Fluid Phase Equilibria 324 (2012) 1316

Contents lists available at SciVerse ScienceDirect

Fluid Phase Equilibria

journal homepage: www.elsevier.com/locate/fluid

Critical point measurement of ferrocene and some of its derivatives

Eugene D. Nikitin , Alexander P. Popov
Institute of Thermal Physics, Ural Branch of Russian Academy of Sciences, Amundsena Street, 106, 620016 Ekaterinburg, Russia

a r t i c l e

i n f o

a b s t r a c t
The critical temperatures and the critical pressures of ferrocene, n-butylferrocene, benzoylferrocene, and 1-acetylferrocene have been measured. All the compounds studied begin to decompose at temperatures below their critical points. A method of pulse heating applicable to thermally unstable compounds has been used. Residence times were from (0.03 to 1.0) ms, which resulted in little decomposition of the substances in the course of measuring. The contributions of the groups Fe CH and Fe C for the estimation of the critical properties of ferrocene derivatives by the method of Marrero and Gani have been determined from the experimental data. 2012 Elsevier B.V. All rights reserved.

Article history: Received 1 February 2012 Received in revised form 13 March 2012 Accepted 14 March 2012 Available online 23 March 2012 Keywords: Critical point Ferrocene n-Butylferrocene Benzoylferrocene 1-Acetylferrocene

1. Introduction Ferrocene and its derivatives belong to the sandwich type organometallic compounds. They are used as fuel additives, in manufacture of polymers with valuable characteristics, in pharmaceutical and material chemistry and so forth [1]. To our knowledge, the critical properties of ferrocene or its derivatives have never been measured before. The paper gives the critical temperatures and pressures of ferrocene, n-butylferrocene, benzoylferrocene, and 1-acetylferrocene. Ferrocene is a stable enough compound. In vacuum, it sustains a temperature of about 500 C without decomposition and even in the air atmosphere ferrocene is stable at temperatures as high as 400 C. However, the derivatives of ferrocene are not so stable. In the course of our experiments, a sufcient decomposition of 1-acetylferrocene and benzoylferrocene was observed (see Table 1). To diminish the degradation of the samples during the measurement of the critical properties we used a pulse-heating method with ultra-low residence times applicable to unstable compounds. 2. Materials and methods 2.1. Materials The samples of the compounds investigated were of commercial origin and used without any further purication. The sources,

purities, and Chemical Abstract Service Registry Numbers (CASRN) are given in Table 1. Before and after the measurements of the critical properties, the purities of the samples were determined by proton magnetic spectroscopy (Bruker DRX 400). Table 1 shows that the purity of ferrocene and n-butylferrocene samples was not changed or a little changed in the course of measuring the critical properties, while the purity of 1-acetylferrocene and benzoylferrocene samples was changed signicantly. 2.2. Method The pulse-heating method was developed by us in the early nineties and since then has been extensively used for measuring the critical properties of thermally unstable compounds. By now the critical temperatures and pressures of about 200 substances have been measured using this technique. Recently the pulse-heating method was recognized as a standard state technique [2]. In comparison with the other modern techniques applicable to thermally unstable compounds [36], the pulse-heating method provides much lesser residence times but in general has larger uncertainties. The method is based on the phenomenon of liquid superheat [79]. In the course of the experiments, the pressure dependence of the temperature of the attainable superheat (spontaneous boilingup) of liquid is measured. The critical point is not only the end point on the vaporliquid equilibrium line, but the end point on the line of the attainable superheat as well. Thus, when the pressure in the liquid approaches the critical pressure, the temperature of the attainable superheat approaches the critical temperature. The pulse-heating apparatus and procedure have been described in detail in previous publications [2,1013], so only a brief outline is given below. A liquid under study lled a thin-walled

Corresponding author. Tel.: +7 343 2678810; fax: +7 343 2678800. E-mail addresses: e-nikitin@mail.ru (E.D. Nikitin), popov@itp.uran.ru (A.P. Popov). 0378-3812/$ see front matter 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.uid.2012.03.021

14

E.D. Nikitin, A.P. Popov / Fluid Phase Equilibria 324 (2012) 1316

Table 1 Purities of compounds used in critical point measurement. Compound CASRNa Supplier Purity (mol%) Before measuring critical constants Ferrocene n-Butylferrocene 1-Acetylferrocene Benzoylferrocene
a

After measuring critical constants 99.9 99.4 96.0 95.0

102-54-5 31904-29-7 1271-55-2 1272-44-2

Fluka ABCR ABCR Acros Organics

99.9 99.7 99.9 99.0

Chemical abstracts service registry number.

