Читать книгу Clathrate Hydrates - Группа авторов - Страница 25

After submitting a thesis entitled Recherches sur les hydrates Sulfhydrés to the Faculté des Sciences de Paris in 1882 [57], Robert Hippolyte de Forcrand (Figure 2.3) spent the next 43 years contributing prolifically to gas hydrate research. Nominally, de Forcrand worked in the organic chemistry laboratory of le Collège de France in Paris under the direction of Marcellin Berthelot, although most of his work was performed in Lyon under A. Loir.

Some 30 years previous, Loir [58] had formed a solid compound from chloroform, hydrogen sulfide, and water at room temperature. Similar compounds were formed when methyl chloride, 1,1‐dichloroethane, ethyl chloride, methyl bromide, or methyl iodide were used instead of chloroform, or hydrogen selenide was used instead of hydrogen sulfide. Since Loir had only measured the relative proportions of chloroform and hydrogen sulfide, the composition of these compound hydrates remained essentially unknown. Since de Forcrand found that the pressure–temperature stability conditions for the compound hydrates (double hydrates, in today's parlance) were more convenient to work with than those of the single component hydrates, he chose the former for analysis.

The term hydrate sulfhydré was coined by de Forcrand to distinguish the compound hydrate of volatile organic liquids with hydrogen sulfide from the simple gas hydrate of hydrogen sulfide itself. Since there is no simple English equivalent of this term, it is best rendered as the awkward “double hydrate with hydrogen sulfide.” de Forcrand pointed out [59] that the compound hydrates appeared to be similar to the double hydrate of phosphine and carbon disulfide, newly formed under pressure by Cailletet and Bordet [42].

Using the production of crystals when hydrogen sulfide was bubbled through an organic liquid layer underlying a layer of water at a temperature near 0 °C as the criterion of hydrate formation, de Forcrand identified double hydrates of the 33 organic compounds, mainly halogenated hydrocarbons, listed in Table 2.1. No doubt, these organic compounds, many of them synthesized by de Forcrand himself, varied greatly in purity and the boiling points (b.p.) given are mid‐points of the observed ranges of the boiling point. Similarly, a number of organic liquids were found to form “hydrates selenhydrés,” double hydrates with hydrogen selenide gas substituted for hydrogen sulfide, such as Loir had found [58] for chloroform. Hydrogen selenide was also found [59] by de Forcrand to form a hydrate by itself, although apparently not soon enough for inclusion in his thesis.

Double hydrates with hydrogen sulfide were formed only by the organic halides which boiled below about 110 °C, a result now known to be due to a rough correlation between boiling point and molecular size (i.e. organic halides with boiling points greater than 110 °C appear to be too large to fit into the clathrate hydrate cages).⁴ The presence of hydrogen sulfide is not required for some of the compounds given in Table 2.1, but the stability of these hydrates is much less in the absence, than in the presence of H₂S.

Three years before, in his first publication [60] on the subject of gas hydrates, de Forcrand had reported the formation of methyl iodide hydrate by the method said by Berthelot [33], to give a snow of CS₂ hydrate: bubbling moist air through the volatile liquid at a rate fast enough to cause substantial cooling. The new methyl iodide hydrate melted at approximately −4 °C and was found to have the composition CH₃I·H₂O. “Analogous hydrates” were said [59] to be formed by chloroform, ethyl bromide, and ethyl iodide. These results were not mentioned in his thesis, and by then, de Forcrand probably had doubts of their authenticity. In retrospect, this method of preparation of hydrates of volatile liquids is very likely to produce hydrates that contain varying amounts of air; thus, they were double hydrates in their own right. In turn, this would give considerable variability to the decomposition temperatures of the double hydrates and thus the difficulties in obtaining reproducible results. Since oxygen and nitrogen hydrates were not reported until 1960, these gases were assumed to be inert with respect to hydrate formation by the hydrate researchers of the day.

Table 2.1 Molecules found by de Forcrand to form double hydrates with H₂S [1, 26].

