Detailed investigation on the oxidation behavior and kinetic triplet of tight oil via TG-DSC-PDSC analyses

3.1.1 TG-DTG analysis

Non-isothermal oxidation TG-DTG curves with three heating rates were shown in Fig. 1. In good accordance with previous literatures (Chen et al., 2019b, Kok, 1993, Kök, 2008, Kok & Bagci, 2004, Pu et al., 2017a), the oxidation process of Mahu tight oil subdivided into three regions, i.e., LTO, FD, and HTO stages, analogous to the heavy oil. However, the mass-loss behaviors, initial and final temperatures at each reaction stage for Mahu light oil were different from the heavy oil. In addition, a more pronounced distillation process, prior to oxygen addition reactions, was presented for the tight oil owing to more light hydrocarbons. Table 2 summarized major characteristic parameters from TG-DTG curves, in some degree, implying the reaction path, main mode, and reaction intensity for the non-isothermal oxidation process (Pu et al., 2017a).

Close

Fig. 1

Mass loss and its mass loss rate curves for the tight oil under different heating rates.

Export to PPT

Under 10 K·min⁻¹, there existed a mild distillation process for oil components (C₃-C₆) at relatively low temperatures (<373 K) in consideration of the boiling temperature of n-C₇H₁₆, 371.45 K. Hence, the first stage, LTO, occurring in 303–642 K, included the distillation/volatilization reactions (<373 K) and oxidation addition reactions (100–642.7 K), during which heterogeneous reactions between the oil and O₂ were dominant. Moreover, vapor-phase oxidation, during the distillation process, was considered as a crucial influence on the fast oxidation, which under actual reservoir condition contributed the spontaneous combustion potentially (Freitag, 2016). While complicated oxidation addition reaction, involving the formation, growth, and disappearance of free radicals, leaded to ∼ 92.60 % mass loss at 642.7 K, and this process only leaded to a mild viscosity increment owing to the mutual influence between the condensation and breakage reactions. This partly supported the potential application of AIP in low permeability reservoirs. It was notable that the leading reaction process was dependent on specific oil properties and oxidation conditions.

A relatively mild fuel deposition (FD), ∼1.382 % mass loss, was detected for this tight oil with a placid fluctuation in DTG curve. Generally, LTO products underwent the re-oxidation, condensation, and aromatization reactions to form coke precursors. For the light oils, the intensity and continuity of coke deposition was not the limiting factor to initiate the combustion process. If LTO exothermic intensity and exothermic continuity were enough to trigger the fast oxidation of light fractions or vapor-phase hydrocarbons, the spontaneous combustion process could be appeared. Besides, the generated net heat under the low temperature region would be reinforced by the high pressure condition of the actual oil reservoir compared with the atmospheric condition (Fan et al., 2015, Huang & Sheng, 2018, Pu et al., 2017b). Especially, no obvious boundary and conversion was detected between LTO and FD stages for the light oil, indicating specific region division in Table 2 was just used to qualitatively depict the tight oil oxidation process. Compared with specific values, it was more significant and practical to focus on the trend variation.

During 689.7–973 K, drastic bond cleavage reaction, namely combustion process, was dominant to generate CO₂, CO, H₂O, and heat. But the mass loss (∼4.446 wt%) and mass loss rate at HTO region were far less than those at LTO stage, which was different from the heavy oil in our previous study and ascribed to abundant light hydrocarbon components in this tight oil (Chen et al., 2018). Naturally, it could be inferred that the thermal effect and in-situ upgrading process were weak and dispensable for the conventional light oil reservoir. However, for the tight light oil, should the gas-phase combustion of the light hydrocarbons and subsequent combustion process be triggered and maintained, the generated exothermic effect would fracture the glutenite matrix and create micro-fractures to in-situ upgrade the permeability and improve the EOR, whose derived technique has been applied in oil shale reservoirs (Kang et al., 2020, Kar & Hascakir, 2017).

3.1.2 DSC-DDSC analysis

DSC-DDSC curve could quantify the thermal effect, including the heat flow (DSC), heat flow rate (the first derivative of DSC curve, DDSC), and cumulative heat release (reaction enthalpy), which was expressed as follows,

where H_t and θ(t) respectively represented the already produced heat and corresponding heat flow at time t (Chen et al., 2019a, Chen et al., 2019b, Pu et al., 2017a, Wang et al., 2018).

