Chemically tuned radiolabeling and membrane affinity in Drug-Functionalized carbon nanotube nanocarriers Authors: Sedigheh Abdollahi, Heidar Raissi, Farzaneh Farzad Published: 2025 ABSTRACT AbstractRadiolabeled nanocarriers that can simultaneously deliver therapeutic and diagnostic agents are promising for targeted cancer therapy. In this study, we employed molecular dynamics and metadynamics simulations to investigate how the chemical identity of drug molecules influences the behavior of technetium-99 m (⁹⁹ᵐTc)-loaded carbon nanotube (CNT) nanocarriers at the atomic scale. Four clinically relevant drugs—methotrexate, diclofenac, ketotifen, and piroxicam—were conjugated to a single-walled CNT (SWCNT) via a bifunctional chelating group. Our results revealed that drug structure significantly affects radionuclide binding strength, molecular diffusion, and membrane affinity. The methotrexate-functionalized nanocarrier exhibited the most stable technetium coordination (− 825.80 kJ/mol), the lowest diffusion coefficient (0.0070 × 10⁻⁵ cm²/s), and the strongest interaction with a model phospholipid membrane (− 1376.02 kJ/mol). Free energy landscapes reconstructed from metadynamics simulations further confirmed the high thermodynamic stability of this configuration. These findings suggest that rational drug selection can effectively tune the physicochemical behavior of CNT-based radiopharmaceuticals, providing a mechanistic foundation for designing multifunctional nanocarriers with improved performance for nuclear imaging and site-specific drug delivery.Keywords: Molecular dynamics, Radiopharmaceutical nanocarriers, Brachytherapy, Technetium-99m, Carbon nanotubes, Targeted drug delivery, MetadynamicsSubject terms: Cancer, Chemistry, Drug discovery Introduction Cancer remains one of the foremost global health challenges, with its incidence and mortality rates continuing to rise due to factors such as population aging, environmental pollutants, and lifestyle-related risks. Conventional treatment modalities—including surgery, chemotherapy, and external beam radiotherapy (EBRT)—often suffer from limited tumor selectivity and undesirable systemic side effects, which hinder therapeutic efficacy and patient quality of life. These shortcomings have underscored the urgent need for more precise and less invasive treatment platforms1. Among modern oncological therapies, brachytherapy (BT)—the localized delivery of high-dose radiation via internal radioactive sources—has demonstrated remarkable clinical success, especially in treating prostate, cervical, and breast cancers2–4. Compared to EBRT, brachytherapy offers superior dose conformity and spares surrounding healthy tissues, improving both efficacy and tolerability2,4–7. However, traditional seed-based BT methods face logistical and biological limitations, such as invasive placement, seed migration, and non-uniform radiation distribution, which restrict their broader clinical deployment8. Recent advances in nanotechnology have opened new avenues for overcoming these limitations. Nanobrachytherapy, an emerging approach that utilizes radioactive nanoparticles instead of solid seeds, offers enhanced spatial control over dose delivery and greater adaptability to tumor microenvironments7. Carbon nanotubes (CNTs), particularly single-walled carbon nanotubes (SWCNTs), have garnered significant interest as potential nanocarriers for drug delivery and imaging applications due to their high surface area, ability to be functionalized with a wide range of chemical groups, and unique structural properties9. However, pristine CNTs are well-known to exhibit cytotoxicity and genotoxicity, which can lead to protein misfolding, DNA disruption, and other cellular disturbances, unless they are appropriately functionalized10. Many studies have highlighted the potential toxic effects of non-functionalized CNTs, emphasizing the need for surface modifications to mitigate these adverse effects. Covalent or non-covalent functionalization of CNTs with various chemical groups can effectively reduce their toxicity, enhance solubility, and improve their biocompatibility, while also enabling selective targeting of biomolecules11. For example, functionalized CNTs have shown promise in drug delivery applications, particularly for chemotherapeutic agents and radiopharmaceuticals, by improving cellular uptake and reducing toxicity both in vitro and in vivo12. Recent experimental and computational studies have provided valuable insights into how functionalization and surface chemistry influence the behavior of CNT-based nanocarriers, including their stability, membrane affinity, and biodistribution13. These studies underscore the critical role of rational design in developing CNT systems that combine therapeutic efficacy with safety. This work is motivated by the need to further explore how the chemical identity of drugs influences the behavior of radiolabeled, functionalized SWCNTs at the molecular level, and aims to contribute to the optimization of CNT-based drug delivery systems for targeted therapy. Recent advances in molecular modeling and simulations have provided a powerful tool to investigate the interactions between CNT-based nanocarriers and drugs, particularly in the context of radiopharmaceuticals such as technetium-99 m. Recent experimental work has demonstrated that carbon-based nanocarriers—including functionalized single-walled and multi-walled carbon nanotubes and mesoporous carbon nanoparticles—can be successfully radiolabeled with technetium-99 m and traced in vitro and in vivo, providing experimental validation of their potential as imaging agents and delivery vehicles14. Comprehensive reviews of radiolabeled nanomaterials summarize the rapid growth of this field and discuss both clinical prospects and biodistribution/toxicity concerns relevant to ⁹⁹ᵐTc-based systems15. On the computational side, atomistic molecular dynamics studies have been widely employed to elucidate drug adsorption and conformational behavior on CNT surfaces and to probe CNT–membrane interactions; these modeling studies complement experimental biodistribution and radiolabeling work by offering molecular-level mechanistic insight16. A central challenge in radiopharmaceutical nanocarrier design lies in achieving stable coordination between the nanomaterial and the radionuclide, without compromising the biological function or integrity of the carrier. The use of bifunctional chelators, which can securely tether radioisotopes like technetium-99 m (⁹⁹ᵐTc) to nanocarrier scaffolds while maintaining targeting capabilities, has emerged as a powerful strategy17. Yet, understanding how such complexes behave in dynamic biological environments—and how the choice of therapeutic or diagnostic drug molecules affects their structural stability, radionuclide retention, and membrane affinity—remains an open question. This study addresses this critical gap by conducting a comprehensive molecular dynamics investigation into the behavior of SWCNT-chelator-drug complexes coordinated with ⁹⁹ᵐTc. Four clinically relevant drugs—methotrexate, diclofenac, ketotifen, and piroxicam—were evaluated for their impact on the physicochemical interactions of the resulting nanocomplexes with ⁹⁹ᵐTc ions and biological membranes. We employed molecular dynamics (MD) and metadynamics simulations to explore several key aspects of system performance, including: Root Mean Square Deviation (RMSD) and structural stability under physiological conditions, Root Mean Square Deviation (RMSD) and structural stability under physiological conditions, Radial Distribution Functions (RDFs) to quantify technetium localization around the nanocarrier surface, Radial Distribution Functions (RDFs) to quantify technetium localization around the nanocarrier surface, Mean Squared Displacement (MSD) and diffusion coefficient analysis to assess radionuclide mobility, Mean Squared Displacement (MSD) and diffusion coefficient analysis to assess radionuclide mobility, Distance-based metrics to track ⁹⁹ᵐTc binding behavior over time, Distance-based metrics to track ⁹⁹ᵐTc binding behavior over time, Membrane interaction energies (van der Waals and electrostatic) to gauge adsorption efficiency, Membrane interaction energies (van der Waals and electrostatic) to gauge adsorption efficiency, And free energy landscapes to map stable binding states and identify thermodynamically favorable configurations. And free energy landscapes to map stable binding states and identify thermodynamically favorable configurations. Our results reveal a strong drug-dependent modulation of ⁹⁹ᵐTc binding