Read without ads and support Scribd by becoming a Scribd Premium Reader.
See Premium Plans
 
NMR/MRI Study of Clathrate Hydrate Mechanisms
Shuqiang Gao,
†
Waylon House,*
,‡
and Walter G. Chapman*
,†
Chemical Engineering Department, Rice Uni
V
ersity, Houston, Texas 77251, and Petroleum Engineering Department, Texas Tech Uni
V
ersity, Lubbock, Texas 79406  Recei
V
ed: April 21, 2005; In Final Form: August 12, 2005
Clathrate hydrates are of great importance in many aspects. However, hydrate formation and dissociationmechanisms, essential to all hydrate applications, are still not well understood due to the limitations of experimental techniques capable of providing dynamic and structural information on a molecular level. NMRhas been shown to be a powerful tool to noninvasively measure molecular level dynamic information. In thiswork, we measured nuclear magnetic resonance (NMR) spin lattice relaxation times (
T 
1
’s) of tetrahydrofuran(THF) in liquid deuterium oxide (D
2
O) during THF hydrate formation and dissociation. At the same time, wealso used magnetic resonance imaging (MRI) to monitor hydrate formation and dissociation patterns. Theresults showed that solid hydrate significantly influences coexisting fluid structure. Molecular evidence of residual structure was identified. Hydrate formation and dissociation mechanisms were proposed based onthe NMR/MRI observations.
Introduction
Gas hydrates are icelike structures in which water molecules,under pressure, form structures composed of polyhedral cagessurrounding gas molecule “guests” such as methane and ethane.Rarely encountered in everyday life, they occur in staggeringabundance under the sea floor and permafrost environmentswhere (
P
,
T 
) conditions ensure their stability. The natural gastrapped in these deposits represents a potential source of energymany times greater than all known natural gas reserves. Hydratescan form as well in undersea piping and above ground gaspipelines where they pose a major problem for gas/oil producers.Detailed understanding of hydrate melting and formationmechanisms on a molecular level is important for successfullytackling all hydrate challenges with accuracy and confidence.
1
However, hydrate growth and dissociation mechanisms stillremain unclear because very few experimental techniques canprovide in situ dynamic information on a molecular scale. Liquidwater structure coexisting with the hydrate phase, especiallynear the water/hydrate interface, is very important in under-standing the hydrate formation and dissociation processes. Theimminent state before guest molecules solidify into the hydratephase and the fluid structure immediately after clathrate hydratedissolves into the liquid state may hold the key to unlock thesecrets of hydrate mechanisms.NMR has been shown to be a powerful tool to noninvasivelymeasure molecular level dynamic information.
T 
1
is an indicatorof local molecular order around the spin-bearing molecules.
2
T 
1
measurement is an effective method to monitor microscopicfluid structure change. In this work, NMR
T 
1
measurements of THF in D
2
O solution were employed to probe the change of water structure around THF during THF hydrate formation anddissociation to understand the role of the water and hydrateinterface. Proton MRI
3
was also utilized to observe the hydrateformation and dissociation patterns.THF molecules become invisible to liquid-state NMR as theyare incorporated into the solid hydrate phase; thus, the
T 
1
’s of THF in the liquid phase can be measured independently of theTHF hydrate. D
2
O is invisible to proton NMR under allconditions, so only THF in the liquid phase is visible to MRI.Results showed that the presence of solid hydrate significantlyinfluences the fluid structure.
T 
1
measurements also indicatedthe existence of residual effects after hydrate dissociation.
Experimental Details
The schematic of the experimental setup is shown in Figure1.
T 
1
measurements of THF (Aldrich, 99
+
%) in D
2
O (Cam-bridge Isotope Laboratories, D 99.9%) and MRI imagingexperiments were performed on an 85 MHz Oxford horizontal31 cm wide bore NMR with imaging capability, using a LITZRF volume coil (with 14 cm internal diameter) from DotyScientific, Inc. Data were acquired and processed using VarianVNMR software and INOVA hardware systems.
T 
1
’s weremeasured using the inversion and recovery technique. VNMRsoftware, given inputs of possible minimum and maximum
T 
1
values, automatically generates standard 180
°
-
τ 
-
90
°
pulsesequences with various values of delay time
τ 
. It took 4 to
∼
6min to take a
T 
1
data point and about 1 h to take an MRI image.An air-jet temperature controller supplied dry and cold air tocontrol the sample temperature. It is capable of controllingtemperature from
-
40 to 100
°
C with 0.1
°
C stability. A glass
* Corresponding authors. E-mail: Waylon.House@ttu.edu; wgchap@rice.edu.
†
Rice University.
‡
Texas Tech University.
Figure 1.
Experimental schematic.
