hydrate; four unit cells viewed along a cubic
crystallographic axis. All cavities are assumed to be filled; after Koh [5]

2
 Just Dummy Figure

Figure 2-2 A 512 pentagonal dodecahedral cavity enclusing methane (left) and a 51262
tetracaidecahedral cavity enclusing ethane (right); after Koh [5]

Structure I are comprised of 46 water molecules and are determined as body-centred

structures. The tetracaidecahedral 51262 cavity and the pentagonal dodecahedral 512
cavity are the two types of cavities of structure I (Error! Reference source not
found.). 12 pentagonal faces and two hexagonal faces are included in the
tetracaidecahedral 51262 cavity and is comparatively larger than the building block of
pentagonal dodecahedral 512. To relieve hydrogen bond strain, the formation of
structure I is linked with additional water molecules with the vertexes of the
pentagonal dodecahedral 512 cavities [1] (Figure 2-6). Thereby, guest molecules
identity significantly influences the kinetics and the stability of hydrate formation as
structure I hydrate comprises of two 512 cavities for every six 51262 cavity [6].
The cavities of structure I hydrates can only encluse smaller diameter gas molecules,
such as methane having 4.36 Å and ethane 5.50 Å [1]. Conversely, the 512 cavity
traps the methane molecules as it have smaller diameter but 51262 cavities trap ethane
with large diameter as it is too large for 512 cavity (Figure 2-2). The hydration
reaction of methane (CH4) can be given in Eq 2-1 as:

!"! + %"#$ "% & ⇌ !"! . %"#$ "% & Eq 2-1

Where %"#$ is the molar ratio of water reacting with methane, normally this number
is 6 [7], however this value can range from 5.75 to 17 [8]

2.2.2 Cubic structure II (Figure 2-3)

Cubic structure II (Figure 2-3) comprises of larger guests molecules of 0.6 to 0.7
nm in process systems, which include; C2, C3 (Figure 2-4) and iC4.

3
 Just Dummy Figure

Figure 2-3 Crystal structures of sII hydrate; two unit cells observed along a face
diagonal. All cavities are supposed to be filled; after [5]

Just Dummy Figure

Figure 2-4 Propane inside a hexacaidecahedral 51264 cavity; after [5]

A cubic framework is entailed in structure II hydrates in which the formation of a

diamond lattice took place, containing 136 water molecules (Error! Reference
source not found.). The pentagonal dodecahedral 512 building block and a
hexacaidecahedral cavity 51264 are the two types of cavity present in structure II
hydrates comprising of four hexagonal and 12 pentagonal faces. A larger free
diameter of 6.66 Å is mostly seen in the 51264 cavity, which consequently forms into
a larger cavity for enclosing guest molecules [1]. In Figure 2-4, the demonstration of
a propane molecule containing a diameter of 6.50 Å, can be placed within the larger
cavity. The formation of structure II hydrate crystal unit requires eight larger 51264
cavities and 16 small 512 cavities. It is evident from the experimental observation that
structure II hydrates are able to enclose propane, I-butane, krypton, nitrogen and
argon [6].

2.2.3 Hexagonal structure H

Type H hydrates is not common as compared to type I and II. In the particular, it
combined with two types of guest molecules, small and large guest’s molecules (0.8
to 0.9 nm) such as C5 – C6.

4
 Just Dummy Figure

Figure 2-5 Crystal structures of sH hydrate. The positions of the hydrogen atoms have
not been encompassed; four unit cells were aligned along the six-fold crystallographic
axis. 51268 has been highlighted by the yellow guests and all cavities assumed to be
filled; after [5].

The observance of structure H hydrates is not commonly perceived in natural gas

environments when comparing with structure I and structure II hydrates. On the
contrary, structure H hydrates possess significant importance in the oil industry. The
formation of hexagonal crystal structure with a large cavity is completed through 34
water molecules of structure H hydrates (Figure 2-6). Ripmeester, John [9] has stated
that additional three square faces are contained in the sH cavities formed.
Additionally, the formation of structure H hydrate crystal unit contains one large
icosahedral 51268 cavity, two small irregular dodecahedral 435663 cavities and three
small pentagonal dodecahedral 512 cavities [10]. Large guest molecules are required
within the formation of structure H hydrates by occupying 51268 cavities and with the
presence of small molecules; for example methane. Therefore, the presence of
structure H hydrates is emerged in denser hydrocarbon mixtures such as condensates
and oil.

Just Dummy Figure

Figure 2-6 Gas hydrates crystal structures; adapted after [11, 12]

The properties of the three common unit crystals are demonstrated in Error!
Reference source not found.. Sloan [13] has indicated that smaller cages are
not able to trap big single guest molecules and; thus, obliged to be filled in

5
 the larger cages even though both small and large cages exist in crystal
structure. Conversely, both cages can be filled by smaller molecules.
Table 2-1 Properties of the three common unit crystals; adapted after Sloan [14]

Hydrate crystal I II H
structure
Cavity Small Large Small Large Small Medium Large
Description 512 51262 512 51264 512 435663 51268
No of cavity 2 6 16 8 3 2 1
per unit cell
Average cavity 3.95 4.33 3.91 4.73 3.91§ 4.06§ 5.71§
radius (Å)
Coordination 20 24 20 28 20 20 36
number*
Number of 46 136 34
water / unit cell
* Number of oxygen at the periphery of each cavity.
§
Estimates of structure H cavities form geometric models.

The production of hydrate cavities took place when there is reduction in the water
temperature and becomes stable when filled with the gas molecules [15]. According
to the experiments, approximately 0.9 is the ratio of the size of guest molecule to the
cavity to become stable [16], whereas the optimal range of stable ratios of 0.86-0.98
was identified by Sloan, Koh [1].
Only one normal guest molecule is present inside each cage of all hydrate structures.
Conversely, it is probable that multiple small guest molecules including noble gases
or hydrogen are occupying a single cage at such conditions of high pressures. It has
been suggested by Mao, Mao [17] that hydrogen atom can form with four occupants in
the large cage and two occupants in the small cage of hydrate structure II at a high
pressure condition. Hydrates are crystalline in nature but non-stoichiometric as some
cavities are left unoccupied and as there is a clear pattern between the diameters of the
cavity and the ratio of the guest molecule [4]. Moreover, same composition of 15
mole% guests and 85 mole% water is found in three hydrate types when all cavities are
occupied [1].

6
When the cavities in the crystal structure are occupied with only one type of gas
molecule, this is called simple hydrates. Methane (CH4), hydrogen sulphide (H2S),
carbon dioxide (CO2) and ethane (C2H6) are examples of structure I natural gas
hydrates. In addition, nitrogen, propane and iso-butane are examples of simple
structure II natural gas hydrates. Moreover, the formation of binary hydrates can be
formed by the clathrate of two gases such as CO2 and CH4 and C2H6 and CH4.
Binary CH4-CO2 mixture forming only structure I hydrate. However, for the
situation of binary CH4-C2H6 mixture (both forms sI individually) the formation of
structure I or structure II might take place, based on the temperature and pressure
conditions [4]. Gases do not occupy all the cavities when forming hydrates. As
stated by Sloan [18], typical hydrate occupancies of large cavities is 50% while
small cavity is 95%. Detailed gas hydrates morphological structures are given
elsewhere [18-20].
Hydrate formation usually took place between the interface of guest molecule and
the aqueous phases because of the availability of the high concentrations of both
guest gases and host cavities which exceed the mutual fluid solubilities. This solid
interface layer prevents further hydrate formation causing an interaction barrier
between the gas-liquid phases, unless fluid surface renewal is activated such as by
agitation or turbulent flow [21, 22].

According to experimental observation of hydrate and ice, there are assorted

different distinctive physical and chemical properties even though they have almost
similar apparent. However, one of the most significant properties is that hydrate can
clathrate at 0 oC or higher temperature whereas cannot be seen in ice. Moreover, the
trapped gases of a gas hydrate can undergo combustion when exposed to extreme
heat (Figure 2-7) while this property cannot be revealed for ice [23].

Just Dummy Figure

Figure 2-7 Reddish methane hydrate flame; after [11, 23]

7
Von Stackelberg [24] has introduced the correlation between the type of hydrate
formed and the size of the molecule. The chart produced by Von Stackelberg [24]
[redrawn by Giavarini and Hester [11] and Carroll [25]] is shown in Figure 2-8,
which indicates the gas hydrate nature relied on the guest molecule size. It is
revealed from the chart that hydrates are not formed with molecules containing
diameters less than 3.8 Å (1 Å = 1 × 10&'( m). From the chart, it is clear that
initial hydrate formers commence with molecules diameters of 4 Å such as nitrogen
and krypton. The formation of a type I or type II hydrate is limited with molecules
size larger than 7 Å. Type H hydrates can be formed through slightly larger
molecules; however, the formation of hydrate is limited by 9 Å. Thereby, molecules
with greater molecules diameters than 9 Å are non-formers, such as hexane, larger
paraffin hydrocarbons and pentane.

