Introduction

Single-stranded deoxyribonucleic acids (ssDNA) are valuable molecules in biomedical research and molecular biology, contributing significantly to disease identification and cellular functions1,2. The accurate isolation and purification of nucleotide sequences, encompassing both ribonucleic acid (RNA) and DNA, holds peerless significance, with broad applicability in curative treatments, genetic diagnosis, and forensic science3,4,5. To meet the escalating demand for precise genetic information, there is a growing necessity for efficient ssDNA extraction methods. In this context, magnetic beads (MB) are notable for their filter-free and time-saving approach. However, conventional MB techniques, reliant on external magnetic forces, may introduce risks of sample contamination during the elution process6,7.

In response to this challenge, a technique utilizing nitrocellulose (NC) membranes has been introduced as a solid extraction method, offering a promising alternative to enhance ssDNA surface assimilation8,9. Nonetheless, NC membranes have certain limitations, particularly in terms of mechanical strength. In their dry state, NC membranes exhibit increased brittleness and fragility, leading to handling difficulties and a heightened risk of structural damage, which significantly reduces their mechanical stability and poses operational challenges, particularly in applications that demand durability and repeated manipulation10,11,12,13. Ongoing advancements in solid-phase extraction methods using NC membranes incorporate surface coatings with graphene derivatives, which improve their stability in the dry state14,15.

In our previous studies, NC membranes modified with both graphene oxide (GO) and polyethylene glycol (PEG) aminated reduced graphene oxide (GA), respectively, have been used as a solid-phase approach for the extraction of oligonucleotides. The developed methodology yielded promising results after a 60-minute incubation in α modified minimal essential medium (αMEM), indicating the extraction of 330–370 pg of total ssDNA for the GO modification and 590–610 pg for the GA modification. However, limitations associated with the use of GO and GA, such as stability, and variability in binding efficiency, prompted us to explore alternative carbon allotropes16,17.

Among the carbon species, single-walled carbon nanotubes (SWCNTs) provide several benefits for improving the NC membrane properties, including structural integrity and chemical stability, which may lead to enhanced performance18,19,20,21. SWCNTs excel as support material for oligonucleotide immobilization across diverse biosensor platforms like fluorescent-based systems22,23, electrochemical-based systems24, and field-effect transistors (FETs)25. Moreover, SWCNTs have already been successfully employed as a quencher in fluorescent-based systems, serving as an electron acceptor by quenching the fluorescence of labeled ssDNA through Förster resonance energy transfer (FRET)26,27. This process, primarily facilitated by surface adsorption and electron transfer mechanisms, involves the transmission of excited-state energy from the donor fluorescent-marked ssDNA donor to the SWCNT acceptor, resulting in the quenching of the fluorescently marked sequence, commonly referred to as the ‘turn-off’ effect. Despite their many advantageous characteristics, SWCNTs face certain limitations, particularly in aqueous environments, where their hydrophobic nature can lead to poor dispersion and reduced interaction efficiency with biomolecules like ssDNA28,29.

In this study, we extend the ssDNA extraction from various complex ionic solutions by exploring the use of a NC membrane as a modified support, leveraging the synergistic effect of SWCNTs and GA, used together as modifying agents. The successful modification of the NC membrane with SWCNTs/GA was validated through scanning electron microscopy (SEM), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, Fourier transform infrared spectroscopy (FTIR), and wettability analysis. In line with these investigations, we evaluated the effectiveness of the hybrid membranes in adsorbing 6-carboxyfluorescein (FAM)-labeled ssDNA by incubating them in three complex solutions that mimic human plasma, each containing different ionic species: magnesium chloride (MgCl2), manganese chloride (MnCl2), and a combination of both. Following incubation, the membranes were immersed in a desorption solution, and fluorescence analysis using the FRET technique validated the effectiveness of the SWCNTs/GA-NC hybrid membranes in extracting ssDNA.

Materials and methods

Materials

Bovine serum albumin (BSA), tris hydrochloride (Tris-HCl), CH3(CH2)11OSO3Na (sodium dodecyl sulfate-SDS), SWCNTs, MgCl2, and αMEM were purchased from Sigma-Aldrich, located in St. Louis, MO, USA. MnCl2 was procured from SILAL (Bucharest, Romania). The NC membranes, with a 47 mm diameter and 8.0 µm pore size, were acquired from Sartorius (Göttingen, Niedersachsen, Germany). The FAM-ssDNA, labeled with 6-carboxyfluorescein at the final primer, and featuring the sequence 5’-TTTCAACATCAGTCTGATAAGCTATCTCCC-3, was acquired from Integrated DNA Technologies, Inc. (Coralville, IA, USA). Additionally, GA, with CAS No.: 7782-42-5, was procured from ACS-Materials (Pasadena, CA, USA).

SWCNTs/GA dispersion preparation

The SWCNTs and GA powders were mixed into a single Berzelius glass containing 20 mL of ultrapure water, forming a dispersion with a final concentration of 1 mg/mL. To obtain the dispersion, a VC×750 sonicator (Sonics & Materials, Inc., Newtown, CT, USA) was used for 2 h while maintaining the dispersion in an ice bath to prevent overheating. The sonicator operated in cycles of 10 s of active pulsing followed by a 5-second rest, with a constant frequency of 20 kHz throughout the process.

Fabrication of SWCNTs/GA-NC hybrid membrane

A commercially available NC membrane was cut into 6 mm diameter discs, each with an approximate mass of 1.1 mg, using a standard paper hole punch. Non-covalent modification of the NC membrane was carried out using a procedure similar to the dot blot technique. As shown in Fig. 1, a diluted dispersion of SWCNTs/GA at 400 µg/mL was prepared, and 5 µL of this solution was drop-cast onto the NC membrane. Before use, the hybrid SWCNTs/GA-NC membranes were rinsed with deionized water and left to air dry overnight.