Molecule	Boiling point (°C)	Molecule	Boiling point (°C)
CH3Cl	−23	C2HCl3	75
CH2Cl2II	40	C2H5BrIII	38
CHCl3III	61	CH3CHBr2III	115
CCl4III	78	C2H3BrIII	20
CH3Br	5	CH2CBr2	91
CH2Br2	80	C2H5IIII	71
CH3III	41	C2H3I	56
CBrCl3II	104	n‐C3H7Cl	46
CCl3NO2III	110	n‐C3H7BrII	71
C2H5ClII	10	i‐C3H7Br	60
CH3CHCl2	64	CH2CHCH2Cl	46
CH3CCl3III	75	CH2CHCH2Br	70
CH2ClCCl3II,H	102	i‐C4H9ClH	67
CH2ClCH2ClIII	83	i‐C4H9BrH	90
C2H3Cl	18	CH3NO2	101
CH2CCl2	40	C2H5NO2	115
CS2	46

The boiling points of the substance are those reported by de Forcrand. Other small halogenated alkanes which did not form clathrate hydrates were listed by de Forcrand.

II Composition determined by de Forcrand by two component analysis.

III Composition determined by de Forcrand by three component analysis.

H Likely structure H hydrate formers synthesized by de Forcrand.

Source: Adapted from Schröder [1], de Forcrand [26].

As evidence for the close similarity of many of the double hydrates with hydrogen sulfide, de Forcrand found that chemical analysis for all three components of the nine hydrates marked “III” in Table 2.1 and for two components (the third being determined by difference) of the seven hydrates marked “II” all gave the same composition, namely M·2H₂S·23H₂O. In hindsight, the compounds listed in Table 2.1 probably represent two different hydrate structures, most of them belonging to the structure II (sII) hydrate family, the compounds flagged with “H” likely are hexagonal structure H (sH or HS‐III) hydrate formers, see Chapter 3 for further discussion.

Another common feature of many of the double hydrates was the morphology of their crystals. The chloroform‐hydrogen sulfide hydrate was found to sublime as well‐defined crystals on the inner surface of the sealed tube. de Forcrand observed of this hydrate [26],

J'ai pu remarquer dans quelques cas des octaèdres presque parfaits. D'ailleurs l'examen au microscope polarisant ne peut laisser aucun doute sut la forme cubique de ces cristaux, qui n'agissent pas sur la lumière polarisée ⁵

Similar observations of several other of the most stable double hydrates showed cubic, cubo‐octahedral, and truncated octahedral forms (Figure 2.5a,b) which had no effect on polarized light when examined with the polarized light microscope. de Forcrand appears to have been the first to argue that the gas hydrate crystals examined belonged to the cubic system and thus were distinguishable from hexagonal ice crystals. The dissociation pressures of nine double hydrates measured by de Forcrand in the presence of liquid water and liquid hydrate former are reproduced in Figure 2.5c. These dissociation pressures and compositions of the gas phase were found to be dependent only on temperature and to be independent of the overall relative amounts of the three components. That these observations were a consequence of the thermodynamic relationship for four‐phase equilibria was not realized until Roozeboom's application of the phase rule a few years later. Figure 4 of Ref. [26], reproduced in Figure 2.5c, shows that the nature of the organic component affects the stability of the double hydrate: the dissociation pressure (mainly contributed by H₂S) is 1 atm at 3 °C, for CH₃CHBr₂. However, the dissociation pressure is reached at a higher temperature of 18 °C for the much more stable double hydrate of CCl₄.

In a first application of calorimetry to gas hydrates, de Forcrand attempted to measure the heat of dissociation of the chloroform‐hydrogen sulfide hydrate. He found that 47 cal g⁻¹ was absorbed in the decomposition of the hydrate into the two liquids and gaseous H₂S. A similar result was found for the double hydrate with ethyl bromide. He concluded that [26],

…les nombres qui en résultant prouvent que la quantité de chaleur produite est assez considérable, mais qu'elle est due surtout au changement d'état de l'eau qui entre dans le composé.⁶

This was a reasonable conclusion since the heat of fusion of ice was 80 cal g⁻¹ and according to de Forcrand's composition, water made up about 70% of the mass of these hydrates. The numbers confirmed the general impression that gas hydrate formation was more akin to a freezing process than to a chemical reaction.