Just Dummy Figure

Figure 2-8 Comparison of guest size, hydrate type, and cavities occupied for various
hydrate formers; after [11]

Gas Hydrate Nucleation

Hydrate nucleation is considered as a process of expansion and dispersion of small
water and gas clusters that accomplish appropriate crystal size for continued growth.
According to Mullin [26], restricted amount of hydrate nucleation experimental
verification is revealed from the involvement of tens to thousands of molecules in the
stochastic and microscopic process. Labile cluster nucleation and local structuring
nucleation are two major hypotheses exist in the current experiments and modelling
[4].

2.3.1 Local structuring nucleation hypothesis

Radhakrishnan and Trout [27] suggested that local structuring nucleation hypothesis
is supported by the formation of carbon dioxide hydrate, indicating that a local

8
structuring model can be used to replace the labile cluster nucleation hypothesis. The
arrangement of guest molecules is caused by thermal functions comparatively to that
hydrate phase in the local structuring model. Conversely, the disruption of water
molecules structure throughout guest molecules is evident, comparing with rest of
the water phase. Afterwards, the arrangement of gas and water phases is structured
closely to the hydrate phase heading towards to form critical size hydrate nucleus
and; thus then growing.

The computation of Landau-Ginzburg free energy for carbon dioxide hydrate

nucleation and Monte Carlo simulations were performed through isobaric and
isothermal experimental method [27]. Their experiments allowed advance analysing
of the nucleation mechanism at the interface level, which concluded that the study of
Christiansen and Sloan [6] on agglomeration of labile clusters might not be favoured
thermodynamically. The free energy required to form clusters is much higher than
the energy required collapsing, thus requiring higher energy to overcome the free
energy barrier. Radhakrishnan and Trout [27] have indicated that it is almost
impossible to form nuclei by labile clusters for carbon dioxide hydrates due to free
energy barrier. They also proposed two major mechanism of nucleation required to
initiate the clathrate phase nucleation based on local structuring hypothesis. The first
mechanism states that a selection of guest molecules is caused with a thermal
fluctuation in order to restructure in a clathrate configuration. The disturbance in the
bulk structures is seen also among the surrounded guest molecules by the water
molecule structures, which indicate a stochastic nature. The second mechanism states
that, if the number of the critical nucleus has been surpassed by the entire extent of
guest molecules in the locally ordered arrangement, this will results in local stability
caused by the relaxation of the surrounding water molecules. The critical nuclei
formation is resultant from the clusters parameters of host-host and guest-guest that
is similar to a clathrate hydrate. Another similar model was simulated by Moon,
Taylor [28] using MD simulations of methane hydrate nucleation. It was
demonstrated that formation of methane hydrates eventually took place at the water
interface as revealed from a steady growth of clathrate clusters of the simulated
system. An evidence for long-range structures is provided in accordance to the local
structuring nucleation hypothesis, observed in the changes in the structures over the
entire simulation. The indication of hydrogen bonds forming hydrate structures is

9
illustrated from the snapshots of ‘hydrate-like’ water lines connecting with methane
molecules, demonstrating the restructure over a longer range of water molecules
instead of creating independent molecules Figure 2-9.

Just Dummy Figure

Figure 2-9 Growth of local structure nucleation lines (with time shown in
nanoseconds) indicate the hydrogen-bond network; after [28]

The investigation was taken place for the water molecule cluster formation,
demonstrating that the faces shared generated stable cages, supporting the hypothesis
of the labile cluster nucleation. The model showing a 512 cavity, sharing a face to
develop a stable cluster is established in Figure 2-10. It has been concluded that
although the methane-water simulations shows similarity to the labile cluster
hypothesis, it consistent with the local order model of nucleation.

Just Dummy Figure

Figure 2-10 Stable sharing of faces in a 512 cavity with methane gas , formed by 6 ns;
after [28]

2.3.2 Labile Cluster Nucleation Hypothesis

Point A in Figure 2-11 illustrates that the labile cluster nucleation hypothesis is relied
on the presumption that hexameter and pentameter labile ring structures organize
pure liquid water molecules [27]. Stillinger [28] has evaluated that the water network
structures are mostly caused by hydrogen bonds. The formation of labile clusters is
immediately reflected at point B, and combined with agglomerate clusters until the
formation of hydrate unit cells (point C) with respect to the guest molecule

10
dissolution in water. A critical size was extended to the point, where unit cells were
combined with agglomerate clusters (at point D). The illustration of labile cluster
nucleation hypothesis is shown when temperature and pressure forms in hydrate
region at initial condition while there is no possibility observed of dissolving the gas
molecules. Formation of labile clusters is based on the size and coordination number
where water molecule number augments with the guest molecule size in each cluster
shell. For instance, natural gas components’ coordination numbers for carbon
dioxide, ethane, propane, I-butane and methane are 24, 24, 28, 28 and 20 respectively
[5].

Just Dummy Figure

Figure 2-11 Labile cluster nucleation after [29]

It is assumed that guest molecules are dissolved in a single cage in contrast to the
local structuring hypothesis where local structuring nucleation cannot be observed
with long range arrangement. The guest moledules were started to dissolve within the
water that has resulted in the formation of clusters arounds the polar guest molecules
(point 1). The labile cluster nucleation process (Figure 2-11) can be linked to the
physical hydrate formation/dissociation process at a PVT cell using isochoric method
(constant volume) as shown in Figure 2-12, before point 1 the gas is not dissolved in
water. It is evident that labile clusters are bounded to different clusters for producing
the hydrate unit cells in the metastable phase of cooling period due to the existence
of labile clusters in subcritical size at both points 1 and 2. The joining of labile
clusters at point 2 becomes evident to extend the critical size of nucleation.
Although, the completion of primary nucleation has been discovered and achieves
quick hydrate growth. A fast pressure drops encompasses gas hydrate growth due to
the encapsulation of gas molecules in the hydrate crystals.
Point 3 is the end of hydrate growth process, where hydrate formation stops. By
moving the structure from point-3 to point A and by heating the system on a higher
temperature commences the hydrate dissociation process. It decomposes the hydrate

11
agglomerates into the liquid and vapour phases. However, quasi-crystalline
metastable cluster structures remained in the liquid form at a certain degree of
superheating [7].
The hydrate formation/dissociation process in a PVT cell using isochoric method has
also been conducted in this thesis as shown in Figure 2-12 and chapter XX.

Just Dummy Figure

Figure 2-12 Schematic of hydrate formation/dissociation on an isochoric method

after [7]

Just Dummy Figure

Figure 2-13: Our work, determination of the hydrate equilibrium point (dissociation
point) From the Pressure-Temperature trend by the intersection point of the cooling
and heating cycles ( the pressure was increased because of gas release, which was the
outcome of increased temperature and hydrate dissociation) [30-33].

The rate of hydrate nucleation is another core focus for labile cluster nucleation.
Christiansen and Sloan [7] have asserted that formation of structure II hydrates
ultimately resultant into hydrate nucleation kinetics [5]. On the contrary, the
excessive contradictions on this assertion and existence of newer experimental
evidence demonstrate the energy restriction for the labile clusters agglomeration in
larger form as compared to the requirement of disintegrated clusters [25].

2.3.3 Nucleation at the interface hypothesis

As discussed by Kvamme [34], the occurrence of the nucleation is emerged on the
vapour side of the interface rather than on the liquid side Figure 2-11. Rodger [35]
conducted a simulation for molecular dynamics indicated that gas molecules causes

12
through attractive dispersion forces captivate the surface of the water interface. The
disruption of water molecule structures is occurred when the molecules are formed
into a layer and resultant into a hydrogen bonding network formation at the interface
in gas hydrates as extend to the surface.

The site of hydrate nucleation was experimentally investigated by Long and Sloan
[36], in which both examined that nucleation for carbon dioxide hydrates are
occurred on the vapour side of the interface at the gas and water interface. Moon,
Taylor [37] have studied the molecular dynamics of hydrate formation in which they
indicated that formation of methane hydrates are intended of the gas and water
interface at the site (Figure 2-14).

Just Dummy Figure

Figure 2-14 Our work, hydrate start formation at interface of a PVT cell. After [38]

Figure 2-15 has shown the adsorption and clustering of the interface on the gas side.
It is assumed that movement of the vapour phase is shown for the vapour-liquid
interface, indicating the preferential placement for hydrate nucleation in this
scenario. The adsorption of the gas molecule reflected the aqueous surface and
afterwards the formation of water molecules are reflected across the gas guest
molecule. The labile clusters agglomerate accomplish a critical size in which the
occurrence of growth is seen on the gas side of the interface, which causes in quicker
hydrate increment [5].

Just Dummy Figure

Figure 2-15 Gas hydrate nucleation at gas-water interface after [5]

13
2.3.4 Morphology of gas hydrate nucleation
Wu and Zhang [39] and Makogon [40] have conducted experiments, which have
shown different types of crystallisation. They identified three morphology types for
hydrate crystals namely massive, whiskery, and jelly. Wu and Zhang [39] observed
by naked eyes that massive, whiskery, and jelly crystals for methane gas hydrates
appeared at 4∼8 oC and 5∼7 MPa.
Figure 2-16 (a-c) present the monopoly massive gas hydrate nucleation formed above
the gas-liquid interface. Figure 2-19 shows the monopoly whiskery gas hydrate
nucleation which appears after complete hydrate formation, it grows upward in the
gas phase.
The monopoly jelly gas hydrate nucleation is illustrated in Figure 2-21. The
formation of jelly crystals is observed in the second hydrate process, which has
followed the dissociation run and primary hydrate formation. Generally, whiskery
crystals augment in a liquid and volume of gas, jelly crystals produced under
particular circumstances in a bulk of water and massive crystals augment in a volume
of gas in regular basis [39, 40].

