Large Language Model Driven Predictive Modeling of Silver, Zinc, Titanium, Magnesium, Gold Doped Hydroxyapatite for Infrared Triggered Drug Delivery

Large Language Model Driven Predictive Modeling of Silver, Zinc, Titanium, Magnesium, Gold Doped Hydroxyapatite for Infrared Triggered Drug Delivery | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Large Language Model Driven Predictive Modeling of Silver, Zinc, Titanium, Magnesium, Gold Doped Hydroxyapatite for Infrared Triggered Drug Delivery OLUMAKINDE CHARLES OMIYALE This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8524202/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 15 You are reading this latest preprint version Abstract In this paper, we report the application of Large Language Models (LLMs) in predictive modeling of infrared (IR)-triggered drug delivery systems, focusing on hydroxyapatite (HA) modified with silver nanoparticles (AgNPs), as well as dopants of zinc (Zn), titanium (Ti), magnesium (Mg), and gold (Au). We predict electronic structure, bandgap reduction, and optical properties utilizing real-world data integrating SMILES, benefiting from Darwin 1.5, a fine-tuned model, and T5, a predictive model, to predict electronic structure, bandgap values, and optical properties. Darwin 1.5 fine-tuned via question-answering and multi-tasking on scientific datasets correlates a mean absolute deviation (MAD) of 0.72 eV to bandgap predictions that are accurate and potentially cut the simulation time by as much as 50-70% compared to the conventional density functional theory (DFT) method. The T5 model enables simulations on optical properties via the computation of absorption spectra at a wavelength of 808 nm and the concentration-dependent roles on absorption and scattering. The predictions obtained from electronic structure calculations with a fine-tuned Darwin 1.5 model, combined with DFT analysis, indicate that Ag-HA's band gap reduction varied from 4.312 eV (0.25 mol% Ag doping) to 3.983 eV (0.75 mol% Ag doping); from about 4.68 eV (0.25 mol% Zn doping) to approximately 4.4 eV (0.75 mol% Zn doping) in praseodymium co-doped-HA; from around 3.8 eV (0.25 mol% Ti doping) to around 3.6 eV (0.75 mol% Ti doping); from about 4.61 eV (0.25 mol% Mg doping) to about 4.39 eV (0.75 mol% Mg doping); and from about 4.3 eV (0.25 mol% Au doping) to about 3.95 eV (0.75 mol% Au doping). Moreover, it was observed that the photothermal efficacies are higher, with a value of 18.8% (internal) to 0.11 L g −1 cm (external) under a concentration of 2% Ag-HA irradiated with a wavelength of 445nm, followed by comparable efficacies observed under similar conditions with a value of 11.6% (internal) to 0.055 L g −1 cm (external) with a concentration of Au-HA. Relatively low cell toxicity are observed in literature, in vitro studies show fairly balanced antibacterial activity, in addition to in vivo studies that reveal promoted bone healing with lowered systemic toxicity. The results are represented in detailed tables, figures, and Python codes. Ultimately, the present study serve to promote the application of LLMs in materials science, particularly in the area of cancer and infectious diseases, as well as considerations of the ethics of AI applications and the dangers of nanomaterials. Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Background on Hydroxyapatite and Light-Responsive Modification Hydroxyapatite (HA), with the chemical formula Ca 5 (PO 4 ) 3 (OH), is the major inorganic material of bones and teeth, characterized by high biocompatibility and bioactivity. Its crystal system of hexagonal symmetry in the P 6 3/m space group, with lattice parameters a = b ∼9.418Å and c ∼6.884Å, has unique ionic substitutions suitable for enhanced applications. (Ressler et al., 2021). In general, HA does not have natural photoresponses to external light. Its traditional application is thus limited in dynamic drug delivery. To overcome these limitations in drug delivery, photoresponsive dopant materials were considered. The idea stemmed from azobenzenes, which are organic compounds capable of reversible transformation from the trans-to-cis isomerization states and vice versa upon irradiation with ultraviolet light. The photoresponse of these compounds is based on reversible conjugation and deconjugation reactions in the molecule (Abdelmohsen et al., 2022; Li et al., 2021; Madau, 2021). HA doping with transition metals like Ag, Zn, Ti, Mg, or Au extends its absorption in the UV and NIR regions, finding use in sunscreens, wound healing, or biomedical implants. For the near-infrared range around 808 nm, optimal for nondestructive penetration of tissue, silver nanoparticles (AgNPs) and gold (Au) have shown efficacy due to surface plasmon resonance for effective photothermal conversion (Debnath et al., 2024, Pasparakis 2022, Wei et al., 2014). Zinc promotes antibacterial properties, while titanium acts as a photocatalytic agent. Magnesium enhances biore-sorbability. This property of photothermal conversion creates a heat source around the material, inducing controlled release of medication. Relative strength among the dopants suggests silver’s economic viability and antimicrobial efficacy, gold’s stability, zinc’s biocompatibility, antimicrobial activity, titanium's photocatalyst properties, or magnesium’s degradation flexibility, potentially using a combination of these materials in hybrid materials (Santos-Coquillat et al., 2021, Omiyale et al., 2023, Omiyale, 2024). Applications of Large Language Models to Materials Prediction Large Language Models (LLMs) have also been identified as significant tools for use in materials science. They have the capacity to handle a large dataset for property predictions without engaging in explicit feature engineering (Lei et al., 2024). Platforms like Darwin 1.5, a free LLM powered by LLaMA-7B, have also been tailored for chemistry and materials studies via fine-tuned approaches that use expert knowledge. LLMs fine-tuned for datasets with SMILES descriptions can predict the structure, band gap, and spectrum, filling the gap between simulation and experimentation. Computed comparisons have also shown that Darwin 1.5 can perform better than certain Graph Neural Networks (GNNs) for certain regression tasks, for instance, mean RMSE of 0.69 for the band gap calculation against 0.85 for the latter, owing to its unique ability to utilize text and graph data (Xie et al., 2024). This research uses Darwin 1.5 for the prediction of the electronic structure and the T5 (Text-to-Text Transfer Transformer) model’s abilities in the simulation of optics, based on real data in line with the Darwin learning materials (NagasawaOPV, band gap, TADF, emission wavelength). The emphasis involves the use of HA with Ag, Zn, Ti, Mg, and Au metals in the design of the IR-based drug delivery system, using a wavelength of 808 nm, and the databases used include the QM9 analogs of the molecules and the results of the DFT analysis from articles in the evaluation of the materials. The methods used (Nikidis et al., 2024) include the synthesis of materials using the agar diffusion technique and the LLM modeling of the materials’ characteristics, including the interaction of the dopants with the HA. The approach of incorporating the LLM in biomaterial design, as the study does, reveals the rapid advancements in AI applications, as well as considerations concerning the data privacy of the patients involved and the bias in the AI analysis. (Buehler, 2023) Materials and Methods Literature Review and Data Compilation The literature searches performed using a large language model were implemented on Consensus (https://consensus.app), a search engine based on AI and providing direct access to the scientific literature. The search engine enables quick and efficient retrieval of brief, answer-driven summaries for research queries. The visualization of literature mapping was done on Litmaps (https://app.litmaps.com) an interactive chart depicting the patterns of citations from a seed paper. Grok (https://grok.com), a generative AI/LLM, enabled processing of varied graphical input like documents, images, graphics, screenshots, and photographs. The compiled results were obtained from databases such as PubChem, SciFinder, and The Materials Project. The search terms used were focused on "doped hydroxyapatite bandgap DFT" for silver, zinc, titanium, magnesium, and gold, and on "fine-tuning LLM SMILES electronic properties." The measured structural properties were gleaned from the Crystallography Open Database (entry COD-9011096), which defined lattice constants a = b = 9.418 Å, c = 6.884 Å, space group P6₃/m (176), and density 3.155 g/cm³. Fine-tuning datasets used were QM9 analogs, composed of small molecules exhibiting quantum characteristics, and material datasets such as NagasawaOPV (band gaps from SMILES representations of molecules), and ESOL (solubility). Synthesis Protocols from Literature Hydroxyapatite (HA) doped with Ag, Zn, Ti, Mg, and Au has been synthesized by agar diffusion reaction and wet precipitation, as reported in various studies. For the agar diffusion reaction described by Eltantawy et al. (2021), inorganic Liesegang rings (LRs) are produced via the diffusion of ions and phase transition of the calcium phosphate. The process of generating hydroxyapatite involves the diffusion of CaCl 2 in N a2 HPO 4 . For this purpose, The HA-Ag composite develops when the inner electrolyte is a solution of Ca(NO 3 ) 2 with 1 wt% AgNO 3 based on the detectable silver nanoparticle peaks in the composition. On the other hand, the HA-AgCl is created when the outer electrolyte is CaCl 2 /AgNO 3 . The process was done in two separate containers using gelled agar and Na 2 HPO 4 . These were stirred and heated until a homogenous mixture was attained. The resulting mixture was added to a Petri dish, and upon cooling to room temperature (25°C), the agar gelled. The dishes were filled with 50 μL of 1 M CaCl 2 solution (outer electrolyte). The samples were allowed to develop inorganic Liesegang rings. Then, samples were harvested, and the agar is removed by rinsing with subsequent reheating. The harvested hydroxyapatite is then dried at 60°C for 3 hours. The hydroxyapatite samples were ground to powder. Then, infrared-sensitive polyelectrolytes were coated on PE-requiring samples. For the wet precipitation process, Ca(NO 3 ) 2 · 4 H 2 O and (NH 4 ) 2 HPO 4 were mixed based on the ratio of 1.67 (Ca/P ratio). The dopant materials (AgNO 3 , Zn(NO 3 ) 2 , or TiCl 4 or Ti(OBu) 4 , Mg(NO 3 ) 2 , AuCl 3 ) were added in a 0.25-0.75 mol% concentration ratio to the amount of Ca. The pH of the mixture is set at 10 with the use of ammonia. Then, it is heated at 70°C for 3 hours. The samples were then further processed by drying at 90°C and subjected to calcination at 450°C. For modified samples doped with Praseodymium, Pr(NO 3 ) 3 is added at 0.25 mol% in addition to the other chemicals. In this process, the samples will be well doped. Substitution at the Ca site will result in modifications at the lattice. The levels of crystallization will vary based upon the type of dopants (Sahin et al., 2024). Characterization Methods X-ray diffraction (XRD) analysis revealed a hexagonal HA structure with less crystallinity (between 94% and 30%) in some samples. Scanning electron microscopy (SEM) imaging indicated the presence of nanoparticles, while FTIR analysis revealed PO4³⁻ and OH⁻ stretching modes. The Ag- and Au-doped samples had absorption bands in the range of 420-430 nm, while variations were observed depending on the doping elements Zn, Ti, or Mg. The photothermal studies used a wavelength of either 445 nm, 532 nm, or 808 nm with a power setting of 90 mW or 1 W/cm², analyzing the temperature increments and conversion efficiencies (Haugen et al., 2022). Predictive Modeling Predicting Electronic Structure by Fine-Tuned Darwin 1.5, an open-source LLM developed using LLaMA-7B, was employed to predict the effects on electronic structures when HA gets doped with Ag, Zn, Ti, Mg, and Au. (Appendix 1) A two-step fine-tuning approach was adopted: First, QA fine-tuning was performed on SciQAG-24D (28,000 QA pairs from scientific literature) to transfer domain knowledge, using questions like “What is the band gap of Ca 5 −xMx(PO 4 ) 3 OH?” where M = Ag, Zn, Ti, Mg, and Au; second, multi-task learning using 21 FAIR datasets, including SMILES datasets like NagasawaOPV (Band Gaps) and ESOL (Solubility). For structured HA, composition formulas were supplied to inputs, adding SMILES descriptions to HA cluster structures, such as SMILES strings explaining simplified (e.g., SMILES for simplified Ca10(PO4)6(OH)2: [Ca+2].[Ca+2].[Ca+2].[Ca+2].[Ca+2].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[OH-].[OH-]) Ca 10 (PO 4 ) 6 (OH) 2 , [Ca +2 ].