behavior. Notably, methotrexate-functionalized SWCNTs demonstrated the highest adsorption energies, lowest diffusion coefficients, and most stable interactions with lipid membranes—underscoring their potential for robust and localized radiopharmaceutical delivery. In contrast, ketotifen-functionalized systems showed more transient interactions and higher technetium mobility, pointing to a possible application in responsive imaging or short-term diagnostics. To the best of our knowledge, this is the first comparative study to systematically connect the chemical identity of the drug moiety to the spatial confinement, mobility, and thermodynamic affinity of a radiolabeled SWCNT complex in a biologically relevant setting. These findings offer valuable design insights for the development of next-generation nanocarriers for targeted nanobrachytherapy, where tuning the chemical microenvironment of the carrier system allows for precise control over therapeutic outcomes. By integrating structural, energetic, and dynamic criteria into a unified framework, this work paves the way for rational design strategies in radiopharmaceutical nanotechnology. The proposed approach not only improves our understanding of drug-modulated radionuclide behavior at the molecular level but also contributes to the broader goal of enhancing the safety, selectivity, and clinical efficacy of nanoparticle-assisted cancer therapies. This study employed well-established molecular dynamics simulations to investigate the behavior of technetium-loaded carbon nanotube drug delivery systems. MD simulations, particularly when using the GROMACS software with CHARMM36 force fields, have been validated in a broad array of studies for accurately simulating the molecular interactions of nanomaterials with biological entities18. Our approach follows rigorous computational protocols that have been used for similar systems and have consistently produced reliable, reproducible results. The results observed in this study are in line with other studies examining nanomaterial-drug interactions, with well-defined methodologies and clearly defined parameters. Computational methods MD simulation In this study, a single-walled carbon nanotube with an armchair configuration (10,10) with a length of 6 nm and a diameter of approximately 1.5 nm was modeled. Folic acid was covalently attached to the nanotube to enable the selective targeting of cancer cells. In the next step, four simulation boxes were created with four nonsteroidal anti-inflammatory drugs (NSAIDs): Methotrexate, diclofenac, ketotifen, and piroxicam, respectively, linked to a bifunctional chelator via an ester bond, which facilitates effective association and drug release at the disease site. Radiolabeling was achieved using technetium-99m (99mTc)17,19,20 a gamma-emitting isotope commonly used for molecular imaging. Each simulation system consists of a single-walled carbon nanotube functionalized with three bifunctional chelators, which were covalently attached to the nanotube surface. This number was selected based on preliminary modeling tests, which indicated that three chelators provide a realistic surface coverage that allows for stable technetium coordination and drug conjugation while maintaining the structural integrity and flexibility of the nanotube. The use of three chelators also helps minimize steric hindrance, ensuring a uniform spatial distribution of the functional groups along the nanotube, which facilitates a meaningful comparison between systems with different drugs. In all systems, three drug molecules (methotrexate, diclofenac, ketotifen, or piroxicam) were conjugated to a bifunctional chelator via an ester bond. The resulting drug-functionalized complexes were used in membrane adsorption simulations. Additionally, nine technetium-99 m particles were placed approximately 1.8 nm from the drug molecules to facilitate tracking of the nanotube-drug complex within the simulated environment. Figures 1 and 2 demonstrate the chemical structure of the CNT, folic acid, 99mTc, the methotrexate, diclofenac, ketotifen, and piroxicam molecules, and the chelator structure. The structure of (A) chelator attached to carbon-nanotube, (B) drugs. The structure of (A) SWCNT, (B) 99mTc, (C) SWCNT-chelator-drug-99mTc complex from the side view, (D) Chelator, (E) Folic acid, (F) SWCNT-chelator-diclofenac-99mTc complex from the front view, (G) SWCNT-chelator-ketotifen-99mTc complex from the front view, (H) SWCNT-chelator-methotrexate-99mTc complex from the front view, (I) SWCNT-chelator-piroxicam-99mTc complex from the front view. The molecular dynamics simulations were performed using GROMACS software (version 2022.5)21, employing the CHARMM3622 force field due to its effectiveness in accurately modeling the interactions within biologically relevant systems, including proteins, lipids, and carbon-based nanostructures. For the simulations involving the carbon nanotubes, we used the CHARMM36 force field, which is well-suited for modeling carbon-based nanomaterials. Specifically, the carbon-carbon bonds within the CNT structure were treated as flexible, with the CNTs allowed to move freely during the simulations. This flexibility refers to the vibrational freedom of the bonds, allowing the nanotube to adopt a range of conformations during the simulation, while the atoms themselves remain fixed in their respective positions along the graphitic network. This reflects the realistic behavior of CNTs in solution and in biological environments. In this study, the CNTs were modeled with flexibility in their bond vibrations, but their overall conformation was restrained to prevent excessive deformation. Specifically, bond length constraints were applied to the CNTs to preserve their structural integrity, while still allowing for bond vibrational freedom. These constraints were necessary to maintain a realistic conformation, as fully free-floating CNTs would have undergone significant deformation. The CNTs were thus allowed to adjust their conformation to the surrounding environment within the limits set by these restraints. The carbon nanotubes were modeled using the CHARMM36 force field with the Table 1 parameters for bonded and non-bonded interactions. Table 1Parameters for CNT.Parameter typeBonded ParametersBond type: C-CBond length: 0.142 nm (default CNT bond length)Bond angle: 120°Non-bonded ParametersAtom type: CA (aromatic carbon)σ (C-C) = 0.3555 nmε (C-C) = 0.29288 kJ/mol Parameters for CNT. However, we acknowledge that these parameters, while standard in CHARMM36, are not always suitable for CNT simulations, as discussed by Hummer et al. (Nature 2001)23. In CNT modeling, a more refined set of parameters is often used for better consistency with experimental data, especially in regard to the water transport properties of CNTs. Since ⁹⁹ᵐTc is not explicitly parameterized in the CHARMM36 force field, we derived the nonbonded parameters for ⁹⁹ᵐTc using the CHARMM General Force Field (CGenFF) protocol24. The Lennard–Jones parameters and partial charges for technetium-99m (⁹⁹ᵐTc) were obtained from the work of Cundari et al. (2000)25, and these parameters were integrated into the CHARMM36 framework to maintain consistency with the rest of the system. For completeness and reproducibility, we have listed the non-bonded parameters for the 99mTc system in Table 2. These parameters ensure consistency with the existing paper and facilitate reproducibility for future studies. Table 2Non-bonded parameters for 99mTc.Atom typeR (Å)ε (kcal/mol) Tc 1.490.241R is the van der Waals radius(Å) and is the energy parameter (kcal/mol). Non-bonded parameters for 99mTc. R is the van der Waals radius(Å) and is the energy parameter (kcal/mol). For the ligands (diclofenac, methotrexate, piroxicam, and ketotifen), we used parameters derived from the SwissParam database26, a well-established resource for parameterization of small organic molecules in molecular dynamics simulations. The chelator used in this study was based on the structure from Das et al. (2013)20, and its parameters were derived using the CHARMM36 force field to ensure compatibility with the CNT and other components in the system. A TIP3P water model was used to represent the biological environment. Na+ and Cl− ions are added at a concentration of 0.15 M to neutralize the system and mimic the biological environment, and the Particle-mesh Ewald was applied for electrostatics, with a cutoff radius of 1.4 nm for short-range interactions. The temperature was maintained at 310 K, mimicking physiological conditions, using the Nose-Hoover thermostat27, while pressure