Just Dummy Figure

Figure 2-16 Massive methane hydrate crystals after [39]

Just Dummy Figure

Figure 2-17 Our work, massive methane hydrate crystals

14
 Just Dummy Figure

Figure 2-18 Our work, massive gas hydrate morphology of methane gas with 20 wt%
MEG at different stages. Similar pattern found by other researchers [41-43]

Just Dummy Figure

Figure 2-19 Whiskery methane hydrate crystals after [39]

Just Dummy Figure

Figure 2-20 Our work, whiskery methane hydrate crystals (side view and top view)

Just Dummy Figure

Figure 2-21 Jelly methane hydrate crystals after [39]

15
2.3.5 Gas Hydrate Memory Effect Phenomenon
Gas hydrate memory phenomenon is the ability of gas hydrate, when melted at
moderate temperature for retaining a memory of their structure [44, 45]. Therefore,
melted hydrate obtain formation of hydrate at shorter induction time of water and
relaxation condition with no previous hydrate history as compared to fresh water
[46].

Makogon [47] has published examination of memory effect, in which the formation
of hydrates are rapidly emerged from the melted hydrate when comparing from
initial time. This surveillance has been considered in the previous forty years, which
includes the work of Schroeter, Kobayashi [48], which successive cooling runs with
same liquid show the decreased formation point at each cycle as seen in Figure 2-22,
which shows how hydrate formation become more relaxed for at each repeated
experiment, S3 formed more easier than S2 and S1, and S2 formed more easier than
S1.

Just Dummy Figure

Figure 2-22 Consecutive hydrate formation cooling curves for several runs modified
after [48]

The gas hydrate memory effect study has gained researchers’ interest, and its
mechanism has been analysed from different aspects [5, 19, 44, 45, 49-54].

The study of Wu and Zhang [39] represented the induction time for gas hydrate
nucleation of methane gas. Once gas hydrate is formed and dissociate; second
formation happens at marginally greater temperatures as compared to earlier
formation (Figure 2-23). It also applies to the 3rd and 4th cycles with the time of
nucleation declining for each successive test. The increment of both dissociation and
nucleation temperatures is determined with successive examination, which
demonstrates the readily formation of hydrates; and therefore, endowing memory
effect consequences. Once the melted hydrate had been warmed to 25°C, memory

16
effect were disturbed as aligned with the consequences of Link, Ladner [55] and
Takeya, Hori [45]. Memory effect phenomena is mostly utilised as a hydrate
promoter for hydrate development projects.

Just Dummy Figure

Figure 2-23 Hydrate formation repetition of same fluid after dissociation after [39]

Lee, Susilo [56] reported the higher assumed induction times, which has resulted in
longer hydrated dissociation before reformation. The hydrate dissociation has shown
that shorter induction time was left for 1 hour as compared to 12 hours.

The outcomes are aligned with the results of Vysniauskas and Bishnoi [46] and
Ohmura, Ogawa [50], in which they reported the influence of thermal history on the
hydrate induction times. The study has examined that induction time for dissociated
hydrate is not greater as compared to the warm water.

Servio and Englezos [57] had focussed on analysing the effects of the macroscopic
crystal morphology of carbon dioxide and methane hydrates made from water
droplets. They showed that memory effect accelerated the hydrate growth. (Figure
2-24). Their analyse showed that, after 30 minutes of hydrate full dissociation of
fresh water droplet with previous hydrate history, the size of 10 minutes hydrate
growth was equivalent to a 25 hours hydrate growth of water droplet with no
previous hydrate history [Figure 2-24 (a) and (b)].
Another distinction was made in their research showed that hydrate surface before
de-composition for 24 hours is exhibiting surface depressions on the basis of water
depletion and slightly pointed on the basis of its hydrate needles [Figure 2-24 (a)].
While, a 10 minutes of hydrate growth on water droplet that’s experienced fully
dissociated for 24 hour before reformation, resulted in irregular and pointed with
numerous needle-like crystals encompassing outward away from the surface [Figure
2-24 (b)]. On other hand, a 10 minutes of hydrate growth on water droplet that’s
experienced fully dissociated for 30 minutes before reformation, resulted in smooth
and shiny surface [Figure 2-24 (c)].
17
 Just Dummy Figure

Figure 2-24 Macroscopic crystal morphology of carbon dioxide hydrate formation

from water droplets. Modified after [57]. The assertion was shown from the
hypothesis indicating that after hydrate dissociation, residual clusters of water
molecules originate the memory effect [50]. While it is obviously not a
thermodynamic effect, the exact cause of the memory effect is unclear. The
examination of hydrate forming systems was completed several times to indicate the
water clustering signs with respect to viscosity, interfacial, refractive index, and
tension after hydrate dissociation. Experimental studies conducted by Ohmura,
Ogawa [50] (using Hydrochlorofluorocarbon HCFC-141b (CH3- CCl2F)), by
Takeya, Hori [45] (using CO2 hydrates from CO2 dissolved water) and by Sloan,
Subramanian [58] (different gases) support the residual water clustering hypothesis.
The consolidation of hypothesis are further conducted through molecular-dynamics
simulation studies in order to illustrate the memory effect mechanism [59-61]. A
distribution of ice-like water molecular structures has been reported by Rodger [60],
showing generation of liquid water by hydrate dissociation rather than in hydrate-
cage structures. It has been further concluded that decomposition of hydrate take
place when dissolved gas remains in solution. Also the experimental studies of Bylov
and Rasmussen [62] and Ohmura, Shigetomi [63] exhibited negative results. [64]
were unable to find any sign of memory effect (continues hydrate crystals post
dissociation) using neutron scattering. It has been further concluded that existence of
structural memory effects is not emerged in comprehensive equilibrated system.
These findings led to a negative view of the above hypothesis (Figure 2-25).
The study has suggested that observations and experiment at lower temperatures and
pressures will become complex for differentiating the attributes of memory effect
and the determination of worst equilibrium condition. Conversely, a conclusive
physical image of the memory effect should be derived through further simulation
analyses.

18
 Just Dummy Figure

Figure 2-25 Snapshots of the residual clathrate (a) and ice (b) structures in the
hydrate melt after 600 ps after [60].

Memory effect can be destroyed once the hydrate system is moved sufficiently far
away from hydrate equilibrium point (formation region) (i.e sufficiently heated) or
enough time is given [11]. For methane gas hydrate at 15 MPa, the memory effect
can disappear, when the solution is heated to approximately 5.5 oC and 8 oC above
the equilibrium point [65]. Furthermore, Table 2-2 below represented researcher’s
comments on memory effect vanishment:

19
Table 2-2 Researcher’s findings on memory effect vanishment.

Authors Findings
Sloan and Koh [5] Evaluated that memory effects are vanished, when the melted water is heated above 24 oC
Lederhos, Long [49] Considered that the residual structure was destroyed for natural gas (with 87.2% C1), after the liquid water
temperature was upraised to 28 oC.
Takeya, Hori [45] Found that the memory effect of CO2/Water destroyed when the meltwater sample was heated to 25 oC.
Wu and Zhang [39] Found that promotion of memory effect is dependent on the dissociation temperature and the memory effect of
natural gas (methane 99.99%) vanished when the dissociation temperature was higher than 25 oC.
[40] Established that there is no residual structure remains to promote hydrate once an upper temperature limit of
about 30 °C is passed.
Becker, Kozielski [66] Concluded that no memory effect exists for experiment conducted using mixtures of tetrahydrofuran and water
Chen, Yan [67] Conducted experiments using Methane/diesel oil/ sorbitan monolaurate and concluded that memory effect cannot
be eliminated if it maintained near the hydrate formation point even for long time (more than 165 hours), while it
will be vanished when the system temperature raised 5 oC above the equilibrium point.
Sefidroodi, Abrahamsen The memory influence of cyclopentane hydrate formation must not always vanish with the superheating of 8.4 oC
[54] for 20 min. it has been suggested that the influence of memory is in the bulk water phase and I possibly because
of residual clathrate hydrate structure unobservable to the naked eye.
Wilson and Haymet [68] Evaluated that the influence of gas hydrates memory effect is damaged when it is heated above the equilibrium
temperature (4 oC). On other hand, they reported that ‘THF’ hydrates do not hold memory effect.
Del Villano and Kelland Reported that memory effect of natural gas with KHI is lost when heating to 8.4 oC above the equilibrium
[53] temperature.