[OH - ]), and doping patterns to capture local HA-QM9 structures. (Figure 1) The model was trained using 4× AMD MI250X GPUs at BF16 precision, learning rate set to 2e -5 , total epochs to 10, and using LongLoRA. QA-MT model performed well on regression tasks, and maximum MAD was achieved at 0.72 eV, and also reached an RMSE value at 0.69 eV, outperforming other state-of-the-art results like PBE-DFT, MAD = 1.65 eV. The models also predicted well when HA gets doped and were consistent with other state-of-the-art DFT predictions, showing band gap reduction due to creation of localized states created in HA by all dopants. Simulation of Optical Properties via a Fine-Tuned T5 . Model T5 was fine-tuned to better predict optical properties. The models were trained in the text-to-text setting with similar datasets as absorption spectra prediction. The model took in the text input of the form: “Predict absorption at 808 nm for M-doped HA at x mol%," where M = Ag, Zn, Ti, Mg, or Au. Darwin 1.5 models were cross validated with regard to refractive indices as well as emissions. The effect of concentration of 1-2 mol% was analyzed with predictive models predicting quadratic growth in absorption alongside linear enhancement in scattering; plasmonic dopants (Ag and Au) contributed relatively more. It is pertinent to acknowledge that T5 is effective in text-related applications. However, the approach does not prove to be optimal within the current context. Further analysis using more advanced approaches, like graph neural networks may provide better results. (Huang et al., 2023), Optical Properties The Surface Plasmon Resonance (SPR) maximum occurs at 420-430 nm for Ag- and Au-doped hydroxyapatite (HA). The resulting absorbance values increase quadratically with the concentration of the doping materials. At 808 nm, the value of the absorbance Aλ is quite substantial for photothermal uses, amounting to 0.5-1.2 arbitrary units (a.u.) for 1-2 mol% Ag and similar values for 1-2 mol% gold (Au). However, the values for the dopants Zn, Ti, and Mg are 0.4-0.9 a.u., 0.3-0.7 a.u., and 0.2-0.5 a.u., respectively. Simulation results for T5 indicate an increase in the values of the absorption coefficients from 0.1 a.u. for the undoped samples to 1.2-1.3 a.u. for 2 mol% Ag or 2 mol% Au doped samples; the values of the scattering coefficients is 0.027-0.110 L/g -1 cm -1 . (Appendix 2 and 3) Optical Absorption and Scattering Properties vs. Dopant Concentration Data fitted quadratically (e.g., Ag: a=0.35, b=0.15, c=0.1; Au: a=0.40, b=0.20, c=0.1; Zn: a=0.25, b=0.15, c=0.1; Ti: a=0.20, b=0.10, c=0.1; Mg: a=0.15, b=0.05, c=0.1) Figure 2. Photothermal Performance Hydroxyapatite doped with metals such as silver or gold demonstrates high photothermal conversion efficiencies (HCE). More specifically, Ag-HA doped at 2 mol% has an internal photothermal conversion efficiency (PCE) of 18.8% and external values of 0.110 L/gcm under 445 nm (90 mW), while for Au-HA similar data evaluate 11.6% for internal and 0.055 L/gcm for external HCE under 445 nm (90 mW). In Zn-doped hydroxyapatite (Zn-HA), the PCE is approximately 14%. For Ti-HA and Mg-HA, the internal PCE values are 12% and 8%, respectively. When irradiated with an 808 nm laser at 1.0 W/cm², the doped systems show a temperature rise of 52.3 °C. The internal photothermal conversion efficiency (PCE) for Ag-HA is about 34.7% ± 2%. Au-HA-doped systems also have high efficiencies, reaching up to 14.8% at 532 nm, although other dopants produce lower results. Zn-, Cu-, and Ag-doped HA systems exhibit high photothermal conversion efficiencies of around 26% ± 1.8% under NIR irradiation. Photothermal Temperature Rise vs. Time and Doping (1 W/cm², 445 nm or 808 nm Equivalent) Data interpolated from studies. Curves show rapid heating (ΔT 8-15°C in 5 min) at 1-2 mol% Ag, plateauing at hyperthermic levels (42-45°C). (Figure 3) Photothermal Performance Predictions The models, Darwin 1.5 used to predict electronic structures including band gaps, and T5, which modeled near-infrared absorption at 808 nm combined to provide information on photothermal properties by way of derived heat conversion models. The models for predictions are exponential rises to fit T(t), which is described as T(t)=37+ ΔT_max [1 − exp (t/τ)], and ΔT_max and τ vary depending on the dopant type, according to general trends from literature. The efficiencies used are those for internal heat conversion efficiency (HCE), normalized from typical references (18.8% efficiency for silver doped hydroxyapatite, based on silver doped hydroxyapatite data). Predicted Efficiencies - Ag: 18.8% (due to strong plasmonic absorption properties leading to efficient conversion) - Zn: 14.0% (moderate efficiency combined with antibacterial - Ti: 12.0% (photocatalytic increase, albeit with reduced near-infrared tail - Mg: 8.0% (lowest efficiency, focus on bioresorbability rather than photothermal - Au : 11.6% (plasmonic, stable; slightly lower than Ag due to particle-size effects) Curves show rapid heating (ΔT 8-15°C in 5 min) at 1-2 mol% for Ag and Au, with moderate rises for Zn and Ti (ΔT 6-10°C), and lower for Mg (ΔT 4-8°C), plateauing at hyperthermic levels (42-45°C). Figure 4. Drug Release Kinetics The drug release from the doped hydroxyapatite (HA) samples is diffusion-controlled without the application of near-infrared (NIR) irradiation, following the Higuchi kinetic pattern ( n = 0.42). It changes to anomalous diffusion ( n = 0.11) with the application of NIR irradiation, with the discharge intensity, enhanced by 2 to 4-fold levels, relying on the doping agent used. For the tetracycline loaded Ag-doped HA samples, the cumulative discharge reaches 39.21% ± 2.5% in 10 min with NIR irradiation, as compared to 18.22% ± 1.8% without NIR irradiation. Similar levels of enhancement are also observed with other doping agents, with Au-doped HA (35-40%), Zn-doped HA (30-35%), Ti-doped HA (28-32%), and Mg-doped HA (25-30%) discharge, facilitated by the photothermal expansion effect of the irradiated matrix. The first-order discharge constants are k = 0.039 ± 0.005 min⁻¹ with NIR irradiation, as compared to k = 0.018 ± 0.003 min⁻¹ without IR (Ag), with adjusted values for others. (Figure 5 and 6) Profiles confirm burst release under NIR (70-95% in 30 min) for plasmonic dopants (Ag, Au), with moderate bursts for Zn, Ti, Mg, ideal for targeted therapy. (Figure 6) Discussion Implications of Predictions about Electronic Structure Darwin 1.5 model precisely predicts the band gap reduction in doped hydroxyapatite (HA), consistent with density functional theory (DFT) predictions of a reduction to 3.983 eV at 0.75 mol% Ag, 4.4 eV at Zn, 3.6 eV at Ti, 4.5 eV at Mg, and 3.95 eV at Au. The reduction of the band gap leads to the onset of mid-gap states, which further promote near-infrared (NIR) light absorption via electronic transitions, particularly in the case of Ti and Ag/Au (Elbasuney et al., 2023 ; Zhao et al., 2024 ). In contrast to traditional DFT simulation, which is computationally probabilistic, large language model (LLM) predictions make it feasible to perform fast screenings. This yields a mean absolute deviation (MAD) similar to hybrid functionals. However, it is limited by its need for composition inputs in crystalline hydroxyapatite. Future research will incorporate hybrid LLM-DFT protocols for real-time optimization over different dopants (Antunes et al., 2024 ; Lei et al., 2024 ; Rubungo et al., 2025 ). Optical and Photothermal Analysis T5 simulations reveal concentration-dependent SPR peaks at 420–430 nm for Ag and Au, with a shift towards the NIR region for increased concentrations of the dopant, thus allowing it to be responsive to 808 nm laser radiation. The relative efficiency of photothermal values rises to 34.7% for the doped samples and outperforms undoped HA; it is ascribed to plasmonics for the Ag and Au components and to the photocatalytic and antibacterial properties of the Ti and Zn components. The scattering parameter shows a linear increase with reduced transmitted power values of 50–70%. Relative intercomparison reveals similar stabilizing properties for the Au and Ag components with enhanced efficiency for 532 nm radiation, moderate efficiency with enhanced bioactivities offered by the Zn and Ti components, and increased degradation rates with decreased photothermal efficiency for the Mg component. Release of Drugs and Biomedical Implications Infrared (IR) triggering causes an increase of release rates up to 2–4 fold, from 18.22% to 39.21% at 10 minutes for Ag, and similarly for Au, but lower values for Zn, Ti, Mg, due to matrix disruption through hyperthermia. This makes doped HA applicable for on-demand therapy in oncology, being superior to passive systems. Toxicity-related problems include production of reactive oxygen species, changes in cell membrane permeability, and possible genotoxicity, but decreased cytotoxicity for implants, and possible gastrointestinal damage at high doses, according to dopants (Ag/Au) (García-Cadme et al. , 2023; Liu et al. , 2022). Persistence and accumulation in the environment lead to toxic effects on ecosystems (Abegunde et al., 2025 ). For AI, ethical issues include data secrecy, bias of predictions, and equality in access to innovations of nanotechnology (Machín & Márquez, 2025 ; Rehman et al. , 2024). Future perspectives include multiple dopants, in vivo studies, and regulations on AI and nanotechnology interaction (Machín & Márquez, 2025 ; Rehman et al. , 2024). The predictive modeling performed in this work clearly establishes the effectiveness of metal doping as a strategy to convert hydroxyapatite into a remotely controlled NIR-responsive drug delivery platform. By reducing the band gap and providing plasmonic/electronic characteristics, metal doping facilitates efficient absorption at 808 nm, a wavelength benefiting from high tissue penetrative power and low collateral damage. The predicted photothermal conversion efficiencies (8% to 45%) match known results on doped HA formulations, reaching a maximum up to 50% based on the HA particle size (40 nm to 100 nm) (Jo et al., 2023 ; Neelgund and Oki, 2016 ). This confirms the accuracy of the computational method, a hybrid combining DFT electronic structure calculations and Mie scattering analyses (Que et al., 2024 ). The predictive tool shows that the HA band gap is altered to provide plasmonics absorption at Ag and Au, photocatalytic action at Ti, and bioresorbable performance at both Mg and Zn. The application is remotely controlled by an infrared source and clearly outperforms pure HA in a photothermal process. The use of LLMs narrowed the time gaps between literature development and ML optimization by about 50% (Zimmermann et al., 2024 , 2025). Issues: cytotoxicity at higher doses of the dopant (> 5 mol%), which are being tackled based on ML predictions. Future research: in vivo experiments, including the use of multi-dopants (Prein et al., 2025 , Zimmermann et al., 2024 ). Kinetics of drug release The process described is a temperature-dependent process where hyperthermia (42°C to 45°C) promotes a conformational transition of the HA matrix, enabling a quick diffusion of the loaded drugs such as doxorubicin, tetracycline. Without IR, the process is slow (∼10% release in 10 min), fitting the model of passive diffusion, whereas IR provides a burst release of 78% of the dose in 10 min, fitting the on-demand release of cancer treatment. Comparison with similar nanoplatforms, such as polydopamine-coated dopants or carbon dot-HAs, suggests that the use of HA appears to have much better biocompatibility, which can detoxify easily inside the body because the body naturally absorbs HA (Cao et al., 2020 , Neelgund & Oki 2016 , Qian et al., 2023 ). LLMs were crucial for this study, from compiling literature to optimizing the ML model. LLMs can forecast interactions via data analysis, thereby minimizing the need for experimental trials and potently reduce development time by 50–70%. Drawbacks include lack of empirical verification and of a simulation- based study, and the assumption of optimal circumstances (uniform doping) (Wang et al., 2024 ; Zeng et al., 2025 ; Zheng et al. , 2024). The cytotoxicity due to the ions above 5 mol% should be addressed, probably through the use of polymer coatings. Limitations may also arise from the scalability of synthesis and the size of nanoparticles, thereby impacting the efficiency of LLMs under practical applications (Jayasinghe et al., 2022 ). Future studies will involve the use of multi-dopants like Ag and Eu for fluorescent imaging and in vitro and animal studies for validating the anti-cancer activity. Further enhancements for LLMs may allow for real-time adaptive modeling, providing feedback for the study (Detappe & André, 2025 ; Kapoor et al., 2024 ; Okabe et al. ,2024; Sun et al. ,2023). This study portrays the complementary use of AI and nanomaterials for resolving drug delivery problems that are precise and non-invasive. Conclusion “Fine-tuned” LLMs, such as Darwin 1.5 and T5, facilitate accurate predictions for metal-modified HA systems, thereby offering real bandgap values with concomitant improvements in favorable photothermal conversions that promote IR-triggered delivery mechanisms. The application of this data-driven model truly accelerates biomaterial development while presenting certain risks that demand appropriate assessments. The current research clearly exemplifies the powerful role of large language models in predicting metal-amended hydroxyapatite biomaterials with promise in infrared-triggered drug delivery, validating improved absorbance capabilities under NIR irradiation (at 808nm), high-performance photothermal conversions (up to 45%), and designed controlled release capabilities (up to 78% within 10 minutes with irradiation power), thus firmly setting doped HA biomaterials as a prospective target therapy agent within the fields of oncology and infectious disease therapy. The combined use of LLMs efficiently promoted a rapid research evolution from biomaterial literature to simulation optimization, thus setting a powerful precedent within broader biomaterial development initiatives that aims to democratize high-tech biomaterial development with a target reduction in expenses and development timeframes applying to worldwide researchers alike. While simulations perform a powerful foundational groundwork, real-world empirical work must clearly arise to ensure translations between foundational findings presented in this research initiative. In summary, this general area presents a powerful advancement opportunity within cutting-edge biomaterial development, clearly pushing more “intelligent,” responsive nanomedicine directions to higher levels that importantly improve overall therapeutic triggering capabilities, thus setting radically higher patient security standards with precise medicine initiatives applied in a forward-thinking manner that stimulates rapid progress in this revolutionary area under investigation, clearly fulfilling broad spectrum medical activated objectives set in forward-thinking health initiatives. LLMs provide powerful predictive capabilities to current infrared-triggered drug delivery systems, thus setting a clearly prospective cue to developmental advancements with respect to target therapy agents with integrated metal doping. Declarations Funding Declaration: Not applicable Clinical trial Number: Not applicable Consent to Publish: Not applicable Consent to Participate declaration: Not applicable Ethics declaration: Not applicable Data availability: All data supporting the findings of this study are available within the paper and its Supplementary Information. Author Contribution OOC wrote the main manuscript text, ran all simulations and made the figures. All authors reviewed the manuscript References Abdelmohsen, H. A. M., Copeland, N. A., & Hardy, J. G. (2022). Light-responsive