control was achieved at 1 bar with the Parrinello-Rahman algorithm28. In these systems, the LINCS algorithm is used to maintain all bonds at their equilibrium positions. The SWCNT-chelator-drug complex was positioned in the center of a 101010 (nm3) simulation box with periodic boundary conditions applied to ensure the continuity of the system. The system underwent an initial energy minimization step to eliminate any unfavorable contacts, followed by a simulation of a 105 ns run, capturing the behavior and dynamics of the nanotube and the radiopharmaceutical agents in a simulated biological environment. Data were collected every 2 fs to record changes in position, energy, and forces. VMD software was used for visualization, enabling detailed inspection of molecular interactions and aiding in the interpretation of simulation results. System Preparation membrane cells Since the drug delivery performance of carbon-based nanomaterials relies on their ability to diffuse to the cell membrane, the interaction of the SWCNT-chelator-drug-99mTc complex with the lipid bilayer is investigated. In this study, four SWCNT-chelator-drug-99mTc complex systems were placed on a lipid membrane to investigate their interaction and adsorption behavior. Each system consisted of a carbon nanotube attached to a bifunctional chelator activated with one of the following drugs: methotrexate, diclofenac, ketotifen, or piroxicam, and technetium particles that were complexed during the adsorption molecular dynamics simulation. The dimensions of this simulation box are considered 11.44 × 11.44 × 24 nm3. A phospholipid bilayer was used to model the cell membrane, replicating the biophysical environment to understand the nanotubes’ potential for membrane adsorption. In this study, we employed a phospholipid bilayer (POPC) model, which is widely used as a representation of mammalian cell membranes29. This model was chosen for its biological relevance, as it reflects the typical lipid composition of cell membranes. The membrane was constructed using the CHARMM-GUI Membrane Builder platform and consisted of 128 POPC molecules (64 per leaflet), fully hydrated with TIP3P water molecules, and supplemented with Na⁺ and Cl⁻ ions at a concentration of 0.15 M to simulate physiological conditions. Each functionalized complex system was initially positioned at a distance of ≈ 2.5 nm above the membrane surface to avoid initial contacts that could introduce bias. Each system was simulated for 200 ns to observe potential membrane adsorption events. We note that in real biological systems, the SWCNT-chelator-drug-⁹⁹ᵐTc complexes could adopt random orientations relative to the membrane. While our simulations used a parallel placement to systematically observe the adsorption process, previous studies (e.g., Gao et al.30; Yazdani et al.31) suggest that the initial orientation can affect the kinetics of approach and the specific contact points. However, the overall trends in adsorption strength, interaction energies, and membrane binding hierarchy are largely preserved across different orientations. This supports the generality of our findings and confirms that the observed drug-dependent interaction patterns are robust across more realistic orientations of the nanotube complexes. Each simulation system was solvated with approximately 23,000 TIP3P water molecules, corresponding to an average hydration level of ~ 35 water molecules per lipid, which is within the standard range recommended for maintaining the structural stability and realistic dynamics of lipid bilayers. This hydration level ensures appropriate lipid mobility, bilayer flexibility, and bilayer thickness while avoiding artificial packing of the lipids or excessive dilution effects that could interfere with the interactions between lipids and other molecules, such as drugs and nanoparticles. This level of hydration is critical to maintain the correct physical organization of the bilayer and ensure that the system’s dynamics are not compromised32–34. The system parameters for the membrane were derived from the CHARMM36 lipid force field, which has been extensively validated for lipid bilayer simulations32. The primary objective of our study was to examine the adsorption interactions between the CNT-drug complexes and the lipid bilayer, which is a crucial aspect of understanding how these complexes behave in biological environments. We focused on the interactions at the membrane interface, where the CNT complexes interact with the lipid head groups. The simulation did not aim to model the penetration or translocation of the CNTs through the membrane, as it has been shown that the process of CNT penetration is energetically costly (ACS Nano, 2010)35. Instead, our goal was to assess the adsorption affinity of the CNT complexes at the membrane surface and their stability at the interface, which plays a significant role in their potential use for targeted drug delivery. While studies using enhanced sampling methods, such as ACS Appl. Mater. Interfaces, 201836, and Journal of Molecular Liquids, 202337, have explored CNT penetration across membranes. Our approach focused on the initial interaction and adsorption of CNTs at the membrane surface. We believe that this aspect is critical for understanding the drug delivery potential of CNT-based systems, as strong membrane adsorption can enhance the retention and effectiveness of the drug carrier at the target site. Metadynamic Metadynamics simulations were implemented using the PLUMED (version 2.5.2)38 plugin in GROMACS software to investigate the free energy landscape of adsorption of SWCNT-chelator-drug-99mTc complex onto the membrane. Well-tempered metadynamics was used to reconstruct the free-energy surface (FES) of the CNT–chelator–drug–⁹⁹ᵐTc complexes during adsorption on the membrane. The collective variable (CV) was defined as the distance between the centers of mass (COM) of the complex and the COM of the lipid bilayer along the z-axis. This coordinate directly reflects the adsorption-desorption pathway and has been widely used in interfacial free-energy studies. To minimize potential distortions of intermediate states, we employed a well-tempered metadynamics scheme with a high bias factor (γ = 20) and a small Gaussian width (σ = 0.25 nm). These settings ensure slow bias deposition and allow the system to relax between bias updates, maintaining physically realistic contact configurations. The Gaussian bias was applied gradually every 500 ps with a small height of 1.0 kJ/mol, ensuring that the energy landscape was only gently perturbed, preserving the physical accuracy of the system dynamics. The resulting adsorption minima were fully consistent with the equilibrium configurations observed in unbiased MD, confirming that the enhanced sampling did not produce artificial states. Metadynamics has been successfully applied to study adsorption and interfacial thermodynamics of nanostructures and biomolecules in several published works, such as Sullivan et al. (2020)39, Fang et al. (2022)40, and Fiorin et al. (2023)41. These studies collectively demonstrate that metadynamics is an effective method for obtaining reliable free-energy profiles even when complete equilibrium sampling is not achievable with conventional MD simulations. This method is particularly useful in systems where free energy barriers hinder the sampling of rare events, allowing for a more thorough examination of the adsorption process in nanoscale carriers. To calculate the free energy surface, Gaussian hills, with an initial height of 1.0 kJ/mol and a width of 0.25 nm, were deposited at intervals of 500 ps to gradually bias the system along this collective variable, effectively driving carbon nanotubes-complex closer to the membrane surface and enabling observation of potential adsorption sites. Over time, the accumulation of these hills allows for a reliable approximation of the free energy landscape, capturing the potential wells and barriers relevant to the nanotube complex with membrane interaction. To assess the adsorption process, two distinct systems were chosen based on their respective adsorption energies from MD simulations: High-adsorption Energy System: Consisting of a CNT connected to a bifunctional chelator and the drug methotrexate, this system represents a strong technetium-binding case. High-adsorption Energy System: Consisting of a CNT connected to a bifunctional chelator and the drug methotrexate, this system represents a strong technetium-binding case. Low-adsorption Energy System: Featuring a CNT connected to a bifunctional chelator and the drug