3.1 Characterization of N-CDs
In unconventional reservoirs, the abundance of micro-nano-scale pore throats necessitates using small NPs to ensure injectability. Furthermore, the type and quantity of functional groups on these particles significantly influence their interaction with salt ions and their dispersion stability, directly impacting their application performance. The analysis of CDs involved several techniques: high-resolution Transmission Electron Microscopy (HR-TEM) is used to determine particle dimensions, crystallinity, and lattice spacing; X-ray Diffraction (XRD) provides insights into CD lattice spacing, crystallinity, and preferred orientations; UV-Vis and Photoluminescence (PL) characterize hydrothermal carbonization, surface states, optical properties, and functional groups; FTIR Spectroscopy reveals CD functionalities; and X-ray Photoelectron Spectroscopy (XPS) determines surface chemistry and elemental composition.
The particle size of E5_L CDs, determined from TEM images analyzed with ImageJ software (supplementary Figure S1), falls within the 3 to 7 nm range. HR-TEM images confirm the graphite nature of E5_L CDs, with a measured lattice fringe distance of 0.24 nm [40]. The XRD patterns reveal diffraction peaks at 21.6°, 24.6°, and 20.6° for E5_L, U5_L, and T5_L CDs, respectively (supplementary Figure S2). These peaks relate to lattice spacings of 0.41 nm, 0.36 nm, and 0.43 nm. Notably, these values exceed those obtained from HR-TEM due to the presence of functional groups on the CD surfaces [41, 42]. UV-Vis and PL spectra of E5_L, U5_L, and T5_L CDs reveal absorption bands at approximately 340 nm and 240 nm, as shown in supplementary Figure S3. These bands correspond to n→ π* and π→ π* transitions of C = O and C = C bonds within the CDs. Also, they exhibit strong, excitation-independent blue emissions around 450 nm. Furthermore, the blue color observed in the CDs solution under a 365 nm UV lamp confirms the N-doping of CDs. The FTIR spectra of E5_L, U5_L, and T5_L CDs are analyzed (supplementary Figure S4) to identify the functional groups present on their surfaces. The findings reveal that CDs exhibit a diverse range of oxygen and nitrogen functionalities, contributing to their excellent solubility in water. Table 2 summarizes the specific FTIR band assignments for the E5_L, U5_L, and T5_L CDs.
Table 2
FTIR band assignment for E5_L, U5_L, and T5_L CDs
| Wavenumber (cm− 1) | Functional group |
| 1000–1400 | C–O, C–S, and C–H stretching vibrations |
| 1188 | C–S stretching modes for T5_L CDs |
| 1378, 1438, 1407 | C–N stretching modes |
| 1559, 1576, 1570 | C = N stretching modes |
| 1636, 1709, 1714 | C = O stretching modes (carboxyl acid groups, –COOH) |
| 2540 | S–H stretching modes for T5_L CDs |
| 3056, 3053, 3060 | N–H stretching vibrations |
| 3212, 3204, 3223 | O–H stretching vibrations |
The XPS spectra (supplementary Figures S5 and S6) support the FTIR spectroscopy analysis and confirm the presence of carbon (C), oxygen (O), nitrogen (N), and sulfur (S) atoms on the surfaces of E5_L, U5_L, and T5_L CDs. The full-scan XPS spectrum exhibits distinct peaks at approximately 284 eV (C1s), 401 eV (N1s), 532 eV (O1s), and 164 eV (S2p). In the figure, high-resolution C1s spectra reveal three peaks at binding energies of approximately 284.5 eV (assigned to C = C/C-C), 286 eV (assigned to C-O/C-N/C-S), and 288.4 eV (assigned to C = O). The N1s spectra exhibit peaks at 399.8 eV (pyrrolic N) and 401.5 eV (graphite N). Additionally, O1s spectra show peaks at 531.4 eV (C = O) and 533.1 eV (C-O). The bar charts in Fig. 2 display the elemental composition of CDs obtained from XPS tests, which can be used to analyze IFT results.
XPS analysis confirms N and O doping in E5_L and U5_L CDs, as well as N, O, and S doping in T5_L CDs. Notably, E5_L CDs exhibit a higher N doping degree (9%) compared to U5_L (about 2.9%) and T5_L (about 3.2%). This suggests that E5_L CDs, rich in pyrrolic N, form long straight-chain hydrocarbon branches with nonpolar properties, enhancing polar-nonpolar phase bonding [40]. Additionally, T5_L CDs have a higher O content (33%) than E5_L and U5_L CDs (both around 17.3%), potentially contributing to greater polarity in water solutions.