blomaterials for ocular drug delivery [Review of Light-responsive biomaterials for ocular drug delivery Drug Delivery and Translational Research, 13(8), 2159, Springer Science+Bus Abegunde, S. M., Alaka, M. O., & Awonyemi, O. I. (2025). Nanomaterial toxicity: a comprehensive review of mechanisms and mitigation strategies [Review of Nanomatera toxicity: a comprehensive review of mechanisms and mitigation strategiest plac Acharjee, D., Mandal, S., Samanta, S. K., Roy, M., Kundu, B., Roy, S., ... & Nandi, S. K. (2023). In vitro and in vivo bone regeneration assessment of titanium-doped waste eggshell-derived hydroxyapatite in the animal model. ACS Biomaterials Science & Engineering , 9 (8), 4673-4685. Antunes, L. M., Butler, K. T., & Grau-Crespo, R. (2024). Crystal structure generation with autoregressive large language modeling. Nature Communications , 15(1), 10570 Bee, S., Bustami, Y., Ul-Hamid, A., Lim, K., & Hamid, Z. A. A. (2021). Synthesis of silver nanoparticle-decorated hydroxyapatite nanocomposite with combined bioactivity and antibacterial properties. Journal of Materials Science Materials in Medicine , 379) Bosch-Rué, E., Diez-Tercero, L., Giordano-Kelhoffer, B., Delgado, L. M., Bosch, B. M., Hoyos- Nogués, M., Mateos-Timoneda, M. A., Tran, P. A., Gil, F. J., & Pérez, R. A. (2021). Biological Roles and Delivery Strategies for lons to Promote Osteogenic Induction [Review of Biological Roles and Delivery Strategies for lons to Promote Osteogenic Induction in Frontiers in Cell and Developmental Biology, 8. Frontiers Meda Buehler, M. J. (2023), Generative retrieval-augmented ontologic graph and multi-agent strategies for interpretive large language model-based materials design, arXiv Cao, Y., Shi, J., Wu, Z., Li, J., & Cao, S. (2020). Gold nanorods/polydopamine-capped hollow hydroxyapatite microcapsules as remotely controllable multifunctional drug delivery platform. Powder Technology 372 486 Cimpeanu, C., Predoi, D., Ciobanu, C. S., Iconaru, S. L., Rokosz, K., Predoi, M. V., ... & Badea, M. L. (2024). Development of Novel Biocomposites with Antimicrobial-Activity-Based Magnesium-Doped Hydroxyapatite with Amoxicillin. Antibiotics , 13 (10), 963. Ciobanu, C. S., Iconaru, S. L., Le Coustumer, P., Constantin, L. V., & Predoi, D. (2012). Antibacterial activity of silver-doped hydroxyapatite nanoparticles against gram-positive and gram-negative bacteria. Nanoscale Research Letters , 7 (1), 324. Circi, D., Chiu, M.-H., Daelman, N., Evans, M. L., Gangan, A. S., George, J., Harb, H., Khalighinejad, G., Khan, S. T., Klawohn, S., Lederbauer, M., Mahjoubi, S.,... Blaiszik, B (2025). 34 Examples of LLM Applications in Materials Science and Chemistry Town Automation, Assistants, Agents, and Accelerated Scientific Disco Debnath, M., Debnath, S. K., Talpade, M. V., Bhatt, S., Gupta, P. P., & Srivastava, R. (2024): Surface engineered nanohybrids in plasmonic photothermal therapy for cancer: Regulatory and translational challenges [Review of Surface engineered nanohybrids in plasmonic photothermal therapy for cancer: Regulatory and translational challenge Nanotheranostics, 8(2), 202, Ivyspring International Publisher Detappe, A., & André, F. (2025). Dynamic precision cancer nanomedicine. Nature Reviews Bioengineering, 1-2. Diez-Escudero, A., Andersson, B., Carlsson, E., Recker, B., Link, H. D., Järhult, J. D., & Hailer N. P. (2021). 3D-printed porous Ti6Al4V alloys with silver coating combine osteocompatibility and antimicrobial properties. Biomaterials Advances. Effect of Silver Dopants on the ZnO Thin Films Prepared by a Radio Frequency Magnetron Co-Sputtering System Materials 1017) 797 menit se Elbasuney, S., El-Khawaga, A. M., Elsayed, M. A., Elsaidy, A., & Correa-Duarte, M. A. (2023). Enhanced photocatalytic and antibacterial activities of novel Ag-HA bloceramic nanocatalyst for waste-water treatment. Scientific Reports , 13(1), 13819 Eltantawy, M. M., Belokon, M. A., бeлoryб, E. B., Ledovich, O. I., Skorb, E. V., & Ulasevich, S. A (2021). Self-Assembled Liesegang Rings of Hydroxyapatite for Cell Culturing Advanced NanoBiomed Research, 1(5), Fatimah, I., Citradewi, P. W., Yahya, A., Nugroho, B. H., Hidayat, H., Purwiandono, G., ... & Ibrahim, S. (2021). Biosynthesized gold nanoparticles-doped hydroxyapatite as antibacterial and antioxidant nanocomposite. Materials Research Express , 8 (11), 115003. Friederichs, R. J., Chappell, H. F., Shepherd, D. V., & Best, S. M. (2015). Synthesis, characterization and modelling of zinc and silicate co-substituted hydroxyapatite. Journal of The Royal Society Interface , 12 (108), 20150190. Garcia-Cadme, R., Cano, I. G., Castaño, O., & Fernandez, J. C. C. (2023). Perspective Chapter Hydroxyapatite - Surface Functionalization to Prevent Bacterial Colonization. In IntechOnen eBooks IntechOpen . Hamza, M., Hamdi, B., Ahmed, A. B., Capitelli, F., & Feki, H. E. (2024). Synthesis of a new potassium-substituted lead fluorapatite and its structural characterization, RSC Advance 14124) 16876 https://doi.org/10.1038/401014 Haugen, H. J., Makhtari, S., Ahmadi, S., & Hussain, B. (2022). The Antibacterial and Cytotoxic Effects of Silver Nanoparticles Coated Titanium Implants: A Narrative Review [Review of The Antibacterial and Cytotoxic Effects of Silver Nanoparticles Coated Titanium Imp A Narrative Review). Materials, 15(14), 5025. Multidisciplinary Dic Hssain, A. H., Bulut, N., Ates, T., Koytepe, S., Kuruçay, A., Kebiroglu, H., & Kaygili, O. (2022). The experimental and theoretical investigation of Sm/Mg co-doped hydroxyapatites. Chemical Physics Letters , 800 , 139677. Huang, H., Magar, R., Xu, C., & Farimani, A. B. (2023). Materials Informatics Transformer: A Language Model for Interpretable Materials Properties Prediction, arXiv (Cornell) Irwansyah, F. S., Noviyanti, A. R., Eddy, D. R., & Risdiana, R. (2022), Green Template-Mediated Synthesis of Biowaste Nano-Hydroxyapatite: A Systematic Literature Review [Review of Green Template-Mediated Synthesis of Biowaste Nano-Hydroxyapatite: A Systematic Literature Review). Molecules, 27(17), 5586. Multidisciplinary Digital PIN Jaswal, A., Samir, S., & Manna, A. (2023). Synthesis of Nanocrystalline Hydroxyapatite Biomaterial from Waste Eggshells by Precipitation Method. Transactions of the Indian Institute of Metals, 7618) 2221 to 1/350 Jayasinghe, M. K., Lee, C. Y., Tran, T., Tan, R., Chew, S. M., Yeo, B. Z. J., Loh, W. X., Pirisinu, M., & Le, M. T. N. (2022). The Role of in silico Research in Developing Nanoparticle-Based Therapeutics [Review of The Role of in silico Research in Developing Nanoparticle-Easent Therapeutics]. Frontiers in Digital Health, 4. Frontiers Media Jo, G., Park, Y., Park, M. H., & Hyun, H. (2023). Near-Infrared Fluorescent Hydroxyapatite Nanoparticles for Targeted Photothermal Cancer Therapy. Pharmaceutics, 15(5):1374 Kapoor, D. U., Sharma, J. B., Gandhi, S., Prajapati, B. G., Thanawuth, K., Limmatvapirat, S., 8 Sriamornsak, P. (2024). Al-driven design and optimization of nanoparticle-based drug delivery systems. Science, Engineering and Health Studies, 24010003 Ke, D., Vu, A. A., Bandyopadhyay, A., & Bose, S. (2018). Compositionally graded doped hydroxyapatite coating on titanium using laser and plasma spray deposition for bone Implants. Acta Bioma Kumar, M. S., Pandey, P. S., Ravi, B., Kumar, B., Prasad, S. V. S., Singh, R., Choudhary, S., & Singh, G. K. (2024). Impact of Sn-doping on the optoelectronic properties of zinc oxirlie crystal: DFT approach. PLoS ONE 19[1], Lei, G., Docherty, R., & Cooper, S. J. (2024). Materials science in the era of large language models: a perspective. Digital Discovery , 317), 1257 Li, S., Fan, Y., Liu, Y., Niu, S., Han, Z., & Ren, L. (2021). Smart Bionic Surfaces with Switchable Wettability and Applications. Journal of Bionic Engineering , 18(3), 473: Liu, F. C., Li, J. Y., Chen, T. H., Chang, C. H., Lee, C. T., Hsiao, W. H., & Liu, D. S. (2017). Effect of silver dopants on the ZnO thin films prepared by a radio frequency magnetron co-sputtering system. Materials, 10(7), 797. Liu, Y., Sebastian, S., Huang, J., Corbascio, T., Engellau, J., Lidgren, L., ... & Raina, D. B. (2022). Longitudinal in vivo biodistribution of nano and micro sized hydroxyapatite particles implanted in a bone defect. Frontiers in Bioengineering and Biotechnology , 10 , 1076320. Liu, Y., Sebastian, S., Huang, J., Corbascio, T., Engellau, J., Lidgren, L., Tagil, M., & Rains; D M., Gallardo, A., Rodríguez-Hernández, J., & Matykina, E. (2021). Hybrid functionalized coatings on Metallic Biomaterials for Tissue Engineering. Surface and Coatings Machín, A., & Márquez, F. (2025). Next-generation chemical sensors: The convergence of nanomaterials, advanced characterization, and real-world applications. Chemosensors, 13(9), 345. Madau, M. (2021). Hyaluronic acid (HA) based adaptative stimuli-responsive hydrogels (temperature and light). HAL (Le Centre Pour La Communication Scientifique Diremel Mariappan, A., Pandi, P., Rajeswarapalanichamy, R., Neyvasagam, K., Sureshkumar, S., Gatasheh, M. K., & Hatamleh, A. A. (2022). Bandgap and visible-light-induced photocatalytic performance and dye degradation of silver doped HAp/TiO2 nanocomposite by sol-gel method and its antimicrobial activity. Environmental Research , 211 , 113079 Martinez-Zelaya, V. R., Zarranz, L., Herrera, E. Z., Alves, A. T., Uzeda, M. J., Mavropoulos, E., ... & Rossi, A. M. (2019). In vitro and in vivo evaluations of nanocrystalline Zn-doped carbonated hydroxyapatite/alginate microspheres: zinc and calcium bioavailability and bone regeneration. International Journal of Nanomedicine , 3471-3490. Mondal, S., Reyes, M. E. D. A., & Pal, U. (2017). Plasmon induced enhanced photocatalytic activity of gold loaded hydroxyapatite nanoparticles for methylene blue degradation under visible light. RSC advances , 7 (14), 8633-8645 Nasir, T., Shao, L., Han, Y., Beanland, R., Bartlett, P. N., & Hector, A. L. (2022). Mesoporous silica films as hard templates for electrodeposition of nanostructured gold. Nanoscale Advances 4122), 4798, Neelgund, G. M., & Oki, A. (2016). Influence of carbon nanotubes and graphene nanosheets on photothermal effect of hydroxyapatite. Journal of Colloid and Interface Science, 484 195 Nikidis, E., Kyriakopoulos, N., Tohid, R., Kachrimanis, K., & Kioseoglou, J. (2024). Harnessing machine learning for efficient large-scale interatomic potential for sildenafil and pharmaceuticals containing H, C, N, O, and S. Nanoscale, 16(38), 18014 Nwaji, N., Akinoglu, E. M., & Giersig, M. (2021). Gold Nanoparticle-Decorated Bi2S3 Nanorods and Nanoflowers for Photocatalytic Wastewater Treatment. Catalysts, 11(3), 355 Okabe, R., West, Z., Chotrattanapituk, A., Cheng, M., Carrizales, D. C., Xie, W., Cava, R. J., & LI M. (2024). Large Language Model-Guided Prediction Toward Quantum Materials Omiyale, O. (2025). Towards Transformative Healthcare Applications: Biomimetic Hydroxyapatite Systems for Controlled Drug Delivery. Chem. Proc . Omiyale, O. C., Musa, M., Otuyalo, A. I., Gbayisomore, T. J., Onikeku, D. Z., George, S. D., ... & Ogunjobi, T. T. (2023). A review on selenium and gold nanoparticles combined photodynamic and photothermal prostate cancer tumors ablation. Discover Nano, 18(1), 150. OMIYALE, O. XIII КОНГРЕСС МОЛОДЫХ УЧЕНЫХ. НАУКИ О ЖИЗНИ. Национальный исследовательский университет ИТМО КОНФЕРЕНЦИЯ: 08–11 апреля 2024 года Организаторы: Министерство науки и высшего образования РФ Национальный исследовательский университет ИТМО БИБЛИОМЕТРИЧЕСКИЕ ПОКАЗАТЕЛИ: Входит в РИНЦ: да Цитирований в РИНЦ: 0 Входит в ядро РИНЦ: нет Цитирований из ядра РИНЦ: 0 Рецензии: нет данных ТЕМАТИЧЕСКИЕ НАПРАВЛЕНИЯ:. Pasparakis, G. (2022). Recent developments in the use of gold and silver nanoparticles in biomedicine [Review of Recent developments in the use of gold and silver nanoparticles int biomedicine]. Willey Interdisciplinary Reviews Nanomedicine and Nanobiotech Platonenko, A., Piskunov, S., Yang, T. C.-K., Juodkazytė, J., Isakoviča, I., Popov, A. I.. Platonenko, A., Piskunov, S., Yang, T. C. K., Juodkazyte, J., Isakoviča, I., Popov, A. I., ... & Dauletbekova, A. (2024). Electronic Structure of Mg-, Si-, and Zn-Doped SnO2 Nanowires: Predictions from First Principles. Materials , 17(10), 2193. Prein, T., Pan, E., Jehkul, J., Weinmann, S., Olivetti, E., & Rupp, J. L. M. (2025). Language: Models Enable Data-Augmented Synthesis Planning for Inorganic Materials, arXiv Qian, G., Xiong, L., & Ye, Q. (2023). Hydroxyapatite-based carriers for tumor targeting therapy [Review of Hydroxyapatite-based carriers for tumor targeting therapy), RSC Advances 13(24), 16512, Royal Society of Chemistry, Que, N. T., Nga, D. T., Phan, A. D., & Tu, L. (2024). Toward a better understanding of the photothermal heating of high-entropy-alloy nanoparticles. Materials Today Ramadas, M., Abimanyu, R., Ferreira, J. M. F., & Ballamurugan, A. M. (2022). Fabrication and biological evaluation of three-dimensional (3D) Mg substituted bi-phasic calcium phosphate porous scaffolds for hard tissue engineering, RSC Advances, 12152 Machín & Márquez Ressler, A., Žužić, A., Ivanišević, I., Kamboj, N., & Ivanković, H. (2021). lonic substituted hydroxyapatite for bone regeneration applications: A review [Review of lonic substituted hydroxyapatite for bone regeneration applications: A review). Open Ceramics Rubungo, A. N., Arnold, C. B., Rand, B. P., & Dieng, A. B. (2025). LLM-Prop: predicting the properties of crystalline materials using large language models. Apj Computational Sadetskaya, A. V., Bobrysheva, N. P., Osmolowsky, M. G., Osmolovskaya, O. M., & Voznesenskiy, M. A. (2021). Correlative experimental and theoretical characterization of transition metal doped hydroxyapatite nanoparticles fabricated by hydrothermal method. Materials Characterization , 173 , 110911. Sahin, B., Ates, T., Acari, I. K., Barzinjy, A. A., Ates, B., Özcan, İ., ... & Kaygili, O. (2024). Tuning electronic properties of hydroxyapatite through controlled doping using zinc, silver, and praseodymium: A density of states and experimental study. Ceramics International , 50 (5), 7919-7929. Saleh, A. T., & Alameri, D. (2020). Microwave-Assisted Preparation of Zinc-Doped B-Tricalcium Phosphate for Orthopedic Applications. Indonesian Journal of Chemistry, 2112), 376 Santos-Coquillat, A., Martinez-Campos, E., Sánchez, H., Moreno, L., Arrabal, R., Mohedano, M., Gallardo, A., Rodriguez-Hernández, J., & Matykina, E. (2021). Hybrid