Fig. 1
figure 1

Fabrication of the SWCNTs/GA-NC hybrid membrane by drop casting a dispersion of SWCNTs/GA on the NC membrane.

Processes for adsorbing, extracting, and detecting FAM-labeled ssDNA in ionic solutions using SWCNTs/GA-NC hybrid membranes

For the adsorption, extraction, and detection of FAM-labeled ssDNA, three distinct ionic complex solutions with a final volume of 300 µL each, based on αMEM, were prepared containing 100 mM MnCl2, 100 mM MgCl2, and a combination of both ions. Additionally, each solution was supplemented with 0.1 mg/mL BSA, 10 mM Tris-HCl (pH 8.0), 0.1% SDS, and 16 nM FAM–ssDNA. The prepared solutions were added to black, flat-bottomed Costar 96-well plates with 100 µL of ionic complex solution per well. In the adsorption procedure, triplicate experiments were conducted in which nine SWCNTs/GA-NC hybrid membranes (Step 1 from Fig. 2) were incubated for 60 min in the prepared ionic complex solutions to promote the binding of FAM-labeled ssDNA (Step 2 in Fig. 2).

Fig. 2
figure 2

(1) Placement of the membrane into Costar 96-well flat-bottom plates; (2) Immersion of the membrane in the solution containing FAM-labeled ssDNA; (3) Removal of unbound FAM-labeled ssDNA from the membrane surface; (4) Desorption of FAM-labeled ssDNA following immersion of the SWCNTs/GA-NC hybrid membrane in Tris-HCl solution; (5) Removal of the membrane and collection of the desorption solution; (6) Measurement of fluorescence emission intensity using the FRET technique.

After incubation, the unbound ssDNA was washed off with distilled water (Step 3 from Fig. 2). The membranes were subsequently immersed in 10 mM Tris-HCl (pH 7.0) for 45 min to promote FAM–ssDNA desorption (Step 4 from Fig. 2). Afterwards, the SWCNTs/GA-NC membranes were removed (Step 5 from Fig. 2), and the fluorescence of the desorption solution was recorded at 535 nm, using a microplate reader (Step 6 from Fig. 2). The measurement was based on FRET technique that detects energy transfer between two light-sensitive molecules when they are in proximity, indicating successful binding or interaction.

Spectrofluorimetry analysis

The intensity of the emitted fluorescence was assessed at 23 °C using a TECAN Spark fluorescence microplate reader (Tecan Trading AG, Männedorf, Switzerland), with five measurements taken per well, at a wavelength of 535 nm.

Membrane characterization

FTIR analyses were conducted to examine the interactions among the three components: SWCNTs, GA, and NC, using the spectrometer SHIMADZU 8900 (Kyoto, Japan). The FTIR spectra were recorded over a range of 400–4000 cm− 1, with a resolution of 4 cm− 1, and each sample was measured 32 times without any additional sample preparation. To verify the consistency of the results, duplicate analyses were performed on the pristine NC membrane, SWCNT/GA-NC membrane, SWCNTs, and GA powders.

Raman analyses were conducted using a Renishaw inVia Raman confocal spectrometer (Renishaw, Brno-Černovic, Czech Republic) with a 100× objective, and the laser power was set to 10% with a 785 nm laser excitation.

XPS measurements were conducted using a K-Alpha spectrometer from Thermo Scientific (Waltham, MA, USA), featuring a monochromatic Al Kα source with an energy of 1486.6 eV. The system operated under high vacuum conditions with a base pressure of 2 × 10⁻⁹ mbar. Each XPS analysis was performed at three different points on the membrane, and the values are reported as averages of these three measurements.

The surface morphology of the NC membranes was analyzed using an FEI Quanta F250 scanning electron microscope. This analysis was performed before and after applying SWCNTs/GA dispersion, as well as before and after immersing the membranes in the complex solution. Prior to SEM imaging, a thin layer of gold-palladium was sputtered onto the membranes to enhance conductivity.

The assessment of the membrane’s hydrophilic/hydrophobic characteristics was conducted using the sessile drop method, employing a Krüss Scientific Drop Shape Analyzer DSA100 (Hamburg, Germany). The CF03 digital camera captured images of a 2 µL deionized water droplet on the membrane surface over a 5-second period to measure static water contact angles at room temperature. The water contact angle values were determined using the DSA3 software (version 2–11), with each value representing the average of three measurements per specimen. The results were analyzed based on the Young–Laplace equation30, which describes the relationship between surface tension forces and the pressure differential across the curved interface of the droplet.

Results and discussion

Evaluation of SWCNTs/GA-NC hybrid membrane

FTIR analysis

FTIR experiments were conducted to assess the modification of the NC membrane with SWCNTs and GA. Figure 3 shows the FTIR spectra of SWCNTs, GA, NC membrane, and SWCNTs/GA-NC hybrid membrane.

The spectrum of pristine SWCNTs typically lacks distinct absorption bands associated with functional groups31,32; however, a peak near 2350 cm−1 appears likely due to CO2 adsorption on the carbon nanotubes, which may have occurred during the calibration process33.

For GA, several characteristic absorption bands are observed at specific wavenumbers. The bands at 2919 cm−1 and 2854 cm−1 correspond to the stretching vibrations of ethylene (C-H) groups, indicating the presence of alkyl chains in the structure. The absorption band at 1565 cm−1 is attributed to the vibrational modes of amine (N-H) groups, confirming the presence of nitrogen-containing functional groups within GA. Additionally, the band at 1089 cm−1 is associated with the vibrational modes of epoxy (C-O-C) groups, reflecting the contribution of oxygen-containing functionalities in the material. Furthermore, a broad absorption band at 3294 cm−1 is indicative of the stretching vibrations of free and unbound hydroxyl (O-H) groups34,35. In the case of the NC structure, three distinct peaks at 838 cm−1, 1280 cm−1, and 1645 cm−1 confirm the presence of NO2 groups, corresponding to their bending vibrations, symmetric stretching, and asymmetric stretching, respectively36.

