Thermochemistry of Eight Membered Ring Hydrocarbons. The Enthalpy of Formation of Cyclooctane

A new value of the enthalpy of formation of cyclooctane (-156.2±1.2 kJ mol) based on heat of combustion measurements is reported. Its solid liquid phase change was investigated by differential scanning calorimetry in both directions revealing an overcooling effect of over 23 K. Our enthalpy of formation of cyclooctane was used together with literature values of heats of hydrogenation of 8 carbon atoms cycloolefins to calculate the enthalpies of formation of the later. The strain energies of the investigated molecules were calculated and discussed.


Introduction
The literature value for the enthalpy of formation of cyclooctane, selected and reanalyzed by Cox and Pilcher [1] (i.e. ∆fHºliq= -169.41.6 kJ mol -1 ) is based on the enthalpy of combustion measured by Spitzer and Huffman in 1947 [2] (ΔcH°liq= -5265.7±0.9 kJ mol -1 ). A relatively more recent determination comes from Kaarsemaker and Coops [3] in 1952 whose value yields ∆fHºliq = -169.0 kJ mol -1 . The values from both references agree within the error limits as well as with that from DIPPR 801 project [4] (-167.74 kJ mol -1 ). An even older value of the enthalpy of combustion was reported by Ruzieka and Schlapfer in 1933 for the solid state [5]. Their value was used by NIST [6] to obtain ∆fHºsolid= -182 kJ mol -1 . Enthalpies of formation of unstrained alkanes were calculated by Zavitsas et al. [7,8] based on the number of different kinds of hydrogen atoms. The enthalpy of formation of cyclooctane is not important only for the study of the energetics of this molecule, but also of that of all eight membered ring cycloalkenes, whose enthalpies of formation are derived from enthalpies of hydrogenation to cyclooctane.
The density measurements at 298.15 K for cyclooctane were carried out by using a density and sound velocity meter Anton Paar DSA-5000 M with an accuracy of ±0.05 kg m -3 . Dried air and deionized ultra-pure water at atmospheric pressure were used as calibration fluids. The sample temperature was maintained constant within ±0.01 K. The experimental data for refractive index and density at T= 298.15 K are shown in Table 1.

Combustion calorimetry
A Parr Instruments model 6200 microprocessor controlled isoperibol oxygen bomb calorimeter was used in combustion experiments. Temperature is measured with a high precision electronic thermometer using a specially designed thermistor sensor. Measurements were taken with 0.0001 K resolution. The jacket temperature is held constant for isoperibol operation. A microprocessor-based controller monitors both the temperatures of the bucket and of the jacket and performs the necessary heat leak corrections that result from differences between these two temperatures. These corrections are applied continuously throughout a test rather than as a final correction based on pre and post test measurements.
The combustion calorimeter was calibrated by means of ten combustion experiments with benzoic acid (CAS 65-85-0) supplied by Parr Company, having the massic energy of combustion under certificate conditions (cuº = -26434.5 J g -1 ). The energy equivalent of our calorimeter was determined as εcalor= (10054.4±1.4) J K -1 . The uncertainty associated with the average value of the energy equivalent is the standard deviation of mean.
The combustion bomb Parr Instruments standard type 1108 is a cylinder, with an internal volume of about 0.35 L, made of stainless steel, as well as its accessories. The samples burned in the oxygen filled bomb were enclosed in Parr 3601 Gelatin Capsules with maximum 0.9 mL volume. A cotton thread was used in order to ignite the samples. The completely equipped bomb (filament, fuse, gelatin capsule, sample, deionized water) was connected to the oxygen cylinder and flushed with oxygen. All combustion experiments were carried out in high purity oxygen 99.998% at 3.546 MPa and at T = 298.15 K and in the presence of 1 cm 3 of deionized water for saturation of the atmosphere.