Teon cup. The pressure outside the cup was created by a press and measured by a dial gauge. The full-scale reading of the gauge was close to the estimated critical pressure of the substance under study. The maximum uncertainty of the gauge was 0.15% of the fullscale reading. Special experiments showed that the pressure drop on the cup walls did not exceed 0.02 MPa. A platinum wire probe, 2 103 cm in diameter and (13) cm in length, was placed in the liquid. The probe served simultaneously as a heater and a resistance thermometer. The probe was included in a low-inductance bridge. Square potential pulses were applied to the bridge from a generator. During a pulse the probe was heated. The amplitude of the pulse was adjusted in such a way that by the end of it the probe and the thin liquid layer near it were heated to the temperature of spontaneous boiling-up (attainable superheat). The time from the start of a pulse to the moment of boiling-up was from (0.03 to 1.0) ms. At the moment of boiling-up a probe temperature perturbation arises from an abrupt change of the conditions of heat transfer from the probe to the liquid. The temperature perturbation is uniquely connected with a voltage perturbation on the probe. This perturbation was isolated with the help of an amplifying lter with a controllable bandwidth. Then the signal which we called the boiling signal was observed on the screens of a digital oscilloscope and a personal computer. The bridge was adjusted in such a way that its balance corresponded to the beginning of the boiling signal. The temperature of the probe at the moment of boiling-up was determined from the condition of the bridge balance and by the well-known temperature dependence of the resistance of platinum. The temperature determined in such a way was the probe-volume average, whereas boiling-up took place on the surface of the probe in its middle part. The temperature of boiling-up (attainable superheat) was calculated taking into account the thermal losses at the probe ends and the radial inhomogeneity of the temperature eld. If the compound under study was liquid at room temperature, the measuring chamber too was at room temperature. If the melting point of the compound under investigation was higher than room temperature, the chamber was at a temperature (1015) K above the melting temperature. For this purpose, there were a furnace and a temperature control system in the apparatus, which maintained the temperature of the compound under study within 1 K. Such a precision was enough. When the measuring chamber was lled with the liquid under study, the pressure established in it was somewhat higher than the atmospheric pressure. Pulses from the generator were periodically applied to the probe. The amplitude of the pulses was increased until the boiling signal was observed on the oscilloscope screen. By changing the box resistance from pulse to pulse the temperature of the attainable superheat at a given pressure was measured. The probe temperature perturbation in the initial stage of boiling-up may be both positive and negative. The sign of the perturbation is determined by the competition of two effects. A vapor bubble growing on the probe screens a part of the probe surface, which leads to a more rapid increase of the probe temperature (positive temperature perturbation). On the other hand, a growing bubble consumes heat for evaporation, which causes a negative temperature perturbation. Our large experience showed that for comparatively

low-molar-mass compounds (nonpolymers) at near-critical pressures the temperature perturbation was negative. As the pressure approaches the critical pressure, the properties of the liquid and vapor phases come closer and the amplitude of the temperature perturbation decreases. The pressure in the liquid increased step by step until the negative temperature perturbation dropped to the level of the apparatus sensitivity (1 103 K). This pressure was taken to be equal to the measured value of the critical pressure pm c , and the temperature of the attainable superheat at this pressure was taken to be equal to m . The values of the measured value of the critical temperature Tc m pm c and Tc are always lesser than the true critical properties and require correction. It is an inherent feature of the pulse-heating method connected with the peculiarities of bubble nucleation and the growth of vapor bubbles near the measuring probe. The true critical constants of a stable compound were calculated from the following equations: pc = pm c
o