Except for bromine, the evidence that simple hydrates could be formed by some substances which are liquid at room temperature, mainly lay in the formation of low temperature frosts, and remained unconvincing until the characterization [61] of a hydrate of chloroform in 1885. Chancel and Parmentier [61] showed this hydrate to decompose above 0 °C and found its composition to be CH₃Cl·18H₂O. Chloroform hydrate was to be recognized as the first of the so‐called “liquid hydrates” which, despite their higher water content possessed many of the same properties as “gas hydrates.”

Figure 2.5 De Forcrand's results showing (a) octahedral and related crystalline forms for the binary clathrate hydrate of H₂S and carbon tetrachloride; (b) the modified octahedral crystal of binary hydrate of H₂S and isopropyl bromide; (c) the dissociation pressures (mmHg) as a function of temperature (degrees Celsius) for nine double hydrates in the presence of liquid water and hydrate former. Source: adapted from: Ref [26], reproduced with permission from the Bibliothéque National de France.

Formally, however, the first liquid hydrate to be reported was that of ethanethiol (ethyl mercaptan). In 1872, Hermann Müller [62] reported that in the distillation of this mercaptan from concentrated aqueous solution of the potassium salt of ethylsulfuric acid and sodium hydrosulfide, the cooling condenser became filled with the mercaptan hydrate which melted at 12 °C to give two liquid layers. From the volume of the two layers, the composition of C₂H₅SH·24H₂O was estimated for the hydrate. Shortly thereafter, similar behavior was noted [63] by Peter Clässon, working at the University of Lund, who used elemental analysis to find the composition C₂H₅SH·18H₂O. Clässon recognized the presence of hydrogen sulfide as an impurity in the mercaptan and attempted to remove it. Nevertheless, since ethanethiol hydrate is now known to decompose below 4 °C, it appears likely that both Müller's and Clässon's hydrates were stabilized by the presence of H₂S. In 1887, Peter Klason reported [64] the formation of methyl mercaptan gas hydrate which decomposed at a temperature far higher than the boiling point of the mercaptan (12 °C).

In 1888, de Forcrand presented a series of short papers on gas hydrates which resulted from collaboration with Paul Villard (Figure 2.3). Villard had graduated from the École Normale in Paris with a teaching certificate in 1884 and became a secondary school teacher in the provinces. It appears that he started to study gas hydrates with de Forcrand in Montpellier in 1887, the year in which the latter became professor of chemistry there. In new measurements of the dissociation pressures of hydrogen sulfide hydrate [65], de Forcrand and Villard closely confirmed the earlier results of de Forcrand [25], in contrast with the higher values since measured by Cailletet and Bordet [42]. They also observed that at low temperatures:

…la présence de la glace amène des perturbations dont il est difficile de tenir compte, et qui nous ont empêché jusqu'ici d'obtenir des résultat concordants.⁷

Roozeboom responded [56] that he had already solved this problem [50] by showing the presence of an ice‐hydrate‐gas equilibrium which differed from the liquid‐hydrate‐gas equilibrium. Roozeboom also pointed out that the discussion by de Forcrand and Villard of Wróblewski's solubility principle for hydrogen sulfide and methyl chloride hydrates merely confirmed what he had himself found for other hydrates [56]; besides, thermodynamics placed no restrictions on the dissolved gas content of the liquid water phase in equilibrium with hydrate. In subsequent studies of the pressure along the methyl chloride hydrate–liquid water–gas equilibrium line [66], the compositions of the hydrates of hydrogen sulfide and methyl chloride (H₂S·7H₂O and CH₃Cl·9H₂O) were found. At the time, Roozeboom's comments were not acknowledged by other researchers.