20
While performing hydrate experiments in this study, the researcher has avoided
hydrate equilibrium shift changes caused by memory effect by starting the first
experiment at the highest pressure and then lowering the pressure as proceeding,
shearing the liquid to the maximum shear stress (at ≈ 1500 RMP), and heating the
sapphire cell to above 30 oC for the succeeding experiments. The memory effect has
significant implications on flow assurance and gas research. It is recommended that
once hydrate is formed in a pipeline, hydrate dissociation process to remain followed
by water removal, as this melted water having residual entity (i.e., dissolved gas,
residual structure and persistent crystallites) will accelerate reformation of gas
hydrate and so plug the pipeline. Equally, memory effect phenomena can be utilised
as a hydrate promoter for hydrate technologies of storage, transportation and
utilization of natural gases in the hydrate form (taking into account that dissociation
temperature not far from equilibrium point and dissociation time is short) [5, 39]

Hydrate Growth
After hydrate nucleation step, the second step to form a solid hydrate mass is the
hydrate growth and coalescence. For this phase, mas and heat transfer plays a big
role. Two dependent factors are emerged on which the rate of hydrate growth is
evaluated; on component mass transfer of growing crystal surface and crystal growth
kinetics at the hydrate surface. Moreover, four categories are major on which the
growth environment conditions classify the process of hydrate crystal growth. Figure
2-26 has shown the first category; single crystal growth. Growth of metastable
phases, multiple crystal growth and hydrate/film shell growth are the other three
states. It is assumed that modelling and hydrate growth data are more acceptable as
compared to nucleation phenomena.

Just Dummy Figure

Figure 2-26 Single Crystal Growth, after [5]

21
Sloan and Koh [5] summarized the modern statements for hydrate growth with below
statements:
• The placement of data can be fitted based on the parameters, which ultimately
reveals the existence of growth model.
• The data was obtained mostly from the high-pressure reactor and; therefore,
formation rates cannot be implicated in a pipeline. Additionally, the
accessibility of flow loop data is beneficial.
• The acceptance of modelling and hydrate growth data is evident as compared
to the nucleation phenomena where the appearance of growth data is linear
for approximately 100 min in Englezos’ data.[69].
• Structure I are dependent for mostly obtained data whereas structure II
reflects number of pipeline hydrates on the basis of propane components of
natural hydrocarbons.
• The formation of metastable phases do not account the models or the
simulations during hydrate growth.
• The effects of heat and transfer can be highly determined in multiphase
systems as compared to intrinsic kinetics.

Hydrate Growth Correlations

Assorted correlations were constructed for crystal process model as several extensive
studies have studied hydrate growth mechanisms. The determination of controlling
the formation rate is essential by acknowledging and representing the formation
process. The classification of key correlations is relied on three existing growth
aspects; heat transfer, growth kinetics and mass transfer.

There are further restrictions to apply each model and at the extent at which actual
hydrate growth is represented; however, the validation of correlations was reflected
among research groups. Hydrate growth is slightly affected through kinetics than the
mass and heat transfer. Therefore, greater application has been practically reflected
among latter models [5].

2.5.1 Hydrate Growth Kinetics

Englezos, Kalogerakis [69] has proposed the hydrate formation in methane and
ethane with the kinetic growth correlations. Crystallisation theory was used to

22
produce the model along with mass transfer phenomena to demonstrate the kinetics
formation at the hydrocarbon-water interface. There are three hydrate formation steps
assumed to condition the work validation. The first step is the commencement of
transport between phases, second is the diffusion through the boundary layer and
finally the water adsorption process [5, 69, 70]; the description of hydrate formation
in mixtures of the gases is extended by focusing on the individual model of methane
and ethane [70].

The focus of growth kinetics model is supposed to react at a boundary layer or

interface and utilizing the core aspect of diffusion. Particle size is considered as a
particle diameter when a minor inconsistency is found from the model of research
group. For future works, the modification for carbon dioxide hydrates is removed
with this error [71]. According to past literature [69, 70], there are two equations of
gas hydrate growth with one adjustable parameter in the kinetic model. The
representation of total consumed gas moles/second by the hydrate with respect to the
!"
extent of growth per particle ! !$#" , is given by Eq 2-2:
%

$%& Eq 2-2
# ' = * ∗ +% ,-&) − -&*+ /
$& %

The surface area of each particle is represented by A, while the fugacity of

component is represented by -&) and -&*+ 1, respectively in the bulk and at
equilibrium. The hydrate formation growth rate constant is represented from * ∗ ,
joining the rate constant for adsorption and transfer processes. The association of
mass transfer coefficient and the reaction rate constant is shown with * ∗ in Eq 2-3.

1 1 1 Eq 2-3
= +
*∗ 3- 3!

It is evident that correlations endow an appropriate basis for future works whereas
several limitations have been shown for the growth kinetics model. By using
experimental data from compounds, the formation of structure I hydrates is revealed
from the limitation of fitting model. However, the accuracy of the model cannot be
proven for structure II and structure H hydrates. The extrapolation of the model and
the sensitivity of the model at the turbidity point are further considered as limitations
[5].

23
The real gas equation is also used to study and analyse growth experiments [72].
56 = 7% 89 Eq 2-4

where P is pressure,
V = gas volume,
z =compressibility factor,
n = number of moles,
R = universal gas constant
T =temperature of the gas
The pressure drop in the gas phase was resultant through the principle of mass
conservation for an isochoric system. In the liquid phase, the approximation of the
amount of formed hydrates is determined by hydrate growth. Thus, Eq 2-4 yields:
6 Eq 2-5
∆ % = ∆ 5
789

Where;
∆ % = Amount of gas consumed during hydrate formation or the amount of hydrates
formed
∆ 5= Measured experimental pressure drop produced due to hydrate formation
.
The approximation of the term /01
as the constant of proportionality cannot be

modified considerably as ∆ % ∝ ∆ 5 indicated that the pressure drop in the gas phase
and the amount of gas consumed in the liquid phase are directly associated to each
other. Growth processes involve fast reactions of coupled mass and heat transfer
especially during the early nucleation stage. Primarily it is limited by mass transfer
of the reactants to the rising crystal and a simultaneous removal of heat away from
the growing crystal. Such coupled heat and mass transfer is a complex process
especially for a multicomponent system. Mass and heat transfer models have been
summarised and may be found at the literature [73-76]. The extent of formation of
gas hydrates can be estimated from the employed concept in order to fill structure I
and II systems cavities, when the appropriate systems of gas hydrate growth is meet
by the systems.

24
 Hydrate Dissociation
Hydrate dissociation enthalpy is a significant attribute for dissociation process and
hydrate formation. The process of hydrates formation is structured; like ice, which
indicated the relaxation in the heat. On the contrary, the pre-requisite for hydrates
dissociation is to overwhelm the activation energy and distribute the intermolecular
and hydrogen bonds of the hydrate structures as hydrates dissociation is further
considered as an endothermic process [11]
In the past 40 years, the hydrate decomposition is proposed through numerous
models [77]. Analytical, theoretical and numerical models are included in the
proposed models with modifying complex degrees. In a past study [78], the earliest
model of hydrate dissociation was used with no restrictions of mass and heat transfer
for kinetics of hydrate dissociation. A two-step dissociation model indicated that the
lattice of the was is destroyed; afterwards, the surface absorbed the guest molecule.
From the proposed model, it is indicated that the difference in fugacity of the guest
molecule is correlated with the decomposition rate. This correlation was compared at
the surface area of the hydrate particles and under decomposition conditions at
equilibrium. This describes the rate of hydrate decomposition as:
$%2 Eq 2-6
− # ' = 3! +3 (-* − -)
$&

Where
+3 = Surface area of the decomposing hydrates,
3! =Decomposition rate constant,
-* = Fugacity of the guest molecule at equilibrium,
- =Fugacity of the guest at the solid surface [79]
It is evident from the model that methane hydrates decomposition was relied on
particle surface area, pressure and temperature. In general, it is important to notify
that it is the first time that the model studied intrinsically the kinetics of hydrate
decomposition regardless of the influence of mass and heat transfer. Figure 2-27 has
shown the proposed process.

25
 Just Dummy Figure

Figure 2-27 Schematic of hydrate dissociation mechanism, after [77, 80]

With the existence of new models, the decomposition of hydrates was also
investigated on the basis of current model, which has the ability to provide reason for
the size of hydrate. Moreover, the hydrate decomposition could be reduced with
particle size estimations, increasing the entire activation energy by 3kj/mol and up to
4 times [81]. Kim et al. had proposed that size of the particles should remain constant
whereas Clarke et al. accepted the difference in the size of particles.The
determination of decomposition rates for carbon dioxide hydrates is executed by
Clarke and Bishnoi [82] and Clarke and Bishnoi [83] showing the mixture of ethane
and methane. Furthermore, the activation energy determination for structure II
hydrates was found to be lower as compared to structure I hydrates, which emphasize
that the dissociation can be faster in structure II hydrates as compared to structure I
hydrates.
Assorted approaches were attempted to separate hydrates. Depressurisation is the
most common method for dissociation. It is examined that dissociation is easily
caused using this method with the requirement of minute energy input as the
conditions of the hydrate will exceed the hydrate stability zone. Hydrates can be
break down for a particular pressure by increasing the temperature of the hydrates.
The second method is mostly used, but the major concern is that it is cost-expensive
as lot of energy is consumed by hydrates, when comparing with depressurisation.
Injection of chemical inhibitors are the third method involved in the hydrate
dissociation.
The contribution of heat transfer is very much accepted for hydrate dissociation than
the intrinsic kinetics. The domination of initial stages of dissociation is revealed from
the intrinsic kinetics in which the gradient temperature among the interface and the
hydrates is very small or non-existent. The increment is shown in the gradient
temperature and then dissociation is controlled by heat transfer as dissociation