Fig. 3
figure 3

FTIR spectra of (a) pristine SWCNTs, (b) GA, (c) NC membrane, and (d) SWCNTs/GA-NC hybrid membrane.

Regarding the SWCNTs/GA-NC hybrid membrane, the FTIR analysis highlights the presence of all the aforementioned absorption bands corresponding to SWCNTs, GA, and NC. Specifically, the peak at 2350 cm−1 is attributed to atmospheric carbon dioxide, a feature observed in the SWCNTs, as was previously discussed. The presence of GA in the hybrid membrane is indicated by the characteristic absorption bands. First, the peak corresponding to N-H groups does not shift from 1565 cm−1; however, its intensity is reduced, likely due to interference from adjacent NO2 functional groups. Additionally, the broad absorption band near 3294 cm−1 also appears, indicating the presence of reduced graphene oxide, while absorption bands at 2919 cm−1 and 2854 cm−1, associated with C–H bending and stretching vibrations, confirm the presence of PEG in the hybrid membrane. Finally, the position of the absorption bands characteristic of NO2 groups observed in the pristine NC membrane remains unchanged after the functionalization process. Overall, the spectral analysis confirms the successful modification of the NC membrane with SWCNTs and GA.

Raman analysis

The Raman analysis (Fig. 4) reveals chemical changes in the SWCNTs/GA-NC hybrid membrane (a) after exposure to complex media containing ssDNA and MnCl2 (b), MnCl2/MgCl2 (c), or MgCl2 (d). Generally, no significant changes were observed between the first three analyzed samples, however shifts in the Raman bands were noticeable for the hybrid membrane incubated in MgCl2 containing medium. The O–NO2 stretching vibrations characteristic of nitrocellulose37 are represented by the band near 850 cm−1 that was seen in all samples, confirming XPS results. Minor changes in the nitrate ester linkages or interactions with the ionic species present in MnCl2 and MgCl2 are suggested by the slight shifts in this band (851 cm−1 in the pristine sample to 849–854 cm−1 in treated samples). This is in line with research on nitrocellulose derivatives, which shows that functional group interactions cause subtle variations in frequency37,38.

In the 1280–1300 cm−1 range, the bands are attributed to C–O ester stretching vibrations or the D-band of SWCNTs and GA, which indicate disorder or defects in the sp2 carbon network. The hybrid membrane exhibits a band at 1287 cm−1, shifting to 1293 cm−1 in the MgCl2-incubated membrane, reflecting increased defect density caused by Mg2+ coordination. Such ionic interactions enhance defect-induced Raman modes, as noted in the literature39. This increase in defects likely contributes to the higher adsorption capacity of ssDNA strands, which preferentially bind to defect sites on the carbon network.

The 1560–1600 cm−1 range, dominated by the G-band (graphitic mode), reflects the sp2-hybridized carbon frameworks in SWCNTs and GA. The hybrid membrane’s G-band at 1568 cm−1 shifts upward to 1590 cm−1 in the MgCl2-incubated membrane, indicating strain effects. This is consistent with charge transfer from Mg2+ ions to the graphenic lattice, which enhances π-π interactions with ssDNA bases, facilitating ssDNA adsorption. Similar strain-induced G-band shifts have been highlighted in prior studies on functionalized carbon materials40.

The band observed in the range 1740–1750 cm−1, corresponding to C = O stretching vibrations in ester groups, shifts from 1744 cm−1 in the hybrid membrane incubated in MnCl2 to 1753 cm−1 in the hybrid membrane incubated in MgCl2 medium, indicating the formation of coordination complexes between Mg2+ ions and carbonyl groups. Furthermore, the 2560–2580 cm−1 band, associated with overtone modes (G′ band) in SWCNTs or graphene, exhibits a subtle shift, highlighting the influence of Mg2+ and Mn2+ ions in modifying the electronic environment and facilitating enhanced interactions with ssDNA.

Notably, in the 100–200 cm−1 range, the observed peaks at 171 and 264 cm−1 for the hybrid membrane, shifting to 161–173 cm−1 and 263–268 cm−1 in the membranes incubated in complex medium, are indicative of radial breathing modes (RBM) of SWCNTs. The distinctive cylindrical structure of SWCNTs, in which the vibrational frequency is inversely proportional to the tube diameter, is characterized by these RBM peaks41. The observed shifts reflect environmental effects, such as charge transfer and local strain induced by Mn2+ and Mg2+ ions. The membrane from the MgCl2 complex medium exhibited the most pronounced changes, consistent with its higher ssDNA adsorption, as RBM shifts and defect formation enhance ssDNA-SWCNT interactions.

Fig. 4
figure 4

Raman spectra of SWCNTs/GA-NC hybrid membrane before (a) and after incubation in ionic complex media containing MnCl2 (b), MnCl2/MgCl2 (c), or MgCl2 (d).

The Raman spectra provide a deeper understanding of the structural and chemical changes in the SWCNTs/GA-NC membrane after exposure to ionic media. Correlating with XPS and FTIR results, the data confirm increased defect density, and enhanced ion-carbon interactions, particularly for the hybrid membrane in the MgCl2 complex medium, which adsorbed the highest quantity of DNA. These findings highlight the role of structural modifications in facilitating DNA adsorption onto the SWCNTs/GA-NC hybrid membrane.