diclofenac, this configuration provides insight into lower-energy interactions between the technetium and the nanocarrier. Low-adsorption Energy System: Featuring a CNT connected to a bifunctional chelator and the drug diclofenac, this configuration provides insight into lower-energy interactions between the technetium and the nanocarrier. Results and discussion 99mTc adsorption on the SWCNT-chelator-drug complex In this section, all analyses—including total energy, root mean square deviation, diffusion coefficient, and radial distribution function —were conducted to examine the adsorption behavior of radioactive technetium atoms onto the complex in the absence of the lipid membrane. Adsorption of radioactive technetium on carrier surfaces The initial and final configurations of the radioactive technetium adsorption process in the SWCNT-chelator-drug complex are presented in Fig. 3. Representative snapshots corresponding to the interaction of SWCNT-chelator-drug complex with 99mTc at t=0 ns and t=105 ns. As depicted in this Figure, the initial configuration (0 ns) was set with radioactive technetium particles positioned approximately 18 Å from the drugs to reduce any bias introduced by their initial orientations. Throughout the simulation time, some technetium particles spontaneously migrate toward the carrier surface, establishing interactions, particularly with carbon and nitrogen groups. By the end of the simulation (105ns), these technetium particles are situated on the drugs, reflecting the formation of adsorption interactions. Energy analysis 99mTc adsorption on the SWCNT-chelator-drug complex To unravel the mechanisms underlying technetium binding and assess the performance of the functionalized nanotube systems, a detailed energy analysis was performed. This section focuses on the quantitative evaluation of non-bonded interaction energies—namely, electrostatic (Coulombic) and van der Waals (Lennard-Jones) forces—between the SWCNT-chelator-drug complexes and the radiotracer ⁹⁹ᵐTc. These energies serve as critical indicators of the thermodynamic favorability, stability, and binding affinity of the system over the simulation time. Figure 4A and B illustrate the time evolution of electrostatic and Lennard-Jones interaction energies, respectively, for the four functionalized SWCNT systems. A clear downward trend in energy—particularly evident in the methotrexate and piroxicam systems—signals increasingly favorable interactions with ⁹⁹ᵐTc over the course of the simulation. Among all systems, methotrexate-functionalized nanotubes exhibit the most pronounced decrease, reaching electrostatic energy values ​​below − 1300 kJ/mol and van der Waals contributions nearing − 60 kJ/mol. This trend underscores the enhanced stability and strong affinity of the methotrexate-functionalized system, likely driven by its molecular structure, electron-rich aromatic rings, and hydrogen bonding capacity. The interaction energy between the SWCNT-chelator-drug complexes and the 99mTc as a function of time. A) electrostatic, B) van der Waals. In contrast, the ketotifen-functionalized system demonstrates a significantly flatter energy profile, stabilizing at much higher (less negative) energy values. This suggests a weaker, less persistent interaction with technetium atoms—an observation consistent across both Coulombic and LJ terms. The electrostatic term dominates the total interaction energy in all systems, aligning with expectations due to the charged nature of the radiolabeled technetium and the polar functional groups on the drug and chelator moieties. However, the van der Waals contributions, although smaller in magnitude, are non-negligible and provide complementary insight into molecular packing and surface complementarity. Figure 5 quantitatively compares the average interaction energies (LJ, Coulomb, and total) for the four systems. The methotrexate-functionalized SWCNT complex emerges as the most stable, with a total interaction energy of −825.80 kJ/mol. This value reflects the synergistic effect of both strong electrostatic attraction (−795.93 kJ/mol) and moderate van der Waals stabilization (−29.87 kJ/mol). The piroxicam-functionalized system follows, displaying a slightly reduced, but still substantial, total interaction energy (−705.15 kJ/mol). Diclofenac exhibits intermediate stability (−374.55 kJ/mol), while ketotifen records the weakest binding, with a total energy of just − 296.89 kJ/mol. Average van der Waals, electrostatic, and total energies of the SWCNT-chelator-drug complexes and the 99mTc. This hierarchy—methotrexate > piroxicam > diclofenac > ketotifen—clearly demonstrates that drug identity plays a decisive role in shaping the thermodynamic landscape of technetium binding. The marked differences in energy values ​​are not random but stem from intrinsic molecular properties such as polar surface area, available donor/acceptor sites, steric accessibility, and electrostatic potential distributions. To provide a more detailed understanding of the energetic differences among the four drug-functionalized SWCNT systems, we analyzed the contributions of specific functional groups and structural regions to the total interaction energies. Methotrexate exhibits the strongest interactions, which can be attributed to its multiple polar and charged groups—including amine, carboxyl, and heterocyclic nitrogen atoms—that form persistent hydrogen bonds with both the chelator and the lipid headgroups, as well as stabilizing π–cation and π–π stacking interactions with the nanotube surface. Piroxicam shows intermediate binding, driven primarily by hydrogen bonding via its sulfonamide group, with additional stabilization from its aromatic rings. Diclofenac interactions are largely dominated by van der Waals forces arising from its hydrophobic aromatic rings, with relatively fewer polar contacts, resulting in a weaker overall adsorption energy. Ketotifen demonstrates the lowest binding affinity due to steric shielding of its polar groups and limited opportunities for hydrogen bonding or π interactions.This detailed analysis provides a mechanistic justification for the observed interaction hierarchy: methotrexate > piroxicam > diclofenac > ketotifen, illustrating how subtle structural differences among the drugs directly influence their binding energetics and spatial behavior in the SWCNT-chelator-drug-⁹⁹ᵐTc systems. The originality of this study lies in its integrative approach: while earlier investigations have examined chelator-based interactions with ⁹⁹ᵐTc or nanotube surface modifications in isolation, our work introduces a composite nanosystem, wherein the drug molecule itself acts as a functional modulator of radiotracer binding. This triple-component strategy (SWCNT + chelator + drug) represents a significant conceptual advancement, offering a new level of control over nanoparticle–radiotracer interactions. By leveraging all-atom molecular dynamics simulations, we provide an atomistically resolved and time-resolved view of how subtle variations in drug chemistry directly impact system energetics and stability. The differential behavior of each system, observed under identical simulation conditions, underscores the potential of rational drug selection as a tunable parameter in the design of next-generation radiolabeled nanotherapeutics. Moreover, the ability of methotrexate-functionalized SWCNTs to sustain highly favorable interaction energies throughout the simulation suggests promising real-world implications. A system that strongly retains ⁹⁹ᵐTc is expected to enhance tumor-targeted accumulation, reduce systemic leakage, and potentially allow for extended therapeutic windows in nanobrachytherapy applications. In contrast, the relatively weak binding observed for ketotifen may be advantageous in scenarios where rapid clearance or transient labeling is preferred. In sum, this energy analysis not only provides a quantitative benchmark for the relative stability of different functionalized nanocarriers but also introduces a new paradigm wherein drug molecules are viewed as active architectural elements, not merely therapeutic payloads. This reframing invites a broader rethinking of how nanomedicines are conceptualized: beyond passive carriers, they can be systematically engineered to co-opt drug structure as a means of enhancing radionuclide capture and control. To the best of our knowledge, this is the first study to establish such a structure–energy correlation in the context of technetium-loaded nanocarriers using full-atom MD simulation. The insights gained here open the door to precision design strategies, where