Except for this brief period of collaboration with de Forcrand, Villard worked entirely by himself. He applied to Debray, shortly before the latter's death in 1888, for permission to work in the chemistry laboratory at his old school, l'École Normale Supérieure, where Cailletet and Wróblewski had done their work on gas hydrates. Within a matter of months, he reported [67] the formation of new gas hydrates by methane, ethane, ethylene, nitrous oxide, and acetylene, having overlooked Cailletet's prior discovery of acetylene [41] hydrate. The hydrate of methane [68] was prepared by compressing methane to about 75 atm over water near 0 °C in a Cailletet tube, followed by détente and recompression. This hydrate and the hydrate of ethylene [68] were particularly interesting since each existed above the critical point of the gas and therefore exhibited a dissociation pressure curve which was not limited at high temperatures by liquefaction of the gas. Nevertheless, from the steep rise in pressure which he observed at relatively high temperatures, Villard wrongly inferred that methane hydrate had a critical temperature of about 21.5 °C at 300 atm and ethylene hydrate had a critical point of about 18.7 °C above 60 atm. In 1890, Villard added the next higher homologue, propane to the list of hydrate formers [69].

In a paper defining [70] the dissociation pressures of new hydrates of methyl and ethyl fluoride, Villard also observed that the hydrate formed when a cold mixture of ethyl chloride or methyl iodide with water was nucleated with ice did not survive heating to much above 0 °C. When prepared in the presence of air, however, these hydrates could be heated to about 5 °C before decomposition occurred. The stabilizing effect of air was later to become a common observation for many gas hydrates and is still a point of interest/caution in hydrate syntheses in open vessels. In testing other “help‐gases,” Villard reported that the decomposition temperature of ethyl chloride hydrate was increased from 4.8 to 5.5 °C under 23 atm of hydrogen or 2.5 atm of oxygen.

Villard next recorded [69] the formation of hydrates by carbon tetrafluoride, fluoroform, methylene fluoride, and ethylene tetrafluoride. Like the methyl and ethyl fluoride previously used [70], these fluorocarbons were synthesized by Villard himself. In the case of carbon tetrafluoride, at least, the purity was highly questionable, as Villard found this hydrate to be stable without applied pressure at 0 °C; however, much later measurements showed that the hydrate requires 40 atm at 0 °C for stability.

With the large number of new gas hydrates available, Villard was encouraged to attempt the measurement of the compositions of these new materials, the first two of which were N₂O and CO₂ [71]. By carefully monitoring the gas pressure of N₂O over water in a sealed tube, he discovered that complete transformation of water to hydrate could take weeks. He developed glass apparatus that could be charged with a weighed amount of water to which mercury was added, which when shaken, would provide mixing. The water was then frozen to ice at −20 °C, at which point the air in the tube was replaced with the hydrate forming gas. The tube was then sealed with wax, and the ice was allowed to melt, forming a hydrate “mash” on the tube walls, see Figure 2.6, with all of the water converted to hydrate. Cooling the tube to −20 °C and melting the wax allowed the excess gas to escape, meanwhile keeping the hydrate intact. Finally, the hydrate was decomposed by warming, and the amount of released gas was measured volumetrically. The procedure was further refined to allow work under pressure at 0 °C. In addition to composition studies, Villard added microscopic observation of the hydrate crystals to obtain their morphology and their interaction with polarized light as well as calorimetric measurements of the heat of decomposition. Based on the regularity of the results obtained for hydrates formed by a number of gaseous hydrate formers, he proposed the following definition of gas hydrates: [72]

Figure 2.6 Villard's apparatus for hydrate formation and characterization. On the left‐hand side, the liquid mercury is at the bottom of the tube, with liquid water above it, and gas in the container on top of the water (labeled 1). The tip of the tube, a, is sealed with wax. The apparatus is inverted (right‐hand side) and shaken with the mercury agitating the water and the gas to form hydrate. After hydrate formation, followed by decomposition, the released gas goes through compartment 2 to a gas measuring device. Source: Adapted from Villard [68], reproduced with permission from the Bibliothéque National de France.