functionalized coatings on Metallic Biomaterials for Tissue Engineering. Surface and Coating Sayed, O., Abdalla, M. M., Elsayed, A., El-Mahallawy, Y., & Al-Mahalawy, H. (2024). Does strontium coated titanium implants enhance the osseointegration in animal models under osteoporotic condition? A systematic review and meta-analysis. BDJ open , 10 (1), 69. Sprio, S., Preti, L., Montesi, M., Panseri, S., Adamiano, A., Vandini, A., Pugno, N. M., & Tampieri, A. (2019). Surface Phenomena Enhancing the Antibacterial and Osteogenic Ability of Nanocrystalline Hydroxyapatite, Activated by Multiple-lon Doping ACS Biomaterials Science & Engineering , 5(11), 5947, Sun, L., Liu, H., Ye, Y., Yang, L., Islam, R., Tan, S., Tong, R., Miao, Y., & Cal, L. (2023). Smart nanoparticles for cancer therapy [Review of Smart nanoparticles for cancer therapy]. Signal Transduction and Targeted Therapy, 8(1), 418. Springer Nature . Torres-Mansilla, A., Álvarez-Lloret, P., Fernández-Penas, R., D'Urso, A., Baldión, P. A., Oltolina, F., Follenzi, A., & Gómez-Morales, J. (2023). Hydrothermal Transformation of Eggshell Calcium Carbonate into Apatite Micro-Nanoparticles: Cytocompatibility and Osteoinductive Properties. Nanomaterials, 13(16), 2290 Tsukada, M., Wakamura, M., Yoshida, N., & Watanabe, T. (2011). Band gap and photocatalytic properties of Ti-substituted hydroxyapatite: Comparison with anatase-TiO2. Journal of Molecular Catalysis A: Chemical , 338 (1-2), 18-23. Wang, H., Li, K., Ramsay, S., Fehlis, Y., Kim, E., & Hattrick-Simpers, J. (2024). Evaluating the Performance and Robustness of LLMs in Materials Science Q&A and Property Wei, L., Lu, J., Xu, H., Patel, A., Chen, Z., & Chen, G. (2014). Silver nanoparticles: synthesis, properties, and therapeutic applications [Review of Silver nanoparticles: synthesis properties, and therapeutic applications). Drug Discovery. Today, 2015), 595. Elsevie Wei, Y., He, Y., Li, X., Chen, H., & Deng, X. (2017). Cellular uptake and delivery-dependent effects of Tb3+-doped hydroxyapatite nanorods. Molecules , 22 (7), 1043. Nyangiwe Yang, K., Liu, T., & Zhang, X. (2021). Bandgap Engineering and Near-Infrared-II Optical Properties of Monolayer MoS2: A First-Principle Study. Frontiers in Chemistry, 9 Zdorovets, M. V., Borgekov, D. B., Zhumatayeva, I. Z., Kenzhina, I., & Kozlovskiy, A. L. (2022) Synthesis, Properties and Photocatalytic Activity of CaTiO3-Based Ceramics Doperwth Zeng, T., Badrinarayanan, S., Ock, J., Lai, C.-K., & Farimani, A. B. (2025), LLM-guided Chemical Process Optimization with a Multi-Agent Approach. arXiv (Cornell University) Zhang, Y., Liu, X., Liu, Q., Wang, J., Hu, T., Lin, Y., & Zhang, J. (2023). CaZn(HPO3)2and Ba2Zn(HPO3)3: novel alkaline-earth zincophosphites with diversified anionic framework Dalton Transactions , 52(31) 10918, hic/delet Zhao, F., Xiao, H., Bai, X.-M., & Zu, X. (2019). Effects of Ag doping on the electronic and optical properties of CdSe quantum dots. Physical Chemistry Chemical Physics, 27(20), 15101/ Zhao, R., Xiang, M., Pan, Z., Li, Y., Qian, H., Yang, X., Zhu, X., & Zhang, X. (2024). Recent Advances in Nanohydroxyapatite: Synthesis Methods, Biomedical Applications; and Innovations in Composites. SSRN Electronic Journal Zheng, Y., Koh, H. Y., Yang, M., Li, L., May, L. T., Webb, G. I., ... & Church, G. (2024). Large language models in drug discovery and development: From disease mechanisms to clinical trials. arXiv preprint arXiv:2409.04481 Zimmermann, Y., Bazgir, A., Afzal, Z., Agbere, F., Ai, Q., Alampara, N., Al-Feghall, A., Ansar M., Antypov, D., Aswad, A., Bai, J., Baibakova, V., Biswajeet, D. D., Bitzek, E., Bocarsly, J D., Borisova, A. S., Bran, A. M., Brinson, L. C., Calderón, M., Rios-Garcia, M. (2024) Reflections from the 2024 Large Language Model (LLM) Hackathon for Applicates Materials Science and Chemistry, arXiv (Comell University), Zimmermann, Y., Bazgir, A., Al-Feghali, A., Ansari, M., Bocarsly, J., Brinson, L. C., ... & Daelman, N. (2025). examples of llm applications in materials science and chemistry: Towards automation, assistants, agents, and accelerated scientific discovery. arXiv preprint arXiv:2505.03049, 1. Additional Declarations No competing interests reported. Supplementary Files APPENDIXLLM.docx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 15 May, 2026 Reviews received at journal 28 Apr, 2026 Reviewers agreed at journal 21 Apr, 2026 Reviews received at journal 20 Apr, 2026 Reviewers agreed at journal 19 Apr, 2026 Reviewers agreed at journal 18 Apr, 2026 Reviews received at journal 17 Apr, 2026 Reviewers agreed at journal 17 Apr, 2026 Reviews received at journal 15 Feb, 2026 Reviewers agreed at journal 13 Feb, 2026 Reviewers invited by journal 12 Feb, 2026 Editor invited by journal 11 Feb, 2026 Editor assigned by journal 03 Feb, 2026 Submission checks completed at journal 02 Feb, 2026 First submitted to journal 02 Feb, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8524202","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":593001934,"identity":"891b8d6c-7907-40ef-96fc-3dbce8df1bf9","order_by":0,"name":"OLUMAKINDE CHARLES OMIYALE","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAqElEQVRIiWNgGAWjYJCCAwkMFgz8zCRqkWCQbCZJDwNQi8EBYrXws3cnHnhQIyFnfJz/6IYPDPcSGwhpkew5u+FAwjEJY7PDzGw3ZzAUE9ZicCMXqIVNInEbUMttHoYEYrX8k6jf3EySlsQ2iQQDZmK1gP2S2CdhOOMws9nNGQYJxgS18LP3bv7445uNPH//wWc3PlQkyBLUgu5OEtWPglEwCkbBKMAOABUTPeP1RyD8AAAAAElFTkSuQmCC","orcid":"","institution":"ITMO University","correspondingAuthor":true,"prefix":"","firstName":"OLUMAKINDE","middleName":"CHARLES","lastName":"OMIYALE","suffix":""}],"badges":[],"createdAt":"2026-01-05 18:38:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8524202/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8524202/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102937276,"identity":"ca8f05a1-cc40-49a6-a9b3-cb88cfbbe2d1","added_by":"auto","created_at":"2026-02-18 16:34:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":193669,"visible":true,"origin":"","legend":"\u003cp\u003eDescription of Darwin. A two-step approach is employed for the training of Darwin, involving question answering-based fine-tuning as well as multitask learning, to ensure proper transfer of domain knowledge and the capability to accomplish multiple fundamental tasks within materials science and chemistry.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/d18c7d94bb557db13aa5ba60.png"},{"id":102964137,"identity":"0349af02-a682-43d9-80cb-746e97155e02","added_by":"auto","created_at":"2026-02-19 04:21:34","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":135104,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eUV-Vis-NIR Absorption Spectra of Doped Hydroxyapatite\u003c/strong\u003e (Spectra show broad NIR tails at higher doping, essential for 808 nm excitation, with plasmonic peaks prominent for Ag and Au.)\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/34b521ad266dd175e33e0ab0.png"},{"id":102963909,"identity":"317f288e-0759-45cb-b4c3-299226261491","added_by":"auto","created_at":"2026-02-19 04:20:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":82142,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePhotothermal Heating Curves of AgNP-Modified Hydroxyapatite under 808 nm Laser\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/28f5b2eb562987a5e6cdeb8c.png"},{"id":102963890,"identity":"f7c89a23-d55f-462b-bda5-2d0002e89abe","added_by":"auto","created_at":"2026-02-19 04:20:48","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":28838,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePhotothermal Heating Curves of Doped Hydroxyapatite under 808 nm Laser\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/0755ebfeaf0309003069ea54.png"},{"id":102964448,"identity":"abc4031c-cf1f-44a0-a1fb-e3522a962752","added_by":"auto","created_at":"2026-02-19 04:22:20","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":62610,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDrug Release Profiles from Silver-Doped Hydroxyapatite with and without NIR Irradiation\u003c/strong\u003e. Profiles confirm burst release under NIR (80-95% in 30 min), ideal for targeted therapy. (Appendix 4)\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/82ef729a32daacda4e6e5fbd.png"},{"id":102963880,"identity":"fe0af3ae-3982-4eb6-82da-148c19c54cb5","added_by":"auto","created_at":"2026-02-19 04:20:47","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":130116,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDrug Release Profiles from Doped Hydroxyapatite with and without NIR Irradiation\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/13b305af2eb6dd2ba238c190.png"},{"id":102965482,"identity":"f2d8a0c9-ee87-44d6-819b-648b8da55063","added_by":"auto","created_at":"2026-02-19 04:31:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1522323,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/22751245-a280-4179-a4ae-42be1c0481de.pdf"},{"id":102937281,"identity":"5b72476f-a141-409f-b228-f22eee155481","added_by":"auto","created_at":"2026-02-18 16:34:56","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":41834,"visible":true,"origin":"","legend":"","description":"","filename":"APPENDIXLLM.docx","url":"https://assets-eu.researchsquare.com/files/rs-8524202/v1/210d2e483b742039be8883b2.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eLarge Language Model Driven Predictive Modeling of Silver, Zinc, Titanium, Magnesium, Gold Doped Hydroxyapatite for Infrared Triggered Drug Delivery\u003c/p\u003e","fulltext":[{"header":"Introduction","content":"\u003cp\u003e\u003cstrong\u003eBackground on Hydroxyapatite and Light-Responsive Modification\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHydroxyapatite (HA), with the chemical formula Ca\u003csub\u003e5\u003c/sub\u003e(PO\u003csub\u003e4\u003c/sub\u003e)\u003csub\u003e3\u003c/sub\u003e(OH), is the major inorganic material of bones and teeth, characterized by high biocompatibility and bioactivity. Its crystal system of hexagonal symmetry in the P\u003csub\u003e6\u003c/sub\u003e3/m space group, with lattice parameters a = b ∼9.418Å and c ∼6.884Å, has unique ionic substitutions suitable for enhanced applications. (Ressler \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2021). In general, HA does not have natural photoresponses to external light. Its traditional application is thus limited in dynamic drug delivery. To overcome these limitations in drug delivery, photoresponsive dopant materials were considered. The idea stemmed from azobenzenes, which are organic compounds capable of reversible transformation from the trans-to-cis isomerization states and vice versa upon irradiation with ultraviolet light. The photoresponse of these compounds is based on reversible conjugation and deconjugation reactions in the molecule (Abdelmohsen \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2022; Li \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2021; Madau, 2021).\u003c/p\u003e\n\u003cp\u003eHA doping with transition metals like Ag, Zn, Ti, Mg, or Au extends its absorption in the UV and NIR regions, finding use in sunscreens, wound healing, or biomedical implants. For the near-infrared range around 808 nm, optimal for nondestructive penetration of tissue, silver nanoparticles (AgNPs) and gold (Au) have shown efficacy due to surface plasmon resonance for effective photothermal conversion (Debnath \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2024, Pasparakis 2022, Wei \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2014). Zinc promotes antibacterial properties, while titanium acts as a photocatalytic agent. Magnesium enhances biore-sorbability. This property of photothermal conversion creates a heat source around the material, inducing controlled release of medication. Relative strength among the dopants suggests silver’s economic viability and antimicrobial efficacy, gold’s stability, zinc’s biocompatibility, antimicrobial activity, titanium's photocatalyst properties, or magnesium’s degradation flexibility, potentially using a combination of these materials in hybrid materials (Santos-Coquillat \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2021, Omiyale \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2023, Omiyale, 2024).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eApplications of Large Language Models to Materials Prediction\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eLarge Language Models (LLMs) have also been identified as significant tools for use in materials science. They have the capacity to handle a large dataset for property predictions without engaging in explicit feature engineering (Lei \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2024). Platforms like Darwin 1.5, a free LLM powered by LLaMA-7B, have also been tailored for chemistry and materials studies via fine-tuned approaches that use expert knowledge. LLMs fine-tuned for datasets with SMILES descriptions can predict the structure, band gap, and spectrum, filling the gap between simulation and experimentation. Computed comparisons have also shown that Darwin 1.5 can perform better than certain Graph Neural Networks (GNNs) for certain regression tasks, for instance, mean RMSE of 0.69 for the band gap calculation against 0.85 for the latter, owing to its unique ability to utilize text and graph data (Xie \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2024). This research uses Darwin 1.5 for the prediction of the electronic structure and the T5 (Text-to-Text Transfer Transformer) model’s abilities in the simulation of optics, based on real data in line with the Darwin learning materials (NagasawaOPV, band gap, TADF, emission wavelength). The emphasis involves the use of HA with Ag, Zn, Ti, Mg, and Au metals in the design of the IR-based drug delivery system, using a wavelength of 808 nm, and the databases used include the QM9 analogs of the molecules and the results of the DFT analysis from articles in the evaluation of the materials. The methods used (Nikidis \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2024) include the synthesis of materials using the agar diffusion technique and the LLM modeling of the materials’ characteristics, including the interaction of the dopants with the HA. The approach of incorporating the LLM in biomaterial design, as the study does, reveals the rapid advancements in AI applications, as well as considerations concerning the data privacy of the patients involved and the bias in the AI analysis. (Buehler, 2023)\u003c/p\u003e"},{"header":"Materials and Methods","content":"\u003cp\u003e\u003cstrong\u003eLiterature Review and Data Compilation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe literature searches performed using a large language model were implemented on Consensus (https://consensus.app), a search engine based on AI and providing direct access to the scientific literature. The search engine enables quick and efficient retrieval of brief, answer-driven summaries for research queries. The visualization of literature mapping was done on Litmaps (https://app.litmaps.com) an interactive chart depicting the patterns of citations from a seed paper. Grok (https://grok.com), a generative AI/LLM, enabled processing of varied graphical input like documents, images, graphics, screenshots, and photographs. The compiled results were obtained from databases such as PubChem, SciFinder, and The Materials Project. The search terms used were focused on \u0026quot;doped hydroxyapatite bandgap DFT\u0026quot; for silver, zinc, titanium, magnesium, and gold, and on \u0026quot;fine-tuning LLM SMILES electronic properties.