26
process continued after implication [5]. Three disassociation methods influence the
hydrate system and leads to an eventual break down and destabilisation of gas
hydrates (Figure 2-28). The initial condition of the hydrate is assumed at temperature
T4 and pressure P4 . Depressurisation method reduces the pressure below the
equilibrium value to P5 and bring in a decomposition driving force of (5*6 − P5 ).
Thermal stimulation method increases the hydrate temperature to T7 to bring in a
decomposition driving force of (5*7 − P4 ). Inhibitor injection shifts the hydrate
phase equilibrium P-T condition to bring in a decomposition driving force of (5*8 −
P4 ).
Where 5*6 and 5*7 are the equilibrium pressure of the temperature T4 and T7
respectively, 5*8 is the equilibrium pressure of temperature T4 .[84]

Just Dummy Figure

Figure 2-28 Driving forces for hydrate decomposition modified after [84]

The latest discovery of the radial model demonstrated more rapid dissociation with
larger surface area; however, hydrate dissociation was first modelled by an axial
dissociation model. It established slower plug dissociation as shown in Figure 2-29,
but. Hydrate dissociation is currently conceptualised by radial dissociation (from the
pipe wall into the centre of the pipe), established in Figure 2-30 surrounded by a
water phase, with a hydrate plug centralised in a pipeline [85].

Just Dummy Figure

27
Figure 2-29 Old axial one sided dissociation model of a hydrate in a pipeline
modified after [86].

Just Dummy Figure

Figure 2-30 Radial dissociation of a hydrate in a pipeline modified after [85]

Figure 2-31 has shown that the pipe radically evolves heat flows in which the pipe
wall initially causes dissociation. After three experiments, the figure has illustrated
three laboratory hydrate plugs [85]. The temperature of hydrate can be reduced
below the temperature of surroundings for pipeline depressurization.

Just Dummy Figure

Figure 2-31 Radial dissociation of Laboratory hydrate plugs in a pipeline, lower part
dissociate faster due to effect of gravity. Modified after [85]

The completion of hydrate formation is experimentally examined to measure the

thermodynamic equilibrium point because of the commencement of dissociation and
formation of water drops. Thermodynamic attributes of the system can potentially
describe the hydrate dissociation along with the determined equilibrium conditions
conducted by dissociation experiments [27].

The significant illustration of hydrate dissociation is radially understand with respect

to the remediation of these issues in pipeline [86]. There are numerous mitigation
techniques applied for hydrates dissociation, including depressurisation, inhibitor
injection and thermal stimulation. [23, 87, 88].

28
Hydrate depressurisation technique can be potentially dangerous as shown in Figure
2-32, hydrate blockage is dissociated by pressurising the pressure downstream of the
hydrate plug Figure 2-32 (b). As pressure is depressurised one side of the pipeline,
the plug can be loosened and would be projected like a bullet laterally the pipeline at
high velocity, Figure 2-32 (c) and Figure 2-33.[23] Xiao, Shoup [89] have
contributed for depressurization with the help of a transient multiphase flow
simulator OLGA for the simulation of hydrate plug velocities. Number of parameters
influence plug movements during simulations, which include plug size, the existence
of oil or condensate, plug location and a plug.

Just Dummy Figure

(a) Hydrate formed at high pressure result in plugging the pipeline

Just Dummy Figure

(b) Depressurisation technique is conducted by open the bleed valve at the

downstream of the hydrate plug, to reduce the pressure and dissociate the
hydrate plug.

Just Dummy Figure

(c) With sudden pressure drop, hydrate plug began to travel at high velocity

29
Figure 2-32 Incorrect and sudden depressurisation of hydrate plug in high pressure
pipeline causing the hydrate plug to being launched like a projectile. Modified after
[11, 23]

Just Dummy Figure

Figure 2-33 Hydrate plug dissociation incident happened due to incorrect single
sided depressurisation procedure after [90] (Courtesy of Chevron Canada Resources,
1992). The application of heat is another hydrate dissociation technique.
Applications of heated bundles, spraying steam on the line, and installation of
electrical heat tracing are different means used to apply heat. The depressurisation
should be conducted at both sides of hydrate plug to minimise the differential
pressure (below 10%) across the hydrate plug . On other hand, if it is not possible to
depressurise both sides of the plug, then step depressurisation of one side should be
applied by releasing and closing the depressurisation line till full plug
dissociation.[23]

The installation of a pipeline is involved in the heated bundles where the flow of
production fluids is executed through internal pipe whereas a heated fluid is flowed
through an outer pipe [2]. On the contrary, individuals must be cautious while
implicating such method. Figure 2-34 has shown a drastic and dangerous situation. It
is examined from the figure that liquid water will be produced and gas will be
released through the dissociated plug. Volume of 1 m3 of methane hydrate discharges
170 Sm3 of gas. The dissociating also produces 51.45 kmol of water as a liquid that
inhabits 0.927 m3 as a liquid. This refers that if the hydrate is dissociated in a limited
space, there is merely 0.073 m3 (1m3 - 0.927 m3) available for the 170 Sm3 of
released gas.[23]

According to [91], the ideal gas law can be used crudely to estimate the pressure of
released gas.

30
 59 69 170 × 101.325
59 69 = 57 67 @A 57 = = = 236 J5K Eq 2-7
62 0.073

We can see From

Eq 2-7 the pressure is independent of the volume dissociated as long as it occurs in a

confined space.
More serious scenario occurs with multiple plugs form in series in the pipeline, as
shown in Figure 2-34 (d). Multiple plugs can trap high intermediate pressure, so
more precautions should be taken by decreasing slowly the pressure of both sides of
the plugs to maintain thermal and hydraulic control of the clearing process. Instead,
if there is movement in hydrate plug, the pressure build-up in the melted section can
result also in a hydrate projectile, with a noteworthy potential for instigating damage.

Just Dummy Figure

(a) Heating technique is applied near the centre of the plug, to increase the
temperature and dissociate the hydrate plug.

Just Dummy Figure

(b) With continues heating, hydrate plug began to dissociate causing rise up in
the pressure.

Just Dummy Figure

(c) With continues pressure build up due to dissociation, the pipeline bursts.

31
 Just Dummy Figure

(d) Multiple hydrate plug that trap intermediate pressure, causing pressure build
up due to hydrate dissociation, resulting in pipeline bursts.
Figure 2-34 Incorrect heating of hydrate plug in high pressure pipeline causing
pipeline rupture. Modified after [11, 23]

Following the above techniques, chapter 3 will present a field case study of various
gas hydrate dissociation/mitigation techniques applied in a gas lift system of a south
field of Oman, which includes:

§ Installation of rock-wool insulators

§ Installation of electrical heat tracing
§ Decreasing the system pressure below hydrate stability point
§ Methanol injection
§ Increasing gas lift temperature.

Thermodynamic Inhibitors
The addition of thermodynamic hydrate inhibitors (THI) effectively shift hydrate
equilibrium curve to the left side towards higher pressures and lower temperatures
relative to normal operating conditions. Therefore, the inhibition of water molecules
participation is formed with the hydrate structures and the performance of inhibition
is greatly relied in aqueous phase on the MEG concentration [93]. The equilibrium
curve can be transmitted to left side of original curve with increased MEG
concentration. Clustering effects are two hydroxyl groups that form hydrogen
bonding with water molecules to represent a molecule of MEG. The formation of
hydrogen bonds is comparatively similar to hydrate formation. THI are added at
relatively high concentration of about 10 wt% to 60 wt% in the aqueous phase. [92]
The hydrate equilibrium data in the presence of different concentration of MEG (0
wt% to 50 wt %) is shown in Figure 2-35, and as shown also by other
researchers.[30, 93, 94]

32
The consequent reduction in the water activity coefficient and the dilution of the
water phase are the primary thermodynamic indicators of the mechanism, mitigating
the impact of hydrate formation [94].

Just Dummy Figure

Figure 2-35 Effect of addition of different concentration of thermodynamic inhibitor

(MEG) on shifting hydrate equilibrium curve of system of natural gas (Methane
79.1%, CO2 2.5%, iso-Pentane 1.7%, n-Pentane 1.7%, iso-Butane 2%, n-Butane 2%,
propane 4%, Ethane 7%) , plotted by Multiflash prediction software (PR equation of
state).