Differential scanning calorimetry
A Perkin Elmer power compensated DSC (model 8500) was used for the measurement of the enthalpies of the processes occurring during temperature scan (fusion and crystallization). The calorimeter was calibrated with high purity indium (Tm= 429.75 K and ΔfusH = 28.46 J g -1 ). The areas of the peaks corresponding to the standard and studied samples were used to calibrate the instrument and calculate the thermal effects for cyclooctane. Aluminum sealed pans were used. The DSC experiments were done at a heating rate of 10 K min -1 using nitrogen with purity > 99.996% (flux 20 mL min -1 ). The heat flow curves were processed by means of Pyris Software, which delivered the thermal effects and the purity.

Combustion energy
The data of the combustion measurements for cyclooctane are given in Table 2. Samples amounting between 0.10.4 g and gelatin capsules with a mass of 0.100.12 g were used in each experiment. The samples were weighed with a Mettler-Toledo microbalance with an accuracy of ±2·10 -6 g. Some runs were rejected because of doubt about combustion completeness. The corrections for the gelatin capsule, the cotton thread and filament used in the combustion experiments were calculated from the corresponding masses and energies of combustion. cu(gelatin capsule)= 19294±3 J g -1 was determined in our laboratory (to be compared with the value certified by the fabricant 19.26 kJ g -1 ), cu(cotton) = 16709±5 J g -1 checked in our laboratory in agrrement with literature value 16240 J g -1 [12] and cu(Ni-Cr) = 5.86 kJ g -1 (certified by the calorimeter's fabricant). In order to bring the experimental values of the energy of combustion to the standard state (T = 298.15 K and P= 0.1 MPa), corrections were made with the Washburn approximate equation [13]: where: P stands for the initial oxygen pressure and −U exp for the experimental energy of combustion; a, b, and c are the numbers of carbon, hydrogen and oxygen atoms from the chemical formula of the compound, CaHb. The Washburn correction  is calculated in percents from the experimental value.  The standard combustion enthalpy cHº of cyclooctane was calculated taking into account the following reaction: C8H16(l)+ 12 O2(g)  8CO2(g) + 8 H2O(l) (2) by using the equation: cHº= cUº + nRT (3) where n is the change in the mole number of gaseous compounds during the combustion reaction: n = nproducts,g -nreactants,g (4) As recommended by Olofsson [14], the uncertainty associated with the molar enthalpy of combustion is twice the overall standard deviation of the mean, and it includes the uncertainties associated with the calibration experiments (uncertainties of massic energy of combustion of benzoic acid and energetic equivalent of the calorimeter) and with the auxiliary materials.

DSC runs
The purity of liquid cyclooctane could be estimated from the DSC curve, based on the shape and the temperature of the DSC melting endotherm, as is shown in Figure. 1.  Our value of the melting temperature of 285.5 K fairly agrees with the range indicated by NIST [17] (285 ± 5 K) or by reference [18] (286 K) but not with that recommended by the CRC Handbook (288 K) [19]. So does our value for the enthalpy of fusion (2.28 kJ mol -1 ) if compared with that reported by Acree and Chickos [20] (2.41 kJ mol -1 ). However these values are very far from that estimated by means of the method of Chickos et al. [21], i.e. 16.25 kJ mol -1 . This discrepancy can be explained as follows: the estimation method yields the enthalpy change between the liquid state and the crystalline state corresponding to 0 K, while the experimental value corresponds to a solid structure https://doi.org /10.37358/Rev. Chim.1949 much more similar to that of the liquid. This explanation is consistent with the value of the enthalpy of sublimation reported by Bondi [22] at 166 K, 58.7 kJ mol -1 .
An important overcooling effect, of over 23 K is observed. The crystallization enthalpy of the overcooled liquid corresponding to a nonequilibrium process is lower in absolute value than that of equilibrium melting.

