Introduction
Crystallization in a solution, particularly wax and asphaltene crystallization in a hydrocarbon system, is a notable phenomenon from the perspectives of both science and technology [1–3]. The precipitation and deposition of wax and asphaltene during the production and transportation of petroleum are common and compatible processes [4–8], which pose considerable challenges to safe and economical operations. Hydrocarbons with a relatively high molecular weight, such as wax and asphaltene in petroleum cuts, as well as their properties, are sensitive to temperature [9,10].
Temperature is related to the structural state, such as the phase transition, of n-alkanes [11–13]. Crystal nucleation from a melt is one of the most comprehensively studied phase transition types [14]. N-alkanes are important components of crude oil, and their nucleation has been the focus of considerable attention [15–17]. The aforementioned transitions exhibit several unusual features, including the existence of intermediate rotator phases between the liquid and crystalline solid phases and surface freezing prior to solidification [18]. In many phase transitions, the critical dynamical questions involve the vibrational modes of a system, as characterized by the vibrational density of states (VDOS). A peak in low-frequency VDOS is associated with a molecular-scale disorder [19,20]. However, the boson peak that is typically observed over a wide temperature range has not previously been related to a strongly temperature-dependent phenomenon, such as liquid–solid phase transition [21–23].
Recent studies have indicated that terahertz (THz) spectroscopy is a promising method for analyzing hydrocarbon and their mixtures with organic solvents. Al-Douseri et al. presented the performance of THz sensing of gasoline products. They used THz spectroscopy to quantitatively determine xylene isomers in gasoline [24]. The far-infrared spectroscopy of molecules is attaining increasing importance in astrochemistry. Cataldo et al. analyzed 33 polycyclic aromatic hydrocarbons (PAHs) and asphaltenes from petroleum fractions, bitumen, and anthracite coal in THz range [25]. Tian et al. investigated the THz spectral features of liquid alkanes (C5–C10), and the results suggested that the refractive index of liquid alkanes would increase with increasing carbon number [26]. Liquid alkanes from pentane to hexadecane at temperatures ranging from 20 °C to 80 °C were measured by Laib et al. via THz spectroscopy [21,23]. Zhan et al. proposed an effective method for qualitatively identifying crude oils from different oil fields based on THz time-domain spectroscopy (THz-TDS) [27]. Lubricating oil, which has considerable industrial significance, consists of several types of hydrocarbons. Naftaly et al. focused on the optical absorption properties of oil composition based on THz-TDS [28].
In the present study, THz-TDS was used to measure the temperature-dependent time-domain spectra of six solid n-alkanes with carbon numbers ranging from 18 to 25. All the samples were heated to a temperature above the melting point and then naturally cooled. THz spectroscopy for online monitoring provides a new experimental tool for characterizing and evaluating the wax crystallization process. This method also provides a new approach for quantitatively distinguishing the content changes of structural order using the THz-TDS noncontact optical method.
Materials and methods
Analytically pure n-alkanes were supplied by Aladdin Chemistry (Shanghai) Co., and the carbon numbers of n-alkanes were 18, 19, 20, 21, 22, and 25. The purity of the samples is over 99.5%; hence, further purification is unnecessary. The samples are all white powder at room temperature. The diagram of the THz-TDS setup with a sample cell is shown in Fig. 1. A conventional transmission THz-TDS system with a mode-locked titanium:sapphire laser (MaiTai, Spectra Physics) was used in this research. The amplitude and phase information of the samples were obtained simultaneously via THz-TDS based on coherent measurement. THz radiation was generated using an emitter, which was composed of a photoconductive antenna. An optical lens was used to focus THz pulses onto a sample, and then the THz beam that carried the information of the sample encountered the probe laser beam at the zinc telluride crystal in the THz detector. A lock-in amplifier was used to amplify the signal. The THz beam path was purged with nitrogen to minimize the absorption of water vapor. Humidity was maintained at less than 3.0%. A quartz vessel with 1 mm wall thickness and dimensions of 40 mm × 10 mm × 45 mm was selected as the sampler to improve signal-to-noise ratio. The minor absorption of THz waves made the quartz vessel an ideal sampler for THz measurement. The temperature of the samples was controlled using a circulating water–heat exchanger with a temperature resolution of 0.1°C. The samples were heated to 80°C and placed in a THz device. Natural cooling and online monitoring were started via THz-TDS. A thermocouple with an accuracy of 0.1°C was inserted into a sample to monitor temperature. The THz spectra of the reference specimen and the samples were obtained by scanning an empty quartz cell and a quartz cell filled with the sample.