both therapeutic and radiochemical performance can be optimized in tandem, guided by atomic-scale simulations. Our results reveal that the methotrexate-functionalized SWCNT complex exhibited the highest binding affinity for technetium-99 m with an adsorption energy of −825.80 kJ/mol, demonstrating the critical role of the drug structure in modulating the interaction with the radionuclide. Additionally, the complex showed the strongest interaction with phospholipid membranes, suggesting its potential for targeted drug delivery applications. In contrast, the ketotifen-functionalized system exhibited weaker interactions, which could be advantageous for transient diagnostic applications. These findings underscore the potential of methotrexate-functionalized nanotubes as a promising candidate for targeted radiopharmaceutical delivery, particularly in nanobrachytherapy applications, where precise tumor targeting and reduced systemic toxicity are crucial. To ensure the reliability of the interaction energies presented in Fig. 5, an uncertainty analysis was conducted based on the temporal fluctuations of the energy profiles throughout the production trajectories. Across all systems, energy variations remained small relative to the mean values, with uncertainties typically within the averaged energy. This narrow range indicates that the interaction energies reached stable plateaus and were not influenced by transient fluctuations. The uncertainty ranges for each system are explicitly included as error bars in Fig. 5, reflecting the statistical confidence in the reported interactions. Root mean square deviation (RMSD) of the SWCNT-chelator-drug complex The root mean square deviation measures the stability and structural integrity of a system over time by quantifying the average displacement of atoms from a reference structure42,43. RMSD analysis in this simulation provides insights into the stability of each functionalized nanotube system when interacting with the radioactive technetium particles. By calculating RMSD values, we can track how much the SWCNT-chelator-drug complex fluctuates or deforms during the simulation time. This metric helps assess the overall stability of the drug-loaded nanotube structure in a simulated biological environment, indicating whether it maintains its shape or undergoes significant structural changes when radioactive technetium particles approach it. As illustrated in Fig. 6, all simulated systems achieved structural equilibrium within the initial 15 nanoseconds of the simulation time. This quick equilibration indicates that each functionalized SWCNT-chelator-drug complex has reached a stable conformation about the radioactive technetium particles, demonstrating the system’s robustness under the specified simulation conditions. RMSD between the SWCNT-chelator-drug complexes and the 99mTc as a function of time. Radial distribution function analysis (RDF): spatial profiling of ⁹⁹ᵐTc around functionalized SWCNTs The radial distribution function, g(r), is a key structural descriptor in molecular dynamics, revealing the probability of finding a specific atom at a given distance from a reference point—in this case, the technetium-99 m (⁹⁹ᵐTc) ion relative to the functionalized single-walled SWCNTs. Understanding the spatial distribution of ⁹⁹ᵐTc around the nanotube-drug complexes is crucial, especially in the context of radiopharmaceutical design, where stability, localization, and retention of the radionuclide directly impact therapeutic efficiency and safety. Figure 7 shows the RDF profiles for four distinct drug-functionalized SWCNTs, each coordinated with ⁹⁹ᵐTc. Despite differences in molecular structure among methotrexate, diclofenac, ketotifen, and piroxicam, all systems consistently exhibit a dominant RDF peak near 0.8 nm. This suggests that, regardless of the drug identity, ⁹⁹ᵐTc tends to accumulate at a characteristic distance from the nanotube surface, likely defined by the steric reach of the chelator and the spatial conformation of the drug linker. The presence of this peak in all cases confirms the formation of a stable technetium coordination shell in the vicinity of the nanotube surface. Radial distribution function (RDF) of the drugs and the 99mTc. Among these, the methotrexate-functionalized SWCNT exhibits the highest and sharpest g(r) peak, indicating a strong and well-localized interaction between ⁹⁹ᵐTc and the methotrexate-nanotube complex. This observation is consistent with its markedly high adsorption energy (−825.80 kJ/mol), as determined from molecular dynamic simulations. Methotrexate, with its multiple aromatic rings, polar groups, and conjugated π systems, likely forms multiple stabilizing interactions—such as π–metal and electrostatic interactions—with technetium, leading to enhanced binding specificity and spatial confinement. In contrast, ketotifen shows the lowest RDF peak intensity and the broadest distribution, pointing to weaker and more diffuse interactions. This aligns well with its lower adsorption energy (–296.89 kJ/mol), suggesting that ketotifen offers fewer or less favorable binding sites for ⁹⁹ᵐTc. The relative lack of electron-rich functional groups capable of forming strong coordination bonds may be responsible for the reduced technetium affinity. Diclofenac and piroxicam display intermediate behavior, with broader yet visible peaks, reflecting moderate technetium localization. These differences underline the important role of drug molecular structure in shaping the microenvironment around the technetium ion. What sets this study apart from prior work is the direct linkage established between RDF behavior and specific drug properties, moving beyond generic nanotube-chelator interactions. To the best of our knowledge, this is the first comparative RDF-based investigation of ⁹⁹ᵐTc localization across a structurally diverse set of drug-functionalized carbon nanotube carriers. Previous studies have often focused either on the chelator’s coordination with technetium or on the nanotube’s electronic properties; Here, we bring the drug molecule itself into the spotlight as a determinant of radionuclide retention behavior. Moreover, the simultaneous interpretation of RDF results alongside adsorption energetics and structural motifs provides a more comprehensive view of technetium binding. This multidimensional approach enables a rational, structure-based selection of drug-linker systems for enhanced radionuclide capture, which is essential for the future design of site-specific radiopharmaceutical agents. Another novel aspect of this work lies in the demonstration of how subtle changes in drug chemistry translate into quantifiable differences in technetium spatial confinement. These findings highlight the importance of tailoring both the carrier and the therapeutic molecule to achieve optimal radiolabeling efficiency—an area that remains underexplored in current nanomedical literature. In conclusion, RDF analysis has proven instrumental in uncovering the nuanced, drug-dependent patterns of technetium localization around functionalized SWCNTs. Methotrexate stands out as the most promising candidate, offering both strong and spatially focused interaction with ⁹⁹ᵐTc. The insights gained here can inform the future design of nanocarrier systems for radiopharmaceutical applications, bridging the gap between molecular-level simulation and clinical translation. 