Fig. 2 showed that the LTO peak heat flow, under 10 K·min⁻¹, was obviously higher (4.45 mW·mg⁻¹) than that (3.32 mW·mg⁻¹) in HTO stage. From the relevant DDSC curve, the trend and intensity of the heat flow rate were more advantageous and violent to reflect the exothermic process at LTO stage. In addition, corresponding lower LTO and higher HTO peak temperatures for Mahu tight light oil, compared with those for Tahe heavy oil in previous references (Chen et al., 2019b, Pu et al., 2017a), indicated that the LTO mode was mainly considered and dominant for the AIP in light oil reservoirs.

Close

Fig. 2

Heat flow and its rate (DSC-DDSC) curves for the tight oil under different heating rates.

Export to PPT

Based on Eq. (6), cumulative heat released curves were presented in Fig S3 under different heating rates, and specific data was summarized in Table 3.

Two obvious stages were identified based on the heat flow fluctuation. As for the down-drift trend during initial LTO process (303–493 K), it was supposedly attributed to the distillation of light components (intercellular water, lighter hydrocarbon, and oxidative products), which was consistent with previous researches on conventional light oils (Kök et al., 2020, Wang et al., 2018). During main oxidation process, a gradual increase trend was presented for the cumulative heat released curves and corresponding reaction enthalpies were respectively 1.76 kJ·g⁻¹ (LTO) and 2.03 kJ·g⁻¹ (HTO) under 10 K·min⁻¹. It was remarkably that, a very slight exothermic process was identified in FD region owing to the coupled effect of intricate thermal cracking and poly-condensation reactions.

3.1.3 Heating rate influence

There was an evident effect for differential heating rates on the non-isothermal oxidation process and corresponding reaction mechanisms, as respectively shown in Fig. 1 and Fig. 2.

With the heating rate increasing, TG/DSC curves deviated to high temperature with higher peak values, while corresponding DTG/DDSC curves were towards a lower position. Specifically, trigging temperatures, peak temperatures, peak heat flows, and region intervals were moved to higher temperature upper limits and narrower temperature ranges, corresponding to greater mass/heat loss rates, and higher peak oxidation rates, as shown in Table 2 and Table 3. This was owing to that, a worse heat provision and conversion efficiency was caused under higher heating rate, which consequently resulted in a less-effective reaction with worse mass loss. In the meantime, light hydrocarbon fractions would be oxidized under a higher temperature, which increased mass loss rate, peak and burnout temperatures owing to a shorter reaction time (Chen et al., 2019b, Pu et al., 2017a). Besides, the temperature shift and increase of DTG curves were also verified via the following Eq. (7) (Vyazovkin et al., 2011).

As for the variation of cumulative released heat under high heating rate, the higher thermal hysteresis and lower heat conversion efficiency would generate a more ungenerous reaction process owing to a shorter reaction time for the same temperature gradient. Consequently, peak and end temperatures were detected at higher temperature zones with a decreasing trend of the cumulative released heat respectively in LTO and HTO regions, as shown in Table 3 (Chen et al., 2019b, Pu et al., 2017a, Yin et al., 2016).

3.2.1 PDSC analysis

Under the heating rate of 10 K·min⁻¹, PDSC curve and contrastive DSC curve were shown in Fig. 3 simultaneously. Two main exothermic stages (303–578 K for LTO, 578–732 K for HTO) were apparently detected for the PDSC, in accordance with the trend in DSC. However, there respectively existed higher peak heat flows and heat enthalpies (8.69 kJ·g⁻¹ for LTO, 4.48 kJ·g⁻¹ for HTO), corresponding to lower threshold and ending temperatures and narrower temperature intervals, compared with those (1.76 kJ·g⁻¹ for LTO, 2.03 kJ·g⁻¹ for HTO) in DSC curve. This implied that, the initiation and progress of the oxidation process, especially for the LTO, were promoted and strengthened under the pressurized condition, which was consistent with previous researches (Dunn & Fuels, 2012, Dunn, 2006, Kök, 2008), and would be analyzed and interpreted in the subsequent section.