3.2 Impact of Salt Types on IFT
To assess the impact of ions on IFT and compare it with that of CDs, the IFT of crude oil in saltwater (without CDs) was measured. NaCl, MgCl2, and CaCl2 were used at concentrations ranging from 0 to 200,000 ppm. Subsequently, the combined effect of different salt concentrations with various CDs on IFT was investigated. The CDs concentration remained constant at 1000 ppm. Additionally, the influence of monovalent and divalent salt ions on CDs IFT was explored using various reaction times and reactant ratios. Each IFT value represents an average of at least three measurements. All tests were conducted until the oil drop reached static equilibrium. Figure 3 illustrates the tree diagram summarizing all IFT tests with different salt ions.
The IFT changes for NaCl, CaCl2, and MgCl2 salts with the crude oil are shown in Fig. 4.
The line graph reveals that as NaCl, CaCl2, and MgCl2 concentrations increase, the IFT generally decreases at low salt concentrations, followed by an increase at high salt concentrations. This behavior can be attributed to two mechanisms. Firstly, at low salt concentrations, the ‘salting-in’ effect leads to increased ion absorption at the phase interface, resulting in reduced IFT. Conversely, at higher salt concentrations, the ‘salt-out’ effect reduces ions at the interface, leading to increased IFT. Secondly, increasing salinity at low concentrations expands the electrical double layer, causing resins and asphaltenes to accumulate at the interface (asphaltenes/resins absorption effect), thereby reducing IFT. However, in high salinity, the electrical double layer condenses, depleting asphaltenes near the interface and increasing IFT (asphaltenes/resins depletion effect) [43]. Figure 5 illustrates this mechanism.
The IFT values of CaCl2 and MgCl2 salts sharply decrease with increasing salt concentration. This reduction is attributed to the high ionic capacity and the ability of Ca2+ and Mg2+ ions to absorb more asphaltenes and resins at the water-oil interface. The graph indicates that the minimum IFT occurs at 50,000 ppm for CaCl2, with an IFT reduction of 13.81 mN/m, and 25,000 ppm for MgCl2, with an IFT reduction of 10.05 mN/m. Previous studies have reported varying IFT reduction values for salts in different crude oils [25, 27]. For this specific crude oil, MgCl2 exhibits a lower IFT value than CaCl2 up to a concentration of 25,000 ppm. This difference may be attributed to the stronger electronegativity of Mg2+, which facilitates the absorption of asphaltenes and resins at the fluid interface, leading to decreased IFT. However, at 50,000 ppm and higher concentrations, CaCl2 demonstrates a lower IFT than MgCl2. This behavior could be due to the higher affinity of Ca2+ ions for asphaltene molecules compared to Mg2+ ions [27]. Asphaltenes and resins exhibit distinct behaviours with increasing salinity, and their interaction within oil further affects IFT. Consequently, Ca2+ ions form stronger bonds with asphaltene molecules, leading to a reduction in IFT. Another alternative theory for evaluating IFT variation is Gibbs’s adsorption isotherm equation. Gibbs related IFT changes to Eq. (1) [44, 45].
$$d\gamma = - RT\sum {{\Gamma _i}d} \ln {\alpha _i}$$
1
Where dγ is the alteration in IFT, Γi represents surface excess concentration, αi shows the chemical activity of ith component in the solution, R is universal gas constant, and T denotes the absolute temperature. This equation reveals that the salt ions adsorb to the interface in low salinity, resulting in a positive surface excess (Γ) and decreased IFT. Conversely, the interface becomes saturated with ions in high salinity, leading to a negative surface excess and increased IFT.
3.3 Combined Impact of Salt Types and N-CDs on IFT
To assess the stability of CDs in the presence of monovalent and divalent salts, 1000 ppm of E5_L CDs was utilized at varying NaCl, CaCl2, and MgCl2 concentrations (Fig. 6). The investigation aimed to evaluate these salts' potential antagonistic or synergistic effects on the IFT behavior of CDs.