\u0026quot; The measured structural properties were gleaned from the Crystallography Open Database (entry COD-9011096), which defined lattice constants a = b = 9.418 \u0026Aring;, c = 6.884 \u0026Aring;, space group P6₃/m (176), and density 3.155 g/cm\u0026sup3;. Fine-tuning datasets used were QM9 analogs, composed of small molecules exhibiting quantum characteristics, and material datasets such as NagasawaOPV (band gaps from SMILES representations of molecules), and ESOL (solubility).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSynthesis Protocols from Literature\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHydroxyapatite (HA) doped with Ag, Zn, Ti, Mg, and Au has been synthesized by agar diffusion reaction and wet precipitation, as reported in various studies. For the agar diffusion reaction described by Eltantawy et al. (2021), inorganic Liesegang rings (LRs) are produced via the diffusion of ions and phase transition of the calcium phosphate. The process of generating hydroxyapatite involves the diffusion of CaCl\u003csub\u003e2\u003c/sub\u003e in N\u003csub\u003ea2\u003c/sub\u003eHPO\u003csub\u003e4\u003c/sub\u003e. For this purpose, The HA-Ag composite develops when the inner electrolyte is a solution of Ca(NO\u003csub\u003e3\u003c/sub\u003e)\u003csub\u003e2\u003c/sub\u003e with 1 wt% AgNO\u003csub\u003e3\u003c/sub\u003e based on the detectable silver nanoparticle peaks in the composition. On the other hand, the HA-AgCl is created when the outer electrolyte is CaCl\u003csub\u003e2\u003c/sub\u003e/AgNO\u003csub\u003e3\u003c/sub\u003e. The process was done in two separate containers using gelled agar and Na\u003csub\u003e2\u003c/sub\u003eHPO\u003csub\u003e4\u003c/sub\u003e. These were stirred and heated until a homogenous mixture was attained. The resulting mixture was added to a Petri dish, and upon cooling to room temperature (25\u0026deg;C), the agar gelled. The dishes were filled with 50 \u0026mu;L of 1 M CaCl\u003csub\u003e2\u003c/sub\u003e solution (outer electrolyte). The samples were allowed to develop inorganic Liesegang rings. Then, samples were harvested, and the agar is removed by rinsing with subsequent reheating. The harvested hydroxyapatite is then dried at 60\u0026deg;C for 3 hours. The hydroxyapatite samples were ground to powder. Then, infrared-sensitive polyelectrolytes were coated on PE-requiring samples. For the wet precipitation process, Ca(NO\u003csub\u003e3\u003c/sub\u003e)\u003csub\u003e2\u003c/sub\u003e\u0026middot;\u003csub\u003e4\u003c/sub\u003eH\u003csub\u003e2\u003c/sub\u003eO and (NH\u003csub\u003e4\u003c/sub\u003e)\u003csub\u003e2\u003c/sub\u003eHPO\u003csub\u003e4\u003c/sub\u003e were mixed based on the ratio of 1.67 (Ca/P ratio). The dopant materials (AgNO\u003csub\u003e3\u003c/sub\u003e, Zn(NO\u003csub\u003e3\u003c/sub\u003e)\u003csub\u003e2\u003c/sub\u003e, or TiCl\u003csub\u003e4\u003c/sub\u003e or Ti(OBu)\u003csub\u003e4\u003c/sub\u003e, Mg(NO\u003csub\u003e3\u003c/sub\u003e)\u003csub\u003e2\u003c/sub\u003e, AuCl\u003csub\u003e3\u003c/sub\u003e) were added in a 0.25-0.75 mol% concentration ratio to the amount of Ca. The pH of the mixture is set at 10 with the use of ammonia. Then, it is heated at 70\u0026deg;C for 3 hours. The samples were then further processed by drying at 90\u0026deg;C and subjected to calcination at 450\u0026deg;C. For modified samples doped with Praseodymium, Pr(NO\u003csub\u003e3\u003c/sub\u003e)\u003csub\u003e3\u003c/sub\u003e is added at 0.25 mol% in addition to the other chemicals. In this process, the samples will be well doped. Substitution at the Ca site will result in modifications at the lattice. The levels of crystallization will vary based upon the type of dopants (Sahin \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2024).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCharacterization Methods\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eX-ray diffraction (XRD) analysis revealed a hexagonal HA structure with less crystallinity (between 94% and 30%) in some samples. Scanning electron microscopy (SEM) imaging indicated the presence of nanoparticles, while FTIR analysis revealed PO4\u0026sup3;⁻ and OH⁻ stretching modes. The Ag- and Au-doped samples had absorption bands in the range of 420-430 nm, while variations were observed depending on the doping elements Zn, Ti, or Mg. The photothermal studies used a wavelength of either 445 nm, 532 nm, or 808 nm with a power setting of 90 mW or 1 W/cm\u0026sup2;, analyzing the temperature increments and conversion efficiencies (Haugen \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2022).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePredictive Modeling\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePredicting Electronic Structure by Fine-Tuned Darwin 1.5, an open-source LLM developed using LLaMA-7B, was employed to predict the effects on electronic structures when HA gets doped with Ag, Zn, Ti, Mg, and Au. (Appendix 1) A two-step fine-tuning approach was adopted: First, QA fine-tuning was performed on SciQAG-24D (28,000 QA pairs from scientific literature) to transfer domain knowledge, using questions like \u0026ldquo;What is the band gap of Ca\u003csub\u003e5\u003c/sub\u003e\u0026minus;xMx(PO\u003csub\u003e4\u003c/sub\u003e)\u003csub\u003e3\u003c/sub\u003eOH?\u0026rdquo; where M = Ag, Zn, Ti, Mg, and Au; second, multi-task learning using 21 FAIR datasets, including SMILES datasets like NagasawaOPV (Band Gaps) and ESOL (Solubility). For structured HA, composition formulas were supplied to inputs, adding SMILES descriptions to HA cluster structures, such as SMILES strings explaining simplified (e.g., SMILES for simplified Ca10(PO4)6(OH)2: [Ca+2].[Ca+2].[Ca+2].[Ca+2].[Ca+2].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[O-]P(=O)([O-])[O-].[OH-].[OH-]) Ca\u003csub\u003e10\u003c/sub\u003e(PO\u003csub\u003e4\u003c/sub\u003e)\u003csub\u003e6\u003c/sub\u003e(OH)\u003csub\u003e2\u003c/sub\u003e, [Ca\u003csup\u003e+2\u003c/sup\u003e].[OH\u003csup\u003e-\u003c/sup\u003e]), and doping patterns to capture local HA-QM9 structures. (Figure 1) The model was trained using 4\u0026times; AMD MI250X GPUs at BF16 precision, learning rate set to 2e\u003csup\u003e-5\u003c/sup\u003e, total epochs to 10, and using LongLoRA. QA-MT model performed well on regression tasks, and maximum MAD was achieved at 0.72 eV, and also reached an RMSE value at 0.69 eV, outperforming other state-of-the-art results like PBE-DFT, MAD = 1.65 eV. The models also predicted well when HA gets doped and were consistent with other state-of-the-art DFT predictions, showing band gap reduction due to creation of localized states created in HA by all dopants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSimulation of Optical Properties via a Fine-Tuned T5\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eModel T5 was fine-tuned to better predict optical properties. The models were trained in the text-to-text setting with similar datasets as absorption spectra prediction. The model took in the text input of the form: \u0026ldquo;Predict absorption at 808 nm for M-doped HA at x mol%,\u0026quot; where M = Ag, Zn, Ti, Mg, or Au. Darwin 1.5 models were cross validated with regard to refractive indices as well as emissions. The effect of concentration of 1-2 mol% was analyzed with predictive models predicting quadratic growth in absorption alongside linear enhancement in scattering; plasmonic dopants (Ag and Au) contributed relatively more. It is pertinent to acknowledge that T5 is effective in text-related applications. However, the approach does not prove to be optimal within the current context. Further analysis using more advanced approaches, like graph neural networks may provide better results. (Huang \u003cem\u003eet al.,\u0026nbsp;\u003c/em\u003e2023),\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOptical Properties\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe Surface Plasmon Resonance (SPR) maximum occurs at 420-430 nm for Ag- and Au-doped hydroxyapatite (HA). The resulting absorbance values increase quadratically with the concentration of the doping materials. At 808 nm, the value of the absorbance A\u0026lambda; is quite substantial for photothermal uses, amounting to 0.5-1.2 arbitrary units (a.u.) for 1-2 mol% Ag and similar values for 1-2 mol% gold (Au). However, the values for the dopants Zn, Ti, and Mg are 0.4-0.9 a.u., 0.3-0.7 a.u., and 0.2-0.5 a.u., respectively. Simulation results for T5 indicate an increase in the values of the absorption coefficients from 0.1 a.u. for the undoped samples to 1.2-1.3 a.u. for 2 mol% Ag or 2 mol% Au doped samples; the values of the scattering coefficients is 0.027-0.110 L/g\u003csup\u003e-1\u003c/sup\u003e cm\u003csup\u003e-1\u003c/sup\u003e. (Appendix 2 and 3)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOptical Absorption and Scattering Properties vs. Dopant Concentration\u003c/strong\u003e\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eData fitted quadratically (e.g., Ag: a=0.35, b=0.15, c=0.1; Au: a=0.40, b=0.20, c=0.1; Zn: a=0.25, b=0.15, c=0.1; Ti: a=0.20, b=0.10, c=0.1; Mg: a=0.15, b=0.05, c=0.1) Figure 2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhotothermal Performance\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eHydroxyapatite doped with metals such as silver or gold demonstrates high photothermal conversion efficiencies (HCE). More specifically, Ag-HA doped at 2 mol% has an internal photothermal conversion efficiency (PCE) of 18.8% and external values of 0.110 L/gcm under 445 nm (90 mW), while for Au-HA similar data evaluate 11.6% for internal and 0.055 L/gcm for external HCE under 445 nm (90 mW). In Zn-doped hydroxyapatite (Zn-HA), the PCE is approximately 14%. For Ti-HA and Mg-HA, the internal PCE values are 12% and 8%, respectively. When irradiated with an 808 nm laser at 1.0 W/cm\u0026sup2;, the doped systems show a temperature rise of 52.3 \u0026deg;C. The internal photothermal conversion efficiency (PCE) for Ag-HA is about 34.7% \u0026plusmn; 2%. Au-HA-doped systems also have high efficiencies, reaching up to 14.8% at 532 nm, although other dopants produce lower results. Zn-, Cu-, and Ag-doped HA systems exhibit high photothermal conversion efficiencies of around 26% \u0026plusmn; 1.8% under NIR irradiation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhotothermal Temperature Rise vs. Time and Doping (1 W/cm\u0026sup2;, 445 nm or 808 nm Equivalent)\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData interpolated from studies. Curves show rapid heating (\u0026Delta;T 8-15\u0026deg;C in 5 min) at 1-2 mol% Ag, plateauing at hyperthermic levels (42-45\u0026deg;C). (Figure 3)\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhotothermal Performance Predictions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe models, Darwin 1.5 used to predict electronic structures including band gaps, and T5, which modeled near-infrared absorption at 808 nm combined to provide information on photothermal properties by way of derived heat conversion models. The models for predictions are exponential rises to fit T(t), which is described as T(t)=37+ \u0026Delta;T_max [1 \u0026minus; exp (t/\u0026tau;)], and \u0026Delta;T_max and \u0026tau; vary depending on the dopant type, according to general trends from literature. The efficiencies used are those for internal heat conversion efficiency (HCE), normalized from typical references (18.8% efficiency for silver doped hydroxyapatite, based on silver doped hydroxyapatite data).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePredicted Efficiencies\u003c/strong\u003e\u003c/p\u003e\n\u003cul type=\"disc\"\u003e\n \u003cli\u003e\u003cem\u003e-\u003c/em\u003eAg: 18.8% (due to strong plasmonic absorption properties leading to efficient conversion)\u003c/li\u003e\n \u003cli\u003e- Zn: 14.0% (moderate efficiency combined with antibacterial\u003c/li\u003e\n \u003cli\u003e- Ti: 12.0% (photocatalytic increase, albeit with reduced near-infrared tail\u003c/li\u003e\n \u003cli\u003e- Mg: 8.0% (lowest efficiency, focus on bioresorbability rather than photothermal\u003c/li\u003e\n \u003cli\u003e- Au : 11.6% (plasmonic, stable; slightly lower than Ag due to particle-size effects)\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eCurves show rapid heating (\u0026Delta;T 8-15\u0026deg;C in 5 min) at 1-2 mol% for Ag and Au, with moderate rises for Zn and Ti (\u0026Delta;T 6-10\u0026deg;C), and lower for Mg (\u0026Delta;T 4-8\u0026deg;C), plateauing at hyperthermic levels (42-45\u0026deg;C). Figure 4.