Close

Fig. 3

DSC and PDSC curve comparison for the tight oil (10 K·min⁻¹).

Export to PPT

Moreover, the first two exothermic peaks (20.35 mW·mg⁻¹, 18.03 mW·mg⁻¹) were recognized in the range of 483–578 K, which may be attributed to the LTO process, while the third lower peak (578–593 K) was considered as the HTO exothermic peak (16.79 mW·mg⁻¹), indicating a more complicated and drastic LTO exothermic process under pressurized condition. Although, the first peak heat flow (20.35 mW·mg⁻¹) was higher than the second (18.03 mW·mg⁻¹), the corresponding released heat (5.32 kJ·g⁻¹) for the second peak was instead greater than the first (3.30 kJ·g⁻¹). It could be explained that, the first LTO exothermic region was mainly contributed to the enhancement of gasification phase oxidation reaction under pressurized condition. Different from the heavy oil, the light oil contained abundant light aliphatic hydrocarbons without oxidation inhibitors (aromatics), which were much more volatile than the lightest aromatic compounds to undergo very rapidly oxidation process in gas phase.

However, with the temperature increasing and the gas-phase reaction terminating, the heat flow trend was decreasing and the liquid-phase oxidation reactions were gradually predominant with a lower heat flow rate (at 503–528 K) to initiate the RO_2• (alkyl-peroxy free radical) generation. Then many reaction pathways, such as hydrogen abstraction, addition to unsaturated molecules, and decomposition to alcohols, aldehydes, and/or ketones, were evolved with an increasing rate of heat release to form the second LTO exothermic peak (Freitag, 2016, Sarma et al., 2002). The difference and intensity of heat flow variations in LTO and HTO processes implied that, if the rapid vapor-phase oxidation, with considerable heat generation, was enough to trigger the subsequent oxidation process, the complete oxidation in low temperature (spontaneous combustion) would become possible for the actual high pressure oil reservoir.

3.2.2 Atmosphere pressure influence

Fig. 4 and Fig. 5 respectively illustrated the influence of different experimental pressures (1.00, 3.00, and 5.00 MPa) on the heat flow and cumulative released heat under the same heating rate of 10 K·min⁻¹.

Close

Fig. 4

Heat flow curves for the tight oil under different pressures (10 K·min⁻¹).

Export to PPT

Close

Fig. 5

Cumulative heat released curves for the tight oil based on PDSC curves.

Export to PPT

With the atmosphere pressure increment, the increasing trends were distinguished for the peak heat flows and cumulative released heat. Specifically, the cumulative generated heat from the oxidation process was respectively 4.32, 10.60, 13.10, and 18.00 kJ·g⁻¹ under the pressure of 0.10, 1.00, 3.00, and 5.00 MPa, exhibiting a positive correlation with the atmosphere pressure (R² = 0.873, in Fig. 6), which was consistent with previous researches (Dunn, 2006, Fan et al., 2015, Sarma et al., 2002).

Close

Fig. 6

The relationship of the cumulative heat generated vs. atmosphere pressure.

Export to PPT

It could be mainly owing to that, the diffusion degree between the oil and O₂ was improved with the elevated pressure to amplify the oxidation progresses in the vaporized and liquid phases. On the other hand, the endothermic effect, caused by the distillation process, was obviously weakened and it resulted in a higher amount of crude oil to undergo the exothermic reactions rather than be blew away by the air flow. Besides, it was inferred that, the released heat would produce much slowly at higher pressure, since total heat release reached the upper limitation when the whole oil sample was completely oxidized.

Furthermore, the increment of released heat in LTO stage was much higher than that in HTO region under pressurized atmosphere. It was in a certain extent indicated that, for the AIP in light oil reservoirs, considerable heat generated in LTO region had a significant contribution on the initiation and duration of the spontaneous combustion or low temperature combustion process, which was a novel method to improve the permeability of low-permeability and tight oil reservoirs. Frankly speaking, great efforts were highly made to study the LTO mechanism and investigate the low temperature combustion potential under elevated pressure condition.