99mTc mobility in the SWCNT-chelator-drug complex interactions The dynamics of the 99mTc molecules are examined with respect to both orientational and translational mobility. The translational dynamics are quantified through the self-diffusion coefficient Di, which is calculated from the mean squared displacement (MSD) using the Einstein relation (Eqs. 1and 2)44: Where ri(0) and ri(t) represent the position of the particle at time 0 and time t, respectively. The resulting MSD profiles (Fig. 8) and corresponding diffusion coefficients (Table 3) reveal distinct mobility patterns driven by the nature of the bound drug molecules. Mean square displacement (MSD) of 99mTc particles versus simulation time for SWCNT-chelator-drug complexes. Table 3Average diffusion coefficient (Di) of 99mTc particles in the SWCNT-chelator-drug complex.Systems Errormethotrexate0.0070 0.0012piroxicam0.0082 0.0006diclofenac0.0127 0.0060ketotifen0.0144 0.0029 Average diffusion coefficient (Di) of 99mTc particles in the SWCNT-chelator-drug complex. It should be noted that the dynamics analyzed here refer specifically to the motion of individual ⁹⁹ᵐTc atoms relative to the functionalized SWCNT-chelator-drug complex, and not to the translational motion of the entire complex. Among the complexes studied, the methotrexate-functionalized SWCNT displayed the lowest diffusion coefficient (0.0070 × 10⁻⁵ cm²/s), suggesting a strong immobilization of technetium on the nanotube surface. This reduced mobility aligns with the highest adsorption energy observed in metadynamics simulations (–825.80 kJ/mol), indicating that methotrexate confers a stable and persistent interaction between ⁹⁹ᵐTc and the nanocarrier—an essential feature for prolonged imaging windows in radiopharmaceutical applications. Conversely, the SWCNT functionalized with ketotifen exhibited the highest diffusion coefficient (0.0144 × 10⁻⁵ cm²/s), indicative of a more dynamic technetium interaction. The comparatively weaker adsorption energy (–296.89 kJ/mol) supports this observation, suggesting ketotifen-based systems may favor rapid ⁹⁹ᵐTc redistribution, which could be advantageous for short-term diagnostics or responsive drug-release scenarios. Intermediate diffusion behavior was observed for SWCNTs functionalized with piroxicam and diclofenac, with diffusion coefficients of 0.0082 and 0.0127 × 10⁻⁵ cm²/s, respectively. These results reflect moderate interaction energies and imply a balance between technetium retention and release, potentially offering a tunable platform for tailored radiotracer delivery. The observed differences underscore the critical influence of the drug moiety on ⁹⁹ᵐTc mobility. Enhanced binding, while beneficial for sustained localization, must be balanced against the need for dynamic distribution depending on the intended clinical application. Thus, the ability to modulate technetium diffusion through rational drug functionalization presents a strategic advantage in designing SWCNT-based nanocarriers for targeted brachytherapy. Ultimately, understanding the translational dynamics of ⁹⁹ᵐTc in these systems provides a vital layer of design insight. Fine-tuning the balance between strong adsorption and sufficient mobility can significantly impact the efficacy of radiopharmaceutical delivery, optimizing both retention at tumor sites and clearance from non-target tissues. Future work will explore the application of these systems in vivo and investigate the use of alternative radionuclides and drug combinations to further optimize therapeutic outcomes in nanomedicine. Drug delivery system proximity to membrane cells This section incorporates the membrane to more accurately represent the physicochemical behavior of the SWCNT-chelator-drug-⁹⁹ᵐTc complexes under biologically relevant conditions. The analyses performed include total energy evaluation, free energy profiling, and detailed examination of the complex membrane interaction dynamics. Energy analysis of the adsorption of SWCNT-chelator-drug-99mTc complex onto the membrane To understand the molecular basis of membrane adsorption, we carried out a detailed energetic analysis of the interactions between the functionalized SWCNT-chelator-drug-⁹⁹ᵐTc complexes and a model lipid bilayer. The primary objective was to elucidate how the chemical nature of the drug moiety influences the adsorption strength and stability at the membrane interface. Particular attention was given to decomposing the total interaction energy into its electrostatic and van der Waals (Lennard-Jones) components, offering mechanistic insight into the driving forces behind adsorption. (see Fig. 9) Average van der Waals, electrostatic, and total energies of the SWCNT-chelator-drug-99mTc complexes and lipid bilayer. Figure 10 presents the temporal evolution of both interaction types throughout the simulation. The results reveal that the adsorption behavior is governed by a synergistic interplay between polar (electrostatic) and nonpolar (dispersion) forces, with notable drug-specific signatures emerging across the four investigated complexes. The interaction energy between the SWCNT-chelator-drug-99mTc complexes and lipid bilayer as a function of time. A) electrostatic, B) van der Waals. Among the studied systems, the methotrexate-functionalized SWCNT complex exhibited the strongest electrostatic interaction with the membrane. This can be attributed to the presence of multiple polar and charged groups in the methotrexate structure, which form stable electrostatic attractions with the phospholipid headgroups. These interactions appear to anchor the complex effectively at the membrane surface, promoting strong and persistent binding. In contrast, while the diclofenac-functionalized system showed more prominent van der Waals interactions—owing to its relatively more hydrophobic and aromatic character—its overall binding affinity remained the weakest. This disparity highlights the critical role of electrostatic complementarity in stabilizing nanotube–membrane interactions. Indeed, a strong van der Waals contribution alone, in the absence of sufficient polar contacts, proved insufficient for establishing robust adsorption. The total adsorption energy, representing the cumulative electrostatic and Lennard-Jones interactions, further clarifies the trends in membrane affinity. The methotrexate complex exhibited the most favorable adsorption energy (− 1375.62 kJ/mol), indicating a thermodynamically stable and strongly bound configuration. The enhanced adsorption is likely the result of an optimal balance between long-range electrostatics and short-range dispersion forces, facilitating firm anchoring without inducing destabilizing repulsion or steric hindrance. In stark contrast, the diclofenac complex displayed the lowest total interaction energy (− 426.24 kJ/mol), pointing to a relatively weak and possibly transient membrane association. In this study, we performed a detailed analysis of the van der Waals (vdW) and electrostatic interaction energies between the CNT-drug complexes and the lipid bilayer. These interactions are critical for understanding the stability of the CNT complexes at the membrane interface, which influences their potential for targeted drug delivery. The vdW interactions, primarily arising from the dispersion forces between the hydrophobic components of the CNTs and the lipid tails, and the electrostatic interactions, which result from the attraction between charged groups on the CNTs and the polar headgroups of the membrane, both play essential roles in the adsorption process. We observed that the methotrexate-functionalized CNT complex exhibited the strongest electrostatic interactions, with values reaching − 1376.02 kJ/mol (Fig. 9). This result is consistent with the presence of multiple polar functional groups in methotrexate, which strongly interact with the zwitterionic phosphocholine groups of the membrane. In contrast, the diclofenac-functionalized CNT complex displayed weaker electrostatic interactions (with a total interaction energy of −426.24 kJ/mol), reflecting the more hydrophobic nature of diclofenac and its weaker interaction with the membrane. To provide a more detailed view of the interactions, we present zoomed-in images of the interaction conformations (Fig. 12), which show the surface contact between the CNT complexes and the membrane. These images clearly depict the molecular arrangement of the CNT complexes at the membrane interface, highlighting the regions where the CNTs make direct contact with the lipid headgroups. In the case of the methotrexate-functionalized CNT complex, the surface contact is extensive, with multiple functional groups forming hydrogen bonds and electrostatic interactions with the membrane surface. This leads to a stable and well-defined adsorption configuration. In contrast, the diclofenac complex shows more dispersed surface contact, with weaker interactions and less stable binding at the membrane surface. Evolution of the progressive insertion of A) SWCNT-chelator-diclofenac-99mTc complex, B) SWCNT-chelator-methotrexate-99mTc complex, C) SWCNT-chelator-ketotifen-99mTc complex, D) SWCNT-chelator-piroxicam-99mTc complex into the lipid bilayer at two times of the simulation, t=0 ns, and t=200 ns. By focusing on these surface interactions, we gain a more nuanced understanding of the adsorption process and the factors that determine the affinity of the CNT complexes for the membrane. The interaction energies, coupled with the detailed surface contact information, allow us to draw conclusions about the stability and performance of these CNT-drug complexes for drug delivery applications. The methotrexate-functionalized complex, with its strong electrostatic and vdW interactions, appears to be the most promising candidate for stable membrane adsorption, which is crucial