Results and discussion
The cooling time-dependent THz waveforms of the six samples are presented in Fig. 2, where the THz signal decreases at different speeds with the increase in cooling time. In addition to cell walls, THz waves penetrate hydrocarbons at a fixed path of 8 mm. The THz waveforms of the six samples shown in Figs. 2(a)–2(f) indicate that THz peaks and time delay change with increasing cooling time. The cooling process for the six samples lasts for half an hour. A significant difference among waveforms is observed, thereby suggesting a significant correlation between the different existential states of n-alkanes and THz waves. In this study, THz signal intensity and time delay for various n-alkanes differ from those of others. The electromagnetic properties of n-alkanes in THz range are closely related to chain lengths and structure-dependent properties. The waveform change of n-pentaccosane during cooling is the most substantial, followed by those of n-docosane, n-heneicosane, n-eicosane, n-nonadecane, and n-octadecane. The peak value (EP) of THz signals depends significantly on cooling time. The peak values (EP) with cooling time of the six samples shown in Figs. 2(a)–2(f) are presented in Fig. 2(g). In particular, the peak value (EP) of n-pentaccosane changes slightly during the first 4 min. Then, a sharp change in Ep from 0.14831 to 0.02285 V is observed in the succeeding 16 min. The peak value remains nearly unchanged in the last 10 min. The change in peak value is more remarkable for hydrocarbons with a long chain than for those with a short chain.
The temperature decline of over 30 min in the samples is plotted in Fig. 3. The temperature measurements indicate that the cooling rate is sensitive to the chain lengths of n-alkanes. For n-octadecane, temperature drops abruptly in the first 10 min of cooling. The initial temperature of n-octadecane is 52.2°C, and then it drops to 27.2°C after 10 min. However, the temperature decrease of n-octadecane is only 1.6°C. To compare the temperature change of the six samples during the same cooling period, the entire cooling process is divided into three stages at every 10 min. As shown in Fig. 3, all the six n-alkanes suffer from a dramatic temperature drop in the first 10 min. In the second 10 min, a moderate temperature drop, with a range that is smaller than that of the first part, occurs. For example, the temperature of n-docosane changes from 51.5°C to 38.2°C after cooling for 10 min, and the temperature difference reaches 13.3°C. In the second 10 min, the temperature difference decreases to 5.3°C. During the last 10 min of cooling, the change in temperature of the six n-alkanes is scarcely observable. The temperature differences are 0.9°C, 2.1°C, 3.3°C, 2.0°C, 3.3°C, and 5.3°C for C18H38, C19H40, C20H42, C21H44, C22H46, and C25H52, respectively. Consequently, the heat release of the alkanes with a high carbon number accelerates. The variations in peak value and temperature of the samples are considered closely related to the crystallization of n-alkanes according to Figs. 2 and 3.