Tc =

m Tc 0

where 1/ 0 and 1/ 0 are correction factors [10]. To calculate the correction factors, the thermophysical properties of the liquid and the vapor phase near the critical point are required. These properties are calculated by the principle of corresponding states using the formulas given in a previous paper [14]. The formulas contain the similarity parameter of the compound under study suggested by Filippov [15]: A = 100 pvp (T/Tc = 0.625) pc (1)

which is related with the acentric factor[16] = 1 log10 pvp T/Tc = 0.7 pc

by the following empirical equation [15]: = 0.401 0.664 log10 A Here pvp is the vapor pressure at reduced temperatures equal to 0.625 or 0.7, respectively. The Filippov parameters and the critical properties of ferrocene and its derivatives were calculated by an iteration method. For the m rst iteration, pm c and Tc were used as the critical constants. The vapor pressure of ferrocene at a reduced temperature T/Tc = 0.625 was calculated using the equations suggested by Barkatin et al. [17] and Niselson et al. [18]. Both equations are based on the measurement of the vapor pressures of ferrocene at temperatures from (451.2 to 523.2) K and (518.8 to 604.2) K, respectively. These equations give a bit different vapor pressures at T/Tc = 0.625; the average of these values was taken. Then the values of A, 0 and 0 , pc and Tc were calculated. For the second iteration, the Filippov parameter and the critical temperature and pressure were calculated using the values obtained after the rst iteration. Two iterations were enough because the values of 0 and 0 are little affected by the variations of Filippovs parameter. For instance, the Filippov parameter and

E.D. Nikitin, A.P. Popov / Fluid Phase Equilibria 324 (2012) 1316

15

the acentric factor of ferrocene determined as described above are equal to A = 1.56 and = 0.273. A variation of the Filippov parameter by 25% causes changes of the critical pressure and the critical temperature approximately by 0.2 and 0.02%, respectively. The well-known data on the vapor pressures of 1acetylferrocene [19] and benzoylferrocene [20,21] have been obtained at sufciently lower temperatures than 0.625 Tc and cannot be used for the estimation of the Filippov parameter. Using the boiling temperature given by Alfa Aesar [22] for n-butylferrocene Tb = 232 C/630 mm Hg leads to a suspiciously high Filippovs parameter A = 2.92 and a low acentric factor = 0.092. Really, it is difcult to imagine that adding the n-butyl side chain to a molecule of ferrocene would cause the decrease of the acentric factor. We used the following way to estimate the Filippov parameters of the ferrocene derivatives. The normal boiling temperature of ferrocene according to the data of Barkatin et al. [17] and Niselson et al. [18] is equal to Tnb = (246 2) C and Tnb = 242 C, respectively. We took the average value Tnb = 244 C. Using this value, we calculated the contribution of the molecule of ferrocene for the estimation of the normal boiling temperature in the groupcontribution method of Marrero and Gani [23]. Then we estimated the normal boiling points of the ferrocene derivatives studied by this method. It is a rough estimation because the contributions of the molecule of ferrocene (C5 H5 )Fe(C5 H5 ) and the radical (C5 H5 )Fe(C5 H4 ) differ from each other. The vapor pressure of the ferrocene derivatives at T/Tc = 0.625 was estimated by the following equation: C T

Table 2 Critical temperatures and pressures of ferrocene and its derivatives: experimental and calculated values. Compound Tc (K) Exptl Ferrocene n-Butylferrocene 1-Acetylferrocene Benzoylferrocene 785 784 847 886 8 12 13 13 Calcul 785 784 794 873 pc (MPa) Exptl 3.61 2.18 3.28 1.59 0.10 0.09 0.13 0.06 Calcul 3.61 2.18 2.71 2.34

Table 3 First-order groups and their contributions for the calculation of the critical temperature and critical pressure of ferrocene derivatives in the method of Marrero and Gani [23]. Property Tc (K) pc (bar) Fe CH 2.98074 7.32203 103 Fe C 2.89660 1.72874 102

