Mechanism for conversion from methane hydrate over

Mechanism for conversion from methane hydrate over to carbon dioxide dominated hydrate during injection of carbon dioxide into methane hydrate filled sediments

International meeting on Petroleum Engineering November 7, 2017, Singapore

*

Bjørn Kvamme, Source: Arthur H. Johnson, Hydrate Energy International

Professor Bjørn Kvamme Department of Physics and Technology, University of Bergen, Bergen, Norway [email protected]

Why are hydrates in porous media so complex • Because:

0.9

20

0.8 40

- dynamics of hydrate phase transitions are implicitly coupled from nano-scale to macro-scale - many significant phases are not well described yet

- click to edit 60

Length(Å)

- hydrates in porous media Normal text can never reach thermodynamic equilibrium

0.7 0.6 0.5

80

0.4 100

0.3

120

0.2 0.1

140 20

40

60 80 Length(Å)

100

120

140

Injection of CO2 into a CH4 hydrate filled pore results in formation of new CO2 hydrate. Released heat assist in dissociating CH4 hydrate. CH4 escapes as bubbles before dissolving into CO2

What governs hydrate formation ? • Possibility to reach equilibrium (Gibbs).

Is there a balance between defined independent thermodynamic variables, conservation laws and conditions of equilibrium? Normal text - click to edit

• Initial distance from equilibrium in all independent thermodynamic variables • Combined First and Second laws of thermodynamics. In terms of Gibbs free energy it implies that all systems will stribe towards minimum free energy as function of temperature, pressure and distribution of masses in the system over possible phases, under constraints of mass and heat transport. Few hydrate systems in industry and mature can reach equilibrium. • Kinetics (combined thermodynamic control, mass transport dynamics and heat transport dynamics) 3

Water adsorbs well on glass and results in efficient nucleation sites. Stainless steel is neutral. Sediment minerals structure water

Which phases are significant for hydrate phase transitions during pipeline transport or flow though porous media? • Starting with a simple system of CO2 hydrate forming during CO2 storage in reservoir with hydrate forming zones

click - Initially there Normal is one activetext phase-since the solid is thermodynamically neutral (just excluded volume)

to edit

- P and T is constant so driving forces here are concentrations (corresponding chemical potentials) and after hydrate forms there are two active phases - Technically (Gibbs) the system can reach equilibrium -In summary: But the neutral solid particles play a Thermodynamically neutral particles do notrole add in to the of because active phases they thenumber kinetics the but confine affect the kinetics of hydrate formationand available space for movement

Hydrate (yellow) forming from dissolved CO2 in fluid (red) in vicinity of neutral solid (black). T=1 C, P=150 bars, initial mole-fraction CO2 is 0.033 (Henry’s law). Hydrate 4 is 0.016 stability limit

What now if the solid surface has distribution of negative and positive atoms?

Kaolinite is a clay mineral

For the tetrahedral cutting direct adsorption of CO2 is feasible (see free energy change for CO2 on right figure) before first maximum for water while secondary adsorption in water density minimums might occur in both cases. Figures from Leirvik, Kvamme & Kuznetsova [1]. From thermodynamics these adsorbed phases are unique phases because density and compositions are unique. Are they significant for hydrate phase transitions? - yes because the lowest chemical potential of water is far lower than hydrate water chemical potential so the solid surfaces are hydrate inhibitors - Yes because they serve as hydrate nucleation sites

Will the same be the case for transport of methane in a rusty pipeline? Water chemical potential in the adsorbed layer will vary proportional to the density and structure so the average chemical potential of four adsobed water layers is what is given below H2O Chemical Potential H2O Klusterv/245K

‐50,7 kJ/mol

H2O Hematittv/245K

‐54,7 kJ/mol

H2O Klusterv/278K

‐52,8 kJ/mol

H2O Hematittv/278K

‐56,2 kJ/mol

Adsorbed Water chemical potential may be in the order of 3.4 kJ/mole lower than liquid water so technically it is a hydrate inhibitor

But direct and indirect (note the dynamic «pockets») adsorption of hydrate formers pluss beneficial heterogeneous hydrate nucleation makes these solid pipeline surfaces very active in hydrate phase transition dynamics

6

A practical consequence – the maximum water content that will be accepted based on water drop-out as liquid (conventional criteria) may be 20 times higher than what will be accepted based on adsorption on Hematite

Normal text - click to edit

Maximum water content before liquid drop-out (left) and adsorption on hematite (right) respectively. Mole fractions 0.01 CO2, 0.001 H2S, and remaining gas being CH4.Curves from top to bottom correspond to pressures of 50, 90, 130, 7 170, 210, and 250 bar.