In the NaCl concentration graph, the addition of 1000 ppm E5_L CDs leads to a sharp reduction in IFT. Remarkably, this IFT reduction of approximately 18.3 mN/m (the base case without CDs and salt) occurs even without the presence of any surfactant, highlighting the synergistic effect between NaCl ions and E5_L CDs. E_CDs exhibit higher N doping levels compared to other N-doped CDs and are rich in pyrrolic N, which facilitates the formation of long straight-chain hydrocarbon branches with nonpolar properties. Additionally, their surface features polar atoms like O, capable of absorbing polar water molecules. Consequently, these CDs function effectively as connectors between two phases, similar to surfactants. XPS and FTIR tests confirm the elevated N content (9%) on the surface of E5_L CDs. Furthermore, the freeze-dried E_CDs exhibit low polarity due to their sticky NP behavior, similar to resin, and prolonged dissolution time in water. When E5_L CDs interact with NaCl, they attract asphaltene and resin from the oil bulk to the phase interface, enhancing bonding with the nonpolar phase and significantly reducing IFT. The CaCl2 curve reveals an initial increase in IFT with rising salt concentration, followed by a subsequent decrease at high salt levels. The elevated IFT in E5_L solution with increasing CaCl2 concentration can be attributed to CD instability upon encountering Ca2+. Previous studies indicate that NP instability disrupts both NP and salt performance, leading to increased IFT [36]. Consequently, E5_L CDs exhibit an antagonistic effect with Ca2+. Furthermore, visual observations confirm that adding calcium salt to E_CDs induces instability, supporting this hypothesis. This instability is evident through the precipitation of CDs at a concentration of 5000 ppm, which significantly increases to 25000 ppm. Beyond 100,000 ppm, substantial CD flocculations are deposited. Removing CDs from the system restores the IFT to its base value. In the case of MgCl2, E5_L CDs exhibit minimal instability, resulting in reduced IFT with increasing salinity. However, a slight increase occurs at concentrations exceeding 50,000 ppm. Figure 7 illustrates the mechanism of IFT change due to the interaction of monovalent and divalent salts with CDs.
Subsequently, the impact of salt ions on IFT for U5_L and T5_L CDs, which differ in precursor atoms (O, S) and molecular structure, was evaluated. Figure 8 displays the IFT measurements using U_CDs across various NaCl, CaCl2, and MgCl2 concentrations.
In the NaCl graph for U5_L CDs, the IFT decreases by adding 1000 ppm of these CDs, from 19.03 to 12.33 mN/m. However, U_CDs do not exhibit as significant an IFT reduction as E_CDs. The underlying cause lies in the synthesis steps of CDs, including dehydrolysis, aromatization, and critical supersaturation, affecting the structure of U_CDs. XPS tests reveal that U5_L CDs have a low N content (2.9%). Unlike EDA, urea molecules have a single C atom and three binding sites, facilitating three-dimensional aggregation during synthesis. These characteristics enhance the polarity of U5_L CDs, restricting their presence at the fluid interface and preventing substantial IFT reduction. Additionally, U_CDs exhibit superior solubility in water compared to E_CDs, emphasizing their polarity. On the other hand, the IFT of U5_L and CaCl2 decreases smoothly up to 10,000 ppm salt concentration, followed by an increasing trend up to 200,000 ppm. This rise in IFT results from U5_L CDs instability at high salt concentrations (similar to E5_L at low salt concentrations). At 25,000 ppm CaCl2, U5_L CDs precipitate visibly (usually occurring at lower salt concentrations but remaining invisible). This precipitated growth of NPs contributes to an increase in IFT. Unlike E5_L CDs at high CaCl2 concentrations, U5_L CD aggregates exhibit minimal growth and do not reach the deposition point due to greater stability. Consequently, the IFT does not decrease at high CaCl2 concentrations. Specifically, at 100,000 ppm CaCl2 concentration, the IFT value for E5_L CDs reaches 5.26 mN/m (see Fig. 6), which is close to the IFT value of the base CaCl2 solution (5.28 mN/m, see Fig. 4). However, at the same CaCl2 concentration, the IFT value for U5_L CDs is 17.20 mN/m, significantly different from the CaCl2 base solution’s IFT. The U5_L CDs exhibit a similar decreasing slope with MgCl2 as NaCl, indicating their stability. Visual observations also confirm that U5_L CDs remain stable with MgCl2, unlike CaCl2. Figure 9 depicts the IFT diagrams for T5_L CDs with NaCl, CaCl2, and MgCl2.