3.3.1 Most probable mechanism function determination

In this section, latent mechanism function variation, under specific oxidation temperate subinterval (323–823 K), was determined via the Popescu method, which was specifically divided into LTO (323–823 K), FD (628–696 K), and HTO (696–823 K) stages. Based on the consensus, Table 4 listed 20 type probable mechanism functions to screen the most probable G(α) (Chen et al., 2019b, Mishra & Bhaskar, 2014). If certain G(α) function was considered as the most probable mechanism function, corresponding correlation coefficient (R²) of the fitting curve should approach towards 1.0 simultaneously with the intercept and standard deviation (r_SD) tending to zero, as briefly exampled in Fig. S5.

In LTO region, three temperature subzones (323–423 K, 423–533 K, 533–628 K) were determined to evaluated corresponding G(α), and Table 5 listed some most probable G(α) with detailed fitting process in Fig. 7.

Close

Fig. 7

Comparisons of different mechanism functions, G(a), employed to reflect the oxidation behavior in sequential oxidation temperature subzones (323–628 K).

Export to PPT

Generally, it was observed that more than one probable mechanisms could depict specific oxidation behavior. Via the common consideration of R² and r_SD, the dominant mechanism in each subzone would be determined via the selection of optimal function models. Briefly, the most probable functions in specific LTO subzones were respectively Jander equation (J-D₃), Power law (P₄), and Valens equation (D₂), indicating the complexity of specific oxidation reactions. As a whole, the LTO process was guided by the nucleation mechanism (P₂), consistent with the leading oxidation addition reaction to form concentrated oxides.

Analogously, specific FD and HTO temperature subzones were determined to evaluate corresponding mechanism functions, as shown in Table 6 and Table 7 with specific fitting processes in Fig. S6 and Fig. S7 respectively.

It was observed that J-D₃ (three-dimensional diffusion mechanism) and R₂ (contracting cylinder mechanism) were the most probable functions corresponding to specific FD temperature subzones, which respectively reflected the addition reaction and coke deposition process. This was in a certain extent illustrated that, the in-depth oxygen addition reaction had a significant acceleration on the progress of dehydrogenation, polymerization, and aromatization reactions to generate coke.

As for HTO stage, the reaction order models (F_1.5 for the first subzone, F_1.0 for the second subzone) were the most probable functions, and the whole HTO process followed the 1.5-order reaction model (F_1.5), indicating that the first subzone was more dominant rather than the second subzone. Moreover, it could be inferred that two different types of cokes were produced and contributed to the combustion with different reactivity and mechanism functions, which was in accord with previous researches and still needed further study (Cinar et al., 2011).

3.3.2 Kinetic parameters (E, A) calculation

Based on T_n, T_m, and corresponding G(α) in specific temperature subzones, activation energy variations could be determined and were presented in Fig. 8.

Close

Fig. 8

Activation energy variation for the Mahu tight light oil during the oxidation process.

Export to PPT

For the conversion rate α < 0.35, a linear rising tendency was observed for the activation energy owing to the distillation process under low temperature region. With the temperature increment, there was a progressive increase trend followed by a mild peak of the activation energy (α = 0.6–0.65) to promote the LTO process with partial oxide generation. Then an obvious peak of the activation energy in α = 0.75–0.94 was observed and mainly ascribed to the negative temperature gradient region (NTGR), which was associated with the upper region of the LTO and not as significant as for heavy oils (Chen et al., 2019b, Freitag, 2016). Subsequently, enormous energy was required to initiate the coke deposition with a sharp peak in α = 0.94–0.98.

Since the amount of deposited coke was slight for tight light oil, few sample had undergone the HTO with slight activation energy variation. Furthermore, in consideration of the coadjutant connection between f(α) and k(T), a positive relationship, namely “compensation effect”, was determined for the E and A as shown in Fig. 9.

Close

Fig. 9

Relationship of the E and lnA for Mahu tight light oil.

Export to PPT

3.3.3 Reaction rate constant, k(T), estimation

Based on the determined mechanism functions in sequential temperature subzones, the corresponding rate constant, k(T), could be calculated via the kinetic parameters in specific oxidation temperature subzones, as shown in Table 8. It was worth noting that, the lower values of the frequency factor in 628–696 K were probably associated with the more difficult and weaker coke deposition process for the tight oil.