So what if the walls are plastic and methane wettening rather than water wettening? Experiments with methane watertoatedit 83 bar and 3 C Normal text -and click (resolution ~ 100 micrometer) Two half cylinders of polypropylene with diameter 4 cm and lenght close to 10 cm separated by a 4 mm spacer with channels for natural fluid flow



Why?

Two primary factors:

1) A methane hydrate film will rapidly form on the water/methane interface and reduce efficiently further growth untill film penetratesdue to local competition based on first and second laws of thermodynamics 2) Methane is the wetting component of the silicone rubber and some methane will migrate along the walls downwards in the chamber due to ill f

Gibbs Phase Rule No. Of deg. Of freedom

No. Of components

  N  2 No. Of Phases

•

Gibbs phase rule is actually very trivial and it is hard to see why it even has a name credited to the statement/equation. It is simply: - Number of independent thermodynamic variables (temperature, pressure and masses in all phases) minus conservation laws minus conditions of equilibrium

And the resulting number is the number of independent thermodynamic variables that must be fixed in order to make equilibrium possible. This is the degrees of freedom that appears in Gibbs phase rule. Local temperature and pressure in a reservoir or in a pipeline is always given so any number of degrees of freedom different than two excludes possibility for full thermodynamic equilibrium

So – if the system cannot reach true thermodynamic equilibrium – then there is no rule that says chemical potential of hydrate formers is the same in all phases

H H,0 i  (T,P,x ) w (T,P)  j ln(1 hkj ) H,i w

j

h

i k j



e

 

( 

i k j



k

 g

i n c l u s i o n k j

What is chemical potential for the guest in the «parent» phase ? What is the resulting free energy of that specific hydrate phase ?

Right: CO2 (enhanced red and grey) adsorbing onto Hematite from water solution. Adsorbed CO2chemical potential: -39.21 kJ/mole at 274 K

12

)

Normal text - click to edit At 150 bar and 274 K the hydrate controlled CO2 solubility is Xco2=0.016. Liquid solubility of CO2 is Xco2=0.033. Hydrate growing from Xco2=0.033 (top) and from Xco2=0.036 (below)

So any change in chemical potential of guest molecules due to changes in concentrations will lead to a new hydrate (composition and density will be different and by definition it is a new phase)

13

Non-equilibrium summary • The minimum number of significant phases for one hydrate former in a pore is:

Hydrate can grow from water super saturated with CH4 with reference to hydrate (in between red and blue) and dissocuate below blue

- Hydrate former phase - Liquid water - Adsorbed phase on mineral surfaces - Initial hydrate forming - Adsorbed phase on initial hydrate - Additional hydrate phases (hydrate former from aquous solution, adsorbed, from gas, water from various phases) So 5+ more phases for 2 components gives a system over specified by min. 3 independent thermodynamic variables when local T and P is given by local flow and statics

10 waters in 2038 methane is a very high water concentration compard to water saturated methane at 100 bar. Water diluted in the gas struggle to totally outcompete methane in adsorption. Buy still manages after 2 ns

15

Secondary adsorption of methane (trapped in adsorbed H2O) Methane on Calcite and a water slabe at 264 K

16

17

Experimental results on adsorbed water gives details on each atom The cusp in the oxygen density between first and second maximum is due to indirect correlations which involves hydrogen. Locations of main peaks and minima is in striking agreement. Also integrals over density profiles are in striking agreements.

P. Geissbühler, P. Fenter, E. DiMasi, G. Srajer, L.B. Sorensen, and N.C. Sturchio. Three-dimensional structure of the calcite– water interface by surface X-ray scattering. Surface Science, 573(2):191–203, 2004

Secondary and primary adsorption of guest molecules Primary adsorption is adsorption directly on solid surface. Secondary adsorption is trapping of hydrate formers in strctured adsobed water

Simple water models like SPC, TIP4P and many other have no short range interaction on H atoms so density profile based on ozygen locations

Note the extreme density of the first adsobed water layer but also the subsequent water density minimum which gives space for trapping CO2. Simulated structure is in accordance with experimental 19 data (IR)

Some of the routes that can lead to CO2 hydrate during pipeline transport of CO2 containing water and impurities CO2 dominated hydrate

CO2 dominated hydrate

Liquid water outside ads. layer Water ads. on rust

Hydrate dominated by Hydrate formers from adsorbed and also from hydrate formers dissolved in outside water