for enhancing drug retention and delivery at target sites. These findings not only validate the qualitative interaction patterns observed in our trajectory analyses but also quantify the binding hierarchy across the drug-functionalized systems. The energetic profiles suggest that methotrexate-functionalized nanotubes may offer superior performance in membrane-targeted drug delivery applications, potentially improving localization, retention time, and therapeutic efficacy. Meanwhile, the weaker interactions observed with diclofenac imply reduced residence time at the membrane surface, which may limit its utility in contexts requiring sustained release or prolonged contact with target cells. Importantly, this study demonstrates that drug identity is not merely a payload detail but a fundamental determinant of system-level behavior in nanocarrier design. By showing that the molecular features of the drug itself modulate the total adsorption energy, our results underscore the need for integrative design strategies in SWCNT-based delivery systems—ones that consider not just bioconjugation feasibility, but also the physicochemical consequences of each drug on the system’s interaction profile with biological membranes. This energy analysis complements our free energy landscape study and collectively paints a detailed thermodynamic and kinetic picture of how SWCNT-chelator-drug-⁹⁹ᵐTc complexes engage with lipid membranes. Together, these insights advance our understanding of radiopharmaceutical behavior at the bio-nano interface and offer guiding principles for rationally designing functionalized nanotubes with improved membrane affinity and delivery potential. To ensure the reliability of the interaction energies presented in Fig. 9, an uncertainty analysis was conducted by examining the statistical fluctuations over the production trajectory. For the systems shown in Fig. 9, the resulting standard deviations were consistently small relative to the corresponding mean interaction energies, indicating that the simulation achieved adequate sampling and that the reported energy trends are robust. These uncertainty ranges have been incorporated into Fig. 9 to provide clearer insight into the reliability of the observed adsorption behavior. Free energy calculation of the adsorption of SWCNT-chelator-drug-99mTc complex onto the membrane Free energy calculations serve as a cornerstone for understanding the thermodynamic feasibility and stability of biomolecular interactions, especially in the context of complex adsorption phenomena45,46. In this study, we employed well-tempered metadynamics simulations to reconstruct the free energy surface of two representative SWCNT-chelator-drug-⁹⁹ᵐTc complexes during their approach and adsorption onto a lipid membrane. This approach enabled us to not only quantify the relative stabilities of these systems but also to explore the mechanistic underpinnings that govern their adsorption behavior at the molecular level. The analysis focused on two systems characterized by markedly different adsorption energies: the methotrexate-containing complex, which exhibited the strongest interaction with the membrane in our earlier binding energy calculations, and the diclofenac-containing complex, which showed the weakest. This selection allowed us to contrast their thermodynamic profiles and gain deeper insights into how drug identity influences membrane binding. The metadynamics simulations revealed distinct free energy minima corresponding to stable adsorption configurations. Notably, the global minimum for the SWCNT-chelator-methotrexate-⁹⁹ᵐTc complex was observed at a distance of approximately 6.3 nm from the membrane center, with a free energy value of − 184.99 kJ/mol. In contrast, the same distance for the diclofenac-containing complex yielded a slightly less negative free energy of − 174.94 kJ/mol. Although both complexes demonstrate favorable thermodynamic profiles for membrane adsorption, the more negative free energy of the methotrexate complex confirms a stronger and more stable interaction, consistent with the system’s higher overall adsorption energy. (see Fig. 11) The 2D free energy profile of adsorption of the SWCNT-Chelator-Drug-99mTc on the lipid bilayer. A) Diclofenac and B) Methotrexate. These differences can be attributed to variations in the physicochemical properties of the drug molecules—particularly their capacity to participate in non-covalent interactions such as hydrogen bonding and electrostatics with membrane lipids. Methotrexate, with its multiple polar functional groups and larger conjugated structure, likely engages in more extensive hydrogen bonding and π–π stacking interactions with both the membrane and the chelator scaffold, thereby stabilizing the adsorbed state more effectively than diclofenac. Importantly, the shape of the free energy surface also revealed kinetic barriers that might influence the dynamics of adsorption. The shallower landscape observed for diclofenac suggests a lower energy barrier for desorption, potentially leading to less stable membrane retention. Conversely, the deeper basin associated with methotrexate indicates a more robust adsorption event, which may have practical implications for improving the residence time and targeting efficacy of nanotube-based radiopharmaceuticals. Together, these results underscore the importance of free energy landscape mapping in the rational design of nanotube-drug systems for biomedical applications. The combination of metadynamics and molecular modeling offers a powerful predictive tool for optimizing drug-membrane interactions. By capturing the subtle interplay between structural features, intermolecular forces, and thermodynamic stability, this work provides a foundation for the targeted engineering of nanocarriers with enhanced adsorption performance and improved in vivo behavior. Interaction of the SWCNT-chelator-drug-99mTc complex with lipid membranes Understanding how functionalized nanocarriers behave at biological interfaces is critical to advancing drug delivery technologies, especially in the context of radiopharmaceutical systems where precise targeting and retention are essential. In this study, we explored the dynamic interaction behavior of four SWCNT-chelator-drug-⁹⁹ᵐTc complexes with a POPC lipid bilayer using long-timescale molecular dynamics simulations. This section integrates visual trajectory analysis with previously discussed energy decomposition and free energy results, offering a unified perspective on interaction strength, spatial behavior, and mechanistic implications. At the onset of each simulation, all complexes were positioned ~ 2.5 nm above the bilayer surface in a parallel orientation, thereby avoiding artificial bias while ensuring an adequate distance for observing natural approach behavior. As time progressed, each system exhibited a spontaneous attraction toward the membrane, with interaction dynamics shaped by both the chemical characteristics of the drug moiety and the electrostatic landscape of the membrane surface. Snapshots taken at 0 ns and 200 ns (Fig. 12) clearly show the progressive movement and eventual stable association of each complex with the membrane, but the extent and nature of these interactions varied significantly across the four systems. The SWCNT-chelator-methotrexate-⁹⁹ᵐTc complex, which previously showed the most negative total interaction energy (− 1376.02 kJ/mol) and lowest global free energy (− 184.99 kJ/mol), also demonstrated the deepest and most stable membrane interaction. By the end of the simulation, it exhibited strong surface alignment and a well-integrated configuration with the polar head groups of the lipid bilayer. This behavior can be attributed to the multiple polar and charged sites on methotrexate, which promote strong electrostatic interactions with the zwitterionic phosphocholine groups on the membrane surface. These polar contacts not only serve as anchoring points but also create a favorable environment for van der Waals interactions to reinforce the complex’s positioning, as confirmed by the local minima on the free energy surface. In contrast, the SWCNT-chelator-diclofenac-⁹⁹ᵐTc complex, which had the least negative adsorption energy (− 420.25 kJ/mol) and higher free energy value (− 174.94 kJ/mol), showed more limited interaction with the membrane. While some surface affinity was observed, the complex remained more loosely associated and retained a relatively superficial orientation. This result is consistent with the lower polar character of diclofenac and its reduced capacity for forming electrostatically