2.3. Uncertainties The uncertainties of the critical constants measured by the pulse-heating method were discussed in detail in our previous papers [2,13,25]. We estimate the uncertainties for ferrocene at 0.03 pc and 0.01 Tc and for its derivatives at 0.04 pc and 0.015 Tc , where Tc is the absolute temperature. The uncertainties for the ferrocene derivatives are higher than for ferrocene because the estimates of the Filippov parameter for these compounds were not as reliable as for ferrocene. 3. Results and discussion The critical temperatures and pressures of ferrocene and its derivatives are given in Table 2. To our knowledge, the critical properties of these compounds have never been measured before. The critical constants of ferrocene were previously presented in the form of a poster at the 21st IUPAC International Conference on Chemical Thermodynamics [26]. The well-known group-contribution methods of calculating the critical properties cannot be used for ferrocene because they do not contain the contribution of ferrum. The critical temperature of ferrocene can be roughly estimated using the Guldberg rule: Tc = 1.5 Tnb = 776 K, which is close enough to our experimental value. Using our experimental critical properties for ferrocene and n-butylferrocene we calculated the contributions of the groups Fe CH and Fe C for the method of Marrero and Gani [23]. The contributions are given in Table 3. Then the critical constants of 1-acetylferrocene and benzoylferrocene were calculated with these contributions (see Table 2). The differences between the experimental and calculated critical properties are great enough, especially for the critical pressure of benzoylferrocene. The possible reason for such a discrepancy may be the fact that the contributions were determined using only two compounds. The estimation of the contributions can be improved when new results of the measurements of the critical properties of ferrocene derivatives appear. 4. Conclusion The critical temperatures and pressures of ferrocene, nbutylferrocene, 1-acetylferrocene, and benzoylferrocene have been measured. The contributions of the groups Fe CH and Fe C for the estimation of the critical properties of ferrocene derivatives by the method of Marrero and Gani have been calculated from experimental data.

ln pvp = B

The parameters B and C were calculated from the values of pm c and m and the estimated normal boiling temperature. At last the FilTc ippov parameters were calculated by Eq. (1). Further the critical parameters were estimated as it was described above for ferrocene. For calculating the correction factors two other quantities are needed: the factor GT ln J/T, where J is the rate of bubble nucleation in a superheated liquid, and the ideal gas heat capacity of the compound under investigation. The factor GT was measured in one experiment with the critical constants as described previously [2,10] and estimated at 10 K1 . The ideal gas heat capacity was estimated using the atomic contribution method by Harrison and Seaton [24]. The apparent critical temperature and pressure of a thermally unstable compound determined as described above may depend on the time from the beginning of a heating pulse to the moment of boiling-up t* due to the decomposition of a compound under study in the course of heating. The critical properties of ferrocene and its derivatives were measured with the help of probes (1, 2, and 3) cm in length at heating times t* = (0.03, 0.06, 0.11, 0.22, 0.46, and 1.0) ms. Three or four samples of each compound were used in the experiments. The compounds investigated by us are unstable at their critical points. In our experiments, ferrocene and n-butylferrocene showed no evidence of decomposition, and no dependence of the apparent critical properties on the heating time t* was found. The behavior of the samples of 1-acetylferrocene and benzoylferrocene was more complex. The decomposition of the samples as a whole occurred in the measuring chamber; as a result the purity of the samples was decreased (see Table 1). The temperatures of the chamber were 375 K and 411 K for the samples of 1-acetylferrocene and benzoylferrocene, respectively. However, the dependence of the apparent critical properties on the heating time was not observed, so for all the compounds investigated, the experimental data were averaged over all the probe lengths, heating times, and samples.