Blue: CO2-phase Green: Liquid water Brown: Rust Yellow: hydrate

Route 5

CO2+H2O +H2S+N2+ Ar

Liq. H2O CO2: CO2+H2S+CH4+N2+Ar

CO2 dominated hydrate

Route 6 H2S, CO2 ads

Routes 9 + 10

Route 8 Liq. H20+H2S+CO2 (CO2: CO2+H2S+CH4+Ar)

Hydrate dominated by dissolved hydrate formers in water (H2S, CO2)

20

Most hydrate evaluation software only focus on route 5 Hydrate from CO2 and liquid water given only T fixed and P dependent variable. (Kvamme & Tanaka,1995)


CO2 hydrate growth on interface

Water Water CO2 plume with with T=1 C, 3,3% 3,3% P=150 bar% saturation CO2 Here is another one for CH4 with CO2of

Route 6 0.8 0.7 0.6 0.5

Route 8

Left: Relative saturation of CO2 in water versus neccesary pressure (horisontal axis) to produce hydrate (upper curve for 6 C and lower for 0 C)

0.4 0.3 0.2 0.1 0 50

100

150

200

250

300

Kvamme, Bjørn. Initiation and growth of hydrate from nucleation theory. International Journal of Offshore and Polar Engineering 2002; 12:256-262 Kvamme, Bjørn. Droplets of dry ice and cold liquid CO2 for self transport to large depths. International Journal of Offshore and Polar Engineering 2003; 13(1):1-8

A practical case of non-equilibrium and motivation for multi-stage modeling strategy Many research groups around the world has been tempted by the possibility of a win-win situations of combined safe storage of CO2 and release of CH4 from in situ CH4 hydrate. The possibilities are easy to see from simple hydrate equilibrium experiments but the mechanisms are more important in order to fully make use of the concept. Hydrate from pure CO2 is more stable than CH4 hydrate over substantial regions. Mixed hydrates in which CO2 dominates large cavities in structure I and CH4 dominates small cavities are more stable than any of the pure hydrates over the whole P,T region.

A production example • Ignik Sikumo – a pilot plant study on injection of CO2/N2 into CH4 hydrate • Mechanisms for conversion of CH4 hydrate over to CO2 dominated hydrate with associated methane release • Non-equilibrium nature of hydrates in porous media • RCB – the first non-equilibrium hydrate simulator

www.pet.hw.ac.uk/

Red is oxygen and grey is hydrogen in water. Ethane in large cavities (green) and methane in small cavities (blue) of structure I is scaled down. Volume of water in hydrate is roughly 10% larger than in liquid 23 water.

Department of Energy, ConocoPhillips and JOGMEC, Japan were interested enough to put money on the table

While different laboratories around the world has investigated the CO2/CH4 The Ignik Sikumo field test was conducted exchange for the last 2 by Conocophillips and JOGMEC only one pilot plant Normal text - clickdecades to edit study has been conducted. Estimated hydrate saturation in the socalled upper C was 75 %, 15 % free water and rest pore bounded water What do we actually know about the conversion mechanisms ?

Various methods to identify hydrate saturation For details on figure: Kvamme, B., Thermodynamic limitations of the CO2/N2 mixture injected into CH4 hydrate in the Ignik Sikumi field trial, 2016, J. Chem. Eng. Data, 2016, 61 (3), pp 1280–1295


Target section for test production

The pilot plant in Alaska was Ignik Sikumo conducted using only one well The first large scale pilot test of CO2 based CH4 hydrate production was in an «huff and puff» method completed April 10, 2012 •

This implies that a period of injection is followed by a reduced pressure production period.


•

Due to the high density of CO2 the neccesary injection pressure would be too high for pure CO2 to enter the formation without geomechanical implications

•

A mixture of 22.5% CO2 by volume and 77.5% N2 was injected. Some nitrogen will enter small cavities of the final hydrate. Rates and timeline in next slide.

•

This mixture will have sufficient relative permeability but nitrogen dilutes the CO2 and reduces thermodynamic driving force for conversion

•

Active section stimulated: A layer between 700 and 800 m below ground

The complete report is open and can be downloaded from DOE web-pages for the hydrate program

These snapshots can also be found in the report summary as presented in short form in the «Fire in the ice» March 2013 issue


Right: production (red) from March 4 to April 9

Below: injection rate (red) for 2 weeks injection (Feb 15 to Feb 29)

A problem with the «huff and puff» for injection of CO2/N2 mix is that the released CH4 will reduce the trapped gas density. In a continuous production in a two well completion this would give a perfect drive from injection well towards production well(s) and reduce associated water and sand production compared to the one well solution, in which the reduced fluid density would enhance a draw of water as well as sand when pressure was reduced again.