favorable interactions with the membrane’s headgroups. Instead, the complex relied predominantly on van der Waals interactions with the lipid tails—interactions which, though present, were insufficient for creating a deeply stable adsorption site. The SWCNT-chelator-ketotifen and SWCNT-chelator-piroxicam complexes displayed intermediate behaviors, suggesting that their amphiphilic properties allow a balanced mode of interaction—leveraging both hydrophilic contacts with headgroups and hydrophobic interactions with the lipid core. These results reinforce the notion that adsorption dynamics are not binary but exist along a spectrum defined by molecular polarity, flexibility, and the spatial distribution of functional groups. Across all systems, the electrostatic landscape of the membrane appeared to be the dominant factor initiating interaction. The polar headgroups acted as molecular beacons, guiding the functionalized nanotube complexes into alignment through charge-based attraction. Once these initial contacts were established, van der Waals forces played a secondary but important role, helping to stabilize the complexes in a membrane-conformant configuration. This interplay was most effective in the methotrexate system, which exhibited a near-optimal balance between electrostatic and dispersion forces, both in interaction energy calculations and in positional behavior during simulation. Importantly, these dynamic findings are strongly supported by our free energy landscapes, which clearly identify favorable adsorption basins for each complex and define the energetic barriers associated with membrane approach. The alignment between free energy minima and final simulation positions confirms that the observed behaviors are not transient or artifactual but rather represent thermodynamically preferred states. From a drug delivery standpoint, these results have far-reaching implications. A higher degree of membrane adhesion not only promotes prolonged residence at the target site but also supports more predictable and controlled release of the therapeutic payload and radiolabel. In this regard, methotrexate-functionalized SWCNTs emerge as the most promising candidate due to their robust binding behavior, consistent with both kinetic trajectories and energy-based descriptors. Conversely, the diclofenac system, while perhaps advantageous in applications requiring rapid clearance or transient contact, may not provide sufficient stability for sustained-release scenarios. These findings suggest that the drug-functionalized nanotube systems, particularly the methotrexate-functionalized complex, hold significant promise for enhancing the efficacy of targeted radiotherapy. The strong adsorption of technetium-99 m onto the nanocarrier could facilitate localized accumulation at tumor sites, reducing systemic toxicity and improving therapeutic outcomes. Future research should explore the application of these systems in vivo, to further assess their potential in clinical settings. Overall, this multidimensional analysis—combining trajectory visualization, interaction energy decomposition, and metadynamics-derived free energy profiling—represents a novel and integrative approach to assessing nanocarrier performance. The findings underscore the importance of rational molecular selection in nanocarrier design, revealing that the drug moiety is not merely a passive component but an active determinant of system-level interaction dynamics. To our knowledge, this is among the first studies to apply such a comprehensive methodological framework to radiolabeled SWCNT systems interacting with lipid membranes, offering valuable insights for both fundamental nanobiophysics and translational drug delivery applications. Conclusion This study provides a comprehensive computational framework for evaluating functionalized SWCNTs as precision carriers in radiopharmaceutical-based cancer therapy, with a specific focus on the delivery and retention of technetium-99 m (⁹⁹ᵐTc). Through a combination of molecular dynamics and metadynamics simulations, we explored the structural, energetic, and dynamic characteristics of SWCNT-chelator-drug complexes interacting with radioactive particles and biological membranes. Our findings reveal a pronounced influence of the conjugated drug molecule on the overall behavior of the nanocarrier. Among the four studied systems—methotrexate, diclofenac, ketotifen, and piroxicam—the methotrexate-functionalized complex emerged as the most promising candidate. It demonstrated the highest technetium adsorption energy (− 825.80 kJ/mol), the most localized RDF peak, and the lowest diffusion coefficient (0.0070 × 10⁻⁵ cm²/s), collectively indicating robust radionuclide immobilization. This behavior is essential for achieving sustained radiopharmaceutical retention at target sites, a critical requirement for effective brachytherapy. Distance analysis further corroborated the affinity of methotrexate-functionalized complexes for technetium, showing a stable decrease in separation over time, consistent with strong, energetically favorable interactions. In contrast, ketotifen-functionalized complexes showed delayed adsorption and higher diffusion coefficients, reflecting weaker technetium confinement. These observations confirm that the molecular structure and binding environment created by each drug-chelator combination can modulate radionuclide behavior in nuanced but predictable ways. Furthermore, interactions between these nanotube complexes and lipid bilayers—used as a model for cell membranes—demonstrated favorable spontaneous adsorption during extended simulations. Methotrexate again outperformed the other systems, displaying the most negative electrostatic and van der Waals interaction energies with the membrane, along with the lowest free energy (− 184.99 kJ/mol) in metadynamics simulations. This highlights its capacity for membrane localization and efficient payload delivery. Diclofenac-functionalized systems, in contrast, showed the weakest membrane affinity and adsorption energetics. Collectively, these results demonstrate for the first time a systematic correlation between drug identity, technetium interaction dynamics, and membrane binding efficiency in a simulated biological context. The original contribution of this work lies in shifting the focus from the chelator alone to the drug-chelator synergy as a determinant of radiopharmaceutical retention and biodistribution behavior. This insight opens a new dimension in nanobrachytherapy design, where therapeutic and targeting properties can be co-optimized through rational drug selection. In conclusion, this study not only underscores the potential of functionalized SWCNTs as high-precision carriers for ⁹⁹ᵐTc but also introduces a structure-driven strategy for enhancing the performance of nanoradiopharmaceutical systems. By tuning the molecular components of the complex, it becomes possible to tailor adsorption strength, mobility, and membrane affinity to suit specific clinical requirements—ranging from long-term retention in tumor tissue to rapid clearance in diagnostic settings. Future research should extend this design strategy to other radionuclides, linker architectures, and cancer-specific ligands, while also validating the simulated results in experimental and biological models to accelerate clinical translation. This study presents a comprehensive approach to designing drug-functionalized carbon nanotube systems for targeted radiotherapy applications. By leveraging molecular dynamics simulations, we have demonstrated how drug selection can influence the performance of radiopharmaceuticals. These findings lay the groundwork for future research aimed at optimizing nanocarrier systems for improved drug delivery and cancer treatment. Author contributions Sedigheh Abdollahi: Devised the computational protocol and prepared the model systems, performed all calculations, analyzed the data, and wrote and edited the original and the revised manuscript.Heidar Raissi: Supervision. Reviewing- Editing, edited the original and the revised version of the manuscript.Farzaneh Farzad: Reviewing-Editing, edited the original and the revised version of the manuscript. Funding This research received no external funding. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Consent for publication All the authors have read the manuscript and have approved this submission. Competing interests The authors declare no competing interests. Conflict of interest The Authors declare no Competing Financial or Non-Financial Interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.