19090
J. Phys. Chem. B
2005,
109,
19090
-
1909310.1021/jp052071w CCC: $30.25 © 2005 American Chemical SocietyPublished on Web 09/28/2005
 
bottle with a cap that has a Teflon liner was used to contain 29mL of THF
-
D
2
O solution (molar ratio 1:17) and was tightlysealed to prevent THF evaporating into the environment. Thesample was weighed, and no THF loss was detected after a longperiod of time. A LUXTRON fluoroptic thermometer wasmounted into the glass container through the cap to monitorthe sample temperature. Its output reading resolution is 0.1
°
C.Since trace amounts of oxygen may alter the
T 
1
of THFsignificantly, pure D
2
O and THF liquids were deoxygenatedseparately in a closed glovebox with a nitrogen environment.D
2
O and THF were contained in two separate Teflon bottles.The gas phase above the liquid phase was periodically flushedwith nitrogen gas, and the bottles were periodically shaken tofacilitate the diffusion of oxygen out of the liquid phase. Afterthe gas phase had been flushed 6 or 7 times over about 12 h,THF and D
2
O were mixed on the molar basis of 1:17 in theglass bottle. This is the same concentration of THF in D
2
O inthe hydrate phase. Therefore, as hydrate forms the liquid-phasecomposition will not change. The sample was then sealed andmoved into the probe for
T 
1
measurements. (The success of deoxygenation was demonstrated by that fact that after deoxy-genation, the
T 
1
of THF in liquid D
2
O increased about 2 scompared to that of the sample without deoxygenation.)In a recent paper,
4
we reported that the
T 
1
of THF in D
2
Oafter THF hydrate dissociation is consistently smaller than thatbefore hydrate formation. It was suggested that the change in
T 
1
is due to the THF
-
D
2
O solution becoming more microscopi-cally homogeneous after hydrate dissociation than before hydrateformation. To investigate residual clathrate hydrate structureafter hydrate dissociation by
T 
1
measurements, it is importantto ensure that the THF
-
D
2
O solution is already homogeneousand in its equilibrium configuration before hydrate formation.Then, if the
T 
1
after hydrate dissociation is different from thatbefore hydrate formation, it would be strong evidence of residualclathrate hydrate structure. Therefore, in this study, the freshlymixed THF
-
D
2
O solution was first turned into hydrate andsubsequently dissociated, to make the sample solution micro-scopically homogeneous before
T 
1
and MRI measurements. Toeliminate any possible residual clathrate structure, the THF
-
D
2
O solution was heated to 35
°
C after hydrate dissociationand equilibrated at room temperature for 24 h.
T 
1
measurementwas then started as the sample was cooled in steps until hydratenucleation, which was indicated by a sudden rise of the sampletemperature. The temperature of the cooling air was thenadjusted to slow the hydrate formation rate for better
T 
1
measurement and imaging experiments. After complete hydrateformation, marked by the disappearance of THF peaks in theNMR spectrum, the temperature was raised to slowly dissociatethe hydrate. Right after complete hydrate dissociation, thetemperature was cooled to form hydrate again, to examine thememory effect. After the reformation was complete, THF wasdissociated and the temperature was raised to room temperaturein steps. Images of hydrate formation and dissociation patternswere taken by the MRI technique using the spin
-
echo multislice(SEMS) pulse sequence. The parameters were adjusted toproduce proton density-weighted images, which depend prima-rily on the density of protons in the imaging volume with theeffects of 
T 
1
and
T 
2
minimized. The orientation of the imageslice was in the horizontal direction.
Results and Discussion
The dependence of ln(1/ 
T 
1
) on 1/ 
T 
(
T 
is temperature in K)during cooling, hydrate formation, hydrate dissociation, andwarming up, is plotted in Figure 2. The relationship is linearduring cooling and warming up. Every
T 
1
data point wasmeasured with a standard deviation less than 0.05 s. It requiredsubstantial subcooling to initiate hydrate formation. The reportedslope change around 8
-
9
°
C during cooling was reproduced.
4
The free induction decay and sign of the slope indicate themotion of THF in D
2
O solution is in the extreme narrowingregion, i.e., the rotational correlation time of THF,
τ 
c
, is lessthan 10
-
9
s. On the basis of NMR theory,
5
the rotationalactivation energy of THF,
E 
a
, can be calculated from the slopeif the relation
τ 
c
)
τ 
o
exp(
 E 
a
 / 
 RT 
) is assumed, where
R
is theuniversal gas constant. The resulting
E 
a
’s for the cooling andwarming up processes in the temperature range of 8
-
25
°
Care about the same, 17.76 and 17.74 kJ/mol, respectively. Therestill is a slight shift in
T 
1
after hydrate dissociation comparedto that before hydrate formation. However, the difference isminuscule compared to the
T 
1
change after dissociating thehydrate that formed from fresh THF
-
D
2
O solution.