As observed by XPS tests, T5_L CDs exhibit a higher O content (33%) and a lower N content (3.2%) on their surface. Additionally, the presence of extra sulfur atoms enhances the CDs’ polarity compared to others. Notably, the molecular structure of thiourea is more similar to urea. Both groups' S and O atoms have similar hybrid orbitals, specifically an sp2 structure. Consequently, T5_L CDs tend to create bonds with oil components, preventing ions from affecting the interface. As a result, the IFT of T5_L CDs remains relatively unaffected by increasing NaCl concentration, showing a reduction of 0.67 mN/m compared to the base case without CDs and salt. However, the CaCl2 curve shows greater IFT reduction (2.53 mN/m compared to the base case without CDs and salt). While divalent Ca2+ ions exhibit the highest ability to destabilize NPs and promote aggregation, T_CDs possess both high polarity and stability. Consequently, Ca2+ ions cannot destabilize these CDs, resulting in minimal changes in the CaCl2 curve and a slight decrease in IFT. The IFT curves for T5_L in MgCl2 solutions also show a more pronounced decreasing trend of 3.37 mN/m. This indicates that divalent Ca2+ and Mg2+ ions can relatively break the surface molecular arrangement of the T_CDs and oil components.
3.4 Impact of Salt and N-CD Reaction Time on IFT
The impact of salts on IFT changes for synthesized CDs in different periods has been investigated. Figure 10 shows the IFT plots with NaCl for different reaction times of E5_L CDs.
Based on the findings, IFTs exhibit a monotonic decrease from approximately 13 mN/m to about 2 mN/m (an 11-unit reduction) for E10_L and E15_L CDs as the NaCl concentration increases from 0 to 200000 ppm. This suggests a favorable synergy between NaCl salt and E10_L/E15_L CDs. However, these CDs still exhibit higher IFT values compared to E5_L CDs at each NaCl concentration, with a difference of approximately 3 mN/m. The decline in pyrrolic N content within E_CDs with longer reaction times leads to an increase in graphite N content [40]. Graphite N resides within the graphene structure (rather than on the surface) and cannot interact with NaCl ions. Consequently, graphite N does not contribute to the observed reduction in IFT. As a result, E_CDs synthesized over an extended time have a minor impact on IFT reduction. Figure 11 depicts the IFT curves of crude oil and NaCl salt, corresponding to different reaction times for U_CDs.
The IFT curves of U_CDs decrease with increasing NaCl concentration. Specifically, IFTs exhibit a downward trend, resulting in an approximately 6 mN/m reduction for U10_L and U15_L CDs as the NaCl concentration rises from 0 to 100,000 ppm. Interestingly, these U_CDs display a shallower slope compared to E_CDs, indicating a weaker synergy effect in the presence of monovalent NaCl salt. Furthermore, the 10-hour IFT graph with NaCl salt is lower than the 5-hour graph, while the 15-hour curve falls between the 5-hour and 10-hour curves. These subtle changes underscore the significance of reaction time in U_CDs when interacting with NaCl ions. Due to their greater polarity compared to E_CDs, longer reaction times are essential to achieve optimal IFT reduction. As reaction time increases, more N atoms incorporate into the graphene layer, resulting in a higher relative amount of graphite N in the U_CDs. Additionally, the overall size of the U_CDs grows with longer reaction times [40]. These combined factors influence the IFT behavior of U_CDs in the presence of NaCl ions. Similarly, supplementary Figure S7 displays the graph comparing NaCl salt and T_CDs synthesized at two different times (5 and 10 hours). Remarkably, T5_L with NaCl exhibits nearly the same IFT value as T10_L with NaCl. Accordingly, the high-polarity T_CDs and NaCl ions do not synergistically interact to reduce IFT beyond their individual effects.
Furthermore, similar tests were conducted for MgCl2 and CaCl2, as shown in Fig. 12 and Fig. 13.