Traditional thermodynamic inhibitors include methanol (molecular weight 62.07)

and ethylene glycols [monoethylene glycol (MEG, molecular weight 62.07),
diethylene glycol (DEG, molecular weight 106.12) and triethylene glycol (TEG,
molecular weight 150.17)], the lower the molecular weight, the better the hydrate
suppression performance.[95] The gas hydrate equilibrium data of natural gas with
25 wt % of different thermodynamic inhibitors, compared to 100 wt% water, are
plotted in Figure 2-36. The hydrate depression temperature and the regression
functions of the fitted data are reported in Table 2-3. For a given pressure, the
hydrate depression value ( ∆ 9! ) was determined as below (Eq 2-8):

∆ 9! = 9*+: (955 <$% <>$*-)@ 9*+:(7A <$% BC 12D) Eq 2-8

Where 9*+: (955 <$% <>$*-) is the hydrate equilibrium temperature of 100 wt% water
and 9*+:(7A <$% BC 12D) is the hydrate equilibrium temperature measured at 25 wt%
of different THI’s. A higher “∆ 9! ” value corresponds to a higher depression (better
inhibition performance).

33
Summarising results from Table 2-3 and Figure 2-36, Methanol showed superior
hydrate inhibition performance in terms of shifting the hydrate curve most to the left
side [with average depression value (∆ 9! ) of 12.9 oC], followed by MEG, DEG and
TEG respectively.

Just Dummy Figure

Figure 2-36 Effect of addition of 25 wt% of different thermodynamic inhibitors on

shifting hydrate equilibrium curve of system of natural gas (Methane 79.1%, CO2
2.5%, iso-Pentane 1.7%, n-Pentane 1.7%, iso-Butane 2%, n-Butane 2%, propane 4%,
Ethane 7%), plotted by Multiflash prediction software (PR equation of state).

Table 2-3 Hydrate depression temperature “∆ 9! ” of Brustad, Loken [95] and of

Figure 2-36, and the regression functions (sorted from highest to poorest inhibitor),
where P is pressure and T is the temperature.

Pressure (bar) versus ∆ 9! (oC) Average

Regression ∆ 9! (oC)
functions of 50 100 150 200 250 300
different THI Multifl- Brusta
ash d,
software Loken
[95]
Methanol:
P (methanol) =
6
0.0004 T – 12.9
0.0167 T5 + 12.8 12.8 12.9 13.0 13.0 12.9 12.3
0.2683T4 - 1.7749
T3 + 5.4311 T2 –
1.4058 T + 36.422
MEG:
P(MEG) = -0.0005
T5 + 0.0235 T4 – 7.1
7.6 8.0 7.9 8.1 8.1 6.1 7.6
0.2596 T3 +
0.9525 T2 +
2.5605 T + 20.902

34
 DEG:
P(TEG) = -0.0006
T5 + 0.0366 T4 – 5.0 4.9 5.0 5.1 5.1 5.0
0.7226 T3 + 6.347 5.0 4.6
T2 – 22.058 T +
48.992
TEG:
P(DEG) = -0.0004
T5 + 0.0219 T4 – 3.9
4.6 4.2 4.1 4.1 4.1 4.0 4.2
0.4192 T3 +
3.4254 T2 –
9.1563 T + 28.593

Analysing Table 2-3, we can see that hydrate depression temperature “∆ 9! ” of

Brustad, Loken [95] followed the same pattern of Multiflash prediction software
(Figure 2-36) with average deviation value of 7%. Thermodynamic inhibitors can be
emerged in the form of ionic salts; such as, sodium chloride that might be exist in the
water formation. The hydrate formation point is suppressed by restricted kinetic
hydrate inhibitors as such inhibitors are accepted at higher pressures [95, 99].
Obanijesu, Barifcani [41] have reported that hydrate inhibitors can be emerged as an
inert gases, including hydrogen and nitrogen. On the contrary, dilution effect reports
the depression of hydrate formation and; therefore, more researches are required to
establish the chemical nature of H2 and N2 that vitally contributes in hydrate
formation depression. These ionic salts can further be used in the organic inhibitor,
deepwater projects and salts inhibitor [96]. Low-Dosage Hydrate Inhibitors (LDHI),
such as Kinetic hydrate inhibitors (KHI) and Anti Agglomerants (AA) becoming
popular in West Africa, UK fields, and the Gulf of Mexico [97, 98]. In contrast,
many limitations were occurred specifically for long distance gas-condensate tie
backs. In general, continuous condensation can be effectively required for long
distance gas tie-backs.

Low-Dosage Hydrate Inhibitors

A simple observation executes the concept of low-dosage hydrate inhibitors reveal

that particular fish did not freeze in sub-zero temperature as microscopic ice crystals
are bounded on the secretion of a protein and consequently it prevent its subsequent

35
growth. The discovery of kinetic hydrate inhibitors is driven through the evidence of
anti-freeze proteins [98, 100]

The presence of low-dosage hydrate inhibitors is comparatively new in the oil and
gas commerce. The inhibition of hydrate growth and nucleation is emerged at very
low extent as low-dosage hydrate inhibitors are chemicals [101]. The classification
of low-dosage hydrate inhibitors is reflected from the inhibition mechanism of
antiagglomerants and kinetic inhibitors.

2.8.1 Kinetic Inhibitors

Kinetic inhibitors are considered as LDHIs, which alter the hydrate formation
kinetics by reacting and increasing the time of hydrate formation to occur by
delaying the initial hydrate nucleation. These inhibitors are generally polymers,
which are water-soluble which suspend the formation of hydrate crystals, such as
Luvicap® EG and Gaffix® VC-713 (Figure 2-37) [101]. Moreover, kinetic inhibitors
can adsorb growing hydrate crystals at the hydrate/water interface, preventing small
hydrate crystals to grow into larger crystals; therefore, it slows down the rate of
growth and prolonging the period of time before catastrophic growth occurs. This
delay in hydrate growth means that one may operate within the hydrate stable area of
the phase diagram for a given length of time without the appearance of hydrates
[102]. Kim, Shin [99] evaluated the synergist function of adding 0.2 wt% of PVCap
with 20 wt% MEG and confirmed that this results in 36% longer delay time and the
MEG concentration can be condensed by 20 wt%. Conversely, there are two major
drawbacks of the kinetic hydrate inhibitors accessible in the recent era. The first
drawback of kinetic hydrate inhibitors is that it is only significant when the sub-
cooling is slightly least than 14°C while the second drawback is the occurrence of
other injected chemicals including corrosion inhibitors [99]. It is evident that the
effectiveness of kinetic inhibitors are shown at higher pressures [95].

Just Dummy Figure

36
Figure 2-37 Chemical structure of Luvicap® EG (a) and Gaffix® VC-713 (b). [103,
104]

2.8.2 Anti-Agglomerants
The clustering of hydrate particles is prevented by anti-agglomerants, that has led to
the plugs formation to process lines and block pipelines. The surface of hydrate
crystals in the agglomeration is prevented through anti-agglomeration, which assists
in stabilizing the crystals. The key to the AA effectiveness is their structures and
surfactant properties. AA surfactants are thought to work by containing polar head
groups that can interact with the lattice of hydrate water molecules, and a
hydrophobic tail group that attracts the hydrocarbon phase [105, 106]. Therefore, the
presence of anti-agglomerants can be determined in an enduring slurry hydrocarbon
stage [101].

Shell described a successful LDHI trial in their Popeye subsea well [98].The well
suffered high watering out outstripped the systems capability to inject the required
methanol rate (250 bpd). The system was only capable to pump 175 bpd which lead
to partial hydrate blockage. As a quick solution, AA was executed at 0.35 gal/bbl
water (0.8% of the water volume) giving a 95% reduction in chemical usage
compared with Methanol. In this trial, AA implementation showed positive results
and the well was opened up resulted in additional of 20 mmscfd of gas production.
Shell estimates a net present value of $8 MM improvement because of the
implementation of the LDHI. [107] AA trial results are illustrated in Figure 2-38.
Kim, Shin [99] stated that the AA have a drawback restraint in constraining
performance with high water cut wells, while Popeye subsea well trial proves it is
efficiency.

Just Dummy Figure

Figure 2-38 Case history of Deepwater Gulf of Mexico where injection of xxx
permits extra gas production in Methanol limited system after [107]

37
LDHIs have a huge application potential to replace the thermodynamic inhibitors. As
of today the way of how LDHIs work at a molecular level is not yet fully understood
or documented, even though they have now been applied in the field.[102] However,
the use of LDHIs is restricted on the Norwegian continental shelf due to their toxicity
and biodegradability. The main concern is their low biodegradability. The work of
developing new and more environmentally friendly LDHIs is currently
ongoing.[108-110]

Hydrates in Natural Gas Production and Transport Systems

Hydrates formation in gas pipeline systems represents a severe flow assurance issue
in the oil and gas industry, especially because gas hydrates can cause flow blockages,
which can severe safety and operational issues [111]. The occurrence of hydrate
formation is emerged formation water is contained in the fluid production [94].
Hydrate crystals will increase and gather into plug in case hydrate formation is not
inhibited and the comprehensive cross-sectional area of the pipeline is blocked
(Figure 2-39).