XPS analysis

The XPS chemical composition survey from Table 1 were used to assess the atomic percentages of C, O, and N in the SWCNTs/GA-NC hybrid membrane before and after incubation in complex media containing ssDNA and MnCl2, MnCl2/MgCl2, or MgCl2. The evaluation revealed that the SWCNTs/GA-NC hybrid membrane before incubation had a high content of carbon atoms (62.2%), followed by oxygen (30.5%) and nitrogen (6.8%).

After the membrane was incubated in the complex medium containing MnCl2, subtle changes were observed in the chemical environment surrounding the carbon atoms. These changes may be attributed to interactions with Mn2+ ions, which likely facilitated the adsorption of ssDNA onto the membrane. This is supported by the observed decrease in C content, which dropped from 62.2 to 49.83%, and the increase in O and N content, which rose from 30.5 to 39.04% and from 6.8 to 9.19%, respectively, compared to the hybrid membrane before exposure to the medium.

The subsequent complex medium, containing both MnCl2 and MgCl2, resulted in an interaction with the hybrid membrane that caused a slight increase in C content from 49.83 to 50.5%, along with decreases in O and N content from 39.03 to 34.13% and from 9.19 to 8.08%, respectively, compared to the hybrid membrane after the incubation in the MnCl2 complex medium. The increase in C content suggests that Mn2+ and Mg2+ ions interacted with the membrane in a way that reorganized or stabilized C-rich areas on its surface. This interaction likely enhanced the prominence of C-containing regions while reducing the relative contribution of O and N groups.

The final medium used, containing MgCl2, led to a decrease in carbon content from 50.5 to 49.99% and an increase in oxygen content from 34.13 to 37.09%, as well as an increase in nitrogen content from 8.08 to 10.15% on the membrane, compared to the previously mentioned medium. These changes may correspond to ssDNA adsorbing onto the membrane through Mg2+-mediated interactions, where its nucleobases, which are rich in nitrogen, become more prominent on the membrane surface. This results in an overall increase in nitrogen content, as reflected in the XPS survey measurements.

Table 1 Atomic percentages before and after SWCNTs/GA-NC hybrid membrane incubation in ionic complex media containing Mn2+, Mg2+/Mn2+, and Mg2+.

To further understand the interactions, Fig. 5 presents the high-resolution XPS spectra of C 1s, Mn 2p, and Mg 1s for the chemical composition of the SWCNTs/GA-NC hybrid membrane before (Fig. 5A) and after incubation in complex media containing ssDNA and MnCl2 (Fig. 5B), MnCl2/MgCl2 (Fig. 5C), or MgCl2 (Fig. 5D). For the SWCNTs/GA-NC hybrid membrane prior to incubation (Fig. 5A), the C 1s spectrum was deconvoluted into four peaks at binding energies of 284.87 eV, 285.6 eV, 287.85 eV, and 288.95 eV, corresponding to C = C, C–C, C–O, and C–ONO2/C = O bonds, which are specific to the NC, SWCNTs, and GA42,43,44. Additionally, before incubation, the atomic percentages in Table 2 indicate that the membrane has a higher proportion of C = C and C–C bonds, while C–O and C–ONO2/C = O bonds are present in lower amounts.

After incubation of hybrid membrane in the complex medium containing MnCl2 (Fig. 5B), significant changes in peak positions, intensities, and areas were observed, indicating chemical modifications. The C1s spectrum displays three peaks at 285.49 eV, 287.92 eV, and 289.22 eV, corresponding to C = C, C–O, and C–O–NO2/C = O bonds, respectively. The higher intensity of C = C, C–O, and C–O–NO2/C = O bonds, suggests that interaction with the ionic complex medium has chemically altered the carbon structure, modifying their binding energies. The shift of C = C single bonds to higher energy, causing their band to overlap with that of the C-C bond, makes the C = C signal more prominent in the XPS spectrum. Moreover, the increase in the binding energy of C–O from 287.85 eV to 288.95 eV and of C–O–NO2/C = O from 288.95 eV to 289.22 eV shows that the interaction with the medium leads to a shift in the electronic environment. This likely corresponds to the adsorption of ssDNA, with its specific bond signals, as reported by Graf et al.45, potentially overlapping with those of the hybrid membrane due to the presence of aromatic rings, oxygen, and nitrogen bases in ssDNA. The overlap signal is further evidenced by increased atomic percentage of C = C, C–O, and C–O–NO2/C = O (Table 2), compared to the hybrid membrane before incubation.

Fig. 5
figure 5

High resolution C1s XPS spectra of SWCNTs/GA-NC hybrid membrane before (A) and after incubation in ionic complex media containing MnCl2 (B), MnCl2/MgCl2 (C), or MgCl2 (D). High resolution Mn 2p XPS spectra of SWCNTs/GA-NC hybrid membrane after incubation in ionic media containing MnCl2 (E) and MnCl2/MgCl2 (F). High resolution Mg1s XPS spectra of SWCNTs/GA-NC hybrid membrane after incubation in ionic media containing MnCl2/MgCl2 (G) or MgCl2 (H).

Table 2 Atomic percentages of the chemical bonds at the surface of the SWCNTs/GA-NC hybrid membrane before and after incubation in ionic complex media containing Mn2+, Mg2+/Mn2+, and Mg2+.