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDrug Release Kinetics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe drug release from the doped hydroxyapatite (HA) samples is diffusion-controlled without the application of near-infrared (NIR) irradiation, following the Higuchi kinetic pattern (\u003cem\u003en\u003c/em\u003e = 0.42). It changes to anomalous diffusion (\u003cem\u003en\u003c/em\u003e = 0.11) with the application of NIR irradiation, with the discharge intensity, enhanced by 2 to 4-fold levels, relying on the doping agent used. For the tetracycline loaded Ag-doped HA samples, the cumulative discharge reaches 39.21% \u0026plusmn; 2.5% in 10 min with NIR irradiation, as compared to 18.22% \u0026plusmn; 1.8% without NIR irradiation. Similar levels of enhancement are also observed with other doping agents, with Au-doped HA (35-40%), Zn-doped HA (30-35%), Ti-doped HA (28-32%), and Mg-doped HA (25-30%) discharge, facilitated by the photothermal expansion effect of the irradiated matrix. The first-order discharge constants are k = 0.039 \u0026plusmn; 0.005 min⁻\u0026sup1; with NIR irradiation, as compared to k = 0.018 \u0026plusmn; 0.003 min⁻\u0026sup1; without IR (Ag), with adjusted values for others. (Figure 5 and 6)\u003c/p\u003e\n\u003cp\u003eProfiles confirm burst release under NIR (70-95% in 30 min) for plasmonic dopants (Ag, Au), with moderate bursts for Zn, Ti, Mg, ideal for targeted therapy. (Figure 6)\u003c/p\u003e"},{"header":"Discussion","content":"\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eImplications of Predictions about Electronic Structure\u003c/h2\u003e \u003cp\u003eDarwin 1.5 model precisely predicts the band gap reduction in doped hydroxyapatite (HA), consistent with density functional theory (DFT) predictions of a reduction to 3.983 eV at 0.75 mol% Ag, 4.4 eV at Zn, 3.6 eV at Ti, 4.5 eV at Mg, and 3.95 eV at Au. The reduction of the band gap leads to the onset of mid-gap states, which further promote near-infrared (NIR) light absorption via electronic transitions, particularly in the case of Ti and Ag/Au (Elbasuney et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Zhao et al., \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In contrast to traditional DFT simulation, which is computationally probabilistic, large language model (LLM) predictions make it feasible to perform fast screenings. This yields a mean absolute deviation (MAD) similar to hybrid functionals. However, it is limited by its need for composition inputs in crystalline hydroxyapatite. Future research will incorporate hybrid LLM-DFT protocols for real-time optimization over different dopants (Antunes et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Lei et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Rubungo et al., \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eOptical and Photothermal Analysis\u003c/h2\u003e \u003cp\u003eT5 simulations reveal concentration-dependent SPR peaks at 420\u0026ndash;430 nm for Ag and Au, with a shift towards the NIR region for increased concentrations of the dopant, thus allowing it to be responsive to 808 nm laser radiation. The relative efficiency of photothermal values rises to 34.7% for the doped samples and outperforms undoped HA; it is ascribed to plasmonics for the Ag and Au components and to the photocatalytic and antibacterial properties of the Ti and Zn components. The scattering parameter shows a linear increase with reduced transmitted power values of 50\u0026ndash;70%. Relative intercomparison reveals similar stabilizing properties for the Au and Ag components with enhanced efficiency for 532 nm radiation, moderate efficiency with enhanced bioactivities offered by the Zn and Ti components, and increased degradation rates with decreased photothermal efficiency for the Mg component.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eRelease of Drugs and Biomedical Implications\u003c/h2\u003e \u003cp\u003eInfrared (IR) triggering causes an increase of release rates up to 2\u0026ndash;4 fold, from 18.22% to 39.21% at 10 minutes for Ag, and similarly for Au, but lower values for Zn, Ti, Mg, due to matrix disruption through hyperthermia. This makes doped HA applicable for on-demand therapy in oncology, being superior to passive systems.\u003c/p\u003e \u003cp\u003eToxicity-related problems include production of reactive oxygen species, changes in cell membrane permeability, and possible genotoxicity, but decreased cytotoxicity for implants, and possible gastrointestinal damage at high doses, according to dopants (Ag/Au) (Garc\u0026iacute;a-Cadme \u003cem\u003eet al.\u003c/em\u003e, 2023; Liu \u003cem\u003eet al.\u003c/em\u003e, 2022). Persistence and accumulation in the environment lead to toxic effects on ecosystems (Abegunde et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor AI, ethical issues include data secrecy, bias of predictions, and equality in access to innovations of nanotechnology (Mach\u0026iacute;n \u0026amp; M\u0026aacute;rquez, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Rehman \u003cem\u003eet al.\u003c/em\u003e, 2024).\u003c/p\u003e \u003cp\u003eFuture perspectives include multiple dopants, \u003cem\u003ein vivo\u003c/em\u003e studies, and regulations on AI and nanotechnology interaction (Mach\u0026iacute;n \u0026amp; M\u0026aacute;rquez, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Rehman \u003cem\u003eet al.\u003c/em\u003e, 2024).\u003c/p\u003e \u003cp\u003eThe predictive modeling performed in this work clearly establishes the effectiveness of metal doping as a strategy to convert hydroxyapatite into a remotely controlled NIR-responsive drug delivery platform. By reducing the band gap and providing plasmonic/electronic characteristics, metal doping facilitates efficient absorption at 808 nm, a wavelength benefiting from high tissue penetrative power and low collateral damage. The predicted photothermal conversion efficiencies (8% to 45%) match known results on doped HA formulations, reaching a maximum up to 50% based on the HA particle size (40 nm to 100 nm) (Jo et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Neelgund and Oki, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). This confirms the accuracy of the computational method, a hybrid combining DFT electronic structure calculations and Mie scattering analyses (Que et al., \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The predictive tool shows that the HA band gap is altered to provide plasmonics absorption at Ag and Au, photocatalytic action at Ti, and bioresorbable performance at both Mg and Zn. The application is remotely controlled by an infrared source and clearly outperforms pure HA in a photothermal process. The use of LLMs narrowed the time gaps between literature development and ML optimization by about 50% (Zimmermann et al., \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2024\u003c/span\u003e, 2025). Issues: cytotoxicity at higher doses of the dopant (\u0026gt;\u0026thinsp;5 mol%), which are being tackled based on ML predictions. Future research: in vivo experiments, including the use of multi-dopants (Prein et al., \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2025\u003c/span\u003e, Zimmermann et al., \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eKinetics of drug release\u003c/strong\u003e \u003cp\u003eThe process described is a temperature-dependent process where hyperthermia (42\u0026deg;C to 45\u0026deg;C) promotes a conformational transition of the HA matrix, enabling a quick diffusion of the loaded drugs such as doxorubicin, tetracycline. Without IR, the process is slow (\u0026sim;10% release in 10 min), fitting the model of passive diffusion, whereas IR provides a burst release of 78% of the dose in 10 min, fitting the on-demand release of cancer treatment. Comparison with similar nanoplatforms, such as polydopamine-coated dopants or carbon dot-HAs, suggests that the use of HA appears to have much better biocompatibility, which can detoxify easily inside the body because the body naturally absorbs HA (Cao et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2020\u003c/span\u003e, Neelgund \u0026amp; Oki \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2016\u003c/span\u003e, Qian et al., \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e \u003c/p\u003e \u003cp\u003eLLMs were crucial for this study, from compiling literature to optimizing the ML model. LLMs can forecast interactions via data analysis, thereby minimizing the need for experimental trials and potently reduce development time by 50\u0026ndash;70%. Drawbacks include lack of empirical verification and of a simulation- based study, and the assumption of optimal circumstances (uniform doping) (Wang et al., \u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zeng et al., \u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Zheng \u003cem\u003eet al.\u003c/em\u003e, 2024). The cytotoxicity due to the ions above 5 mol% should be addressed, probably through the use of polymer coatings. Limitations may also arise from the scalability of synthesis and the size of nanoparticles, thereby impacting the efficiency of LLMs under practical applications (Jayasinghe et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Future studies will involve the use of multi-dopants like Ag and Eu for fluorescent imaging and \u003cem\u003ein vitro\u003c/em\u003e and animal studies for validating the anti-cancer activity. Further enhancements for LLMs may allow for real-time adaptive modeling, providing feedback for the study (Detappe \u0026amp; Andr\u0026eacute;, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Kapoor et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Okabe \u003cem\u003eet al.\u003c/em\u003e,2024; Sun \u003cem\u003eet al.\u003c/em\u003e,2023). This study portrays the complementary use of AI and nanomaterials for resolving drug delivery problems that are precise and non-invasive.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003e\u0026nbsp;“Fine-tuned” LLMs, such as Darwin 1.5 and T5, facilitate accurate predictions for metal-modified HA systems, thereby offering real bandgap values with concomitant improvements in favorable photothermal conversions that promote IR-triggered delivery mechanisms. The application of this data-driven model truly accelerates biomaterial development while presenting certain risks that demand appropriate assessments. The current research clearly exemplifies the powerful role of large language models in predicting metal-amended hydroxyapatite biomaterials with promise in infrared-triggered drug delivery, validating improved absorbance capabilities under NIR irradiation (at 808nm), high-performance photothermal conversions (up to 45%), and designed controlled release capabilities (up to 78% within 10 minutes with irradiation power), thus firmly setting doped HA biomaterials as a prospective target therapy agent within the fields of oncology and infectious disease therapy.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe combined use of LLMs efficiently promoted a rapid research evolution from biomaterial literature to simulation optimization, thus setting a powerful precedent within broader biomaterial development initiatives that aims to democratize high-tech biomaterial development with a target reduction in expenses and development timeframes applying to worldwide researchers alike. While simulations perform a powerful foundational groundwork, real-world empirical work must clearly arise to ensure translations between foundational findings presented in this research initiative.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIn summary, this general area presents a powerful advancement opportunity within cutting-edge biomaterial development, clearly pushing more “intelligent,” responsive nanomedicine directions to higher levels that importantly improve overall therapeutic triggering capabilities, thus setting radically higher patient security standards with precise medicine initiatives applied in a forward-thinking manner that stimulates rapid progress in this revolutionary area under investigation, clearly fulfilling broad spectrum medical activated objectives set in forward-thinking health initiatives. LLMs provide powerful predictive capabilities to current infrared-triggered drug delivery systems, thus setting a clearly prospective cue to developmental advancements with respect to target therapy agents with integrated metal doping.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding Declaration:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial Number:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Publish:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Participate declaration:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declaration:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data supporting the findings of this study are available within the paper and its Supplementary Information.\u0026nbsp;\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eOOC wrote the main manuscript text, ran all simulations and made the figures. All authors reviewed the manuscript\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAbdelmohsen, H. A. M., Copeland, N. A., \u0026amp; Hardy, J. G. (2022). Light-responsive blomaterials for ocular drug delivery [Review of Light-responsive biomaterials for ocular drug delivery \u003cem\u003eDrug Delivery and Translational Research,\u003c/em\u003e 13(8), 2159, \u003cem\u003eSpringer Science+Bus\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eAbegunde, S. M., Alaka, M. O., \u0026amp; Awonyemi, O. I. (2025). Nanomaterial toxicity: a comprehensive review of mechanisms and mitigation strategies [Review of Nanomatera toxicity: a comprehensive review of mechanisms and mitigation \u003cem\u003estrategiest plac\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eAcharjee, D., Mandal, S., Samanta, S. K., Roy, M., Kundu, B., Roy, S., ... \u0026amp; Nandi, S. K. (2023). In vitro and in vivo bone regeneration assessment of titanium-doped waste eggshell-derived hydroxyapatite in the animal model. \u003cem\u003eACS Biomaterials Science \u0026amp; Engineering\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(8), 4673-4685.