Just Dummy Figure Just Dummy Figure

Figure 2-39Gas hydrate plug in a pipeline after [112, 113]

Figure 2-40 showed the common locations for hydrates formation in a subsea
petroleum production system and include the wellhead, flowline and riser. Hydrate
formation was occurred during the onshore fields as the gas as operated at high
pressure (above hydrate equilibrium point). Chapter 3 will present hydrate formation
at onshore gas lift system. Hydrate formation can occur in different places in the oil
and gas industry, depending on many factors; the operating pressure and temperature,
the nature of the hydrate forming component (i.e. type of gaseous guest molecule,
single- or multiple component gas), the composition of the water phase (pure water
or water with condensate or dissolved salts/inhibitors).[114]

38
 Just Dummy Figure

Figure 2-40 Probable locations of hydrate formation in an offshore system after [11]

Just Dummy Figure

Figure 2-41: Hydrate formation during winter season at Gas lift manifold caused by
drop in ambient temperature and high differential pressure across the control valve
(Joule –Thompson effect) (PDO custody)

The correlation of plug formation is revealed with the events and following elements.
Thereby, the elements and events must be neglected [115].

• Start-ups following emergency shut-in

• An uninhibited water phase
• Sudden reduction in pressure is influenced from Joule-Thompson at orifices
specifically include short radius elbows, open control valves, sudden
enlargement in pipelines.

One of the greater risk is the long distance transmission lines of hydrate formation
because of their raised pressures of about 7 MPa established by compression to
preserve optimum operating conditions for transmission. [116]

Locations of water accumulations, such as “S” shapes in flowlines (Figure 2-42),

should also be minimized. Pipeline topography which provokes water accumulations
are particularly vulnerable and may require pig or inhibitor injection to prevent
hydrate formation. For many fields it is practically unenforceable to design and
operate hydrate-free systems. In many cases this is due to seabed topography. If a
flow line requires hydrate inhibition, it is very likely that hydrates will form during
the operational lifetime. This emphasizes the importance of identifying the likely

39
points of hydrate formation so that hydrate prevention and dissociations can be
addressed.[115]

Just Dummy Figure

Figure 2-42 Hydrate plug formations in "s" shapes

Mono Ethylene Glycol

The appearance of ethylene glycol is viscous, odourless, clear and colourless liquid
along with a pleasant taste through which dramatic toxicity can be produced. The
direct oxidation of ethylene is the currently effective technology for creating ethylene
oxide. At the starting of 1930s, the direct oxidation of ethylene was developed by a
French firm and afterwards restructured by union carbide. Ethylene glycol is
commonly found in automotive cooling systems and antifreeze systems; therefore, it
is emerged as a raw material in industrial settings. Ethylene glycol is further used in
different processes such as gas hydrate inhibition. Union carbide in 1920s commence
its commercial production even though ethylene glycol was manufactured in 1856.
New technologies have been developed by shell and the engineering firm scientific
design in 1940s on the basis of ethylene oxidation [117]. The call for MEG is
expected to be 17 million tons per year internationally with the expected growth rate
of almost 7%/year [118].

Table 2-4 compares some of the selected physical properties of the most communally
used thermodynamic inhibitors, MEG and Methanol. Monoethylene glycol is a diol
(alcohols that have two hydroxyl groups in each molecule) with the chemical
formula L7 ME (NM)7 . MEG is entirely is hygroscopic and miscible in water, which is
able to absorb 200 percent of its weight in water at 100 percent relative
humidity.[119]

Table 2-4 Physical properties of MEG and Methanol [120]

Mono-ethylene Glycol Methanol

40
 Family Glycol Alcohol
Representation

Chemical C2H4(OH)2 CH3OH

Formula
Appearance Colourless liquid Colourless liquid
Molecular Weight 68.068 g/mol 32.04 g/mol
Viscosity (cp) @ 21 centipoise 0.55 centipoise
20 oC
Density (g/cc) @ 1.1135 g/cm3 0.9715 g/cm3
20 oC
Freezing Point , -12.9 oC -97 oC
o
C
Boiling Point, oC 197.3 oC 64.7 oC
Flash Point 111 oC 11 oC
Solubility in Fully miscible Fully miscible
Water
NFPA 704 rating
and GHS
pictograms

- Flammable
Toxic when ingested - Toxic when ingested

2.10.1 Hydration of Ethylene Oxide to Produce Ethylene Glycol

Ethylene oxide produces MEG that is often represented in the form of catalytic or
thermal process. Liquid-phase hydration is the most common method of ethylene
oxide hydrolysis [121]. The occurrence of ethylene oxide and feedstock is
determined through multi-tubed catalytic reactor where oxygen and ethylene are

41
combined across a solid-bed of silver catalyst at high purity. The exact specifies of
the reactor operated is dependent on the reaction of temperatures such as 1-3 MPa,
and 210 ℃ to 285 ℃ [122]. EO production selectivity is boosted by the use of silver
catalyst. Silver is acted as a catalyst that adsorbs the silver ion surface in order to
form a ionized superoxide and highly commendable ethylene [123].

The purification of the glycol products was exceeded by distillation columns for the
product stream, which operated glycol products at a relatively lower pressure. The
energy generated from the conversion of EO were used for heating the distillation
columns to MEG. MEG, DEG and TEG. Then these were separated using vacuum
distillation as shown in Figure 2-43. It has been reported that higher polyethylene
glycols are habitually not separated because of their low concentrations. The
occurrence of a non-catalytic liquid stage hydrolysis reaction can be determined
through commercial production. A non-catalytic liquid hydrolysis will exceed at
temperatures 140 - 230℃ with a volume of water existing inside the reactor. The
construction of higher glycols is prevented with the use of water in the reactor as a
medium. Therefore, the formation of ethylene oxide and ethylene glycol is combined
together because of water usage [124].
In this conventional reaction system, glycols production reaction mechanisms shown
in below equations, Eq 2-9 (EO + Water = MEG), Eq 2-10 (EO + MEG = DEG) and
Eq 2-11 (EO + DEG = TEG) [118].
L7 ME N + M7 N → L7 ME (NM)7 Eq 2-9

L7 ME N + L7 ME (NM)7 → LE M95 N8 Eq 2-10

L7 ME N + LE M95 N8 → LF M9E N8 Eq 2-11

Figure 2-43 is the block flow diagram showing the statistically revelation of
successive glycol reactions, which is identified through the feed ratio of water and
EO. The selectivity of MEG is increased with the dilution of EO and large excess of
water. For instance, 89 percent selectivity of MEG needs 20 mol of excess water to 1
mol of EO. [118]

42
 Just Dummy Figure

Figure 2-43 Block flow diagram of conventional ethylene oxide (EO) to MEG
process. After [118]

Numerous patents were filed for the production and obtaining of MEG with few
processes operational activities due to financial restraints and technologic
developments.[125-131]. Appropriate catalysts are used to attain high selectivity;
therefore, it is considered as one of the simplest ways of technological developments.
This concept was based on the lot of efforts. For instance, catalyst were investigated
with some metal complex anions [128]. van Hal, Ledford [121] has aimed to explore
different types of catalysts of EO to MEG for catalytic hydration including Salen
compound catalysts, amine and bi-function catalysts. It has been examined that the
consistency of acid or base catalysed reactions mechanisms is dependent on the
acidity and basicity of the preferred catalyst. The mechanisms is described in
reaction schemes as shown in Figure 2-44. A proton first attacks the nucleophilic
oxygen of an EO molecule in order to create an intermediate species, LME (NMG )LM7
which subsequently changes to a more stable one,G LME (NM)LM7 for acid catalyzed
reaction. H2O molecule is reacted with the second intermediate species to create a
Monoethylene glycol. The reaction rate has been increased with both strongly acidic
and basic catalysts. Conversely, the selectivity of MEG formation was not increased
using the same catalysts. MEG and the higher glycols are the two mechanisms that
lead to the sequences of desired and undesired products. The core intermediate
species are distinctive due to the difference in basicity and acidity of the catalysts
used. It releases a proton at the same time that maintains the reaction solution
concerning its constant acidity. Another intermediate species could react with
another EO in order to produce Diethylene glycol. Higher glycols and Triethylene
might be produced with the same mechanism based on the base catalysed followed
by the same reaction as shown in Figure 2-44.

43
 Just Dummy Figure

Figure 2-44 Schematic diagram of reaction mechanisms of acid and base catalyzed
hydration of ethylene oxide (L7 ME N ) to ethylene glycols. After [121]. The OMEGA
process was developed by Shell Global Solutions with the partnership of Mitsubishi
chemical and utilised catalysts of CRI Catalyst Company for catalysing the ethylene
conversions to EO to MEG. 99% of EO were converted into MEG, which were
capable of converting it effectively.

It is not the production of TEG, DEG or different other less preferential polyethylene
glycols. Shell has demanded that the OMEGA processes, which are around 30% less
wastewater and 20% less steam is utilized by the company, justifying the greenhouse
emission of gases in contrast with the traditional EO/MEG processes [132],
additionally with the higher MEG yields of between 14.7-27.5% per tonne of
ethylene. Ethylene carbonate is encompassed in OMEGA process as it is created with
existence of phosphorus halide catalyst and carbon dioxide. A small quantity of
water is added for the hydrolysis reaction into the ethylene carbonate, which
establishes MEG and carbon dioxide. A little quantity of water is then added in to the
ethylene carbonate, not the 22:1 by mole basis required conventionally.
Carbondioxide is then recycles and added into the feed stream. Additionally, to the
enhanced process effectiveness, the cost of OMEGA plant is less than in both
operational and capital expenditure [121].