Figure 4 is plotted to provide a deeper insight into the variation of the peak value (EP) with respect to natural cooling. The peak value (EP) does not exhibit a linear change with the temperature of the samples, thereby indicating that the chain length and existence state of the n-alkanes cause the change in peak value. At the beginning of the THz test, the six samples were melted by heating. When the temperature of the samples is above the melting point, the samples are maintained in liquid state. The lines marked in Fig. 4 show the melting points of each n-alkane. The melting points of the alkanes, with their corresponding curves located from right to left in Fig. 4, significantly increase. Crystallization commences when temperature reaches the melting point. Solid and liquid phases coexist before the end of the crystallization process. In addition, the solid fraction in liquid significantly affects the peak values of the samples. The dramatic decline in EP occurs at a temperature that approaches the melting point because the THz wave is sensitive to the size and structure of the particles in liquid. When the crystallization of the n-alkanes starts, a crystal nucleus initially appears in liquid phase; then, the number of crystal nuclei continues to increase. Finally, the n-alkanes crystallize thoroughly and become white solid materials. Consequently, significant changes with a sharp slope in EPare observed, as shown in Fig. 4.
A preliminary exploration of inducing the signal attenuation (SA) equation is conducted by combining the experimental results of the THz tests and the exothermic test to estimate the cooling process as follows:
where is the peak value of the THz signal for the sample measured at any time during the cooling period, and is the peak value of the THz signal for the same sample measured prior to cooling. The temperature-governed SA shown in Fig. 5 identifies the phase-based change in the cooling process. Such SA model is adopted to numerically distinguish structural change in the alkane samples to the maximum extent. For the cases in which SA is nearly 100%, the sample is considered in the state of complete disorder, whereas for the cases in which SA is approximately 0%, the state of the sample is considered in order. As shown in Fig. 5, liquid samples exist within a temperature range that exceeds the melting point, and accordingly, SA is ~100%. As temperature drops, the liquid sample changes into solid, accompanied by the recovery of molecular order and the increase in SA value. Thus, the SA value is small for solid samples. The melting points are 28.18°C, 31.9°C, 36.8°C, 39°C, 44.4°C, and 53.3°C for C18H38, C19H40, C20H42, C21H44, C22H46, and C25H52, respectively. At the same cooling time, the final temperatures reach 25.6°C, 26.2°C, 27.9°C, 27.1°C, 29.6°C, and 26.9°C, respectively. In addition, is defined as the difference between the melting point and the final temperature. Figure 5 also indicates that not all the sample curves reach 0% because of the limited temperature range monitored via THz-TDS. For example, when SA is ~0% for C25H52, reaches 26.4°C and becomes sufficiently high to achieve perfect crystallization. The exothermic process slows down because of the small (2.58°C) for C18H38. Only some parts of the liquid sample have changed into solid, and therefore, the SA value does not reach 0%.
Gaber and Peticolas proposed a parameter to describe the degree of lateral inter-chain order and localized intra-chain conformational order based on Raman spectral intensity [29]. On the basis of previous works, information, such as the intensity of Raman spectroscopy, is obtained for the quantitative analysis of the phase transition process of alkanes. Octane, dodecane, and hexadecane were used as models to describe portions of the Raman spectrum of PE, and the effect of conformation on the spectral intensities was studied [30–32]. Meier’s theory states that the definition of order parameter fails to provide a precise description of the change in the amount of ordered fraction within alkane samples and only provides the relative amount of ordered fraction [33].
The cooling of n-alkanes is regarded as an extremely complicated process accompanied by a variety of complex phase changes. A number of factors work on peak values. In this study, SA is measured and analyzed qualitatively. However, the mechanism is inexplicit and a detailed research is necessary in the future. The SA empirical equation obtained in this research, which targets six types of alkane, also requires correction. Nevertheless, the proposed THz-TDS technology provides a novel method for characterizing change in structural order that is applicable to evaluating wax crystallization.
Conclusions
In this work, THz-TDS was used to monitor phase change, particularly the crystallization, of six types of n-alkanes. The peak values of the THz signal were found to be associated with the cooling temperature of n-alkanes because the THz wave was sensitive to the size and structure of the particles in a liquid. In addition, an empirical equation based on SA was proposed to describe the extent to which molecular structural order changes numerically. The present study suggests a novel noncontact optical method for characterizing the crystallization of wax using THz-TDS.