4
It isprobably because in this experiment THF was already homo-geneously distributed before hydrate formation, and its distribu-tion did not change much after hydrate dissociation.One interesting feature in Figure 2 is that the
T 
1
values of THF in the liquid phase during hydrate formation and dissocia-tion fall far off the linear lines for the cooling and heatingprocesses in the single liquid phase. Since the concentration of THF in the liquid phase remains unchanged during hydrateformation/dissociation, this reveals that the presence of solidhydrate strongly influences its coexisting fluid structure. Theliquid/hydrate interface probably played an important rolethrough the O
-
H bonds of the partial cages on the hydratesurface sticking into and structuring the liquid phase. Figure 3shows the change of 
T 
1
as a function of peak height, i.e., theamount of remaining liquid, during hydrate formation anddissociation. As the amount of liquid diminishes and morehydrate is present, the
T 
1
of THF gets higher. One possibility
Figure 2.
T 
1
behavior of THF at various temperatures and conditions.
Figure 3.
T 
1
as a function of NMR peak height, i.e., the amount of liquid, during hydrate formation and dissociation.
Clathrate Hydrate Mechanisms
J. Phys. Chem. B, Vol. 109, No. 41, 2005
19091
 
is a higher percentage of the water being structured by thehydrate surface, as more hydrate is present. This structuredinterface hypothesis is in agreement with the “reaction film”proposition by Clarke and Bishnoi
6
for hydrate growth. Figure3 also reveals that
T 
1
’s during hydrate dissociation are higherthan those during hydrate formation under the same conditionsand phase composition. This difference in
T 
1
between formationand dissociation transitions indicates a difference in fluidstructure. This difference can be argued as the evidence of residual hydrate structure. However, this residual structure doesnot necessarily mean clathrate aggregates remaining in solution.
7
It could be residual hydrogen bonds from collapsing of the solidclathrate structures. After complete hydrate dissociation,
T 
1
returned to normal values within 16 h without further raisingthe temperature.MRI images of hydrate formation and dissociation patternswere also taken for better understanding of hydrate growth anddissociation mechanisms. A higher degree of subcooling and alonger induction time were required to form THF hydrate forthe first time. Once hydrate started to form, formation was rapidbecause the sample temperature was lower than the equilibriumtemperature. MRI images of hydrate formation are shown inFigure 4. It was demonstrated that hydrate nucleated homoge-neously along the perimeter of the sample and formed fromthe outside inward.After hydrate formation was complete, the temperature wasraised to melt the hydrate. As shown in Figure 5, hydrate startedto melt along the perimeter. The dissociation progressed alongthe hydrate/liquid interface toward complete dissociation, whichwas indicated by the NMR peak heights returning to the originalvalues before hydrate formation. To examine the memory effectand hydrate reformation pattern, the sample was cooled to formhydrate again right after the dissociation was complete. Thistime it required shorter induction time and less subcooling. Thehydrate formation was also gradual compared to last time. Theformation kinetics could be well controlled by changing thecooling air temperature. The MRI images of hydrate reformationare displayed in Figure 6. The nucleation took place at a pointnear the wall of the sample bottle instead of homogeneouslyalong the sample perimeter as it formed hydrate for the firsttime. This may be due to a hydrate residual structure mentionedearlier, which functioned as a nucleation seed. Both
T 
1
and MRIexperiments were repeated several times to confirm theirreproducibility.In summary,
T 
1
measurements of THF in the liquid phaseduring hydrate formation and dissociation strongly suggest thatthe presence of solid hydrate coexisting with the liquid phaseinfluences the fluid structure, probably through hydrogenbonding between the liquid and the solid at the interface, whichcreates a structured liquid layer. As expected, MRI imagesshowed that once hydrate nucleates, hydrate grows only at thehydrate/water interface, even though the THF concentration ishomogeneous in the liquid phase. Hydrate only dissociates atthe interface even though the temperature is uniform in the solidhydrate. Tohidi et al.
8
discovered that hydrates isolated fromthe liquid phase require higher than equilibrium temperature todissociate, which demonstrated the important role of the liquid/ hydrate interface for hydrate dissociation. Ikeda-Fukazawa andKawamura
9
simulated ice melting using molecular dynamics.It was concluded that the dangling motion of the free O
-
H
Figure 4.
THF hydrate formation pattern from THF
-
D
2
O solution. The white area is the liquid phase. THF hydrate is invisible to MRI.
Figure 5.
Hydrate dissociation pattern as the sample is warmed fromthe perimeter.
Figure 6.
Reformation pattern of THF hydrate from the THF
-
D
2
O solution that formed hydrate before. Hydrate formation progresses from (a) to(c).
19092
J. Phys. Chem. B, Vol. 109, No. 41, 2005
Gao et al.
Search
Search History:
Searching...
Result 00 of 00
00 results for result for
  • p.
    • Notes
    Load more