The graph reveals that Ca2+ ions in E_CDs solution cause significant precipitation. Within the precipitation range (1000 to 25000 ppm), the IFT exceeds the expected value. Subsequently, particle deposition occurs with higher salinity (50000 ppm and above), leading to the IFT returning to the standard value observed for CaCl2 solution and crude oil without CDs. This decrease in IFT occurs because the CDs exit the bulk aqueous solution and the interface between the two phases. Additionally, the graphs related to U_L CDs and CaCl2 initially decrease and then increase due to reaching the CDs’ precipitation point (10000 to 100000 ppm). Meanwhile, the T_L CDs graphs exhibit higher IFT values and consistently follow a downward trend with a slight slope. The results demonstrate that Ca2+ and Mg2+ ions can more significantly occupy the sites created at the interface of T_CDs solutions and the oil phase components. Consequently, increasing the reaction time enhances the ability of Ca2+ and Mg2+ to reduce the IFT for T_CDs (e.g., from 3.37 to 6.4 mN/m for MgCl2 compared to the base case without CDs and salt). Overall, E_CDs consistently exhibit the lowest IFT values across different reaction times in the presence of various salts, showing the best synergy impact. Notably, the IFT value for U_CDs is also lower than that of T_CDs, except for CaCl2 solutions, where instability leads to increased IFT values.
3.5 Impact of Salt and Reactant Ratio on IFT
To explore the influence of salt ions on the IFT of CDs with different reactant ratio, the N sources (EDA, urea, and thiourea) and the C source (citric acid) were used at mole fractions of 3, 1, and 0.5. Subsequently, their IFT was analyzed in the presence of varying concentrations of NaCl, CaCl2, and MgCl2. Figure 14 illustrates the impact of NaCl on E_CDs with varying reactant ratios.
As NaCl concentration increases, the IFT of E5_M and E5_H CDs remains relatively stable. Notably, the IFT graph for E5_H CDs is slightly lower than that of E5_M CDs by approximately one unit. The IFT values exhibit a downward trend with minor fluctuations, resulting in a reduction of less than 1.5 mN/m as NaCl concentration varies from 0 to 200,000 ppm (5.63 mN/m compared to the base case without NaCl and CDs). Interestingly, E5_M and E5_H CDs exhibit a weaker synergy effect in the presence of NaCl compared to E5_L CDs. Unlike the reaction time, the reactant ratio significantly influences the interaction between NaCl and E_CDs. The IFT impact of the reactant ratio for U_CDs was evaluated, considering mole fractions of 1 and 0.5 in addition to the previously tested value of 3 (as shown in Fig. 15).
The IFT of U5_M and U5_H CDs remains relatively stable as NaCl concentrations increase. Although their diagrams exhibit slight fluctuations, NaCl does not significantly impact the IFT of these U_CDs. Notably, CDs synthesized from urea with a reactant ratio of 3 (which are more N-doped) show a steeper decrease in IFT with rising NaCl concentration, similar to E_CDs. Overall, the reactant ratio plays a significant role in the NaCl and U_CD interaction affecting IFT. Additionally, the diagrams representing different reactant ratios for T_CDs exhibit no further changes with increasing salt concentration, as depicted in supplementary Figure S8. Figure 16 depicts the impact of CaCl2 on IFT variations in EDA_5h CDs with different molar ratios.
At high CaCl2 concentrations, the E5_M and E5_H CD curves remain stable, unlike E5_L CDs. According to the observation method, E5_M and E5_H CDs exhibit greater stability in CaCl2 solutions and do not deposit. Consequently, the CD precipitate remains soluble, increasing IFT behavior stability. Sharp IFT reduction does not occur at high CaCl2 concentrations because Ca²+ ions are unable to occupy the sites created by E5_M and E5_H CD solutions at the oil phase interface. As a result, these curves do not converge to the IFT value of the CaCl2 solution without CDs. The IFT diagram of E5_M shows a steady increase as CaCl2 concentration rises. This behavior is attributed to the relative stability of the CDs against CaCl2. As CaCl2 concentration increases, the instability levels of E5_M CDs gradually increase, leading to a steady rise in IFT. Therefore, the polarity of E5_M CDs renders them ineffective in altering IFT when exposed to CaCl2 salts. These CD aggregations create sites with oil components that Ca2+ ions cannot occupy. The IFT graph of E5_H initially rises with increasing CaCl2 concentration (until 1000 ppm). However, beyond this point (from 5000 to 200000 ppm), a downward trend emerges, indicating lower precipitation and greater stability compared to E5_M CDs. The enhanced stability of E5_H CDs (compared to E_L CDs) may result from reduced N atoms and eliminated linear hydrocarbon branches. This behavior suggests that some Ca2+ ions can occupy the sites created by E5_H CD solutions at the oil phase interface. Due to its higher citric acid content in the reactant ratio, E5_H CDs exhibit greater polarity and stability, resulting in the observed downward IFT trend with increasing CaCl2 concentration. Figure 17 illustrates the IFT curves for EDA with different reactant ratios in the presence of MgCl2 salt.