The hybrid membrane incubated in complex medium containing both MnCl2 and MgCl2 (Fig. 5C), exhibits four peaks at 284.85 eV, 285.76 eV, 287.59 eV, and 288.74 eV, corresponding to C = C, C–C, C–O, and C–ONO2/C = O bonds, respectively. For the C = C bonds, the peaks shifted to a lower binding energy, from 285.49 eV to 284.85 eV, as well as the C–O and C–ONO2/C = O bonds which also shifted to lower binding energies, from 287.92 eV to 287.59 eV and from 289.22 eV to 288.74 eV, respectively, compared to the hybrid membrane incubated in MnCl2 complex medium. The lower binding energy for the C = C, C–O and C–ONO2/C = O bonds suggests increased electron density due to electron donation or reduced oxidation states from these ions. These observations are in strong agreement with the atomic percentage corresponding to the aforementioned bonds (Table 2). In comparison to the MnCl2 complex medium, which exhibited higher atomic percentage for C = C and C–O bonds, the presence of both Mn2+ and Mg2+ ions results in a notable shift in the relative content, with the combined complex medium showing lower atomic percentage for C = C and C–O. Moreover, a significantly higher atomic percentage for C–ONO2/C = O is observed, which may be consistent with bonds characteristic of ssDNA. This suggests that the interaction of the hybrid membrane with the chemical environment formed by both Mn2+ and Mg2+ ions in the complex medium may lead to increased adsorption of ssDNA compared with the previous medium used, as the distinctive bonds signals of nitrogen bases of ssDNA are reflected in the enhanced C–ONO2/C = O peak area.

The hybrid membrane incubated in MgCl2-containing complex medium (Fig. 5D), exhibited four peaks at 285.11 eV, 286.05 eV, 287.66 eV, and 288.88 eV, corresponding to C = C, C–N, C–O, and C–ONO2/C = O, respectively. The emergence of a new C–N peak at 286.05 eV is accompanied by an increase in nitrogen content to 10.15%, as revealed by the chemical composition survey of the membrane, compared to the other media. This suggests that Mg2+ ions are particularly effective at coordinating with nitrogen-containing groups and promoting the formation of C–N bonds by activating nearby carbon atoms. The presence of C–N bond could indicate a reorganization of the surface chemistry, with Mg2+ ions facilitating the interaction with nitrogen-containing species of ssDNA. These findings align closely with the survey of the atomic percentage of the C = C, C–N, C–O, and C–ONO2/C = O chemical bonds listed in Table 2. A notable reduction in the C = C bond content was observed, along with a shift in the binding energy from 285.76 eV to 285.11 eV, when compared to the previous membrane incubated in the complex medium containing both MnCl2 and MgCl2. As the C = C bond content decreased, there was a corresponding increase in the C–ONO2/C = O bonds. This suggests that the interaction with the MgCl2 complex medium adsorbed a significantly higher amount of ssDNA, as evidenced by the aforementioned emergence of the C–N bond, with an atomic percentage of 28.16% indicates a high content of these bonds.

Figure 5E-H presents high-resolution Mn 2p and Mg 1s XPS spectra of the SWCNTs/GA-NC hybrid membrane after incubation in ionic complex media, with Mn 2p spectra corresponding to media containing MnCl2 (Fig. 5E) and MnCl2/MgCl2 mixture (Fig. 5F), and Mg 1s spectra corresponding to medium containing MnCl2/MgCl2 mixture (Fig. 5G) and MgCl2 (Fig. 5H). In Fig. 5 (E and F), the Mn 2p spectra exhibit three distinct peaks corresponding to Mn 2p3/2, Mn 2p1/2, and a satellite peak, which arises from shake-up transitions and reflects the presence of Mn2+ on the surface46. For the hybrid membrane incubated in MnCl2 complex media (Fig. 5E), the binding energies of these peaks are recorded at 643.28 eV, 654.51 eV, and 646.34 eV, respectively. Upon incubating the hybrid membrane in the MnCl2/MgCl2 complex media (Fig. 5F), these peaks shift to lower binding energies: 642.82 eV, 654.14 eV, and 646.98 eV, respectively. The decrease in binding energy of the Mn 2p3/2, and Mn 2p1/2 peaks suggests an increase in electron density around Mn, likely due to interactions with oxygen or nitrogen groups from the ssDNA structure. The presence of Mg2+ ions may further enhance this effect by shielding Mn ions and stabilizing them in a lower oxidation state, such as Mn2+, which is characterized by lower binding energies. Additionally, the increased binding energy of the satellite band under the MnCl2/MgCl2 condition indicates stronger electronic interactions and possibly enhanced stabilization of Mn ions within the hybrid membrane.

The Mg 1s spectra in Fig. 5 (G and H) reveal a single peak corresponding to Mg–O bonding interactions47. When the membrane is incubated in MnCl2/MgCl2 complex media (Fig. 5G), the Mg 1s binding energy is observed at 1304.87 eV, which is slightly higher than the binding energy of 1304.70 eV after the incubation of the membrane in MgCl2 complex media (Fig. 5H). Despite this slight shift in energy, the Mg 1s peak demonstrates stable binding energy across the two conditions, indicating that Mg2+ ions maintain consistent interactions with the oxygen-containing groups of the membrane, regardless of the presence of Mn ions.

From the three media used, the MgCl2 complex medium has emerged as the most effective in facilitating the interaction with ssDNA. The results demonstrate a significant increase in nitrogen content, reaching 10.15%, along with the emergence of C–N bonds, which are indicative of nitrogen-rich ssDNA nucleobases interacting with the hybrid membrane. Additionally, the MgCl2 medium facilitated the formation of more prominent C–ONO2/C = O peaks, further highlighted by the Mg 1s spectra, which demonstrate consistent Mg–O bonding interactions across different conditions, indicating the strong adsorption and stabilization of ssDNA on the membrane. These findings are well-correlated with the Raman results, which consistently indicate enhanced molecular interactions and structural stabilization of ssDNA within the hybrid membrane under the influence of Mg2+ ions.

Morphological analysis

Morphological analyses were performed on NC and SWCNTs/GA-NC membranes before and after the adsorption experiments. The SEM micrograph in Fig. 6A shows the NC membrane prior to the experiment, characterized by a sponge-like structure and surface irregularities with interconnected open pores and high porosity.