Conversion mechanisms: 1. Solid state conversion

Solid state conversion is slow, with diffusivity coefficient in the order of 10^‐16 m^2/s


The process in entropy dominated (see next slide) This mechanism has been verified experimentally by Ripmeester et.al. The relative impact of this mechanism will increase with lower free water in pores

-0.2

-1.25

-0.25

-1.3

Normal text - click to edit F re e e n e rg y c h a n g e (k J / m o le )

E n t h a lp y c h a n g e (k J / m o le )

Free energy and enthalpy changes for conversion from pure methane hydrate to pure carbon dioxide hydrate Blue: 43 bar, Green: 83 bar, Red: 120 bar

-0.3

-0.35

-0.4

-0.45

-0.5

-0.55 273

-1.35

-1.4

-1.45

-1.5

-1.55

274

275

276 277 278 Temperature (K)

279

280

281

-1.6 273

274

275

276


280

281

282

Free energy chage for the water in the structure is not large. CO2 change is also limited since it comes from a fairly dense phase and have reasonable filling. But CH4 will benefit from the entropy change of getting released.

283

CH4 Production Rates & Amounts from Hydrate Sw=0.5, T=3.5oC, P=1200 psi (8.3 MPa) 0.6

November 2004 Methane Production Experiment


1st Flush

0.4

MRI intensity

MRI intensity almost directly translated to fraction converted

0.5

Flush - March Expt.

Inc. T [3.5 to 6.8 C] Dec. P [1200 to 800 psi]

0.3

0.2

2nd Flush

0.1

3rd Flush Norm alized Initial Intensity = 1.0

4th Flush

0 0

48

96

144

192

240

288

336

384

432

480

Time after CO2 Flush (hours)

• So what is the thermodynamics involved in this transformation and what is the conversion mechanism?

528

So what makes the exchange of CH4 hydrate over to CO2/CH4 mixed hydrate possible? •

Phase Field Theory predictions of Molecular modeling can verifiy the solid CO2 conversion of CH4 (blue) state exchange mechanism experimentally hydrate X(t) versus experiments (red) detected by several reserach groups but maybe most detailed by Ripmeester et.al. But the process is very slow – with

Normal text - click to edit diffusivities in the order of 10^-16 square meter per second

These are predictions without adjustable parameters

• Phase Field Theory modelling of the experiments for exchange between CO2 and CH4 indicates an average diffusivity of 10^-12 square meter persecond and obviously composite mechanisms. A closer look at the systems with present generation of our PFT theory gives deper insight

2- Second mechanism • New CO2 hydrate forms from injected CO2 and textvolume. - click to edit free waterNormal in the pore • Heat released will contribute to dissociation of in situ CH4 hydrate. • This also explains why some of the injected N2 was trapped in the CO2 hydrate since N2 may assist in stabilization of the hydrate by filling small cavities.

Three component Phase Field Theory with implicit hydrodynamics


Parameters ε and w can be fixed from the interface thickness and interface free energy. ε ij set equal to ε

277.25

20

277.2

40

277.15

Length(Å)

60

277.1

80

It may not be easy to read from these figures but heat released from new CO2 hydrate formation mainly goes inwards through liquid water and hydrate while local cooling during in situ CH4 hydrate is smeared more out on interface towards CO2 due to low heat transport capacity of CO2.

Normal text click to edit The thicker initial liquid water film around 277.05

100

277

120

276.95

140 20

40

60 80 Length(Å)

100

120

140

hydrate the longer period of fast exchange. I.e.: Free pore water is the key 5

0

Water layer = 5 Å Water layer = 50 Å Water layer = 70 Å

0.9

20

-1

0.7

60

0.6 0.5

80

0.4 100

0.3

120

0.2 0.1

140 20

40

60 80 Length(Å)

100

120

140

CH4 dissociation rate (mmol/m2s)

0.8 40

Length(Å)

x 10

-2 -3

0 -1000

-4

-2000 -5

-3000

-6

-4000

-7

-5000

0

1

2

3

4

5

6

7 -10

x 10

-8 -9

0

0.5

1

1.5 Time (s)

2

2.5 -8

x 10

Adding large amounts of N2 has several drawbacks and is not the right direction forward on combined CO2 storage and clean energy production. -49