The IFT remains constant as MgCl2 concentration increases. In summary, the IFT values of E5_M CDs slightly increase, while E5_H CDs slightly decrease with the addition of CaCl2 or MgCl2 salt. Next, the IFT values for U_CDs with varying reactant ratios in the presence of CaCl2 salt are calculated and presented in Fig. 18.
The IFT graphs for U5_M and U5_H in the presence of CaCl2 exhibit a low slope and demonstrate stable behavior. In contrast, the graphs for U5_L initially show a downward trend followed by an upward trend due to CD precipitation. Supplementary Figure S9 presents the IFT graphs for MgCl2 and U_CDs with different reactant ratios. The U5_M CDs curve remains relatively unchanged across various MgCl2 concentrations. However, the U5_H CDs curve decreases by approximately 5 mN/m as the MgCl2 concentration increases from 0 to 200000 ppm. Additionally, the IFT graphs for CaCl2/MgCl2 salts and T_CDs with different reactant ratios are shown in supplementary Figures S10 and Fig. 19.
The data reveal that T5_M CDs exhibit a more pronounced reduction in IFT with increasing concentrations of CaCl2 and MgCl2 compared to T5_L CDs. The IFT reductions are 3.86 mN/m for CaCl2 and 7.44 mN/m for MgCl2 by increasing salinity from 0 to 200000 ppm (5.3 mN/m and 8.88 mN/m compared to the initial case without salt and CDs, respectively). Overall, CDs synthesized from EDA demonstrate a significant synergistic effect with NaCl salt, suggesting that their use in EOR operations could be advantageous. The straightforward synthesis methods for these CDs make them a potentially cost-effective option for EOR applications when used with the appropriate salt. Additionally, the primary materials required for CD preparation, such as EDA at low concentrations (e.g., 1000 ppm), are relatively inexpensive. Nonetheless, further research and field trials are necessary to confirm the performance and economic viability of CDs in large-scale oil recovery under reservoir conditions. Table 3 and Table 4 summarize the stability and IFT changes resulting from the interaction of different salts and CDs in this study.
Table 3
Polarity, stability, and instability potential of N-CDs and other parameters
| Parameter | Polarity | Stability | Instability Potential |
| E-CDs | + | + | |
| U-CDs | ++ | ++ | |
| T-CDs | +++ | +++ | |
| NaCl | + | | + |
| MgCl2 | +++ | | ++ |
| CaCl2 | ++ | | +++ |
| Increasing Reaction Time | ↑ | ↑ | |
| Increasing CA as Reactant | ↑↑ | ↑↑ | |
Table 4
Summary of different salt ions Impact on stability and IFT behavior of various CDs (√: Stable, ×: Unstable, √×: Initially stable, then unstable, ↗: Increasing, ↘: Decreasing, ↘↗: Initially increasing, then decreasing, -: No change)
| Parameters | Stability by Increasing … |
| Salinity | Reaction time | CA as Reactant |
| Salts | NaCl | MgCl2 | CaCl2 | NaCl | MgCl2 | CaCl2 | NaCl | MgCl2 | CaCl2 |
| E-CDs | √ | √× | ×× | √ | × | ×× | √ | ×√ | ×√ |
| U-CDs | √ | √ | √× | √ | √ | × | √ | √ | √ |
| T-CDs | √ | √ | √ | √ | √ | √ | √ | √ | √ |
| Parameters | IFT Changing by Increasing … |
| Salinity | Reaction time | CA as Reactant |
| Salts | NaCl | MgCl2 | CaCl2 | NaCl | MgCl2 | CaCl2 | NaCl | MgCl2 | CaCl2 |
| E-CDs | ↘↘↘ | ↘↗ | ↗↘ | ↗ | - | - | ↗ | ↗↘ | ↗↘ |
| U-CDs | ↘↘ | ↘↘ | ↘↗ | ↘↗ | ↘↗ | - | ↗ | ↗↘ | ↗ |
| T-CDs | - | ↘ | ↘ | - | ↘ | ↘ | - | ↘↘ | ↘ |
Data availability
All data generated or analyzed during this study are included in this published article.