16

E.D. Nikitin, A.P. Popov / Fluid Phase Equilibria 324 (2012) 1316 [2] E.D. Nikitin, P.A. Pavlov, Method GSSSD ME 163-2010, Method of experimental determination of the critical temperature and critical pressure of pure substances, The Russian Scientic and Technical Centre for Information on Standardization, Metrology and Conformity Assessment. FSUO STANDARTINFORM, Deposit document No 865a-2010 kq. Moscow, 2010 (in Russian). [3] A.S. Teja, M. Gude, D.J. Rosenthal, Fluid Phase Equilib. 52 (1989) 193200. [4] L.C. Wilson, W.V. Wilding, H.L. Wilson, G.M. Wilson, J. Chem. Eng. Data 40 (1995) 765768. [5] D.M. VonNiederhausern, G.M. Wilson, N.F. Giles, J. Chem. Eng. Data 45 (2000) 157160. [6] W.V. Steele, R.D. Chirico, S.E. Knipmeyer, A. Nguyen, J. Chem. Eng. Data 41 (1996) 12551268. [7] V.P. Skripov, Metastable Liquids, Wiley, New York, 1974. [8] V.P. Skripov, E.N. Sinitsyn, P.A. Pavlov, G.V. Ermakov, G.N. Muratov, N.V. Bulanov, V.G. Baidakov, Thermophysical Properties of Liquids in the Metastable (Superheated) State, Gordon and Breach, London, New York, 1988. [9] P.G. Debenedetti, Metastable Liquids: Concepts and Principles, Princeton University Press, Princeton, 1996. [10] E.D. Nikitin, P.A. Pavlov, P.V. Skripov, J. Chem. Thermodyn. 25 (1993) 869880. [11] E.D. Nikitin, P.A. Pavlov, A.P. Popov, Fluid Phase Equilib. 149 (1998) 223232. [12] E.D. Nikitin, P.A. Pavlov, M.G. Skutin, Fluid Phase Equilib. 161 (1999) 119134. [13] E.D. Nikitin, A.P. Popov, Y.G. Yatluk, V.A. Simakina, J. Chem. Eng. Data 55 (2010) 178183. [14] E.D. Nikitin, P.A. Pavlov, N.V. Bessonova, J. Chem. Thermodyn. 26 (1994) 177182. [15] L.P. Filippov, Podobie Svoistv Veshchestv (Similarity of Substance Properties), Moscow University Press, Moscow, 1978. [16] K.S. Pitzer, D.Z. Lippmann, R.F. Curl, Ch.M. Huggins, D.E. Petersen, J. Am. Chem. Soc. 77 (1955) 34333440. [17] A.A. Barkatin, I.L. Gaidym, A.K. Baev, Khim. Khim. Tekhnol. (USSR) 12 (1977) 4448. [18] L.A. Niselson, T.D. Sokolova, R.K. Nikolaev, Vest. Mosk. Univ. Khim. (USSR) 4 (1972) 432434. [19] M. Pelino, M. Tomassetti, V. Piacente, G. DAscenzo, Thermochim. Acta 44 (1981) 8999. [20] M. Pelino, R. Gigli, Thermochim. Acta 61 (1983) 301305. [21] V.N. Emelyanenko, S.P. Verevkin, O.V. Krol, R.M. Varushchenko, N.V. Chelovskaya, J. Chem. Thermodyn. 39 (2007) 594601. [22] http://www.alfa-chemcat.com. [23] J. Marrero, R. Gani, Fluid Phase Equilib. 183-184 (2001) 183208. [24] B.K. Harrison, W.H. Seaton, Ind. Eng. Chem. Res. 27 (1988) 15361540. [25] E.D. Nikitin, A.P. Popov, Y.G. Yatluk, J. Chem. Eng. Data 51 (2006) 13261330. [26] E. Nikitin, A. Popov, Program & Abstracts of 21st IUPAC Int. Conf. Chem. Thermodyn. ICCT-2010, Tsukuba, Japan, July 31 to August 6, 2010, p. 121.

List of symbols Filippovs similarity parameter A J rate of bubble nucleation p pressure T temperature time from the beginning of a heating pulse to the moment t* of boiling-up Greek symbols 1/ 0 correction factor for the critical pressure correction factor for the critical temperature 1/ 0 acentric factor Subscripts c critical state nb normal boiling vp vapor Superscript m measured value Acknowledgements The study was performed in the framework of the Program N 2 of the Presidium of the Russian Academy of Sciences under the support of the Ural Branch of RAS (Project 12-P-2-1008). We are grateful to Prof. Yuri Yatluk from Institute of Organic Synthesis of the RAS (Ekaterinburg) for analysis of the samples purities. References
[1] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 86th ed., CRC Press, Boca Raton, FL, 2005.