Formation of new CO2 dominated hydrate is only possible if hydrate water chem.pot. (solid) is lower than liquid water chem.pot. (dash)

-49.1

• CO2 dissolves in water, • CO2 adsorbs on mineral surfaces • Creation of new hydrate will extract CO2 from the mixture • There must be a lower limit for CO2 in the gas for creation of new hydrate and feasibility for the fast mechanism

Water Chemical potential (kJ/mole)

Normal text - click to edit -49.2 -49.3 -49.4 -49.5 -49.6 -49.7 -49.8

80% CO2

-49.9 -50 264

•

1% CO2

266

268


276

278

280

Estimated water chemical potential in hydrate (solid) and liquid water (dash) as function of temperature for 80 bars and CO2 mole-fractions of 0.80, 0.6, 0.4, 0.2, 0.1, 0.05, 0.02, 0.01, with 0.80 mole-fraction curve lowest and 0.01 molefraction curve on top.

But CH4 hydrate will still dissociate towards N2 which is undersaturated on CH4 -24

- click to edit Methane Chemical potential (kJ/mole)

• But the kinetic rates are Normal text low ! • Hydrate will also dissociate towards water undersaturated with methane so during flow and fluid exchange in the pores many different scenarios are possible

-25

-26

-27

-28

-29

-30 264

266

268


276

278

280

Estimated chemical potential for CH4 in equilibrium hydrate (-) versus chemical potential for 10 mole % CH4 mixed into a gas phase originally containing 1mole % CO2 and rest N2 (--) at 80 bars pressure. Pressures on the solid curve are the estimated equilibrium pressures for pure CH4 hydrate

Investigating competing phase transitions in porous media (read: solid material surfaces) requires multiscale modeling approach Charge distribution for hematite by Gaussian03 MD for Studies of mechanisms, thermodyn, interface properties and parametrisation Simulation of hydrate x =0.033 CO2 growth dynamics on interface of a CO2 plume using PFT

from quantum (characterisation of charge distribution in model molecules, from below nano in scale) to nano (Molecular Dynamics simulations, MD) and micro (Phase Field Theory)

Pure CO2

RCB – a new hydrate reservoir simulator • As discussed - Hydrates in porous media cannot reach thermodynamic equilibrium and stationary situations depends on sealing mechanisms (clay, shale) • There is a need for a hydrate reservoir simulator which can take into account competing phase transitions of formation and dissociation under constraints of mass and heat transport

• In a VERY compact form: •

RCB handles every hydrate phase transition as a pseudo mineral reaction

•

Formation of hydrate from gas and liquid is one unique “reaction”, formation from gas dissolved in water is another one and so on.

•

Since all thermodynamic properties in all phases are described by residual thermodynamics then minimizing free energy locally is easy and provides local distribution of phases as well as compositions.

•

Hydrate formation is considered as “mineral deposition” and reduces available volume for fluids. Opposite for dissociation Associated changes in available volume for flow affects geomechanical properties

•

RetrasoCodeBright is the only CO2 storage simulator with appropriate handling of hydrate

Normal text click to edit modules (similar to EQ3/6 and with - State of the art reactive transport

options for PHREEQ and other data bases) - Adequate description of fractures as hydrodynamic channels (not funded yet) - Several geomechanical models built in and easy incorporation of alternative models - Userfriendly graphical interface for easy setup of new systems and presentation of output (files, figures, animations)

Hydrate formation is a possible effect in aquifer reservoirs which can be candidates for CO2 storage •

Impact of hydrate formation during aquifer storage of CO2 has been a dedicated effort in FME-SUCCESS

• Illustrations Normal text - click to edit from a

•

This has also been a general discussion in the CO2 aquifer storage community, including publications from Kvamme’s group

•

And regardless of how deep the CO2 injection is (except extreme conditions where density of CO2 is higher than groundwater density) the CO2 plume may enter hydrate formation regions.

previous presentation

Simulation of CO2 injection in a system with two small fractures and hydrate forming conditions above the fractures. After one year (top) flow through fractures is slow. After two years hydrate is growing but still not blocking (next) while after 4 years CO2 flux is reduced to a low value (3rd) and after 42 years (bottom) the system is very stational with very low CO2 flux.. Without hydrate formation situation is similar to fig2 also after 42 years.

The fast mechanism for conversion of CH4 hydrate involves formation of new CO2 dominated hydrate from the free pore water. If formation is substantially faster than dissociation we expect reduction in available pore volume locally •

The formation of a new CO2 hydrate is the primary mechanism in Kvammes US patent, as well as a new patent idea, and as such this part is not special.