Determination of purity by DSC
The Perkin Elmer Pyris DSC purity analysis software permits the assessment of purities for samples which undergo simultaneous degradation during melting. In Figure 2 is plotted the purity using Pyris software for cyclooctane. In the T vs. 1/FS curve, the corrected data points are shown as black circles and the line drawn through them shows how the data fit the van't Hoff relationship Eq. (5). The black boxes represent the uncorrected values calculated for 1/FS (fraction melted) at given temperatures ( Figure 3).   Our value of the enthalpy of formation in liquid state (-156.2±1.2 kJ mol -1 ) is less negative than literature data. This value corresponds to very large ring strain energy, of 49.6 kJ mol -1 , if the group additivity scheme is used [16]. This value is closer to those of cyclononane and cyclodecane that to that of cycloheptane.

TS-sample
Different literature sources [3,23] report for the standard enthalpy of vaporization values between 43.3 and 44.7 kJ mol -1 . We obtain a value of -112.8 kJ mol -1 for the ideal gas state enthalpy of formation, corresponding to a ring strain of over 52 kJ mol -1 . Strain energies for cyclooctane ranging between 40 and 50 kJ mol -1 were reported in literature according to Zavitsas et al. [8] and of almost 60 kJ mol -1 by Rogers et al. [26].
For the solid state enthalpy of formation of cyclooctane a value of -182 kJ mol -1 was reported by Ruzieka and Schlapfer [5]. The calculated value in this state is -235.28 kJ mol -1 [16] and -238.4 kJ mol -1 [24]. A ring strain of 55 ± 3 kJ mol -1 is resulting from these data. The very low experimental value of the heat of fusion does not allow us to evaluate a ring strain in this state.
Most values of enthalpies of formation of eight membered ring cycloalkenes come from heats of hydrogenation and are based on the value of the enthalpy of formation of cyclooctane. In Table 5 are listed the experimental and calculated enthalpies of formation of such molecules, together with the corresponding strain energy.
The case of cyclooctene was more thoroughly investigated. The values of fHº of the cis isomer of this compound are calculated by means of heats of hydrogenation ranging between -96 and -103 kJ mol −1 . [25][26][27][28] The resulting strain of this molecule in liquid state (32-39 kJ mol −1 ) is to be compared with the gas phase values resulting from data of authors [25,26,30] i.e. between 24 and 43 kJ mol −1 . It is considered that cyclooctane contains considerable strain from van der Waals repulsion and unfavorable torsion, which is partly relieved in cis-cyclooctene. The much larger strain energy of trans-cyclooctene (71 -93 kJ mol −1 ) is consistent with the instability of trans isomers of cycloalkenes with less than 10 atoms. Although the most stable cyclooctadiene isomer is observed to be 1, 3-cyclooctadiene, (Z,Z)-a high value of strain energy is calculated for it. In fact this comes partially from a lack of conjugation of the two double bonds, conjugation taken into account in the calculated value. For 1,5-cyclooctadiene, three values of the enthalpy of formation based on heats of combustion were reported too [32][33][34]. The first two are assigned to the -ZZ stereoisomer, while for the last it is not specified. The last two values of fHº, 18 kJ mol −1 [33] and 24 kJ mol −1 [34] would correspond to an improbable quasi-zero ring strain, while the value of authors [32] (57.7 kJ mol −1 ) is over 10 kJ mol −1 lower than the value obtained from heats of hydrogenation.
The isomer with conjugated structure is again the most stable in the case of cyclooctatrienes. It is observed that for samples of unspecified stereoisomers the enthalpy of hydrogenation has usually less negative values. This is probably due to a mixing effect.
An important value of the strain energy (75 kJ mol −1 ) is calculated for cyclooctatetraene. As in the case of the other conjugated molecules, the calculated value of fHº includes a cconjugative stabilization contribution (group parameter [Cd-CdH]) [16]. In fact in such a non-planar molecule such a contribution should be largely suppressed. [28,35,36] Two values of fHº(l) obtained from heats of combustion are available 254.5 kJ mol −1 [37] and 248 kJ mol −1 [38]. The agreement between this first value and that calculated by us is excellent.

Conclusions
The ring strain energy of cyclooctane is quite large which explains the difficulty of synthesizing compounds with 8-membered rings.
The thermodynamic parameters of melting indicate that premelting processes occur at lower temperatures.