The morphology of the NC membrane changed after the incubation procedure in the utilized media (FAM-ssDNA, αMEM, BSA, SDS, and different ionic species, i.e., MgCl2, MnCl2/MgCl2, or MnCl2), as can be observed in Fig. 6B. Specifically, two distinct differences in morphology have been identified; first, the surface of the NC membrane has flattened, suggesting moderate stability in aqueous media, and second, a noticeable alteration in the shape and dimension of the pores, indicating less uniformity and larger pore diameters. This change suggests that the ionic species in the media might have led to alterations of the membrane’s surface, potentially collapsing some pores and causing the membrane to swell.

Fig. 6
figure 6

SEM images of the NC membrane morphology (A) prior and (B) following the experiments conducted in a complex media containing ionic species. The SWCNTs/GA-NC membrane is depicted (C) prior and (D) following the experiments in the previously described media. All images are presented with a 4 µm scale bar, while Figures C’ and D’ display a 1 μm scale bar.

Figure 6C indicates that the membrane was successfully covered with SWCNT and GA, displaying a smoother surface compared to the control. The partially visible morphology of the pristine NC membrane suggests that the surface is coated with a thin network of SWCNTs and GA layers. Moreover, the topography of the hybrid membrane, as shown in Fig. 6C’, reveals that the SWCNTs exhibit a homogeneous distribution without agglomeration across the surface, and a network-like structure, which may help preserve the structural integrity of the underlying hybrid membrane.

Following the use of the hybrid membrane in the adsorption media (Fig. 6D), the morphology of the SWCNTs/GA-NC remains well-preserved, indicating that both the structural integrity and stability of the membrane are maintained. While the surface still maintains its general structure, larger particle aggregates are observed, likely due to interactions between the SWCNTs/GA and the ionic species in the media. At higher magnification, as shown in Fig. 6D’, the SWCNTs/GA-NC hybrid membrane maintains surface features, despite some visible impurities from the used media. These observations indicate that the modification with SWCNTs and GA significantly enhances the resilience and functionality of the hybrid membrane under challenging conditions.

Wettability characteristics

Figure 7 illustrates the wettability properties of the NC and SWCNTs/GA-NC membranes, as determined by water contact angle measurements, showing distinct surface wettability characteristics for each membrane. According to Fig. 7A, the surface of the NC membrane exhibits water contact angles of approximately 26.7°, indicating a hydrophilic surface, and thus water droplet spreads across the membrane. This strong affinity for water may be attributed to the presence of polar NO2 groups (through dipole-dipole interactions with H2O molecules), and to the porous structure of the membrane, which increases the surface area.

In contrast, the water contact angles of the SWCNTs/GA-NC hybrid membrane, as shown in Fig. 7B, are measured at around 110°, indicating a hydrophobic surface. This hydrophobicity can be attributed to the nonpolar and highly structured surface characteristics of the SWCNTs and GA, resulting in water repellency. The observed water-repellent behavior shows that the surface modification has changed the membrane’s interaction with water, confirming the successful modification of the NC membrane with SWCNTs and GA.

Fig. 7
figure 7

Water contact angles of (A) NC and (B) SWCNTs/GA-NC hybrid membrane.

Adsorption, detection, and extraction of oligonucleotides in various ionic complex solutions utilizing NC and SWCNTs/GA-NC hybrid membranes

The results presented in Fig. 8 illustrate the fluorescence intensity of the extracted ssDNA after incubating the NC and SWCNTs/GA-NC membranes for 60 min in complex solutions, with free of ionic compounds as a control and with various ionic compounds: MnCl2, MgCl2, or a combination of both. As shown in Fig. 8, the fluorescence intensity of the NC membrane remained consistent across all ionic complex solutions, with values ranging from 245 r.f.u. for MnCl2 to 250 r.f.u. for the MnCl2/MgCl2 mixture and 265 r.f.u. for MgCl2, showing no significant differences compared to the control. In contrast, the SWCNTs/GA-NC hybrid membrane demonstrated a higher binding affinity even in the complex solution without ionic compounds, with fluorescence intensity values of 335 r.f.u. Moreover, after incubation in ionic complex solutions, it can be observed that the various divalent ions modulate the interaction of ssDNA with the SWCNTs/GA-NC hybrid membrane, with fluorescence intensity values measured at 420 r.f.u. when incubated with MnCl2, 510 r.f.u. after incubation in the MnCl2/MgCl2 solution, and 670 r.f.u. when incubated with MgCl2. Although the hybrid membrane exhibits a lower binding affinity in the MnCl2 solution compared to the MgCl2 solution, it still demonstrates a consistently stronger interaction with ssDNA than the NC membrane overall.

Fig. 8
figure 8

The fluorescence intensity (r.f.u.) of SWCNTs/GA-NC and NC membranes in complex solutions: without ionic compounds (control) and with different ionic compounds, including MnCl2, MnCl2/MgCl2, and MgCl2.

The synergistic effect between SWCNTs and GA may enhance the hybrid membrane interaction with divalent cations, such as Mn2+ and Mg2+, by creating specific binding sites through coordination chemistry. Carbon atoms on the SWCNT surface oscillate around their equilibrium positions, allowing flexible adaptation for interactions with adjacent molecules. The inherent flexibility of SWCNTs is key to their ability to interact with larger and more complex biomolecules like ssDNA. The oscillatory behavior facilitates the accommodation of ssDNA by providing a dynamic and adaptable binding surface, which is important for stable adsorption33,48,49. Physisorption, characterized by weak van der Waals forces, is the primary mode of interaction between ssDNA and SWCNTs. Unlike chemisorption, which involves the formation of strong chemical bonds, physisorption relies on the physical attraction between the π-electron systems of the SWCNTs and the nucleobases of ssDNA48,50,51,52,53,54,55. Moreover, density functional theory (DFT) calculations reported in the literature indicate that the binding energies of Mn2+ and Mg2+ cations differ significantly when binding to SWCNTs. Specifically, binding energies range from 4 eV for Mn2+ to 16 eV for Mg2+, which explains the observed differences in binding affinity of the two cations to the surface modified with SWCNTs. As a result, the higher binding energy of Mg2+ indicates a greater affinity for more stable adsorption of ssDNA compared to Mn2+, highlighting the importance of cation type in enhancing ssDNA interactions with the SWCNT surface56,57,58,59,60.