•

The released heat from CO2 hydrate formation assist in dissociating CH4 hydrate and releasing CH4 gas for energy if the CO2 is injected in the CH4 hydrate formation or close enough.

•

If injected in aquifer below CH4 hydrate a formed CO2 hydrate will establish below CH4 hydrate and provide extra CO2 storage sealing.

CO2 flux (above) 4.5 years of CO2 injection into a model CH4 hydrate reservoir, and changes in apparent porosity (below) as available pore volume fractions for fluid filling.

Conclusions • Number of active phases involved in hydrate phase transitions in porous media are too many to secure thermodynamic equilibrium • It does not help if the number of components increase – rather the opposite since the most stable hydrates (minimum free energy) will form first, and selective adsorption followed by liquid side water interface concentrations will dictate dynamically available masses for hydrate formation

A simple illustration using a 2D adsorption model (Kvamme, Thermodynamic limitations of the CO2/N2 mixture injected into CH4 hydrate in the Ignik Sikumi field trial, J. Chem. Eng. Data, 2016, 61 (3), pp 1280–1295). If we consider 10 mole % CO2 in a CO2/N2 gas mixture this result in estimated 32% CO2 adsorbed on liquid water surface at 30 bar and 273 K.

Conclusions cont. • And no natural gas mixture have a perfect relative ratio of gas molecules, as compared to available cavities in structures I and II. So when the most stable hydrates have consumed the «best» hydrate formers a variety of structure I and II hydrates will form. Methane hydrate growing from a gas/water interface

- These hydrates will have varying composition and unique free energies, depending on what molecules that are available to extract from the hydrate former phase. - Often it ends up with methane as the final excess gas molecule, which will the typically form a structure I hydrate.

In the ultimate limit an infinite number of hydrate phases can form – with infinitely small changes in free energies. from lowest free energy and up to (typically) free energy of methane hydrate structure I.

In a non-equilibrium situation it is very rare to find unconditional phase transitions in one direction. Competition between formation and dissociation phase transitions under constraints of mass- and heat transport is the more typical situation 0.9

20

0.8 40

0.7

60 Length(Å)

Initial CH4 hydrate core (blue) surrounded by CO2 (brown). New CO2 hydrate forms around core while leaving some regions heated and dissociated CH4 from hydrate escape

0.6 0.5

80

0.4 100

0.3

120

0.2 0.1

140 20

60 80 Length(Å)

100

120

140

277.25

20

277.2

40

277.15

60 Length(Å)

Note the temperature distribution due to formation of new CO2 hydrate and dissociation of CH4 hydrate

40

277.1

80

277.05

100

277

120 140 20

40

44

60

276.95 80

100

120

140

• Solid surfaces are significant phases for hydrate phase transitions because: - they give rice to confinements (non-thermodynamic effect) for efficient nucleation - they serve as regions of hydrate former enrichment through primary or secondary adsorption. - they serve as efficient (2D masstransport) nucleation sites. - the first layers of water adsorbed are efficient hydrate inhibitors so no real “hydrate cementing” of grains.

Conclusions cont.

Excluded regions (minerals) have kinetic effects and physical interaction characteristics have thermodynamic effects

• The use of N2/CO2 mixtures for producing hydrate while at the same time storing CO2 is tempting bacause even flue gas might be stored directly in hydrate. • But: - Hydrate CH4/CO2 swap is possible through direct solid state conversion (extremely slow) and - a second mechanism in which CO2 hydrate forms from injection gas and water. Associated heat release dissociate in situ CH4 hydrate. Too much N2 deactivates the fast mechanism

Conclusions cont.

Liquid water slab exposed to CO2 at 83 bars and 276 K. CO2 phase to the right (but hard to see). Fairly thick (roughly 1.2 nm) and dynamic interface

• Hydrates in real sediments are in a situation of stationary flow ranging from diffusion (very close reservoir) to various levels of hydrodynamics depending on fracture systems that brings in ground water from above and/or hydrocarbons from below. • There is a need for a multi-stage modeling strategy to build up realistic non-equilibrium description of hydrates in sediments • And in the macro-end there is a need for a reservoir simulator that can adopt this new physical description (read: dynamic non-equilibrium with competing phase transitions) • Implicit geomechanics is a must since hydrate phase transitions can be very fast

Conclusions cont.

Simulated production of gas from Bjørnøya hydrate using pressure reduction. Results not readable here but papers available !