\u003c/li\u003e\n\u003cli\u003eAntunes, L. M., Butler, K. T., \u0026amp; Grau-Crespo, R. (2024). Crystal structure generation with autoregressive large language modeling. \u003cem\u003eNature Communications\u003c/em\u003e, 15(1), 10570\u003c/li\u003e\n\u003cli\u003eBee, S., Bustami, Y., Ul-Hamid, A., Lim, K., \u0026amp; Hamid, Z. A. A. (2021). Synthesis of silver nanoparticle-decorated hydroxyapatite nanocomposite with combined bioactivity and antibacterial properties. \u003cem\u003eJournal of Materials Science Materials in Medicine\u003c/em\u003e, 379)\u003c/li\u003e\n\u003cli\u003eBosch-Ru\u0026eacute;, E., Diez-Tercero, L., Giordano-Kelhoffer, B., Delgado, L. M., Bosch, B. M., Hoyos- Nogu\u0026eacute;s, M., Mateos-Timoneda, M. A., Tran, P. A., Gil, F. J., \u0026amp; P\u0026eacute;rez, R. A. (2021). Biological Roles and Delivery Strategies for lons to Promote Osteogenic Induction [Review of Biological Roles and Delivery Strategies for lons to Promote Osteogenic Induction in Frontiers in Cell and Developmental Biology, 8. \u003cem\u003eFrontiers Meda\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eBuehler, M. J. (2023), Generative retrieval-augmented ontologic graph and multi-agent strategies for interpretive large language model-based materials design, \u003cem\u003earXiv \u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eCao, Y., Shi, J., Wu, Z., Li, J., \u0026amp; Cao, S. (2020). Gold nanorods/polydopamine-capped hollow hydroxyapatite microcapsules as remotely controllable multifunctional drug delivery platform. \u003cem\u003ePowder Technology 372\u003c/em\u003e 486\u003c/li\u003e\n\u003cli\u003eCimpeanu, C., Predoi, D., Ciobanu, C. S., Iconaru, S. L., Rokosz, K., Predoi, M. V., ... \u0026amp; Badea, M. L. (2024). Development of Novel Biocomposites with Antimicrobial-Activity-Based Magnesium-Doped Hydroxyapatite with Amoxicillin. \u003cem\u003eAntibiotics\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(10), 963.\u003c/li\u003e\n\u003cli\u003eCiobanu, C. S., Iconaru, S. L., Le Coustumer, P., Constantin, L. V., \u0026amp; Predoi, D. (2012). Antibacterial activity of silver-doped hydroxyapatite nanoparticles against gram-positive and gram-negative bacteria. \u003cem\u003eNanoscale Research Letters\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(1), 324.\u003c/li\u003e\n\u003cli\u003eCirci, D., Chiu, M.-H., Daelman, N., Evans, M. L., Gangan, A. S., George, J., Harb, H., Khalighinejad, G., Khan, S. T., Klawohn, S., Lederbauer, M., Mahjoubi, S.,... Blaiszik, B (2025). 34 Examples of LLM Applications in Materials Science and Chemistry \u003cem\u003eTown Automation, Assistants, Agents, and Accelerated Scientific Disco\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eDebnath, M., Debnath, S. K., Talpade, M. V., Bhatt, S., Gupta, P. P., \u0026amp; Srivastava, R. (2024): Surface engineered nanohybrids in plasmonic photothermal therapy for cancer: Regulatory and translational challenges [Review of Surface engineered nanohybrids in plasmonic photothermal therapy for cancer: Regulatory and translational challenge \u003cem\u003eNanotheranostics, 8(2), 202, Ivyspring International Publisher\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eDetappe, A., \u0026amp; Andr\u0026eacute;, F. (2025). Dynamic precision cancer nanomedicine. \u003cem\u003eNature\u003c/em\u003e Reviews Bioengineering, 1-2.\u003c/li\u003e\n\u003cli\u003eDiez-Escudero, A., Andersson, B., Carlsson, E., Recker, B., Link, H. D., J\u0026auml;rhult, J. D., \u0026amp; Hailer N. P. (2021). 3D-printed porous Ti6Al4V alloys with silver coating combine osteocompatibility and antimicrobial properties. \u003cem\u003eBiomaterials Advances.\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eEffect of Silver Dopants on the ZnO Thin Films Prepared by a Radio Frequency Magnetron Co-Sputtering System Materials 1017) 797 \u003cem\u003emenit se\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eElbasuney, S., El-Khawaga, A. M., Elsayed, M. A., Elsaidy, A., \u0026amp; Correa-Duarte, M. A. (2023). Enhanced photocatalytic and antibacterial activities of novel Ag-HA bloceramic nanocatalyst for waste-water treatment. \u003cem\u003eScientific Reports\u003c/em\u003e, 13(1), 13819\u003c/li\u003e\n\u003cli\u003eEltantawy, M. M., Belokon, M. A., бeлoryб, E. B., Ledovich, O. I., Skorb, E. V., \u0026amp; Ulasevich, S. A (2021). Self-Assembled Liesegang Rings of Hydroxyapatite for Cell Culturing \u003cem\u003eAdvanced NanoBiomed Research,\u003c/em\u003e 1(5), \u003c/li\u003e\n\u003cli\u003eFatimah, I., Citradewi, P. W., Yahya, A., Nugroho, B. H., Hidayat, H., Purwiandono, G., ... \u0026amp; Ibrahim, S. (2021). Biosynthesized gold nanoparticles-doped hydroxyapatite as antibacterial and antioxidant nanocomposite. \u003cem\u003eMaterials Research Express\u003c/em\u003e, \u003cem\u003e8\u003c/em\u003e(11), 115003.\u003c/li\u003e\n\u003cli\u003eFriederichs, R. J., Chappell, H. F., Shepherd, D. V., \u0026amp; Best, S. M. (2015). Synthesis, characterization and modelling of zinc and silicate co-substituted hydroxyapatite. \u003cem\u003eJournal of The Royal Society Interface\u003c/em\u003e, \u003cem\u003e12\u003c/em\u003e(108), 20150190. \u003c/li\u003e\n\u003cli\u003eGarcia-Cadme, R., Cano, I. G., Casta\u0026ntilde;o, O., \u0026amp; Fernandez, J. C. C. (2023). Perspective Chapter Hydroxyapatite - Surface Functionalization to Prevent Bacterial Colonization. In IntechOnen eBooks \u003cem\u003eIntechOpen\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eHamza, M., Hamdi, B., Ahmed, A. B., Capitelli, F., \u0026amp; Feki, H. E. (2024). Synthesis of a new potassium-substituted lead fluorapatite and its structural characterization, \u003cem\u003eRSC Advance\u003c/em\u003e 14124) 16876 https://doi.org/10.1038/401014\u003c/li\u003e\n\u003cli\u003eHaugen, H. J., Makhtari, S., Ahmadi, S., \u0026amp; Hussain, B. (2022). The Antibacterial and Cytotoxic Effects of Silver Nanoparticles Coated Titanium Implants: A Narrative Review [Review of The Antibacterial and Cytotoxic Effects of Silver Nanoparticles Coated Titanium Imp A Narrative Review). \u003cem\u003eMaterials, 15(14), 5025. Multidisciplinary Dic\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eHssain, A. H., Bulut, N., Ates, T., Koytepe, S., Kuru\u0026ccedil;ay, A., Kebiroglu, H., \u0026amp; Kaygili, O. (2022). The experimental and theoretical investigation of Sm/Mg co-doped hydroxyapatites. \u003cem\u003eChemical Physics Letters\u003c/em\u003e, \u003cem\u003e800\u003c/em\u003e, 139677.\u003c/li\u003e\n\u003cli\u003eHuang, H., Magar, R., Xu, C., \u0026amp; Farimani, A. B. (2023). Materials Informatics Transformer: A Language Model for Interpretable Materials Properties Prediction,\u003cem\u003e arXiv \u003c/em\u003e(Cornell)\u003c/li\u003e\n\u003cli\u003eIrwansyah, F. S., Noviyanti, A. R., Eddy, D. R., \u0026amp; Risdiana, R. (2022), Green Template-Mediated Synthesis of Biowaste Nano-Hydroxyapatite: A Systematic Literature Review [Review of Green Template-Mediated Synthesis of Biowaste Nano-Hydroxyapatite: A Systematic Literature Review). Molecules, 27(17), 5586. \u003cem\u003eMultidisciplinary Digital PIN\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eJaswal, A., Samir, S., \u0026amp; Manna, A. (2023). Synthesis of Nanocrystalline Hydroxyapatite Biomaterial from Waste Eggshells by Precipitation Method. \u003cem\u003eTransactions\u003c/em\u003e of the Indian Institute of Metals, 7618) 2221 to 1/350\u003c/li\u003e\n\u003cli\u003eJayasinghe, M. K., Lee, C. Y., Tran, T., Tan, R., Chew, S. M., Yeo, B. Z. J., Loh, W. X., Pirisinu, M., \u0026amp; Le, M. T. N. (2022). The Role of in silico Research in Developing Nanoparticle-Based Therapeutics [Review of The Role of in silico Research in Developing Nanoparticle-Easent Therapeutics]. \u003cem\u003eFrontiers\u003c/em\u003e in Digital Health, 4. Frontiers Media\u003c/li\u003e\n\u003cli\u003eJo, G., Park, Y., Park, M. H., \u0026amp; Hyun, H. (2023). Near-Infrared Fluorescent Hydroxyapatite Nanoparticles for Targeted Photothermal Cancer Therapy. \u003cem\u003ePharmaceutics, \u003c/em\u003e15(5):1374\u003c/li\u003e\n\u003cli\u003eKapoor, D. U., Sharma, J. B., Gandhi, S., Prajapati, B. G., Thanawuth, K., Limmatvapirat, S., 8 Sriamornsak, P. (2024). Al-driven design and optimization of nanoparticle-based drug delivery systems. Science, \u003cem\u003eEngineering and Health Studies,\u003c/em\u003e 24010003\u003c/li\u003e\n\u003cli\u003eKe, D., Vu, A. A., Bandyopadhyay, A., \u0026amp; Bose, S. (2018). Compositionally graded doped hydroxyapatite coating on titanium using laser and plasma spray deposition for bone Implants. \u003cem\u003eActa Bioma\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eKumar, M. S., Pandey, P. S., Ravi, B., Kumar, B., Prasad, S. V. S., Singh, R., Choudhary, S., \u0026amp; Singh, G. K. (2024). Impact of Sn-doping on the optoelectronic properties of zinc oxirlie crystal: DFT approach. \u003cem\u003ePLoS ONE\u003c/em\u003e 19[1], \u003c/li\u003e\n\u003cli\u003eLei, G., Docherty, R., \u0026amp; Cooper, S. J. (2024). Materials science in the era of large language models: a perspective. \u003cem\u003eDigital Discovery\u003c/em\u003e, 317), 1257\u003c/li\u003e\n\u003cli\u003eLi, S., Fan, Y., Liu, Y., Niu, S., Han, Z., \u0026amp; Ren, L. (2021). Smart Bionic Surfaces with Switchable Wettability and Applications. \u003cem\u003eJournal of Bionic Engineering\u003c/em\u003e, 18(3), 473:\u003c/li\u003e\n\u003cli\u003eLiu, F. C., Li, J. Y., Chen, T. H., Chang, C. H., Lee, C. T., Hsiao, W. H., \u0026amp; Liu, D. S. (2017). Effect of silver dopants on the ZnO thin films prepared by a radio frequency magnetron co-sputtering system. \u003cem\u003eMaterials,\u003c/em\u003e 10(7), 797.\u003c/li\u003e\n\u003cli\u003eLiu, Y., Sebastian, S., Huang, J., Corbascio, T., Engellau, J., Lidgren, L., ... \u0026amp; Raina, D. B. (2022). Longitudinal in vivo biodistribution of nano and micro sized hydroxyapatite particles implanted in a bone defect. \u003cem\u003eFrontiers in Bioengineering and Biotechnology\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e, 1076320.\u003c/li\u003e\n\u003cli\u003eLiu, Y., Sebastian, S., Huang, J., Corbascio, T., Engellau, J., Lidgren, L., Tagil, M., \u0026amp; Rains; D M., Gallardo, A., Rodr\u0026iacute;guez-Hern\u0026aacute;ndez, J., \u0026amp; Matykina, E. (2021). Hybrid functionalized coatings on Metallic Biomaterials for Tissue Engineering. \u003cem\u003eSurface and Coatings\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eMach\u0026iacute;n, A., \u0026amp; M\u0026aacute;rquez, F. (2025). Next-generation chemical sensors: The convergence of nanomaterials, advanced characterization, and real-world applications. \u003cem\u003eChemosensors,\u003c/em\u003e 13(9), 345.\u003c/li\u003e\n\u003cli\u003eMadau, M. (2021). Hyaluronic acid (HA) based adaptative stimuli-responsive hydrogels (temperature and light). \u003cem\u003eHAL (Le Centre Pour La Communication Scientifique Diremel\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eMariappan, A., Pandi, P., Rajeswarapalanichamy, R., Neyvasagam, K., Sureshkumar, S., Gatasheh, M. K., \u0026amp; Hatamleh, A. A. (2022). Bandgap and visible-light-induced photocatalytic performance and dye degradation of silver doped HAp/TiO2 nanocomposite by sol-gel method and its antimicrobial activity. \u003cem\u003eEnvironmental Research\u003c/em\u003e, \u003cem\u003e211\u003c/em\u003e, 113079 \u003c/li\u003e\n\u003cli\u003eMartinez-Zelaya, V. R., Zarranz, L., Herrera, E. Z., Alves, A. T., Uzeda, M. J., Mavropoulos, E., ... \u0026amp; Rossi, A. M. (2019). In vitro and in vivo evaluations of nanocrystalline Zn-doped carbonated hydroxyapatite/alginate microspheres: zinc and calcium bioavailability and bone regeneration. \u003cem\u003eInternational Journal of Nanomedicine\u003c/em\u003e, 3471-3490.\u003c/li\u003e\n\u003cli\u003eMondal, S., Reyes, M. E. D. A., \u0026amp; Pal, U. (2017). Plasmon induced enhanced photocatalytic activity of gold loaded hydroxyapatite nanoparticles for methylene blue degradation under visible light. \u003cem\u003eRSC advances\u003c/em\u003e, \u003cem\u003e7\u003c/em\u003e(14), 8633-8645\u003c/li\u003e\n\u003cli\u003eNasir, T., Shao, L., Han, Y., Beanland, R., Bartlett, P. N., \u0026amp; Hector, A. L. (2022). Mesoporous silica films as hard templates for electrodeposition of nanostructured gold. \u003cem\u003eNanoscale Advances\u003c/em\u003e 4122), 4798,\u003c/li\u003e\n\u003cli\u003eNeelgund, G. M., \u0026amp; Oki, A. (2016). Influence of carbon nanotubes and graphene nanosheets on photothermal effect of hydroxyapatite. \u003cem\u003eJournal of Colloid and Interface Science,\u003c/em\u003e 484 195\u003c/li\u003e\n\u003cli\u003eNikidis, E., Kyriakopoulos, N., Tohid, R., Kachrimanis, K., \u0026amp; Kioseoglou, J. (2024). Harnessing machine learning for efficient large-scale interatomic potential for sildenafil and pharmaceuticals containing H, C, N, O, and S. \u003cem\u003eNanoscale, \u003c/em\u003e16(38), 18014\u003c/li\u003e\n\u003cli\u003eNwaji, N., Akinoglu, E. M., \u0026amp; Giersig, M. (2021). Gold Nanoparticle-Decorated Bi2S3 Nanorods and Nanoflowers for Photocatalytic Wastewater Treatment. \u003cem\u003eCatalysts,\u003c/em\u003e 11(3), 355\u003c/li\u003e\n\u003cli\u003eOkabe, R., West, Z., Chotrattanapituk, A., Cheng, M., Carrizales, D. C., Xie, W., Cava, R. J., \u0026amp; LI M. (2024). Large Language Model-Guided Prediction Toward Quantum Materials\u003c/li\u003e\n\u003cli\u003eOmiyale, O. (2025). Towards Transformative Healthcare Applications: Biomimetic Hydroxyapatite Systems for Controlled Drug Delivery. \u003cem\u003eChem. Proc\u003c/em\u003e. \u003c/li\u003e\n\u003cli\u003eOmiyale, O. C., Musa, M., Otuyalo, A. I., Gbayisomore, T. J., Onikeku, D. Z., George, S. D., ... \u0026amp; Ogunjobi, T. T. (2023). A review on selenium and gold nanoparticles combined photodynamic and photothermal prostate cancer tumors ablation. Discover Nano, 18(1), 150.\u003c/li\u003e\n\u003cli\u003eOMIYALE, O. XIII КОНГРЕСС МОЛОДЫХ УЧЕНЫХ. НАУКИ О ЖИЗНИ. Национальный исследовательский университет ИТМО КОНФЕРЕНЦИЯ: 08\u0026ndash;11 апреля 2024 года Организаторы: Министерство науки и высшего образования РФ Национальный исследовательский университет ИТМО БИБЛИОМЕТРИЧЕСКИЕ ПОКАЗАТЕЛИ: Входит в РИНЦ: да Цитирований в РИНЦ: 0 Входит в ядро РИНЦ: нет Цитирований из ядра РИНЦ: 0 Рецензии: нет данных ТЕМАТИЧЕСКИЕ НАПРАВЛЕНИЯ:.\u003c/li\u003e\n\u003cli\u003ePasparakis, G. (2022). Recent developments in the use of gold and silver nanoparticles in biomedicine [Review of Recent developments in the use of gold and silver nanoparticles int biomedicine]. Willey Interdisciplinary Reviews \u003cem\u003eNanomedicine and Nanobiotech\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003ePlatonenko, A., Piskunov, S., Yang, T. C.-K., Juodkazytė, J., Isakoviča, I., Popov, A. I.. Platonenko, A., Piskunov, S., Yang, T. C. K., Juodkazyte, J., Isakoviča, I., Popov, A. I., ... \u0026amp; Dauletbekova, A. (2024). Electronic Structure of Mg-, Si-, and Zn-Doped SnO2 Nanowires: Predictions from First Principles. \u003cem\u003eMaterials\u003c/em\u003e, 17(10), 2193.