Les combinaisons dissociables, susceptibles d'exister seulement à l'état solide, formées par l'eau avec divers gaz, sont isomorphes entre elles, cristallisent dans le système cubique, et leur constitution est exprimée par formule générale M, 6H₂O, M représentant une molécule du gaz considéré.⁸

The liquid hydrates were treated as being similar to the gas hydrates except that their decomposition temperatures were all close to 0 °C, thus distinguishing them from the gas hydrates which were stable to higher temperatures. It was noted that under a pressure of a “helper” gas, the liquid hydrate decomposition temperatures increased markedly, with the decomposition of ethyl chloride hydrate rising from 4.8 to 5.5 °C under 23 atm of hydrogen or 2.5 atm of oxygen, as mentioned above. Mistakenly, Villard asserted that the “helper” gases did not participate in hydrate formation and proposed a thermodynamically untenable explanation for these results.

It did not take long for “Villard's rule” regarding the M·6H₂O composition of the hydrate phases to be challenged, as it is clear that the direct compositional analysis of gas hydrate depends on the purity of the hydrate material prepared. Values deviating from Villard's rule easily could be attributed to either excess water or excess hydrate former associated with or trapped in the solid hydrate. W. Hempel and J. Seidel's [73] experiment on the determination of the composition of CO₂ hydrate is worthy of note. CO₂ hydrate was prepared by sealing water and “carbonic acid” (CO₂) in a sealed tube at −79 °C, and allowing the tube to warm to room temperature. When the two liquid layers that formed were cooled to 0 °C, the hydrate formed readily. The sealed tube was again cooled to −79 °C, the tube broken open and fitted to a capillary delivery pipe, and the contents were allowed to warm slowly. The evolved CO₂ gas was collected in a gasometer over mercury. After the non‐bound CO₂ escaped, gas evolution ceased almost completely at −25 °C only to start again at −2 °C. Vigorous effervescence was then evident again between 0 and 15 °C. The hydration number derived from the known amount of water and the evolved gas depended on whether all of the gas released between −25 and –2 °C was attributed to hydrate decomposition. Thus, the experiment was not conclusive in testing Villard's rule. However, the slow release of gas above −25 °C may well be attributed to a “self‐preservation” effect, as the pressure over CO₂ hydrate reaches 1 atm at −55 °C, and it would have been expected that most of the hydrate would have decomposed well below −2 °C. The self‐preservation effect of hydrates has been active topic of research and is discussed in Chapter 13.

In 1897, de Forcrand and Thomas initiated new studies on double hydrates to see if other help gases in addition to H₂S and H₂Se could be found that might stabilize known hydrates [74]. Starting with initial success with acetylene, he also found that ethylene, carbon dioxide, and SO₂ could perform that function.

In the new century, de Forcrand initiated a new approach to the determination of hydrate compositions in recognition of the fact that direct determinations were difficult and prone to errors. He generalized Trouton's rule, proposed in 1887, that the entropy of vaporization for various kinds of liquids at their boiling points is almost the same value, about 85–88 J K⁻¹ mol⁻¹ [75]. The entropy of vaporization is defined as the ratio between the enthalpy of vaporization and the boiling temperature. de Forcrand calculated compositions for all of the known hydrates, first improving doubtful data when necessary. The results of his calculations are shown in Table 2.2 [28], of which about half of the entries appear to support Villard's rule. Except for outliers Ar and Br₂, for the other entries, both the heats of dissociation to form ice Q(ice) and water Q(water), and the hydration number generally increased with molecular weight to give up to eight waters/guest.