On the other hand, the amine groups on the GA surface can align themselves in response to interactions with Mn2+ and Mg2+, and due to their high flexibility, they form multiple coordination bonds and hydrogen bonds with biomolecules such as ssDNA. The PEG chains and amine groups undergo significant conformational changes, adapting to the binding sites created by the cations. The dynamic behavior of GA enhances the efficient binding and stabilization of biomolecules by optimizing the orientation of the amine groups61,62. The interaction between amine functionalities and divalent cations, such as Mn2+ and Mg2+, involves the formation of coordinate covalent bonds, wherein the nitrogen atoms of amine groups act as Lewis bases by donating lone pairs of electrons. Conversely, the divalent cations function as Lewis’s acids, accepting these electron pairs to form stable coordinate covalent bonds. This bonding process significantly alters the electronic environment of GA, modifying its charge distribution and electronic density, while the coordination of Mn2+ and Mg2+ with amine groups enhances molecular interactions, thereby improving overall binding efficiency52,63,64. Thus, the synergy between SWCNTs and GA arises from their complementary properties, with SWCNTs providing a flexible, adaptive surface for strong interactions with ssDNA, while GA offers dynamic, chemically active sites that facilitate multiple bonding interactions. Another important factor in both adsorption and desorption processes is the pH of the complex solution, as it affects the electrostatic interactions between ssDNA and the hybrid membrane. In a Tris-HCl buffer with pH 8, the mildly alkaline conditions lead to partial deprotonation of both ssDNA and the NH2 groups on the hybrid membrane surface. This results in the exposure of negatively charged phosphate groups on the ssDNA and positively charged NH2 groups on the hybrid membrane, which enhances the electrostatic attraction between the two. Consequently, this increased interaction improves the adsorption of ssDNA onto the hybrid membrane. In contrast, a pH of 7 in the Tris-HCl buffer introduces neutral conditions that cause partial protonation of the phosphate groups on the ssDNA and the NH2 groups on the hybrid membrane. This protonation reduces the net charge on both molecules, weakening the electrostatic interactions. As a result, ssDNA is more likely to detach or desorb from the hybrid membrane, leading to lower adsorption efficiency compared to the slightly alkaline environment65,66,67,68.

Multiple components of the complex medium, including SDS, BSA, and certain molecules from αMEM, enhance the attachment of ssDNA to the membrane while protecting it from non-specific binding. Glucose and other carbohydrates in αMEM increase osmotic pressure, causing water to move from the surrounding medium into the membrane, which brings ssDNA closer to the SWCNTs/GA-NC surface69. This movement improves contact between ssDNA and the membrane, leading to better adsorption. BSA prevents non-specific binding of negatively charged ssDNA molecules by covering surfaces and reducing repulsive electrostatic interactions with the SWCNTs/GA-NC membrane, thus facilitating more effective adhesion70,71. Additionally, SDS acts as a surfactant that inactivates nucleases, preserving the integrity of the ssDNA, and helps control non-specific adsorption by creating a more controlled environment on the SWCNTs/GA-NC hybrid membrane surface72,73.

Therefore, the immobilization of ssDNA on the SWCNTs/GA-NC membrane surface is dependent on several factors that demand precise control, such as the composition of the ionic complex solution, ionic valence, and pH, as highlighted by the events discussed.

Quantification and analysis of FAM-ssDNA desorption from NC and SWCNTs/GA-NC hybrid membranes

To quantify the amount of FAM-ssDNA desorbed from the surface of pristine and hybrid membranes, Eq. (1) was used to convert the fluorescence data into units of mass, with the obtained results shown in Table 3.

$$\:{M}_{f}\:\:\:=\frac{{F}_{f}{*M}_{i}}{{F}_{i}}$$

where, Fi represents the fluorescence of the FAM–ssDNA sequence in the adsorption solution, with a value of approximately 3800 r.f.u. Mi stands for the initial mass of FAM–ssDNA in the adsorption solution, which is around 4420 pg, calculated using the ‘DNA molecular weight and conversion’ tool from ThermoFisher74. Ff denotes the final fluorescence in r.f.u. after the desorption process, while Mf refers to the final mass of the released FAM–ssDNA expressed in pg.

According to Table 3, the NC membrane desorbed approximately 220 pg in ion-free complex solutions. The addition of divalent ions resulted in a slight increase, though not significant, with values ranging from 290 pg for MnCl2 to 300 pg for the MnCl2/MgCl2 mixture, and 320 pg for MgCl2. In contrast, the SWCNTs/GA-NC hybrid membrane exhibits more variability, even in the complex solution without ionic species, with a desorbed mass of approximatively 390 pg. The amount increased when incubated in complex medium with divalent cations, reaching around 490 pg in the MnCl2 solution, followed by 590 pg after incubation in the MnCl2/MgCl2 solution, and peaking at 780 pg when incubated in the MgCl2 complex medium. The observed variability, including a slightly higher standard deviation, is likely due to occasional formation of FAM-ssDNA aggregates in solution.