MEG Regeneration and Reclamation Systems

MEG is expensive and used in large amounts; therefore, it is essential to recycle it
[133]. Maintaining a higher reliability of MEG supply is profoundly reliant on good
MEG regeneration and reclamation system. If the only thing to worry about the
system is separating water from MEG, then the design of an MEG Regeneration Unit
(MRU) would be simple. However, there is a complexity involved in the design of
the production facility and MRU to mitigate the impact of these contaminants on

44
MRU operation due to the presence of various contaminants in the pipeline [134].
MEG reclamation and regeneration systems may be described as a closed loop
systems. Where, ‘Rich MEG’ is shaped from the wellhead and conceded through a
MEG regenerator and/or reclaimer where ‘Lean MEG’ is injected and produced at
the wellhead to constrain hydrate formation.[135]
The closed loop MEG regeneration system comprises of three operational areas; feed
blending, Pre-treatment and Regeneration and Reclamation [136, 137]. These
operational areas are illustrated in Figure 2-45 for the recently constructed MEG
pilot plant by the Curtin Corrosion Engineering Industry Centre (CCEIC) [138]. This
pilot plant will be used in Chapter 7 to study the efficiency of thermodynamic
hydrate inhibition of MEG solutions that collected from both the regeneration and
reclamation outlet.

Just Dummy Figure

1) Feed blending 2) Pre-treatment 3) Regeneration

Figure 2-45 MEG pilot plant operational areas

There are three options available in the design of MEG regeneration and reclamation
plants: slip-stream reclamation, full reclamation, and conventional regeneration;
and.[139, 140]

2.11.1 Convention Recovery Model

The process of convention regeneration is the least preferred and least employed
approach due to the requirement of degrading MEG and water separation. Mutual
applications for the procedures occur in the conditions, having decreased level of
total dissolved solids, inside the recovery stream of MEG [141]. The procedure
mechanism over the water removal by distillation, with similar pressure reserved as
atmospheric pressure and the temperature of the system is considered relied on the
specifications for the manufactured lean Mono-Ethylene Glycol stream purity. A

45
reboiler would be intricated in a typical system, linked to a kettle or recycle heater,
which usually flows into a column of distillation.
Another issue with the application of traditional process of regeneration is the part of
contaminants in the product stream of MEG. While, in the distillation column, the
water has been boiled off. MEG stream will be only left with the pipeline corrosion
salts and chemicals within the distillation column. Degradations will occur in the
system where the stages of contaminants will be low with respect to cycle, even in
the systems that needs cleaning and replacements of MEG to make sure the
effectiveness. Conventionally, many systems are not created to endure the high
volume of the water formation, particularly from wells. It will commonly need a
water separator before re-injecting MEG [134]. Shell Mensa plant has gained much
importance, which is situated in the Gulf of Mexico. The procedure or the system
utilizes the MEG for both dehydration and hydrate control and has experienced the
operational issues linked with the MEG regeneration plants. The challenges have
been discovered from the applications of the system as a result of trails based on
operational plants system. Conventional Reclamation method was used by three
plants along with some drawbacks. The drawbacks included the Statoil/Gassco
Kollsnes, Total South Pars (National Iranian Oil Company), and Shell Mensa. As
completion fluids and formation water entered into the stream of recovery, the
traditional regeneration process transfers the higher level of plugging and scaling
within the system. Issues were also suffered through the contamination of Mono-
Ethylene Glycol stream product along with some blockages created within the subsea
injection lines [140].

2.11.2 Full-stream Reclamation Model

Full reclamation process is one of the most employed process (Figure 2-46), which is
used for regeneration within industry along with addition of slipstream method if
required. The procedure is more focused on rich Mono-Ethylene Glycol monitored
by boiling and distillation in a flash separator to gain appropriate lean Mono-
Ethylene Glycol product. It discourses the major issue of the traditional reclamation
procedure, due to this it can lodge higher rates of water formation and deal suitably
with the dissolved solids [141].

46
 Just Dummy Figure

Figure 2-46 Full-stream Reclamation. Modified after [142]

An effective and typical reclamation process commences with a stage of pre-

treatment, where MEG is heated and depressurized within a separator having three-
phase for elimating hydrocarbons from the mixture. Vacuum operated (typically
0.10-0.15 bara) flash separator takes the flow of MEG stream within it in order to
augment the stream purity and eliminate contaminants. It has been evaluated that the
flash separator usually vaposises MEG by using heat at lower temperature for
preventing the process of decomposition and elimination of the contaminants. The
contaminants removal may include non-volatile chemicals, particles, and salt. The
rich MEG exists as vapours from the Flash Separator. The outgoing flows of the
product stream is concerned with the vaporized MEG and some flow of water into
distillation column for separation. The liquid product stream flows into a centrifugal
decanter for the elimination of contaminants from the cycle [140].

Just Dummy Figure

Figure 2-47 Full-stream MEG reclaimer in Gulf of Mexico after [143]

2.11.3 Slip-stream Reclamation Model

The salt removal process of slipstream is an edition of traditional regeneration
procedure that uses an ion exchanges for eliminating the flow slipstream for the
purpose of re-use (Figure 2-48). Although, a traditional process of regeneration has a
reclaimer applied to the complete elimination of salts and non-volatile chemicals. pH
inhibitors and stabilizers are also removed through salt removal procedure for re-

47
using it. It may advantage the requirements of energy because of lower level of
contaminants exist before the distillation within the Mono-Ethylene Glycol stream.
When operators need the highest recovery of Mono-Ethylene Glycol, the systems of
slipstream are usually structured supposing that operators can endure the higher
manufacture price of the system in contrast with the full reclamation model. These
particles will exist over time an accumulation inside the Mono-Ethylene Glycol loop
of the system. The most prominent issue observed in the industry with the slipstream
salt removal process is the pH stabilizing and inhibition of particles that are being re-
used. It postures important issues for the scaling issues and equipment safety, if
blockages occur within the injections pipeline. A slip-stream model will be used to
investigate the efficiency of thermodynamic hydrate inhibition of MEG solutions
collected from a MEG pilot plant, simulating six scenarios of the start-up and clean-
up phases of a typical gas field.

Just Dummy Figure

Figure 2-48 Slip-stream MEG reclamation modified after [139]

The comparative merits and demerits of MEG reconcentration and full/slip-stream

MEG reclamation are compared in table 3.2.

Table 2-5 Comparative merits and demerits of MEG reclamation systems

Operating model Disadvantages Advantages

Convention It is unable to handle the Least expensive of options
Recovery limited usage of
continuous formation
water production. Non-
volatile chemicals and
salts gather in the closed
group

48
Full- Higher capital The non-volatile chemicals and salts
Scream expenditure & larger size are removed.
reclamati It may control the higher water rates
on formation

Slip- MEG’s viscosity and Reconcentration is not completely

Scream density influenced by relied on the MEG reclaimer
reclamati impurities and salt. operation.
on Improved chance of Superior flexibility in administering
corrosion and plugging MEG reclaimer
Capital expenditure and condensed
size
pH stabilisers and chemical extracts
may be reused and reserved
Lower cooling and heating
responsibilities with less MEG
vaporised.

49
 Gaps in Literature
It is important to minimize the oxygen level and thermal exposure inside the closed
loop MEG regeneration/reclamation structure to prevent possible degradation of
MEG [140, 141, 144-146]. Furthermore, the rate of MEG degradation is accelerated
by high temperature and metal ions of solutions. The study has suggested that formic
acid and glycolic are main products of MEG degradation [147, 148]. Before this
study, the accessible courses on MEG degradation were concentrating primarily on
MEG’s influence on the aspects of chemical identification of MEG degradation
product [145] and the influence on increasing corrosion of metallic components. No
research is conducted about the effect on inhibition performance. AlHarooni,
Barifcani [149] bridged this gap by identifying the influence of thermally degraded
pure MEG on gas hydrate inhibition; while AlHarooni, Pack [135] studied the effects
of thermally degraded MEG with film formation and methyl diethanolamine
corrosion inhibitor on gas hydrate inhibition. They further identified the analytical
techniques that can be utilized to recognise the severity of thermal MEG degradation
and developed a novel MEG thermal degradation scale. Moreover, this scale also
provided evaluation of used MEG to review the flow assurance strategies in terms of
MEG doses and corrosion protection strategies [38]. As there is a knowledge gap in
evaluating MEG inhibition performance; it has been regenerated and reclaimed,
further to investigate the efficiency of thermodynamic hydrate inhibition. Their work
opened a new area of research interest on thermal MEG degradation-hydrate
relationship; and the association between the final product of MEG regeneration and
reclamation with gas hydrate inhibition performance. In the following chapters, the
study will follow:
• Gas Hydrate in Gas Lift system
• Influence of inhibition concerning thermally degrade MEG on hydrate
formation for gas systems
• Effects of thermally degraded Monoethylene Glycol with Methyl
Diethanolamine and Film-Forming Corrosion Inhibitor on Gas Hydrate
Kinetics
• Thermally Degraded Monoethylene Glycol with Methyl Diethanolamine
and Film Formation Corrosion Inhibitor should be analysed with
analytical techniques

50
• Influence of regenerated Monoethylene Glycol on natural gas hydrate
formation

51