Further progress in determining hydrate compositions virtually ground to a halt, as neither direct, nor indirect, methods were able to give a convincing explanation of the apparent complexity of the variable hydrate compositions. For instance, de Forcrand, from previously obtained data, calculated the composition of the chloroform–H₂S hydrate to be CHCl₃·2H₂S·19H₂O. This composition was explained by de Forcrand in terms of the formula (CH₃Cl·7H₂O) + 2(H₂S·6H₂O); thus, two hydrates present in a 1 : 2 ratio. This formula was then taken to be common to all sulfhydrated hydrates (binary hydrates with H₂S). The composition found that for chloroform hydrate, CHCl₃·18H₂O in 1885 by Chancel and Parmentier [61] was ascribed to the presence of a large excess of water, although CHCl₃·17H₂O is the true formula.

More complexities arose from de Forcrand's efforts to investigate hydrate formation by the noble gases [76], in particular argon hydrate after it having been reported by Villard [28, 77]. He was able to make krypton hydrate, and from the dissociation behavior and heats of formation, he arrived at a composition of Kr·5.08H₂O, and a redetermination of the value for Ar hydrate led to a composition of Ar·5.5H₂O. Eventually, he was able to form Xe hydrate and determined its composition to be Xe·6.6H₂O [78]. Rounding off, Ar and Kr then have a hydration number of 5 or 6; however, xenon's value then would be 6 or 7, which again led to speculation why these rather similar noble gases would have different hydration numbers. There were further efforts made to confirm or refute Villard's rule, but without much success either way. The formation of hydrates of noble gas indicated that the chemists of the day realized that the water–gas interactions in hydrates were not chemical in nature.

Table 2.2 De Forcrand's hydrate compositions obtained using calorimetric data [1, 28].

Guest	Tboiling (K)	Tdissoc. (K)	Q(ice) (cal)	Q(water) (cal)	Calculated formula	Probable formula
Ar	86	229.2	13.30	6.87	Ar + 4.5H2O	4/5H2O
CH4	109	244	16.35	7.32	CH4 + 6.31H2O	6H2O
CO2	194.8	251.8	16.16	7.55	CO2 + 6.02H2O	6H2O
N2O	185	253.7	16.29	7.61	N2O + 6.06H2O	6H2O
C2H2	188	257.6	15.92	7.73	C2H2 + 5.73H2O	6H2O
C2H6	188	257.2	17.71	7.71	C2H6 + 6.99H2O	7H2O
C2H4	169	259.6	18.34	7.76	C2H4 + 7.37H2O	7H2O
PH3	188	266.6	16.44	8.00	PH3 + 5.90H2O	6H2O
H2S	211	273.35	16.34	8.20	H2S + 5.69H2O	6H2O
C2H5F	241	276.7	20.12	8.30	C2H5F + 8.27H2O	8H2O
SO2	263	280	19.83	8.40	SO2 + 8.06H2O	8H2O
CH3Cl	250	280.5	18.83	8.41	CH3Cl + 7.28H2O	7H2O
H2Se	231	281	16.82	8.43	H2Se + 5.87H2O	6H2O
Cl2	238.4	282.6	18.36	8.48	Cl2 + 7H2O	7H2O
Br2	332	>282.6	NA	NA	Br2 + 10H2O	10H2O

Note the rounding off of the actual calculated compositions.

Source: Adapted from Schröder [1], de Forcrand [28].

It is interesting to examine a summary of the interpreted results on gas hydrates as reported by A. Bouzat in the 1920s [79], some one hundred years after Davy's description of the first gas hydrate.

1 Gas hydrates are stable dissociable bodies that form solids with low heats of formation from their constituents in the solid state.

2 They have the same composition and similar properties even though they are formed by vastly different substances, inorganic, organic, elements, or compounds.

3 They are able to combine to form hydrates of a more complex nature nM·n′M′·(n + n′) 6H2O.

4 The six water molecules are bonded to the molecule M through “subvalences” and can be assumed to take on the form of a regular octahedron around the guest molecule.

5 Since the two bodies in a hydrate combine with very small affinity, they retain in the compound a molecular state that is as similar as possible to their separated state.

We can take note of the hydration number – rounded off to a whole number, and the reluctance to recognize a set of hydrates with high hydration numbers.