The differences observed among the ionic complex solutions can be attributed to the distinct properties of the Mg2+ and Mn2+ ions. Specifically, Mg2+ ions are smaller, with a radius of 0.72 Å, and possess a significantly higher charge density of 1.66, compared to Mn2+ ions, which have a radius of 0.83 Å and a lower charge density of 0.93. The higher charge density of Mg2+ ions may reverse the unfavorable orientation of ssDNA on SWCNTs/GA surfaces, promoting better binding by reducing repulsive forces and encouraging a more favorable and stable interaction75,76. This interaction reduces the repulsive Coulomb forces between nucleotides, leading to a phenomenon known as charge screening. The screening effect diminishes the electrostatic repulsion and allows the ssDNA strand to contract. This contraction effectively decreases the ssDNA’s overall length and enables it to adopt a more compact and condensed structure. As a result, the ssDNA becomes more compliant to binding with hydrophobic or electrostatically active surfaces, such as those provided by SWCNTs-GA/NC hybrid membrane77,78,79.

Table 3 The mass (in pg) of ssDNA and the standard deviation (s.d.) recorded following a 60-minute incubation of NC and SWCNTs/GA-NC membranes in a complex solution containing different ionic species (MnCl2, MnCl2/MgCl2, or MgCl2).

In contrast to the GA-NC hybrid membrane used in our earlier research, which extracted approximately 610 pg of ssDNA when incubated in the MgCl2 complex solutions16,17, the SWCNTs/GA-NC hybrid membrane in the current study demonstrates a higher extraction yield of over 780 pg of ssDNA. This improvement can be attributed to the addition of SWCNTs, which mitigate steric hindrance within GA-NC membranes. By alleviating the crowded environment created by the dense network of aminated groups, SWCNTs facilitate easier access for ssDNA molecules to active sites by creating nanoscale gaps between themselves and the GA matrix. This is particularly beneficial when ssDNA adopts a coiled or bulky conformation that complicates interactions with functional groups. These gaps effectively reduce the steric interference that ssDNA might encounter while accessing binding sites on the GA surface, promoting better diffusion and movement toward active sites.

Overall, modifying the NC membrane with SWCNTs and GA represents a promising approach that not only enhances binding efficiency by providing a more accessible and stable environment for ssDNA interactions, but also increases the available surface area for binding. This synergy of increased surface area, improved stability, and reduced steric hindrance ultimately contributes to the superior extraction yield observed with the SWCNTs/GA-NC membrane compared to the GA-NC membrane alone.

Although the SWCNTs/GA-NC hybrid membrane extracted a greater amount of ssDNA from complex media containing various ionic species, we acknowledge the need for further refinement. Continued research will be essential to enhance the efficiency and reliability of this extraction technique, to broaden its application in isolating diverse biomolecules from various samples, and to advance molecular biology and related fields that require precise biomolecule extraction.

Conclusions

In this study, we modified a NC membrane with SWCNTs and GA for the extraction of ssDNA from various ionic complex solutions. The modification of the hybrid membrane was confirmed both by structural and morphological analyses using FTIR, Raman spectroscopy, XPS, and SEM, as well as by wettability investigations. The FTIR spectra confirmed the successful modification of the NC membrane with SWCNTs and GA by revealing the characteristic absorption bands of all components. The Raman analysis shows structural and chemical changes in the SWCNTs/GA-NC hybrid membrane, mainly after incubation in the MgCl2-containing medium. These changes are consistent with higher ssDNA adsorption, as RBM shifts and defect formation enhance ssDNA-SWCNT interactions, highlighting the role of Mg2+ ions in facilitating ssDNA adsorption. The XPS analysis demonstrated that Mn2+ and Mg2+ ions influence the chemical composition and interactions of the SWCNTs/GA-NC hybrid membrane with the ssDNA. Upon exposure to divalent cation media, an increase in oxygen and nitrogen content is observed, with the highest nitrogen content (10.15%) occurring in the MgCl2 complex media, corresponding to a greater amount of ssDNA adsorbed on the membrane.

SEM images indicated the modification of the NC membrane with SWCNTs and GA, unveiling a smoother surface with a uniform distribution of these components. Post-adsorption experiments revealed that the structural integrity and stability of the membrane’s morphology are maintained, with only minor aggregation likely due to interactions with ionic species from complex media.

The wettability assay showed that the SWCNTs/GA-NC hybrid membrane exhibited significantly higher hydrophobicity, with water contact angles of approximately 110°, compared to the pristine NC membrane’s contact angle of around 27°, confirming the successful modification of the NC membrane with SWCNTs and GA.

Furthermore, the SWCNTs/GA-NC hybrid membrane was tested for ssDNA extraction efficiency in three ionic complex solutions: MnCl2, MnCl2/MgCl2, and MgCl2. The results indicate that the hybrid membrane consistently exhibited superior adsorption and desorption capabilities compared to the control NC membrane. Specifically, the NC membrane maintained relatively stable fluorescence intensities (245–265 r.f.u.) and ssDNA desorption levels (290–320 pg) across all solutions. In contrast, the SWCNTs/GA-NC membrane demonstrated both increased binding affinity and variability, with fluorescence intensities rising from 420 r.f.u. in MnCl2 to 510 r.f.u. in the MnCl2/MgCl2 mixture, and peaking at 670 r.f.u. in MgCl2. Correspondingly, ssDNA desorption levels increased from 490 pg for MnCl2 to 590 pg for the MnCl2/MgCl2 solution, reaching 780 pg for MgCl2, underscoring the SWCNTs/GA-NC hybrid membrane’s superior extraction efficiency and selective binding capacity, particularly in MgCl2 environments, and highlighting its potential for targeted ssDNA isolation across varied ionic conditions.

Therefore, our findings suggest that modifying the NC membrane with SWCNTs/GA greatly enhances its ability to immobilize oligo-DNA on its surface. This notable improvement in oligo-DNA immobilization highlights the advantages of using SWCNTs/GA for membrane modifications, which could have significant implications for applications that require effective immobilization and extraction of oligo-DNA.