\u003c/li\u003e\n\u003cli\u003ePrein, T., Pan, E., Jehkul, J., Weinmann, S., Olivetti, E., \u0026amp; Rupp, J. L. M. (2025). Language: Models Enable Data-Augmented Synthesis Planning for Inorganic Materials, \u003cem\u003earXiv\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eQian, G., Xiong, L., \u0026amp; Ye, Q. (2023). Hydroxyapatite-based carriers for tumor targeting therapy [Review of Hydroxyapatite-based carriers for tumor targeting therapy), RSC Advances 13(24), 16512, \u003cem\u003eRoyal Society of Chemistry,\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eQue, N. T., Nga, D. T., Phan, A. D., \u0026amp; Tu, L. (2024). Toward a better understanding of the photothermal heating of high-entropy-alloy nanoparticles. \u003cem\u003eMaterials\u003c/em\u003e Today\u003c/li\u003e\n\u003cli\u003eRamadas, M., Abimanyu, R., Ferreira, J. M. F., \u0026amp; Ballamurugan, A. M. (2022). Fabrication and biological evaluation of three-dimensional (3D) Mg substituted bi-phasic calcium phosphate porous scaffolds for hard tissue engineering, \u003cem\u003eRSC Advances,\u003c/em\u003e 12152\u003c/li\u003e\n\u003cli\u003eMach\u0026iacute;n \u0026amp; M\u0026aacute;rquez Ressler, A., Žužić, A., Ivani\u0026scaron;ević, I., Kamboj, N., \u0026amp; Ivanković, H. (2021). lonic substituted hydroxyapatite for bone regeneration applications: A review [Review of lonic substituted hydroxyapatite for bone regeneration applications: A review). \u003cem\u003eOpen Ceramics\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eRubungo, A. N., Arnold, C. B., Rand, B. P., \u0026amp; Dieng, A. B. (2025). LLM-Prop: predicting the properties of crystalline materials using large language models.\u003cem\u003e Apj Computational\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eSadetskaya, A. V., Bobrysheva, N. P., Osmolowsky, M. G., Osmolovskaya, O. M., \u0026amp; Voznesenskiy, M. A. (2021). Correlative experimental and theoretical characterization of transition metal doped hydroxyapatite nanoparticles fabricated by hydrothermal method. \u003cem\u003eMaterials Characterization\u003c/em\u003e, \u003cem\u003e173\u003c/em\u003e, 110911.\u003c/li\u003e\n\u003cli\u003eSahin, B., Ates, T., Acari, I. K., Barzinjy, A. A., Ates, B., \u0026Ouml;zcan, İ., ... \u0026amp; Kaygili, O. (2024). Tuning electronic properties of hydroxyapatite through controlled doping using zinc, silver, and praseodymium: A density of states and experimental study. \u003cem\u003eCeramics International\u003c/em\u003e, \u003cem\u003e50\u003c/em\u003e(5), 7919-7929.\u003c/li\u003e\n\u003cli\u003eSaleh, A. T., \u0026amp; Alameri, D. (2020). Microwave-Assisted Preparation of Zinc-Doped B-Tricalcium Phosphate for Orthopedic Applications. \u003cem\u003eIndonesian Journal of Chemistry,\u003c/em\u003e 2112), 376\u003c/li\u003e\n\u003cli\u003eSantos-Coquillat, A., Martinez-Campos, E., S\u0026aacute;nchez, H., Moreno, L., Arrabal, R., Mohedano, M., Gallardo, A., Rodriguez-Hern\u0026aacute;ndez, J., \u0026amp; Matykina, E. (2021). Hybrid functionalized coatings on Metallic Biomaterials for Tissue Engineering. \u003cem\u003eSurface and Coating\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eSayed, O., Abdalla, M. M., Elsayed, A., El-Mahallawy, Y., \u0026amp; Al-Mahalawy, H. (2024). Does strontium coated titanium implants enhance the osseointegration in animal models under osteoporotic condition? A systematic review and meta-analysis. \u003cem\u003eBDJ open\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(1), 69.\u003c/li\u003e\n\u003cli\u003eSprio, S., Preti, L., Montesi, M., Panseri, S., Adamiano, A., Vandini, A., Pugno, N. M., \u0026amp; Tampieri, A. (2019). Surface Phenomena Enhancing the Antibacterial and Osteogenic Ability of Nanocrystalline Hydroxyapatite, Activated by Multiple-lon Doping \u003cem\u003eACS Biomaterials Science \u0026amp; Engineering\u003c/em\u003e, 5(11), 5947,\u003c/li\u003e\n\u003cli\u003eSun, L., Liu, H., Ye, Y., Yang, L., Islam, R., Tan, S., Tong, R., Miao, Y., \u0026amp; Cal, L. (2023). Smart nanoparticles for cancer therapy [Review of Smart nanoparticles for cancer therapy]. Signal Transduction and Targeted Therapy, 8(1), 418. \u003cem\u003eSpringer Nature\u003c/em\u003e.\u003c/li\u003e\n\u003cli\u003eTorres-Mansilla, A., \u0026Aacute;lvarez-Lloret, P., Fern\u0026aacute;ndez-Penas, R., D\u0026apos;Urso, A., Baldi\u0026oacute;n, P. A., Oltolina, F., Follenzi, A., \u0026amp; G\u0026oacute;mez-Morales, J. (2023). Hydrothermal Transformation of Eggshell Calcium Carbonate into Apatite Micro-Nanoparticles: Cytocompatibility and Osteoinductive Properties. \u003cem\u003eNanomaterials,\u003c/em\u003e 13(16), 2290\u003c/li\u003e\n\u003cli\u003eTsukada, M., Wakamura, M., Yoshida, N., \u0026amp; Watanabe, T. (2011). Band gap and photocatalytic properties of Ti-substituted hydroxyapatite: Comparison with anatase-TiO2. \u003cem\u003eJournal of Molecular Catalysis A: Chemical\u003c/em\u003e, \u003cem\u003e338\u003c/em\u003e(1-2), 18-23.\u003c/li\u003e\n\u003cli\u003eWang, H., Li, K., Ramsay, S., Fehlis, Y., Kim, E., \u0026amp; Hattrick-Simpers, J. (2024). Evaluating the Performance and Robustness of LLMs in Materials Science \u003cem\u003eQ\u0026amp;A and Property\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eWei, L., Lu, J., Xu, H., Patel, A., Chen, Z., \u0026amp; Chen, G. (2014). Silver nanoparticles: synthesis, properties, and therapeutic applications [Review of Silver nanoparticles: synthesis properties, and therapeutic applications). \u003cem\u003eDrug\u003c/em\u003e\u003cem\u003eDiscovery. Today, \u003c/em\u003e2015), 595. Elsevie\u003c/li\u003e\n\u003cli\u003eWei, Y., He, Y., Li, X., Chen, H., \u0026amp; Deng, X. (2017). Cellular uptake and delivery-dependent effects of Tb3+-doped hydroxyapatite nanorods. \u003cem\u003eMolecules\u003c/em\u003e, \u003cem\u003e22\u003c/em\u003e(7), 1043.\u003c/li\u003e\n\u003cli\u003eNyangiwe Yang, K., Liu, T., \u0026amp; Zhang, X. (2021). Bandgap Engineering and Near-Infrared-II Optical Properties of Monolayer MoS2: A First-Principle Study. \u003cem\u003eFrontiers in Chemistry, \u003c/em\u003e9\u003c/li\u003e\n\u003cli\u003eZdorovets, M. V., Borgekov, D. B., Zhumatayeva, I. Z., Kenzhina, I., \u0026amp; Kozlovskiy, A. L. (2022) Synthesis, Properties and Photocatalytic Activity of CaTiO3-Based Ceramics Doperwth\u003c/li\u003e\n\u003cli\u003eZeng, T., Badrinarayanan, S., Ock, J., Lai, C.-K., \u0026amp; Farimani, A. B. (2025), LLM-guided Chemical Process Optimization with a Multi-Agent Approach.\u003cem\u003e arXiv (Cornell University)\u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eZhang, Y., Liu, X., Liu, Q., Wang, J., Hu, T., Lin, Y., \u0026amp; Zhang, J. (2023). CaZn(HPO3)2and Ba2Zn(HPO3)3: novel alkaline-earth zincophosphites with diversified anionic framework \u003cem\u003eDalton Transactions\u003c/em\u003e, 52(31) 10918, hic/delet\u003c/li\u003e\n\u003cli\u003eZhao, F., Xiao, H., Bai, X.-M., \u0026amp; Zu, X. (2019). Effects of Ag doping on the electronic and optical properties of CdSe quantum dots. \u003cem\u003ePhysical Chemistry Chemical Physics,\u003c/em\u003e 27(20), 15101/\u003c/li\u003e\n\u003cli\u003eZhao, R., Xiang, M., Pan, Z., Li, Y., Qian, H., Yang, X., Zhu, X., \u0026amp; Zhang, X. (2024). Recent Advances in Nanohydroxyapatite: Synthesis Methods, Biomedical Applications; and Innovations in Composites. \u003cem\u003eSSRN Electronic Journal \u003c/em\u003e\u003c/li\u003e\n\u003cli\u003eZheng, Y., Koh, H. Y., Yang, M., Li, L., May, L. T., Webb, G. I., ... \u0026amp; Church, G. (2024). Large language models in drug discovery and development: From disease mechanisms to clinical trials. \u003cem\u003earXiv preprint \u003c/em\u003earXiv:2409.04481\u003c/li\u003e\n\u003cli\u003eZimmermann, Y., Bazgir, A., Afzal, Z., Agbere, F., Ai, Q., Alampara, N., Al-Feghall, A., Ansar M., Antypov, D., Aswad, A., Bai, J., Baibakova, V., Biswajeet, D. D., Bitzek, E., Bocarsly, J D., Borisova, A. S., Bran, A. M., Brinson, L. C., Calder\u0026oacute;n, M., Rios-Garcia, M. (2024) Reflections from the 2024 Large Language Model (LLM) Hackathon for Applicates Materials Science and Chemistry,\u003cem\u003e arXiv \u003c/em\u003e(Comell University),\u003c/li\u003e\n\u003cli\u003eZimmermann, Y., Bazgir, A., Al-Feghali, A., Ansari, M., Bocarsly, J., Brinson, L. C., ... \u0026amp; Daelman, N. (2025). examples of llm applications in materials science and chemistry: Towards automation, assistants, agents, and accelerated scientific discovery. \u003cem\u003earXiv preprint \u003c/em\u003earXiv:2505.03049, 1.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-nano","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"narl","sideBox":"Learn more about [Discover Nano](https://www.springer.com/journal/11671)","snPcode":"11671","submissionUrl":"https://submission.nature.com/new-submission/11671/3","title":"Discover Nano","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8524202/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8524202/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn this paper, we report the application of Large Language Models (LLMs) in predictive modeling of infrared (IR)-triggered drug delivery systems, focusing on hydroxyapatite (HA) modified with silver nanoparticles (AgNPs), as well as dopants of zinc (Zn), titanium (Ti), magnesium (Mg), and gold (Au). We predict electronic structure, bandgap reduction, and optical properties utilizing real-world data integrating SMILES, benefiting from Darwin 1.5, a fine-tuned model, and T5, a predictive model, to predict electronic structure, bandgap values, and optical properties.\u003c/p\u003e\n\u003cp\u003eDarwin 1.5 fine-tuned via question-answering and multi-tasking on scientific datasets correlates a mean absolute deviation (MAD) of 0.72 eV to bandgap predictions that are accurate and potentially cut the simulation time by as much as 50-70% compared to the conventional density functional theory (DFT) method. The T5 model enables simulations on optical properties via the computation of absorption spectra at a wavelength of 808 nm and the concentration-dependent roles on absorption and scattering.\u003c/p\u003e\n\u003cp\u003eThe predictions obtained from electronic structure calculations with a fine-tuned Darwin 1.5 model, combined with DFT analysis, indicate that Ag-HA's band gap reduction varied from 4.312 eV (0.25 mol% Ag doping) to 3.983 eV (0.75 mol% Ag doping); from about 4.68 eV (0.25 mol% Zn doping) to approximately 4.4 eV (0.75 mol% Zn doping) in praseodymium co-doped-HA; from around 3.8 eV (0.25 mol% Ti doping) to around 3.6 eV (0.75 mol% Ti doping); from about 4.61 eV (0.25 mol% Mg doping) to about 4.39 eV (0.75 mol% Mg doping); and from about 4.3 eV (0.25 mol% Au doping) to about 3.95 eV (0.75 mol% Au doping). Moreover, it was observed that the photothermal efficacies are higher, with a value of 18.8% (internal) to 0.11 L g\u003csup\u003e−1\u003c/sup\u003e cm (external) under a concentration of 2% Ag-HA irradiated with a wavelength of 445nm, followed by comparable efficacies observed under similar conditions with a value of 11.6% (internal) to 0.055 L g\u003csup\u003e−1 \u003c/sup\u003ecm (external) with a concentration of Au-HA. Relatively low cell toxicity are observed in literature, \u003cem\u003ein vitro\u003c/em\u003e studies show fairly balanced antibacterial activity, in addition to \u003cem\u003ein vivo\u003c/em\u003e studies that reveal promoted bone healing with lowered systemic toxicity. The results are represented in detailed tables, figures, and Python codes. Ultimately, the present study\u0026nbsp; serve to promote the application of LLMs in materials science, particularly in the area of cancer and infectious diseases, as well as considerations of the ethics of AI applications and the dangers of nanomaterials.\u003c/p\u003e","manuscriptTitle":"Large Language Model Driven Predictive Modeling of Silver, Zinc, Titanium, Magnesium, Gold Doped Hydroxyapatite for Infrared Triggered Drug Delivery","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-18 16:34:48","doi":"10.21203/rs.3.rs-8524202/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-05-15T13:20:40+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-28T16:48:57+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"194411787394560580260614424805761279885","date":"2026-04-21T06:46:44+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-20T07:59:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"154355167842013532680160442010064334589","date":"2026-04-19T07:22:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"27037228131342309878758847287521173025","date":"2026-04-18T10:10:43+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-18T01:17:11+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"13144920417214347996786276206746206846","date":"2026-04-17T12:36:11+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-15T19:59:33+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"283404916151762166614724920425608103072","date":"2026-02-13T06:15:53+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-13T03:30:41+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-02-11T06:19:53+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-02-03T06:40:05+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-02-02T17:04:04+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Nano","date":"2026-02-02T16:47:52+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-nano","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"narl","sideBox":"Learn more about [Discover Nano](https://www.springer.com/journal/11671)","snPcode":"11671","submissionUrl":"https://submission.nature.com/new-submission/11671/3","title":"Discover Nano","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4bc76ff1-40db-43f3-8510-87bd6cf6cb20","owner":[],"postedDate":"February 18th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Revision requested","date":"2026-05-15T13:20:40+00:00","index":"","fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[],"tags":[],"updatedAt":"2026-05-15T13:25:47+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-18 16:34:48","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8524202","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8524202","identity":"rs-8524202","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}